A method and system for on-line monitoring of connector contact impedance

Abstract

This invention relates to the field of online monitoring technology, and more particularly to an online monitoring method and system for connector contact impedance. The method includes: acquiring initial material property descriptions of the connector surface and internal layers; constructing a test signal propagation model; determining the correlation curves between penetration capability and frequency, and resolution and frequency; extracting key inflection points from the correlation curves; adjusting simulation parameters and optimizing the frequency response feature set according to the attenuation values ​​of the corresponding frequency bands; separating frequency band subsets sensitive to degradation processes; grouping the response modes within the frequency band subsets; identifying sets of similar response modes; and generating a dynamic frequency band combination scheme; performing signal injection and reflection coefficient measurement based on the dynamic frequency band combination scheme to obtain microscopic change indicators of contact impedance. This method solves the problems of simultaneously balancing signal penetration capability and detection resolution, and the inability to dynamically adapt to changes in connector structural impedance, achieving adaptive online precise monitoring of connector contact impedance.

Description

Technical Field

[0001] This invention relates to the field of online monitoring technology, and in particular to an online monitoring method and system for connector contact impedance. Background Technology

[0002] In modern electronic devices, communication systems, and automotive and aerospace fields, connectors are key components for signal and power transmission, and their performance stability directly affects the reliability and safety of the entire system. During long-term service, connectors are susceptible to mechanical wear, environmental corrosion, and material aging, leading to gradual deterioration of the contact interface and consequently changes in contact impedance.

[0003] However, existing technologies for connector contact impedance monitoring still have shortcomings. Most detection methods rely on offline testing or periodic inspections, requiring equipment disassembly or system interruption. This not only fails to meet the needs of online monitoring but also cannot continuously acquire dynamic changes during the degradation process, making it difficult to detect early anomalies in a timely manner. Secondly, connectors typically have complex surface structures and multi-layered internal plating structures. The electrical characteristics of different layers of materials differ significantly, resulting in complex path distribution and attenuation characteristics of signals during propagation. In this context, traditional detection methods based on a single frequency or fixed frequency band cannot simultaneously consider signal penetration and detection resolution, failing to comprehensively reflect the microscopic degradation of the connector's interior and surface. Thirdly, changes in contact impedance exhibit distinct frequency response characteristics, with varying sensitivities to degradation at different frequencies: low-frequency signals have strong penetration but insufficient ability to distinguish minute defects; while high-frequency signals, although possessing higher resolution, are susceptible to structural complexity and noise interference, making it difficult to stably reflect deep degradation information. This makes it difficult for fixed-frequency detection methods to accurately capture the microscopic changes in contact impedance. Furthermore, in actual online monitoring scenarios, the working state and environmental conditions of connectors are constantly changing, and a single detection model is difficult to adapt to various working conditions, which limits the stability and reliability of the detection results.

[0004] Therefore, how to construct an online monitoring method that can dynamically select the detection frequency band, take into account both penetration capability and resolution, and realize real-time sensing of microscopic changes in contact impedance without interrupting system operation has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] This invention provides an online monitoring method and system for connector contact impedance, which enables accurate and adaptive online monitoring of microscopic changes in connector contact impedance.

[0006] In a first aspect, the present invention provides an online monitoring method for connector contact impedance, characterized in that the method comprises: Step S1: Obtain the initial material property description of the connector surface and internal layer data; based on the initial material property description, construct the test signal propagation model using the finite element simulation method, and obtain the propagation path and attenuation distribution of the test signal, and determine the correlation curves between penetration capability and frequency and resolution and frequency, respectively. Step S2: Extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the frequency response feature set after each frequency band is converted, separate the frequency band subset that is sensitive to the degradation process. Step S3: Group the response patterns in the frequency band subsets through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes; Step S4: Perform signal injection and reflection coefficient measurement according to the dynamic frequency band combination scheme to obtain the microscopic change index of contact impedance.

[0007] Secondly, the present invention provides an online monitoring system for connector contact impedance, used to implement the above-described method, the system comprising: The data acquisition unit is used to obtain initial material property descriptions of the connector surface and internal layers. The model building unit is used to construct a test signal propagation model using the finite element simulation method based on the initial material property description, and to obtain the propagation path and attenuation distribution of the test signal, and to determine the correlation curves between penetration capability and frequency and resolution and frequency, respectively. The frequency band extraction unit is used to extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the frequency response feature set after each frequency band is converted, a subset of frequency bands that are sensitive to the degradation process is separated. The dynamic scheme generation unit is used to group the response patterns in the frequency band subset through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes. The impedance measurement unit is used to perform signal injection and reflection coefficient measurement according to the dynamic frequency band combination scheme to obtain the microscopic change index of contact impedance.

[0008] Thirdly, the present invention provides an online monitoring device for connector contact impedance, characterized in that the device comprises: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the online monitoring device for connector contact impedance to perform the above-described method.

[0009] Fourthly, the present invention provides a computer-readable storage medium storing instructions, characterized in that the instructions, when executed by a processor, implement the above-described method. The beneficial effects of the present invention are as follows: This invention obtains material property descriptions of the connector surface and internal layers, and establishes a test signal propagation model based on finite element simulation. It generates correlation curves between penetration capability and frequency, and resolution and frequency, laying the foundation for subsequent frequency band selection and solving the problem of difficulty in quantifying signal propagation patterns in complex multi-layered structures. Based on these correlation curves, key inflection points are extracted, and model parameters are iteratively optimized using attenuation values ​​to make the frequency response feature set more closely match actual working conditions. Then, a support vector machine classifier is used to separate frequency band subsets sensitive to degradation processes, achieving precise focusing from broadband response to sensitive frequency bands, effectively filtering redundant information, and providing efficient data support for online monitoring. Cluster analysis is performed on the sensitive frequency band subsets, and the structure is identified based on the grouping results of the response patterns. The complexity is assessed by extracting key pattern features and constructing priority rules, ultimately generating a dynamic frequency band combination scheme. This scheme deeply couples the selected sensitive frequency bands with the actual structural features of the connector, achieving adaptive dynamic matching of the detection frequency bands. This allows for the coordinated allocation of high-frequency and low-frequency resources, ensuring penetration capability into deep internal interfaces while maintaining resolution for minor surface defects. This directly addresses the core contradiction of balancing penetration depth and resolution in online monitoring scenarios. Finally, signal injection and reflection coefficient measurements are performed based on the dynamic frequency band combination scheme. The microscopic change index of contact impedance is calculated through reflection coefficient inversion, generating a traceable degradation feature description. Through the cooperation of these steps, accurate and adaptive online monitoring of microscopic changes in connector contact impedance is achieved. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating an online monitoring method for connector contact impedance in one embodiment. Figure 2 This is a comparison diagram of the connector contact impedance before and after degradation in the embodiment; Figure 3 This is a structural diagram of an online monitoring system for connector contact impedance in one embodiment. Detailed Implementation

[0012] This invention provides an online monitoring method and system for connector contact impedance. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0013] For ease of understanding, the specific process of the embodiments of the present invention will be described below, such as... Figure 1 As shown, an online monitoring method for connector contact impedance in an embodiment of the present invention includes: Step S1: Obtain the initial material property description of the connector surface and internal layer data; based on the initial material property description, construct the test signal propagation model using the finite element simulation method, and calculate the propagation path and attenuation distribution of the test signal, respectively determining the correlation curves between penetration capability and frequency and resolution and frequency; In step S1, the material property description of the connector surface and internal layer data is obtained, including: Non-contact data acquisition of the connector is performed using a 3D optical profilometer, laser confocal microscope, or X-ray tomography to obtain surface texture data, contour data, and internal layer data, which are then used as multi-source data. Coordinate registration and scale normalization are performed on the multi-source data to extract the number of layers, the thickness of each layer, and their geometric arrangement. Structural complexity information is constructed based on the number of layers, thickness variation coefficient, and the number of discontinuous interfaces. Based on this structural complexity information and coating material annotation data, the distribution of coating types and thickness variation characteristics are analyzed. The multi-source data, structural complexity, number of layers, thickness of each layer, geometric arrangement, coating distribution, and thickness variation characteristics are then standardized to generate material property description data.

[0014] Specifically, during use, connectors are prone to wear, corrosion, and other degradation phenomena at their contact interfaces, leading to changes in contact impedance. Furthermore, connectors typically possess complex surface textures, multi-layered plating structures, and diverse materials, resulting in significant differences in electrical characteristics between different layers. This makes the propagation path and attenuation distribution of test signals extremely complex. Traditional detection methods based on single or fixed frequency bands struggle to simultaneously consider signal penetration and resolution, failing to accurately reflect the microscopic degradation information within the connector. Therefore, for the connectors to be monitored, a three-dimensional optical profilometer, laser scanning confocal microscope, or X-ray microtomography is used to non-contactly acquire surface texture data, contour data, and internal layer data. Surface texture data characterizes the microscopic morphological features of the contact interface, including roughness, waviness, and surface defect distribution. Contour data characterizes the external geometry of the connector, including the shape, size, and mating relationship of the contact terminals. Internal layer data characterizes the multi-layered structure of the connector in the thickness direction, including the layer sequence and thickness distribution of the substrate, intermediate transition layer, and surface plating. By performing coordinate registration and scale normalization on the above multi-source data, a complete geometric and material structure model covering the connector surface to its interior is constructed.

[0015] Based on this, the number of layers and geometric arrangement information are extracted from the aforementioned internal hierarchical data. Specifically, image segmentation algorithms are used to identify the interface boundaries between different material layers, count the number of layers, and record the sequential position of each layer in the thickness direction and its spatial geometric arrangement, such as whether the coating is continuous, whether there are local gaps, or interface overlaps. Subsequently, based on the aforementioned number of layers and geometric arrangement information, structural complexity information is calculated. This structural complexity information is used to quantify the irregularity and heterogeneity of the connector's internal structure in spatial distribution. The calculation method is as follows: using the number of layers as the first dimension, and... The thickness variation coefficients of each layer are used as the second dimension, and the number of discontinuous interfaces in the geometric arrangement is used as the third dimension. A structural complexity index is constructed through weighted normalization to characterize the complexity of reflection, refraction, and scattering paths that the signal may encounter during propagation. Based on the above structural complexity information and the obtained coating material type labeling data, the distribution of coating types and thickness variation characteristics are analyzed. The coating type distribution includes the types of coating materials such as gold plating, silver plating, and tin plating, and their distribution patterns in different areas of the connector. The thickness variation characteristics include the thickness fluctuation range and standard deviation of the same coating at different locations.

[0016] Finally, the multi-source data obtained are integrated into a unified data structure, including the number of layers, geometric arrangement information, structural complexity index, coating type distribution, and thickness variation characteristics. The dimensions of each parameter are standardized and the numerical range is normalized to form an input file that can be directly called by the finite element simulation module. Through the above technical solution, the obtained material property description data not only fully records the multi-dimensional information of the connector in terms of geometry, layer composition, and material distribution, but also provides a structured physical basis for subsequent test signal propagation path calculation, attenuation distribution analysis, and frequency response characteristic optimization, ensuring a high degree of consistency between the simulation process and the actual state of the connector.

[0017] Further, in step S1, the correlation curves between penetration capability and frequency and between resolution and frequency are determined, including: A test signal propagation model is constructed using the finite element method; simulation parameters are set based on the material property description data; the propagation path of the test signal in different frequency bands is calculated using the simulation parameters; the attenuation distribution characteristics of the signal in each frequency band are analyzed for the propagation path; and the correlation curves between penetration capability and frequency and resolution and frequency are generated based on the attenuation distribution characteristics.

[0018] Specifically, to accurately monitor minute changes during connector degradation, especially in connectors with complex multi-layered plating structures, and to provide high-sensitivity and high-resolution degradation detection capabilities, a test signal propagation model is constructed using the finite element method. Specifically targeting the specific scenario of online monitoring of connector contact impedance, a three-dimensional full-wave electromagnetic simulation test signal propagation model is built using the frequency domain finite element method. The geometric solution domain of the model completely covers the contact terminal substrate, multi-layered metal plating, contact mating interface, and surrounding air domain of the connector under monitoring. During mesh generation, a refined tetrahedral mesh is used for areas with drastic changes in signal field, such as thin plating layers and contact interfaces, while a conventional-scale mesh is used for the substrate and surrounding air domain. This approach ensures simulation accuracy while controlling computational resource consumption, guaranteeing that the model can accurately reproduce the test signal within the connector. The actual propagation behavior of the signal was analyzed. After completing the framework construction of the test signal propagation model, simulation parameters were set according to the material property description data. Specifically, the core simulation parameters of the model were set sequentially based on this data: In terms of electromagnetic parameters, the conductivity, relative permeability, and relative permittivity of each plating material (such as gold plating, nickel plating, copper substrate, etc.) were set according to the distribution of plating types, and the equivalent conductivity of the contact interface was corrected according to the surface texture data; In terms of structural parameters, the thickness, layer sequence, and interface geometric features of each plating layer were accurately set according to the number of layers and geometric arrangement information; In terms of boundary conditions, a characteristic impedance wave port matching the actual online monitoring equipment was set at the signal injection end of the connector, impedance boundary conditions were set at the contact interface, and a perfect matching layer was set on the periphery of the model to eliminate boundary reflection interference, thus completing the configuration of all parameters and model verification.

[0019] After configuring the simulation parameters, the propagation paths of the test signal in different frequency bands are calculated using these parameters. The test signal is a swept-sinusoidal excitation signal for online monitoring of the connector. The preset frequency bands cover a preset wideband range, divided into three intervals: low frequency, mid frequency, and high frequency. Each frequency band has at least N uniform sampling frequency points, where N is greater than or equal to 20, to ensure the continuity of the frequency response calculation. For each sampling frequency point, a full-wave electromagnetic simulation is performed based on the configured test signal propagation model. The spatial distribution of the electric and magnetic fields inside the connector at that frequency point is obtained through numerical solutions of Maxwell's equations in the frequency domain. Based on the field distribution, the complete propagation path of the signal is traced and analyzed, including the reflection and refraction paths at each plating interface, the scattering path at the contact interface, and the transmission path penetrating multiple plating layers into the substrate. Simultaneously, the propagation components of direct waves, multiple reflected waves, and scattered waves are distinguished to clarify the differences in the propagation paths of signals in different frequency bands. The propagation path of the low-frequency signal is characterized by penetration... The multi-layer coating primarily extends into the substrate, and the propagation path of high-frequency signals mainly follows the surface of the coating and the interlayer interface. After calculating the propagation path across the entire frequency band, the attenuation distribution characteristics of the signal in each frequency band are analyzed based on the propagation path of the test signal in each frequency band. Specifically, based on the propagation path of the test signal in each frequency band, the incident wave power, reflected wave power, and transmitted wave power of the signal at each sampling frequency point are extracted, and the total attenuation during the signal propagation process at that frequency point is calculated. At the same time, combined with the spatial distribution of the propagation path, a spatial distribution curve of the attenuation along the connector thickness direction, such as from the surface coating to the internal substrate, is generated, forming a complete attenuation distribution characteristic. On this basis, the proportion of different attenuation types is further decomposed and quantified, including ohmic attenuation caused by Joule heat loss of the material, interface reflection attenuation caused by interlayer impedance mismatch, and scattering attenuation caused by surface roughness and interface defects. At the same time, the penetration depth corresponding to the signal attenuation to the preset attenuation threshold at each frequency point, as well as the smallest identifiable structural defect size at that frequency point, are recorded, completing a full-dimensional quantitative analysis of the attenuation distribution characteristics.

[0020] Based on the attenuation distribution characteristics across the entire frequency band, correlation curves between penetration capability and frequency, and between resolution and frequency, are generated. Specifically, for the correlation curve between penetration capability and frequency, the frequency of the test signal is plotted on the x-axis, and the penetration depth corresponding to the preset attenuation threshold of the signal at the corresponding frequency in the attenuation distribution characteristics is plotted on the y-axis, creating a continuous numerical variation curve. This curve can intuitively characterize the penetration capability of different frequency signals through the multi-layered structure inside the connector. Similarly, for the correlation curve between resolution and frequency, the frequency of the test signal is plotted on the x-axis, and the smallest identifiable structural defect size of the signal at the corresponding frequency in the attenuation distribution characteristics is plotted on the y-axis, creating a continuous numerical variation curve. This curve can intuitively characterize the resolution capability of different frequency signals for microscopic defects and coating thickness variations at the connector contact interface. After plotting the curves, the least squares method is used to smoothly fit the two curves. This method eliminates numerical noise anomalies generated during simulation calculations, ensuring the continuity and accuracy of the curves. The two related curves and the frequency response corresponding to each frequency band are used as the output of the test signal propagation model. The frequency response includes signal attenuation, penetration depth, resolution values, and phase shift, i.e., the phase shift of the attenuated signal relative to the test signal. Through this technical solution, the propagation characteristics and attenuation laws of the test signal in the complex multi-layer structure of the connector within a wide frequency range can be accurately quantified. The quantitative correspondence between signal frequency and penetration capability and resolution is clarified, providing a reliable quantitative basis for subsequent sensitive frequency band screening and dynamic frequency band combination scheme generation. This fundamentally solves the core pain point that traditional fixed frequency band detection methods cannot simultaneously take into account deep penetration capability and high detection resolution, effectively ensuring the comprehensiveness and accuracy of online monitoring of connector contact impedance.

[0021] Step S2: Extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the optimized frequency response feature set, separate the frequency band subset that is sensitive to the degradation process. In step S2, adjusting the simulation parameters of the test signal propagation model and optimizing the frequency response feature set of each frequency band includes: By analyzing the correlation curves between penetration capability and frequency and between resolution and frequency, key inflection point locations are identified; for each key inflection point location, the attenuation value of the corresponding frequency band is obtained; if the attenuation value exceeds a preset attenuation value, the structural parameters of the test signal propagation model are adjusted; based on the adjusted structural parameters, the frequency response characteristics are recalculated; through the recalculated frequency response characteristics, an optimized frequency response feature set for each frequency band is generated, wherein the correlation curves include the correlation curves between penetration capability and frequency and the correlation curves between resolution and frequency.

[0022] Specifically, to adapt to the actual working conditions of online monitoring of connector contact impedance and eliminate frequency response deviations caused by the mismatch between initial simulation parameters and the complex structure and multi-layer plating characteristics of the connector, it is necessary to adjust the simulation parameters and optimize the frequency response feature set based on the simulation curves. Specifically, continuous numerical analysis and slope difference calculation are performed on the correlation curves of penetration capability and frequency and resolution and frequency obtained from the above simulations. By comparing the numerical change amplitude of adjacent frequency points on the curves, the key inflection point where the signal propagation characteristics of the curve change abruptly is accurately identified, that is, the maximum value point of the curve, which is also the core feature point characterizing the abrupt change in the signal attenuation law inside the connector. After determining the location of the key inflection point, the attenuation value of the test signal in the corresponding frequency band is extracted. The attenuation value is the test signal. The total energy loss value after propagation through the connector surface plating, internal hierarchical structure, and contact interface is quantitatively compared with a preset attenuation value set based on the connector's historical online monitoring data. This preset attenuation value is the maximum attenuation value at key turning points corresponding to various historical monitoring conditions, including normal and abnormal situations. When the attenuation value exceeds the preset attenuation value, it is determined that the currently constructed test signal propagation model cannot fully cover the actual scenario of the connector. Based on this judgment, it is possible that the initial material characteristic description collected above is not accurate enough, causing errors in the structural parameters when constructing the model. Therefore, by repeating step S1 to re-obtain the initial material characteristic description, and repeating the above steps to re-obtain the structural parameters of the model, effective coverage of the actual monitoring scenario is achieved.

[0023] After adjusting the structural parameters, a new full-band simulation was performed based on the updated model parameters to accurately calculate the frequency response characteristics of the test signal in each frequency band. These frequency response characteristics cover multi-dimensional monitoring indicators such as signal attenuation distribution, penetration depth, resolution value, and phase shift. The recalculated full-dimensional, standardized frequency response characteristics were integrated and abnormal data was removed to generate an optimized frequency response feature set. The relevant curves specifically include the correlation curves between penetration capability and frequency and resolution and frequency. Through the above parameter iterative optimization and feature reconstruction process, the deviation between the initial simulation and the actual connector structure can be effectively corrected, improving the authenticity and adaptability of the frequency response characteristics. This provides reliable data support for the accurate separation of subsequent degradation-sensitive frequency bands and ensures the stability and accuracy of the online monitoring process.

[0024] Furthermore, in step S2, a subset of frequency bands sensitive to the degradation process is separated, including: A support vector machine classifier is used to perform feature separation on the optimized frequency response feature set of each frequency band. Through feature separation, a subset of frequency bands sensitive to degradation processes is extracted. For the subset of frequency bands, their quantity distribution characteristics are analyzed. Based on the quantity distribution characteristics, statistical data of the subset of frequency bands is generated. The validity of the subset of frequency bands is determined by the statistical data. The valid subset of frequency bands is used as input data for subsequent group analysis.

[0025] Specifically, based on the optimized frequency response feature set, a support vector machine (SVM) classifier is used to separate the frequency response feature set, thereby extracting a subset of frequency bands that are sensitive to the degradation process, thus enhancing detection accuracy and providing reliable degradation monitoring data.

[0026] In the implementation process, a Support Vector Machine (SVM) classifier is used to perform feature separation on the optimized frequency response feature set. The SVM classifier is a supervised learning method that constructs a decision boundary by maximizing the classification margin, thereby classifying high-dimensional feature points in the frequency response data into different categories. Its training sample data consists of historical feature vectors and supervision labels for each frequency band of the connector. The supervision labels include sensitive and insensitive categories. The historical feature vectors include the signal attenuation distribution in the frequency response data of the test signal in each frequency band. The signal attenuation distribution characterizes the energy loss value and spatial distribution of the signal in each frequency band, and is used for subsequent identification of degradation-sensitive frequency bands. Frequency bands where the sum of the attenuation differences between each frequency band and its adjacent frequency bands, or the signal attenuation value corresponding to the key position, is greater than a set threshold are marked as sensitive; the remaining frequency bands are marked as insensitive. During training, the SVM classifier is trained based on the historical feature vectors and corresponding labels. It constructs the decision boundary between sensitive and insensitive classes by maximizing the classification margin, and optimally selects... A kernel function is used to process the nonlinear relationship in the frequency response features to obtain the optimal hyperplane for frequency band feature separation. After training, the feature vector corresponding to each frequency band in the optimized frequency response feature set is input into the classifier to extract a subset of frequency bands sensitive to degradation. These frequency bands usually exhibit significant signal attenuation and can reflect the microscopic changes in degradation, especially under the influence of factors such as coating thickness variation and material aging. For the extracted frequency band subset, its quantity distribution characteristics are further analyzed. Specifically, by statistically analyzing the number of each frequency band in the frequency band subset and their distribution ratio in the low-frequency, mid-frequency, and high-frequency ranges, detailed characteristics of the frequency band distribution are obtained. Specifically, by calculating the total number of different frequency bands in the subset and the number of frequency bands in each range, the proportion of different frequency bands in the total frequency bands can be quantified, further revealing the degradation performance of the connector in different frequency bands. For example, if the frequency band subset contains 12 frequency bands and the statistics show that the low frequency accounts for 30% and the high frequency accounts for 50%, it indicates that the degradation process of the connector is mainly concentrated in the high-frequency band, which reflects the characteristics of the complex connector structure.

[0027] Based on the distribution of frequency band subsets, statistical data of these subsets are generated. This data includes average location, variance, and interval proportions, which helps quantify the characteristics of the frequency band distribution. The effectiveness of the frequency band subsets is then assessed using this statistical data. The criteria typically include the number of subsets and a preset variance value. The number of frequency band subsets characterizes their coverage of degradation-sensitive areas, while the variance of the frequency band distribution characterizes its concentration and stability in the frequency domain. When the number of frequency band subsets exceeds a preset threshold, it indicates that the subsets have sufficient coverage in the frequency domain, avoiding misjudgments caused by only a few accidental sensitive frequencies. When the variance of the frequency band subsets is less than a preset variance threshold, it indicates that the frequency band distribution within the subsets is relatively stable. Concentration can characterize a relatively stable degradation response region. For example, if the number of subsets exceeds a set number, such as 5, and the variance is less than a preset variance value, it is considered a valid frequency band subset. Based on the joint judgment of the above quantity constraints and variance constraints, valid frequency band subsets with both coverage capability and statistical stability can be screened, thereby improving the accuracy of connector degradation monitoring results. The above technical solution, by optimizing the frequency response feature set, feature separation, and frequency band validity judgment, can accurately identify the micro-changes in the connector degradation process. Especially in connectors with complex multi-layer plating structures, it can effectively capture signal attenuation and contact impedance changes, which not only improves the sensitivity and accuracy of detection, but also ensures a complete characterization of the degradation process.

[0028] Step S3: Group the response patterns in the frequency band subsets through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes; In step S3, generating a dynamic frequency band combination scheme includes: For the frequency band subset data, cluster analysis is used to classify the response patterns and obtain preliminary pattern grouping results. Based on the preliminary pattern grouping results, the frequency band division characteristics within each group are analyzed to determine the complexity of the identification structure. By analyzing the complexity of the identification structure, key pattern features in dynamic combinations are extracted to obtain a set of combined patterns. Based on the set of combined patterns, frequency band division priority rules are constructed, and the combined patterns are sorted according to the priority rules to determine the final dynamic frequency band combination scheme.

[0029] Specifically, after extracting and validating the degradation-sensitive frequency band subset data, to adapt to the complex internal hierarchy, plating distribution, and structural variation characteristics of the connector and achieve adaptive dynamic matching of the detection frequency band, cluster analysis is used to classify the response patterns of the aforementioned frequency band subset data to obtain preliminary pattern grouping results. This frequency band subset data contains multiple frequency points and their corresponding response characteristics. The response characteristics of each frequency point include signal attenuation, penetration depth, resolution value, and phase shift at that frequency point. These characteristics collectively reflect the physical process of the interaction between the test signal and the contact interface during propagation within the connector. Considering that multiple degradation mechanisms may exist simultaneously during connector degradation, such as increased high-frequency attenuation due to contact surface oxidation and low-frequency phase distortion caused by internal plating peeling, different degradation mechanisms will exhibit differentiated response patterns in the frequency domain. The K-means clustering algorithm is used to perform unsupervised classification on the above multidimensional feature vectors. Specifically, each frequency point is treated as an independent sample, and a multidimensional feature vector containing signal attenuation amplitude, penetration depth, resolution value, and phase shift is constructed. By calculating the Euclidean distance between samples, the cluster centers are iteratively updated, dividing all frequency points into several clusters with intrinsic similarity. For example, in practical applications, the clustering results may group frequency points with high attenuation values ​​and stable phase changes in the high-frequency band into one cluster, while frequency points with drastic attenuation rate changes and unstable phases in the low-frequency band into another cluster. Each cluster corresponds to a pattern with similar frequency response characteristics. The preliminary pattern grouping results obtained thus include at least a surface oxidation sensitive pattern and an interface wear sensitive pattern, providing a structured data foundation for subsequent development of differentiated frequency band usage strategies for different degradation characteristics.

[0030] Secondly, based on the preliminary pattern grouping results, the frequency band division characteristics within each group are analyzed to determine the complexity of the identification structure. For each pattern group obtained from clustering, its internal frequency band division characteristics are further extracted. These characteristics specifically include the continuous frequency range covered by the group, i.e., the frequency band defined by the minimum and maximum values ​​of the frequency points within the group; the fluctuation amplitude of the attenuation value within the frequency band, reflecting the stability of signal attenuation; and the variance of the phase change, reflecting the consistency of the signal phase. By comprehensively analyzing the above characteristics, the complexity of the connector structure corresponding to the pattern can be identified. For example, if a certain pattern group covers a wide frequency range, spanning multiple frequency bands from low to high frequencies, and its attenuation value fluctuates drastically, the complexity of the connector structure can be identified. A large positional variance indicates that the response pattern originates from a structurally complex region within the connector, such as a multi-layered plating of different materials, abrupt geometric changes, or contact areas with multiple interfaces. Conversely, if a group covers a narrow frequency band and exhibits relatively stable attenuation and phase response, it corresponds to a relatively uniform region within the connector, such as a substrate composed of a single material. By quantifying the structural complexity into high, medium, and low levels, the subsequent priority rule construction can be differentiated based on the actual complexity of the physical structure. This achieves a mapping from abstract frequency response characteristics to specific physical structural complexity, ensuring that the dynamic frequency band combination scheme can accurately employ appropriate detection strategies for different structural regions.

[0031] Next, by analyzing the complexity of the aforementioned identification structures, key pattern features in the dynamic combination are extracted to obtain a set of combined patterns. Specifically, the aforementioned key pattern features include at least the preferred frequency band interval corresponding to the combined pattern, the typical attenuation trend, the phase response stability index, and the level of identification structure complexity associated with the physical source of the pattern. The preferred frequency band interval is selected within each pattern group by filtering out continuous frequency bands that simultaneously meet the preset resolution threshold and the preset penetration depth threshold based on the resolution value and penetration depth index in the frequency response features. The typical attenuation trend is characterized by calculating the first-order difference mean of the attenuation values ​​of all frequency points within the group, which is used to reflect the sensitivity of the pattern to the degradation process. The phase response stability index is quantified by calculating the standard deviation of the phase values ​​within the group, which is used to evaluate the repeatability of the pattern under repeated measurement conditions. The aforementioned preferred frequency band interval, typical attenuation trend, phase response stability index, and identification structure complexity are structurally encapsulated to form a combined pattern corresponding to each cluster. The combined patterns corresponding to all clusters together constitute a set of combined patterns.

[0032] Finally, based on the above set of combined modes, frequency band allocation priority rules are constructed. The specific construction of these priority rules needs to be combined with the specific scenario objectives of the current online monitoring. In this embodiment, different working modes are selected according to user needs or preset strategies. For example, in depth-priority mode, if it is necessary to focus on monitoring the degradation of the deep interface inside the connector, modes with low structural complexity (i.e., stable response and low low-frequency attenuation) are given higher priority because the areas corresponding to these modes allow the use of lower frequency bands to obtain greater penetration depth, thereby deeply probing the internal interface state. In resolution-priority mode, if it is necessary to focus on identifying minute wear or corrosion points on the surface coating, modes with high structural complexity (i.e., drastic response fluctuations and rich high-frequency components) are given higher priority because the areas corresponding to these modes require higher frequency bands to obtain higher spatial resolution, thereby distinguishing subtle structural changes. After sorting all modes in the combined mode set from high to low priority, the corresponding modes are selected sequentially from each mode according to the sorting results. The optimal frequency band range is determined, and the time occupancy ratio of each frequency band in a single signal injection cycle is determined. For example, the highest priority mode is allocated more detection time, and the lower priority mode is allocated less detection time. Thus, the final dynamic frequency band combination scheme is determined. This scheme not only clearly specifies the types of frequency bands to be used in this monitoring, i.e., which specific continuous frequency band ranges and quantities, but also specifies the excitation order and duration of each frequency band. Through the above technical solution, the generated dynamic frequency band combination scheme can be dynamically regenerated as the connector degradation process evolves, i.e., the frequency band subset data is updated in real time. This achieves adaptive adjustment of the test frequency band, enabling the online monitoring system to intelligently allocate detection resources according to the current actual response characteristics of the connector. While ensuring the penetration capability of deep interfaces, it significantly improves the resolution of minor degradation of complex coating surfaces, thereby providing the optimal excitation signal configuration for accurately measuring the microscopic changes in contact impedance in subsequent steps. This effectively overcomes the technical defects of traditional fixed-frequency band detection methods, which struggle to accommodate different structural regions and different degradation types.

[0033] Step S4: Using the aforementioned dynamic frequency band combination scheme, signal injection and reflection coefficient measurement are performed to obtain the microscopic variation index of contact impedance; specifically including: According to the dynamic frequency band combination scheme, a signal injection device is configured; a test signal is injected into the connector through the signal injection device; reflection coefficient data is collected in real time for the test signal; based on the reflection coefficient data, a microscopic change index of contact impedance is calculated; a characteristic description of the degradation process is generated based on the microscopic change index; and the characteristic description is stored as traceable detection result data.

[0034] Specifically, based on the generated dynamic frequency band combination scheme, the real-time quantification of microscopic changes in contact impedance and the visualization of degradation characteristics are achieved through precise signal excitation and response acquisition.

[0035] During implementation, a signal injection device is configured according to the aforementioned dynamic frequency band combination scheme. This scheme specifies the frequency band ranges, quantities, excitation order, and time occupancy ratios to be used in this monitoring. Based on this scheme, the signal injection device is parameter-configured. This device includes a programmable radio frequency signal generator and its front-end amplification circuit. The specific configuration process involves sequentially writing the specified frequency band ranges (e.g., several consecutive frequency ranges) from the dynamic frequency band combination scheme into the sweep sequence of the signal generator, and setting the dwell time of each frequency band according to the specified order and time ratio. Simultaneously, based on the signal attenuation characteristics of each frequency band, the corresponding output power level is set to ensure sufficient signal-to-noise ratio in different frequency bands. For example, for higher-priority frequency bands in the dynamic scheme, their dwell time can be appropriately extended and their output power increased to obtain richer response data. Test signals are injected into the connector using the signal injection device. After configuration, the signal injection device applies the test signal to the connector's test contact point via a coaxial cable and radio frequency probe according to the set sweep sequence. The test signal injection uses a reflection method. The measurement configuration involves transmitting the signal to the connector via a directional coupler, ensuring good impedance matching between the injection point and the connector contact interface to guarantee effective signal entry into the test area. During injection, the signal generator outputs continuous wave signals point-by-point according to a preset frequency step, maintaining a preset dwell time at each frequency point to ensure the signal reaches a steady state within the connector before data acquisition. Next, for the aforementioned test signal, reflection coefficient data is acquired in real-time. Simultaneously with signal injection, a vector network analyzer or an equivalent reflection coefficient measurement module is used to acquire the reflected signals at each frequency point. This reflection coefficient data includes both amplitude and phase dimensions. Amplitude reflects the degree of energy reflection at the contact interface, while phase reflects the time delay characteristics of the signal transmission path. During acquisition, the frequency point sequence and time window specified in the dynamic frequency band combination scheme are used as the synchronization reference, and the reflection coefficient is recorded point-by-point. To ensure data reliability, multiple samples are taken within the dwell time of each frequency point, and the average value is calculated to suppress random noise interference. Through this method, a complete reflection coefficient curve is obtained, covering the entire frequency band range specified by the dynamic frequency band combination scheme.

[0036] Then, based on the aforementioned reflection coefficient data, the microscopic variation index of the contact impedance is calculated. A definite functional relationship exists between the reflection coefficient and the contact impedance, specifically calculated using the reflection coefficient formula in transmission line theory. For each frequency point, using the acquired reflection coefficient amplitude and phase, combined with the known characteristic impedance (usually a system-preset standard value, such as 50 ohms), the corresponding contact impedance value is calculated. The calculation formula is as follows: ,in, The reflection coefficient, The aforementioned characteristic impedance is used to obtain a set of contact impedance curves that vary with frequency. To extract micro-change indicators, time-domain analysis or differential processing is further performed on these impedance curves. For example, by point-by-point differencing the impedance curve measured at the current moment with the impedance curve measured at the previous moment, the distribution curve of impedance change with frequency is obtained. The portion of this change exceeding a preset noise threshold is considered a micro-change indicator, used to characterize the subtle degradation of the contact interface within the monitoring time window. For example, if the impedance change continues to increase in a certain high-frequency band, it may indicate that the surface coating corresponding to that frequency band has undergone oxidation or wear; if the impedance change fluctuates in the low-frequency band, then... This may indicate that the internal interface has become loose or peeled off. Using the aforementioned micro-change indicators, a characteristic description of the degradation process is generated. The calculated micro-change indicators are then integrated with the corresponding frequency band information, timestamps, and mode grouping information in the dynamic frequency band combination scheme to form a structured degradation feature vector. This feature vector includes at least the frequency band interval where the change occurred, the impedance change amount including amplitude and polarity, the duration of the change, and the level of structural complexity corresponding to the mode group to which the frequency band belongs. Based on this, a feature fusion method, such as logistic regression, is used to generate a comprehensive characteristic description of the degradation process. The logistic regression model is based on historical data. The historical data, obtained through pre-training, consists of multiple online monitoring samples accumulated during historical monitoring processes. Each sample includes an input feature vector, such as impedance change, frequency band range of the change, duration of the change, and the complexity level of the corresponding identification structure for that frequency band. Simultaneously, professional inspectors or offline verification methods label each sample with the corresponding degradation type, such as surface oxidation, coating wear, or interface loosening, as output labels. During training, the samples are divided into training and testing sets. Maximum likelihood estimation is used to fit the model parameters, and iterative optimization minimizes the cross-entropy loss between the model's predicted class probabilities and the true labels. The classification accuracy is then verified on the test set. The training process ultimately yields a logistic regression model. When used, the aforementioned feature vectors are taken as input, and the output is the probability value of the current monitored state belonging to a preset degradation type, such as surface oxidation, coating wear, or interface loosening. Based on the degradation type corresponding to the maximum probability, a qualitative description of the degradation process is generated, such as "increased impedance in the high-frequency band indicates mild oxidation of the surface coating" or "intensified impedance fluctuations in the low-frequency band indicate loosening of the internal interface." This method achieves automated conversion from raw reflection coefficient data to degradation state information with clear physical meaning, providing an intuitive and reliable basis for subsequent state assessment and maintenance decisions.

[0037] Finally, the above-mentioned characteristic descriptions are stored as traceable test result data. The microscopic change indicators, degradation process characteristic descriptions, corresponding dynamic frequency band combination schemes, and corresponding reflection coefficient data obtained from each monitoring are all encapsulated into a structured data file and stored in a local database or cloud storage platform. During storage, an index is created according to the time series and associated with the connector's unique identification information to form a complete traceable test record. To more intuitively demonstrate the changes in contact impedance before and after degradation and their correspondence with degradation types, Figure 2 A comparison graph of connector contact impedance degradation before and after is provided, such as... Figure 2 As shown, the horizontal axis represents frequency (0.1~10 GHz), and the vertical axis represents contact impedance (Ω). The solid black line represents the contact impedance curve under normal conditions before degradation, and the dashed black line represents the contact impedance curve after degradation (e.g., surface coating oxidation). The gray shaded area indicates the frequency band with the most significant impedance change (6~9 GHz). It can be clearly seen from the figure that in the 6~9 GHz frequency band, the contact impedance after degradation shows a significant increase and more severe fluctuations, while the changes in other frequency bands are smaller, indicating that this frequency band is highly sensitive to surface coating oxidation. The results of this figure verify that the method of the present invention can effectively capture the microscopic changes in contact impedance and accurately locate the sensitive frequency band through the dynamic frequency band combination scheme, thus providing an intuitive verification basis for the identification of degradation type. Through the above technical solution, the monitoring process can adaptively utilize the optimal frequency band resources, capture the subtle changes in contact impedance in real time, and transform them into degradation feature descriptions with clear physical meaning, significantly improving the sensitivity, accuracy, and traceability of connector online monitoring.

[0038] This invention also provides an online monitoring system for connector contact impedance, used to implement the above-described method, such as... Figure 3 As shown, the system includes: The data acquisition unit is used to obtain initial material property descriptions of the connector surface and internal layers. The model building unit is used to construct a test signal propagation model using the finite element simulation method based on the initial material property description, and to obtain the propagation path and attenuation distribution of the test signal, and to determine the correlation curves between penetration capability and frequency and resolution and frequency, respectively. The frequency band extraction unit is used to extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the frequency response feature set after each frequency band is converted, a subset of frequency bands that are sensitive to the degradation process is separated. The dynamic scheme generation unit is used to group the response patterns in the frequency band subset through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes. The impedance measurement unit is used to perform signal injection and reflection coefficient measurement according to the dynamic frequency band combination scheme to obtain the microscopic change index of contact impedance.

[0039] The present invention also provides an online monitoring device for connector contact impedance, characterized in that the device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the online monitoring device for connector contact impedance to perform the above-described method.

[0040] The present invention provides a computer-readable storage medium storing instructions, wherein the instructions, when executed by a processor, implement the above-described method.

[0041] In summary, this invention obtains material property descriptions of the connector surface and internal layers, establishes a test signal propagation model based on finite element simulation, and generates correlation curves between penetration capability and frequency, and resolution and frequency, laying the foundation for subsequent frequency band selection and solving the problem of difficulty in quantifying signal propagation laws under complex multi-layer structures. Based on the aforementioned correlation curves, key inflection points are extracted, and model parameters are iteratively optimized by judging attenuation values, making the frequency response feature set more closely match actual working conditions. Then, a support vector machine classifier is used to separate frequency band subsets sensitive to degradation processes, achieving precise focusing from broadband response to sensitive frequency bands, effectively filtering redundant information, and providing efficient data support for online monitoring. Cluster analysis is performed on the sensitive frequency band subsets, and the grouping results of response patterns are used to identify… Based on the varying degrees of structural complexity, key pattern features are extracted and priority rules are constructed to ultimately generate a dynamic frequency band combination scheme. This scheme deeply couples the selected sensitive frequency bands with the actual structural features of the connector, achieving adaptive dynamic matching of the detection frequency bands. This allows for the coordinated allocation of high-frequency and low-frequency resources, ensuring penetration capability into deep internal interfaces while maintaining resolution for minor surface defects. This directly addresses the core contradiction of balancing penetration depth and resolution in online monitoring scenarios. Finally, signal injection and reflection coefficient measurements are performed based on the dynamic frequency band combination scheme. The microscopic change index of contact impedance is calculated through reflection coefficient inversion, generating a traceable degradation feature description. Through the coordination of these steps, accurate and adaptive online monitoring of microscopic changes in connector contact impedance is achieved.

[0042] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0043] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0044] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for online monitoring of connector contact impedance, characterized in that, The method includes: Step S1: Obtain the initial material property description of the connector surface and internal layer data; based on the initial material property description, construct the test signal propagation model using the finite element simulation method, and obtain the propagation path and attenuation distribution of the test signal, and determine the correlation curves between penetration capability and frequency and resolution and frequency, respectively. Step S2: Extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the frequency response feature set after each frequency band is converted, separate the frequency band subset that is sensitive to the degradation process. Step S3: Group the response patterns in the frequency band subsets through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes; Step S4: Perform signal injection and reflection coefficient measurement according to the dynamic frequency band combination scheme to obtain the microscopic change index of contact impedance.

2. The method according to claim 1, characterized in that, In step S1, the initial material property description of the connector surface and internal layer data is obtained, including: Non-contact data acquisition of the connector is performed using a 3D optical profilometer, laser confocal microscope, or X-ray tomography to obtain surface texture data, contour data, and internal layer data, which are then used as multi-source data. Coordinate registration and scale normalization are performed on the multi-source data to extract the number of layers, the thickness of each layer, and their geometric arrangement. Structural complexity information is constructed based on the number of layers, thickness variation coefficient, and the number of discontinuous interfaces. Based on this structural complexity information and coating material annotation data, the distribution of coating types and thickness variation characteristics are analyzed. The multi-source data, structural complexity, number of layers, thickness of each layer, geometric arrangement, coating distribution, and thickness variation characteristics are then standardized to generate material property description data.

3. The method according to claim 1, characterized in that, In step S1, the correlation curves between penetration capability and frequency and between resolution and frequency are determined, including: A test signal propagation model is constructed using the finite element method; simulation parameters are set based on the material property description data; the propagation path of the test signal in different frequency bands is calculated using the simulation parameters; the attenuation distribution characteristics of the signal in each frequency band are analyzed for the propagation path; and the correlation curves between penetration capability and frequency and resolution and frequency are generated based on the attenuation distribution characteristics.

4. The method according to claim 1, characterized in that, In step S2, the simulation parameters of the test signal propagation model are adjusted and the frequency response feature set is optimized, including: By analyzing the correlation curves between penetration capability and frequency and between resolution and frequency, key inflection point locations are identified; for each key inflection point location, the attenuation value of the corresponding frequency band is obtained; if the attenuation value exceeds a preset attenuation value, the structural parameters of the test signal propagation model are adjusted to cover more scenarios; based on the adjusted structural parameters, the frequency response characteristics are recalculated; through the recalculated frequency response characteristics, an optimized frequency response feature set is generated, wherein the correlation curves include the correlation curves between penetration capability and frequency and the correlation curves between resolution and frequency.

5. The method according to claim 4, characterized in that, In step S2, a subset of frequency bands sensitive to the degradation process is separated, including: A support vector machine classifier is used to perform feature separation on the optimized frequency response feature set of each frequency band. Through feature separation, a subset of frequency bands sensitive to degradation processes is extracted. For the subset of frequency bands, their quantity distribution characteristics are analyzed. Based on the quantity distribution characteristics, statistical data of the subset of frequency bands is generated. The validity of the subset of frequency bands is determined by the statistical data. The valid subset of frequency bands is used as input data for subsequent group analysis.

6. The method according to claim 1, characterized in that, In step S3, a dynamic frequency band combination scheme is generated, including: For the aforementioned frequency band subset, cluster analysis is used to classify the response patterns and obtain preliminary pattern grouping results. Based on the preliminary pattern grouping results, the frequency band division characteristics within each group are analyzed to determine the complexity of the identification structure. By analyzing the complexity of the identification structure, key pattern features in dynamic combinations are extracted to obtain a set of combined patterns. Based on the set of combined patterns, frequency band division priority rules are constructed, and the combined patterns are sorted using the priority rules to determine the final dynamic frequency band combination scheme.

7. The method according to claim 1, characterized in that, In step S4, the microscopic variation indices of contact resistance are obtained, including: According to the dynamic frequency band combination scheme, a signal injection device is configured; a test signal is injected into the connector through the signal injection device; reflection coefficient data is collected in real time for the test signal; based on the reflection coefficient data, a microscopic change index of contact impedance is calculated; a characteristic description of the degradation process is generated based on the microscopic change index; and the characteristic description is stored as traceable detection result data.

8. An online monitoring system for connector contact impedance, used to implement the method as described in any one of claims 1-7, characterized in that, The system includes: The data acquisition unit is used to obtain initial material property descriptions of the connector surface and internal layers. The model building unit is used to construct a test signal propagation model using the finite element simulation method based on the initial material property description, and to obtain the propagation path and attenuation distribution of the test signal, and to determine the correlation curves between penetration capability and frequency and resolution and frequency, respectively. The frequency band extraction unit is used to extract key inflection points from the relevant curves, obtain and adjust the simulation parameters of the test signal propagation model and optimize the frequency response feature set according to the attenuation value of the corresponding frequency band; for the frequency response feature set after each frequency band is converted, a subset of frequency bands that are sensitive to the degradation process is separated. The dynamic scheme generation unit is used to group the response patterns in the frequency band subset through cluster analysis, identify sets of similar response patterns, and generate dynamic frequency band combination schemes. The impedance measurement unit is used to perform signal injection and reflection coefficient measurement according to the dynamic frequency band combination scheme to obtain the microscopic change index of contact impedance.

9. An online monitoring device for connector contact impedance, characterized in that, The device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the online monitoring device for connector contact impedance to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the method as described in any one of claims 1-7.

Images