A high-voltage line mirror frame split type safety link structure testing method and system

Abstract

The application relates to the technical field of high-voltage line mirror frame safety testing, and discloses a high-voltage line mirror frame split type safety link structure testing method and system. The method performs plugging and bending tests on the high-voltage line mirror frame split type link structure through a mechanical arm, and collects interface pressure and structure deformation image data in real time. A pre-trained wear detection model is used to analyze interface wear and structure fatigue damage characteristics, and real-time wear degree evaluation results are output. After the test cycle ends, multi-modal data and evaluation results are integrated to generate a three-dimensional visual detection report, the concentrated wear position is located, abnormal wear points are marked, and the structure safety level is quantified. The application solves the problems of incomplete testing and inaccurate evaluation in the prior art, and improves the product safety and reliability.

Description

Technical Field

[0001] This invention relates to the field of high-voltage line mirror frame safety testing technology, specifically to a testing method and system for a split-type safety connection structure of a high-voltage line mirror frame. Background Technology

[0002] Ultra-micro ion glasses utilize ultra-micro ion technology to improve the intraocular environment and alleviate eye strain caused by excessive eye use. With the widespread application of ultra-micro ion glasses in the medical field, the reliability of their high-voltage wire frame-type safety connection structure has become crucial. However, traditional testing methods often focus only on single-dimensional performance evaluations, such as performing only plug-and-play tests or observing only the structural appearance, failing to comprehensively reflect the overall performance of the connection structure during actual use. Furthermore, existing technologies have limited ability to identify potential risks such as interface wear and structural fatigue damage, making it difficult to detect and warn of these risks in their early stages, thus increasing the risk of product failure during use. Therefore, developing a comprehensive and accurate testing method and system for evaluating the reliability of high-voltage wire frame-type safety connection structures is of paramount importance. Summary of the Invention

[0003] To address the aforementioned technical shortcomings, the purpose of this invention is to provide a testing method and system for a split-type safety connection structure of high-voltage line mirror frames, thereby solving the problems of incomplete reliability testing and inaccurate evaluation of split-type connection structures of high-voltage line mirror frames in the prior art.

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0005] In a first aspect, the present invention provides a testing method for a split-type safety connection structure of a high-voltage line mirror frame, the method comprising:

[0006] Step S100: Perform a preset number of insertion and removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion and removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, and the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range.

[0007] Step S200: Real-time acquisition of pressure data of the interface during the plug-in / plug-out test and deformation image data of the link structure during the bending test, wherein the pressure data is acquired by a pressure sensor integrated into the end of the robotic arm and the deformation image data is acquired by an image sensor;

[0008] Step S300: Input the pressure data and deformation image data into the pre-trained wear detection model, analyze the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and output the real-time wear degree assessment results.

[0009] Step S400: After the test cycle is completed, integrate the multimodal data and wear assessment results within the test cycle to generate an inspection report. This inspection report locates the concentrated wear positions, marks abnormal wear points, and quantifies the structural safety level.

[0010] Preferably, in one possible embodiment of the first aspect, the split-type connection structure includes a male connector and a female connector, wherein the male connector has a built-in spring pin that forms an electrical connection with the conductive groove of the female connector; an elastic buckle is nested on the outside of the male connector, and its protrusion is mechanically interlocked with the limiting groove of the female connector; and an insulating sleeve covers the joint between the male connector and the female connector.

[0011] Preferably, in one possible implementation of the first aspect, the wear detection model includes an input layer, a dual-channel convolutional layer, a multimodal fusion layer, and an output layer;

[0012] The input layer normalizes the pressure data sequence and deformation image data, respectively;

[0013] The dual-channel convolutional layer employs parallel temporal and spatial convolutional kernels. The temporal convolutional kernel extracts pressure fluctuation features, while the spatial convolutional kernel identifies image crack features.

[0014] The multimodal fusion layer correlates wear correlations through an adaptive feature weighting module and a time-space dual-flow analysis module;

[0015] The output layer outputs the wear level through a fully connected layer and a Softmax classifier.

[0016] Preferably, in one possible implementation of the first aspect, the adaptive feature weighting module performs the following operations:

[0017] Calculate the pressure characteristic weighting coefficient:

[0018]

[0019] in For pressure data energy entropy, Image structure entropy;

[0020] Generate fused feature vectors :

[0021]

[0022] and These are the pressure feature vector and the image feature vector, respectively.

[0023] Preferably, in one possible implementation of the first aspect, the time-space dual-stream analysis module includes:

[0024] The time-domain analysis stream uses an LSTM network to extract time-dependent features of the stress sequence:

[0025]

[0026] Spatial domain analysis streams use a ResNet-18 network to extract spatial damage features from images:

[0027]

[0028] Finally, the damage coefficient is output through a gating fusion mechanism:

[0029]

[0030] in For time step The fused feature vector, For LSTM networks, For ResNet-18 networks, , For network parameters, For the Sigmoid function, for The transpose of .

[0031] Preferably, in one possible implementation of the first aspect, step S400 specifically comprises:

[0032] After completing a preset number of test cycles, the pressure data from the insertion and removal test and the deformation image data from the bending test are integrated.

[0033] A 3D visualization inspection report is generated based on the wear assessment results. The heat map is used to locate the wear concentration area of ​​the interface and mark the spatial distribution coordinates of structural fatigue cracks.

[0034] By combining the historical safety database with the current wear level and remaining life threshold, the system outputs a structural integrity assessment conclusion that includes a quantified safety factor and markers of abnormal wear points.

[0035] Preferably, in one possible implementation of the first aspect, the security level quantification employs a dynamic weighting function:

[0036]

[0037] in The crack density factor is determined by the product of the number of cracks per unit area and the average crack length. The pressure decay factor is a weighted sum of the peak decay rate and the variance of the pressure data. The spatial correlation weight coefficient is calculated using the following formula:

[0038]

[0039] In the formula The gradient norm of the image. For the differential entropy of stress data, These are the calibration parameters for the equipment.

[0040] Secondly, the present invention provides a test system for a split-type safety connection structure of a high-voltage line mirror frame, the system comprising:

[0041] The test execution module performs a preset number of insertion and removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion and removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, while the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range.

[0042] The data acquisition module is used to acquire pressure data of the interface during plug-in and plug-out tests and deformation image data of the link structure during bending tests in real time. The pressure data is acquired by a pressure sensor integrated into the end of the robotic arm, and the deformation image data is acquired by an image sensor.

[0043] The wear analysis module is used to input pressure data and deformation image data into a pre-trained wear detection model, analyze the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and output real-time wear degree assessment results.

[0044] The report generation module is used to integrate the multimodal data and wear assessment results within the test cycle to generate an inspection report after the test cycle is completed. This inspection report locates the concentrated wear positions, marks abnormal wear points, and quantifies the structural safety level.

[0045] The beneficial effects of this invention are as follows: This invention performs a preset number of insertion and removal tests and bending tests by a robotic arm, collects data in real time by combining pressure sensors and image sensors, and uses a pre-trained wear detection model to perform in-depth analysis of multimodal data, thereby achieving accurate quantitative assessment of the wear characteristics of the interface and the fatigue damage characteristics of the high-voltage line frame split safety link structure.

[0046] This method not only improves the comprehensiveness and accuracy of testing but also enables the identification of potential risks at an early stage, providing strong support for product optimization and quality control. Simultaneously, by generating a 3D visualization inspection report, it visually displays the locations of concentrated wear and abnormal wear points, providing clear guidance for repair and replacement, effectively enhancing product safety and reliability. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This application provides a flowchart of a test method for a split-type safety connection structure for high-voltage line mirror frames.

[0049] Figure 2 This application provides a structural diagram of a high-voltage line mirror frame split-type safety link structure test system.

[0050] Figure labels: 1-Test execution module, 2-Data acquisition module, 3-Wear analysis module, 4-Report generation module. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1: As Figure 1 As shown, the present invention provides a testing method for a split-type safety connection structure of a high-voltage line mirror frame, comprising:

[0053] Step S100: Perform a preset number of insertion / removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion / removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, while the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range.

[0054] In this embodiment, the split connection structure of the high-voltage line frame includes a male connector and a female connector. The male connector has a built-in spring pin that forms an electrical connection with the conductive groove of the female connector. An elastic buckle is nested on the outside of the male connector, and its protrusion is mechanically interlocked with the limiting groove of the female connector. An insulating sleeve covers the joint between the male connector and the female connector.

[0055] The insertion and removal test uses a robotic arm to simulate the connection and separation operations in actual use, ensuring the electrical connection stability between the male spring pin and the female conductive groove. At the same time, it tests the interlocking reliability of the elastic buckle protrusion and the limiting groove, avoiding deformation of the spring pin or wear of the conductive groove due to repeated insertion and removal, which would affect the transmission performance of the ultra-micro ion glasses.

[0056] The male end of the robotic arm, a split-type connecting structure, performs a preset number of linear reciprocating motions along the interface axis. The motion speed is controlled at once per second, and the preset number of cycles is 50,000 to cover the interface wear risk under long-term use. The bending test addresses fatigue damage to the connecting structure under bending conditions. The robotic arm grasps the insulating sleeve-covered joint area and performs reciprocating bending motions within a preset angle range. The angle range is set as follows: The test simulates accidental bending or external impact during the wearing of eyeglasses to assess the mechanical interlocking strength of the elastic buckle and limiting groove, as well as the bending resistance of the insulating sleeve, preventing structural fatigue-induced fracture or electrical connection failure. During the test, the robotic arm's motion parameters are adjusted in real-time through a closed-loop control system to evaluate the stress distribution of the male spring pin during insertion and removal of the split-type connection structure and the deformation response of the female conductive groove during bending.

[0057] Step S200: Real-time acquisition of pressure data of the interface during the plug-in / plug-out test and deformation image data of the link structure during the bending test, wherein the pressure data is acquired by a pressure sensor integrated into the end of the robotic arm and the deformation image data is acquired by an image sensor.

[0058] In this embodiment, the pressure sensor adopts a six-dimensional force sensor array, which is deployed at the contact interface between the end effector of the robotic arm and the male clamp to monitor the axial positive pressure and radial shear force of the contact surface between the spring pin and the conductive groove in real time during the insertion and removal process. The sampling frequency is set to 1kHz.

[0059] The image sensor uses a high-speed industrial camera with a ring-polarized light source to capture the dynamic deformation sequence of the link structure during bending tests at a frame rate of 500Hz. The camera's optical axis is perpendicular to the mating surface of the male and female connectors. The displacement trajectory of the elastic buckle protrusion and the limiting groove is extracted in real time through a sub-pixel edge detection algorithm, and the expansion morphology of the micro-cracks on the surface of the insulating sleeve is recorded.

[0060] Step S300: Input the pressure data and deformation image data into the pre-trained wear detection model, analyze the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and output the real-time wear degree assessment results.

[0061] In this embodiment, the wear detection model is based on a deep neural network architecture, achieving accurate quantification of interface wear and structural damage through multimodal data fusion. The model input layer first normalizes the real-time pressure data sequence collected by the six-dimensional force sensor at the end of the robotic arm, linearly mapping it to... Within the range, channel separation and grayscale normalization are performed simultaneously on the deformed images captured by the image sensor to eliminate illumination interference.

[0062] The dual-channel convolutional layer employs parallel temporal and spatial convolutional kernels: the temporal convolutional kernel acts on the pressure data stream, extracting the pressure fluctuation characteristics of the contact surface between the spring pin and the conductive groove during insertion and removal through three layers of dilated convolution (dilation factors of 1, 2, and 4), and identifying the periodic pressure decay caused by metal fatigue; the spatial convolutional kernel uses a 5×5 deformable convolutional kernel to process the displacement trajectory of the elastic buckle protrusion and the limiting groove in the deformation image, enhancing the ability to extract the geometric features of micro-cracks on the surface of the insulating sleeve.

[0063] The multimodal fusion layer dynamically associates the importance weights of stress and image features through an adaptive feature weighting module, specifically performing the following operations: calculating the energy entropy of stress data. (FFT spectral entropy based on a 1024-point sliding window) and image structure entropy (Extracting edge gradient histogram entropy using the Sobel operator) to generate pressure feature weight coefficients. When the insulating sleeve develops large-area cracks, Significantly increased, Reduce the weighting of image features; when a sudden pressure change occurs during the insertion / removal process. Enhanced pressure characteristics. Fusion of feature vectors. By weighted formula Generate, where This is the 128-dimensional feature vector output by the temporal convolution. This is a 256-dimensional feature tensor output by spatial convolution.

[0064] The time-space dual-stream analysis module further correlates cross-modal damage characteristics: the time-domain analysis stream is processed using a bidirectional LSTM network. The network structure contains two 128-unit hidden layers to extract the time-dependent features of the stress sequence in the insertion and removal test. The pressure decay pattern corresponding to the increase in spring pin contact resistance is captured; the spatial domain analysis flow uses a ResNet-18 network to extract spatial damage features from bending test images. Its residual block outputs the plastic deformation area of ​​the positioning elastic buckle protrusion. Among them... For time step The fused feature vector, , For network parameters, This is the Sigmoid function. The damage coefficient is ultimately calculated using a gated fusion mechanism. When the pressure abrupt change point coincides with the image crack region in space and time, As the inner product value increases, Approaching 1 triggers a high-risk warning.

[0065] The model output layer consists of a fully connected layer and a Softmax classifier. The fully connected layer compresses the fused features to 64 dimensions, and the Softmax classifier outputs five wear levels (0: no wear, 1: slight wear, 2: moderate wear, 3: severe wear, 4: critical failure).

[0066] The wear detection model was trained using a multi-stage transfer learning and cross-modal feature alignment strategy. The training data came from the historical test database of high-voltage line frames, which included 10,000 sets of pressure time-series data from insertion and removal tests and 10,000 sets of image sequences from bending tests.

[0067] In the data preprocessing stage, the raw signal from the pressure sensor is normalized using a sliding window:

[0068]

[0069] in for The six-dimensional force sensor readings at any given time. and The mean and standard deviation are respectively the values ​​within a 1024-point sliding window to eliminate the influence of differences in device range. The deformation image data is corrected for shooting angle deviation through a perspective transformation matrix and uniformly scaled to a resolution of 512×512 pixels using a bicubic interpolation algorithm.

[0070] The transfer learning pre-training of the model is conducted in two stages: the first stage trains a dual-channel convolutional layer on a general dataset of industrial connectors. The temporal convolutional kernels use three layers of convolutions with increasing dilation rates (d=1,2,4) to learn the ability to extract periodic features of pressure fluctuations. The loss function is:

[0071]

[0072] in To reconstruct the feature vector, the spatial convolution kernel uses a deformable convolution module, which adaptively captures the geometric features of the crack through offset learning.

[0073]

[0074] Here The offset tensor for deformable convolution is used, and a regularization term enhances the stability of deformation modeling. This stage freezes the parameters of the multimodal fusion layer, focusing on optimizing the low-level feature extractor.

[0075] The second stage uses a proprietary dataset of high-voltage line mirror frames for domain adaptation fine-tuning. A gradient inversion layer is introduced to achieve cross-device feature alignment.

[0076]

[0077] Among them, the discriminator Distinguish between device source domains and generators For feature extraction networks, adversarial training forces the feature vectors to... The distribution is consistent across device domains. Additionally, a feature decoupler is embedded at the end of the ResNet-18 network.

[0078]

[0079] orthogonal constraint loss Characteristics of noise and inherent damage associated with separation equipment.

[0080] The historical safety threshold database was constructed using hierarchical clustering technology. The validation set samples were processed using 128-dimensional fusion features extracted via a network. The OPTICS clustering algorithm is used to generate five security level feature clusters. The center vector of each cluster is... As a baseline threshold, key statistics for samples within the cluster are calculated:

[0081] Pressure decay rate ;

[0082] Crack propagation rate ;

[0083] During inference, cosine similarity is used to match the current feature with the cluster center:

[0084]

[0085] Combining the historical failure cycle distribution of the matching cluster Calculate the remaining life prediction value .

[0086] After inputting real-time pressure data and deformation image data into the pre-trained wear detection model, the analysis and evaluation are performed according to the following process:

[0087] First, the model's dual-channel processing architecture synchronously parses multimodal data. The temporal analysis channel receives real-time pressure sequences acquired by a six-dimensional force sensor at the end of the robotic arm, and extracts dynamic pressure features from the insertion / removal test using a three-layer dilated convolution kernel: the first layer identifies the impact peak at the moment the spring pin contacts the conductive groove; the second layer captures the attenuation trend of pressure fluctuations caused by periodic insertion / removal; and the third layer detects abnormal pressure abrupt changes caused by metal fatigue. The spatial analysis channel processes deformation images captured by a high-speed camera, employing deformable convolution kernels to enhance adaptability to microscopic features: on the one hand, it locates the displacement trajectory offset of the elastic buckle protrusion during bending tests; on the other hand, it quantifies the propagation length and branching morphology of cracks on the insulating sleeve surface, especially performing pixel-level edge analysis on the stress concentration area at the edge of the limiting groove.

[0088] Subsequently, a multimodal fusion layer establishes a cross-domain damage correlation mechanism. The adaptive feature weighting module dynamically allocates weights based on the real-time ratio of image structural entropy to pressure energy entropy: when a large-area crack appears in the insulating sleeve, the contribution of image features to the fusion vector is significantly increased; when an abnormal pressure change is detected during the insertion / removal process, the analysis priority of pressure features is strengthened. The temporal-spatial dual-stream analysis module further extracts the time-dependent patterns in the pressure sequence through a bidirectional LSTM network, identifying the pressure decay period corresponding to the increase in spring pin contact resistance, while the spatial feature tensor output by the ResNet-18 network locates the plastic deformation area of ​​the elastic buckle. The gated fusion mechanism aligns the two types of features spatiotemporally; when the pressure change time point coincides with the spatial coordinates of the image crack, a high-risk damage coefficient threshold is triggered.

[0089] Finally, the model combines transfer learning with a historical threshold database to generate quantitative assessments. It utilizes cross-device universal features established during pre-training to eliminate the impact of range biases between different testing devices; and decouples device-related noise and intrinsic damage features through gradient inversion layer adversarial training. During online inference, it dynamically matches the current fused feature vector with the historical safety database: using cosine similarity to retrieve the closest safety level feature cluster, combined with the statistical distribution parameters of the cluster's historical failure cycles (such as pressure decay rate). Crack propagation rate It outputs real-time assessment results including five wear classification levels (0 to 4), and simultaneously generates three-dimensional coordinates of the damage location and predicted remaining life. During the test, it continuously provides feedback on the risk of insulation sleeve crack propagation and early warning of conductive groove contact failure.

[0090] Step S400: After the test cycle is completed, integrate the multimodal data and wear assessment results within the test cycle to generate an inspection report. This inspection report locates the concentrated wear positions, marks abnormal wear points, and quantifies the structural safety level.

[0091] In this embodiment, based on the real-time evaluation results output by the wear detection model, the system extracts the pressure data of the insertion and removal test and the deformation image data of the bending test within the test cycle.

[0092] Statistical characteristics of pressure data include the peak pressure decay rate and fluctuation variance, which are used to calculate the pressure decay factor. The deformed image is analyzed using a sub-pixel edge detection algorithm to extract the number of cracks per unit area and the average crack length, generating a crack density factor. .

[0093] Based on this dataset, the system constructs a 3D visualization inspection report. Using an OpenGL rendering engine, it maps the damage features of the male-female connector joint area to a 3D coordinate system: a thermal layer is used to locate areas of concentrated interface wear, with the chromaticity gradient driven by local wear level values. Red highlight areas correspond to conductive groove contact surfaces and elastic buckle protrusions with wear levels greater than or equal to 3. Simultaneously, a spatial annotation layer marks the distribution of structural fatigue cracks, accurately calibrating the crack initiation and termination coordinates (such as the coordinates of the limiting groove edge) on the 3D model surface. The coordinates of the center of the microcrack clusters on the surface of the insulating sleeve.

[0094] The security level quantization module executes a dynamic weighting function. Spatial correlation weight coefficient Determined through real-time calculation: In the formula Given the image gradient norm, the Sobel operator is used to perform a convolution operation on the current deformed image to extract the sum of squared edge intensities. of The pressure data differential entropy is calculated from the entropy difference between adjacent sampling points in a six-dimensional force sensor sequence, and is used for equipment calibration parameters. Retrieved by matching from historical security database.

[0095] In this embodiment, the historical safety database is constructed by accumulating historical test data, containing over 50,000 test records of high-voltage line frame split-type connection structures. Each record stores multimodal data features (such as pressure attenuation factor). Crack density factor Wear level labels and corresponding actual failure cycle data. The database uses a hierarchical index structure, categorized by security level. A model that associates three-dimensional damage coordinates, material fatigue parameters, and remaining life using a primary key. It supports real-time feature vector matching via cosine similarity. With the center of the historical cluster .

[0096] When large-area cracks appear in the insulating sleeve, the image gradient norm increases significantly. The contribution of the crack density factor approaching 1; when abnormal pressure decay is detected during insertion / extraction testing, the differential entropy increases, leading to... Decrease and increase the weight of the pressure attenuation factor.

[0097] The system incorporates a failure threshold library from the historical security database (containing thresholds corresponding to each security level). (threshold range), will be the currently calculated Values ​​are mapped to five security levels (Level A: Grade B: Grade C: Grade D: Class E: (and correlate the historical remaining lifetime distribution of that level) Predict remaining service life. The final test report includes a quantified safety factor. The system provides the coordinates of the wear concentration area marked with values ​​and three-dimensional thermal maps, the location markers of abnormal wear points, and the structural integrity assessment conclusions, giving the "pass / fail" safety certification result and maintenance suggestions for key failure locations.

[0098] Example 2: This invention provides a high-voltage line mirror frame split-type safety connection structure testing system, comprising:

[0099] Test execution module 1 performs a preset number of insertion and removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion and removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, while the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range.

[0100] Data acquisition module 2 collects pressure data of the interface during plug-in / plug-out testing and deformation image data of the link structure during bending testing in real time. The pressure data is acquired by a pressure sensor integrated into the end of the robotic arm, and the deformation image data is acquired by an image sensor.

[0101] Wear analysis module 3 inputs pressure data and deformation image data into a pre-trained wear detection model, analyzes the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and outputs real-time wear degree assessment results.

[0102] The report generation module 4, after completing the test cycle, integrates the multimodal data and wear assessment results within the test cycle to generate a test report. This test report locates the concentrated wear positions and quantifies the structural safety level, while also marking abnormal wear points.

[0103] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A test method for a split-type safety connection structure of a high-voltage line mirror frame, characterized in that, The method includes: Step S100: Perform a preset number of insertion and removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion and removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, and the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range. Step S200: Real-time acquisition of pressure data of the interface during the plug-in / plug-out test and deformation image data of the link structure during the bending test, wherein the pressure data is acquired by a pressure sensor integrated into the end of the robotic arm and the deformation image data is acquired by an image sensor; Step S300: Input the pressure data and deformation image data into the pre-trained wear detection model, analyze the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and output the real-time wear degree assessment results. Step S400: After the test cycle is completed, integrate the multimodal data and wear assessment results within the test cycle to generate an inspection report. This inspection report locates the concentrated wear positions, marks abnormal wear points, and quantifies the structural safety level.

2. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 1, characterized in that, The split-type connection structure includes a male connector and a female connector. The male connector has a built-in spring pin that forms an electrical connection with the conductive groove of the female connector. An elastic buckle is nested on the outside of the male connector, and its protrusion is mechanically interlocked with the limiting groove of the female connector. An insulating sleeve covers the joint between the male connector and the female connector.

3. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 1, characterized in that, The wear detection model includes an input layer, a dual-channel convolutional layer, a multimodal fusion layer, and an output layer; The input layer normalizes the pressure data sequence and deformation image data, respectively; The dual-channel convolutional layer employs parallel temporal and spatial convolutional kernels. The temporal convolutional kernel extracts pressure fluctuation features, while the spatial convolutional kernel identifies image crack features. The multimodal fusion layer correlates wear correlations through an adaptive feature weighting module and a time-space dual-flow analysis module; The output layer outputs the wear level through a fully connected layer and a Softmax classifier.

4. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 3, characterized in that, The adaptive feature weighting module performs the following operations: Calculate the pressure characteristic weighting coefficient: in For pressure data energy entropy, Image structure entropy; Generate fused feature vectors : and These are the pressure feature vector and the image feature vector, respectively.

5. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 4, characterized in that, The time-space dual-stream analysis module includes: The time-domain analysis stream uses an LSTM network to extract time-dependent features of the stress sequence: Spatial domain analysis streams use a ResNet-18 network to extract spatial damage features from images: Finally, the damage coefficient is output through a gating fusion mechanism: in For time step The fused feature vector, For LSTM networks, For ResNet-18 networks, , For network parameters, For the Sigmoid function, for The transpose of .

6. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 1, characterized in that, Step S400 specifically involves: After completing a preset number of test cycles, the pressure data from the insertion and removal test and the deformation image data from the bending test are integrated. A 3D visualization inspection report is generated based on the wear assessment results. The heat map is used to locate the wear concentration area of ​​the interface and mark the spatial distribution coordinates of structural fatigue cracks. By combining the historical safety database with the current wear level and remaining life threshold, the system outputs a structural integrity assessment conclusion that includes a quantified safety factor and markers of abnormal wear points.

7. The test method for a split-type safety connection structure of a high-voltage line mirror frame as described in claim 6, characterized in that, The security level quantification uses a dynamic weighting function: in The crack density factor is determined by the product of the number of cracks per unit area and the average crack length. The pressure decay factor is a weighted sum of the peak decay rate and the variance of the pressure data. The spatial correlation weight coefficient is calculated using the following formula: In the formula The gradient norm of the image. For the differential entropy of stress data, These are the calibration parameters for the equipment.

8. A test system for a split-type safety connection structure of a high-voltage line mirror frame, characterized in that, The system includes: The test execution module performs a preset number of insertion and removal tests and bending tests on the split connection structure of the high-voltage line frame using a robotic arm. The insertion and removal test controls the robotic arm to perform linear reciprocating motion along the interface axis, while the bending test controls the robotic arm to perform reciprocating bending motion within a preset angle range. The data acquisition module is used to acquire pressure data of the interface during plug-in and plug-out tests and deformation image data of the link structure during bending tests in real time. The pressure data is acquired by a pressure sensor integrated into the end of the robotic arm, and the deformation image data is acquired by an image sensor. The wear analysis module is used to input pressure data and deformation image data into a pre-trained wear detection model, analyze the interface wear characteristics caused by the insertion and removal process and the structural fatigue damage characteristics caused by the bending process, and output real-time wear degree assessment results. The report generation module is used to integrate multimodal data and wear assessment results from the test cycle to generate an inspection report after the test cycle is completed. The inspection report locates the concentrated wear position, marks abnormal wear points, and quantifies the structural safety level.

Images