Automatic cable testing method and system based on multi-modal data fusion

Abstract

The invention relates to the technical field of cable detection, and discloses an automatic cable test method and system based on multi-modal data fusion, and the method comprises the steps: synchronously collecting an electrical performance test signal and structural integrity image data of a cable, and generating a time-aligned synchronous test data set; respectively extracting electrical parameter characteristics and structural form characteristics from the synchronous test data set, and generating a multi-modal characteristic set by establishing an association mapping relation between the electrical parameter characteristics and the structural form characteristics; constructing a dynamic association network between features based on historical fault data, identifying a degradation feature mode representing progressive degradation of performance from the multi-modal feature set, and generating a feature degradation track describing a performance degradation rate and path; according to the method, dynamic risk assessment and grading are completed by integrating the life prediction result and the feature abnormal points in the real-time test, and finally a quality assessment and early warning report with both accuracy and timeliness is output.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable detection, and in particular to a cable automatic testing method and system based on multi-modal data fusion. BACKGROUND

[0002] As the core connecting component in power systems and communication networks, the performance reliability of cables is directly related to the normal operation and safety of the entire system. During manufacturing, installation and long-term operation, cables may have quality problems such as insulation performance degradation, conductor damage or structural deformation due to material defects, process fluctuations or external stress. To ensure system reliability and prevent equipment downtime or safety accidents caused by cable failure, it is necessary to conduct comprehensive quality detection and screening of cables through automatic testing means, so as to effectively distinguish between qualified products and unqualified products with hidden dangers.

[0003] The existing cable automatic testing method has the following limitations: first, most testing systems rely on independent analysis of a single type of data, such as only detecting electrical parameters or performing appearance inspection, which is difficult to fully capture the comprehensive performance state of cables under the action of multiple physical fields; second, the existing method uses fixed thresholds for qualification determination, which has insufficient recognition ability for gradual performance degradation, and cannot effectively warn products that are currently qualified but have potential rapid degradation risks; in addition, in the traditional testing process, long-term life prediction and real-time anomaly detection are disconnected, and there is a lack of dynamic correlation and fusion analysis of testing data, which limits the accuracy and timeliness of risk assessment, affecting the accurate judgment and grading control of cable quality. SUMMARY

[0004] The present application provides a cable automatic testing method and system based on multi-modal data fusion to solve the problems raised in the background.

[0005] To achieve the above purpose, the present application provides a cable automatic testing method based on multi-modal data fusion, which comprises: S1, synchronously collecting electrical performance test signals and structural integrity image data of the cable to generate time-aligned synchronous test data sets; S2, extracting electrical parameter features and structural morphology features from the synchronous test data sets, and generating a multi-modal feature set by establishing a correlation mapping relationship between the electrical parameter features and the structural morphology features; S3, constructing a dynamic correlation network between the features based on historical failure data, identifying a degradation feature mode representing gradual performance degradation from the multi-modal feature set, and generating a feature degradation trajectory describing the performance degradation rate and path; S4. Running the cable under preset aging conditions, synchronously tracking the coordinated evolution of the electrical parameter features and the structural morphology features, deducing the fault development path based on the feature degradation trajectory, and generating the remaining service life prediction of the cable; S5. Integrating the remaining service life prediction and the feature abnormal points in real-time testing, performing risk assessment and grade division, and outputting the cable quality assessment and early warning report.

[0006] Preferably, the electrical performance test signals and the structural integrity image data of the cable are synchronously collected to generate a time-aligned synchronous test dataset, including: identifying the type identifier of the cable, and determining the corresponding electrical test parameter set and image acquisition parameter set from the pre-defined parameter set based on the type identifier; using the electrical test parameter set to configure the collection mode of the electrical performance test equipment, and simultaneously using the image acquisition parameter set to configure the collection area and resolution of the structural integrity image acquisition equipment; generating a hardware synchronization trigger signal and simultaneously sending it to the electrical performance test equipment and the structural integrity image acquisition equipment to start synchronous data collection; collecting electrical performance test signals and structural integrity image data, and attaching a unified time identifier to each data sample based on the hardware synchronization trigger signal to generate a time-aligned synchronous test dataset.

[0007] Preferably, the electrical parameter features and the structural morphology features are extracted from the synchronous test dataset respectively, and a multi-modal feature set is generated by establishing a correlation mapping relationship between the electrical parameter features and the structural morphology features, including: analyzing the cable type identifier in the synchronous test dataset, retrieving the pre-configured electrical feature rule library, and applying the rule to extract the time-domain features and frequency-domain features of the electrical performance test signals; analyzing the cable type identifier in the synchronous test dataset, retrieving the pre-configured structural feature rule library, and applying the rule to extract the geometric features and texture features of the structural integrity image data; based on the cable type identifier, loading the pre-defined correlation mapping table, matching the electrical features and the structural features, and verifying the feature consistency; integrating the verified consistent electrical features and structural features to generate a multi-modal feature set.

[0008] Preferably, the correlation mapping relationship between the electrical parameter features and the structural morphology features includes: loading the pre-defined correlation rule table based on the cable type identifier to verify the consistency of the electrical parameter feature set and the structural morphology feature set; generating a feature mapping index according to the verification result to establish a dynamic correlation relationship between the electrical parameter features and the structural morphology features; The electrical parameter features and the structural morphology features are integrated based on a dynamic correlation relationship to generate a multi-modal feature set.

[0009] Preferably, a dynamic correlation network between features is constructed based on historical failure data, a degradation feature mode representing performance progressive degradation is identified from the multi-modal feature set, a feature degradation trajectory describing performance degradation rate and path is generated, including: Based on the cable type identifier in the multi-modal feature set, a pre-stored historical failure feature database is activated, and a hierarchical retrieval strategy is used to locate a failure feature group matching the current test condition; The multi-modal feature set is segmented by time window, the segmented feature sequence is compared with the matched failure feature group in mode, and the feature change point meeting the progressive degradation rule is marked; Based on the marked feature change point, a time axis sequence is established, the feature evolution direction is determined, the feature change points are connected to form a degradation path, and a feature degradation trajectory is generated.

[0010] Preferably, the time axis sequence based on the marked feature change point includes: Identify the spatiotemporal correspondence between the electrical feature mutation point and the structural feature deformation point, and mark the electrical-structural co-occurrence degradation point; Connect the electrical-structural co-occurrence degradation points in chronological order to construct a spatiotemporal evolution path; Based on the spatiotemporal evolution path, the electrical performance degradation rate and the structural deformation expansion trend are fused to generate a feature degradation trajectory with multi-dimensional degradation features.

[0011] Preferably, the cable is operated under a preset aging condition, the synergistic evolution of the electrical parameter features and the structural morphology features is tracked synchronously, the failure development path is deduced based on the feature degradation trajectory, and the remaining service life prediction of the cable is generated; Apply rated current load to the cable in a constant temperature environment to trigger stable temperature rise of the cable, and synchronously collect electrical parameter drift data and structural morphology change data; Compare the spatiotemporal correspondence of the electrical parameter drift trajectory and the structural morphology change trajectory to identify the key turning point of synergistic degradation; Divide the cable aging stages based on the key turning point, determine the duration of each stage, and generate the remaining service life prediction.

[0012] Preferably, comparing the spatiotemporal correspondence of the electrical parameter drift trajectory and the structural morphology change trajectory to identify the key turning point of synergistic degradation includes: Compare the matching degree of the electrical parameter drift rate and the structural morphology change rate to mark the rate mutation period; Extract the feature correlation mode of the rate mutation period to verify the causal relationship between electrical degradation and structural degradation; Based on the verified causal relationship, a key turning point of synergistic degradation is determined to divide the cable aging stage boundary.

[0013] Preferably, the remaining useful life prediction and the feature abnormal points in real-time testing are integrated to perform risk assessment and grade division, and a cable quality assessment and early warning report is output, including: The remaining useful life prediction is associated with the feature abnormal points in real-time testing data to establish a life-abnormality correlation mapping. Based on the life-abnormality correlation mapping, a risk grade is determined to divide the early warning level. The risk grade and the early warning level are integrated to generate a cable quality assessment and early warning report.

[0014] To solve the above problems, the application also provides a cable automatic testing system based on multi-modal data fusion, which comprises: A data acquisition module is used to synchronously acquire electrical performance test signals and structural integrity image data of the cable to generate time-aligned synchronous test data sets. A feature extraction module is used to extract electrical parameter features and structural morphology features from the synchronous test data sets, and to generate a multi-modal feature set by establishing an association mapping relationship between the electrical parameter features and the structural morphology features. A degradation trajectory generation module is used to construct a dynamic association network between the features based on historical failure data, to identify degradation feature patterns representing performance progressive degradation from the multi-modal feature set, and to generate feature degradation trajectories describing performance degradation rate and path. A fault evolution module is used to run the cable under a preset aging condition, to synchronously track the synergistic evolution of the electrical parameter features and the structural morphology features, to deduce a fault development path based on the feature degradation trajectories, and to generate a remaining useful life prediction of the cable. A quality assessment module is used to integrate the remaining useful life prediction and the feature abnormal points in real-time testing to perform risk assessment and grade division, and to output a cable quality assessment and early warning report. Advantages

[0015] Compared with the prior art, the application has the following advantages: 1. By synchronously collecting multi-modal data of electrical performance and structural integrity, a time-aligned test data set is constructed, realizing comprehensive coverage of detection dimensions. The correlation mapping relationship between electrical parameter features and structural morphology features is established, forming a multi-modal feature set that can comprehensively represent the state of the cable; a dynamic correlation network is constructed based on historical failure data, which can accurately identify performance progressive degradation patterns, generate feature degradation trajectories describing the degradation path, and effectively warn those products that are currently qualified in parameters but have potential rapid degradation risks; the synchronous tracking of the coordinated evolution of electrical and structural features under preset aging conditions realizes the accurate deduction of the failure development path, and forms a reliable remaining useful life prediction. By integrating the life prediction results and the feature abnormal points in real-time testing, dynamic risk assessment and grading are completed, and finally the quality assessment and warning report with accuracy and timeliness are output.

[0016] 2. The cable automatic test method can significantly improve the accuracy of identifying the performance degradation trend of the cable through deep fusion and dynamic analysis of multi-modal data, making the remaining useful life prediction results more consistent with the actual operation of the cable; the entire test process can complete the full-link automatic operation of data acquisition, feature processing, life prediction and risk warning without human intervention, greatly improving the efficiency and consistency of cable testing; at the same time, through the accurate assessment and grading warning of the cable quality, data support can be provided for the whole life cycle management of the cable, potential problems caused by cable failure can be avoided in advance, and the stable operation of the downstream system relying on the cable is further ensured. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flowchart of a cable automatic test method based on multi-modal data fusion provided by an embodiment of the present application is shown; Figure 2 A functional module diagram of a cable automatic test system based on multi-modal data fusion provided by an embodiment of the present application is shown; The implementation, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0019] The embodiment of the application provides a cable automatic test method based on multi-modal data fusion. The execution subject of the cable automatic test method based on multi-modal data fusion includes but is not limited to at least one of electronic devices such as a server, a terminal and the like which can be configured to execute the method provided by the embodiment of the application. In other words, the cable automatic test method based on multi-modal data fusion can be executed by software or hardware installed in a terminal device or a server device. The server includes but is not limited to a single server, a server cluster, a cloud server or a cloud server cluster and the like. The server can be a stand-alone server, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content distribution networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0020] Embodiment 1, refer to Figure 1 As shown in the figure, it is a flowchart of the cable automatic test method based on multi-modal data fusion provided by an embodiment of the application. In this embodiment, the cable automatic test method based on multi-modal data fusion includes: S1, synchronously collecting electrical performance test signals and structural integrity image data of the cable to generate time-aligned synchronous test data sets; S2, extracting electrical parameter features and structural morphology features from the synchronous test data sets respectively, and generating a multi-modal feature set by establishing an associated mapping relationship between the electrical parameter features and the structural morphology features; S3, constructing a dynamic association network between the features based on historical failure data, identifying a degradation feature mode representing performance progressive degradation from the multi-modal feature set, and generating a feature degradation trajectory describing the performance degradation rate and path; S4, running the cable under a preset aging condition, synchronously tracking the cooperative evolution of the electrical parameter features and the structural morphology features, deducing a failure development path based on the feature degradation trajectory, and generating a remaining useful life prediction of the cable; S5, integrating the remaining useful life prediction and feature abnormal points in real-time testing, performing risk assessment and grade division, and outputting a cable quality assessment and early warning report.

[0021] As a preferred embodiment of the application, synchronously collecting electrical performance test signals and structural integrity image data of the cable to generate time-aligned synchronous test data sets includes: identifying a type identifier of the cable, and determining a corresponding electrical test parameter set and an image acquisition parameter set from a predefined parameter set based on the type identifier; Use the electrical test parameter set to configure the acquisition mode of the electrical performance test equipment, and use the image acquisition parameter set to configure the acquisition area and resolution of the structural integrity image acquisition equipment. A hardware synchronization trigger signal is generated and simultaneously sent to the electrical performance testing equipment and the structural integrity image acquisition equipment to initiate synchronous data acquisition; Electrical performance test signals and structural integrity image data are acquired, and a unified time identifier is attached to each data sample based on the hardware synchronization trigger signal to generate a time-aligned synchronization test dataset.

[0022] Specifically, the type identifier is identified by reading the model code printed on the cable sheath, and the corresponding set of electrical test parameters and image acquisition parameters are retrieved from the predefined parameter library. The verification is completed by checking the consistency between the type label of the parameter set and the cable type identifier.

[0023] According to the requirements in the electrical test parameter set, adjust the acquisition mode knob of the electrical performance test equipment to the specified position. At the same time, according to the image acquisition parameter set, set the acquisition area of ​​the structural integrity image acquisition equipment to the full length of the cable and the joint part, and the resolution to a fixed pixel specification through the equipment operation panel. Confirm through the equipment display screen that the configuration parameters are completely consistent with the requirements of the set.

[0024] A standard hardware synchronous trigger signal is generated using a signal generator and simultaneously connected to the trigger interfaces of the electrical performance testing equipment and the structural integrity image acquisition equipment via dual signal transmission lines. Successful signal transmission is confirmed by observing that the trigger indicator lights on both devices illuminate simultaneously.

[0025] Two devices are started to collect data. During the acquisition process, for each electrical performance test signal sample and structural integrity image data sample acquired, the time code corresponding to the hardware synchronization trigger signal is appended to the end of the sample data as a unified time identifier. Finally, all samples with unified time identifiers are integrated to form a time-aligned synchronization test dataset. Verification is completed by checking that the time identifiers of all samples in the dataset are completely consistent.

[0026] In summary, this solution breaks through the limitations of single-data detection, simultaneously covering the two core dimensions of electrical performance and structural integrity, comprehensively capturing the overall state of cables under the influence of multiple physical fields, and avoiding the one-sidedness of detection caused by relying on only single data.

[0027] By matching specific test and acquisition parameters with cable type identifiers, and configuring device modes, acquisition areas, and resolutions accordingly, the accuracy of data acquisition is improved and errors caused by general parameters are reduced.

[0028] By using hardware synchronization triggering and a unified time identifier, strict time alignment of the two types of data is achieved, laying a reliable data foundation for subsequent association mapping and collaborative analysis of electrical and structural features, and ensuring the effectiveness of multimodal data fusion.

[0029] In a preferred embodiment of the present invention, electrical parameter features and structural morphology features are extracted from the synchronous test dataset, and a multimodal feature set is generated by establishing a correlation mapping relationship between the electrical parameter features and the structural morphology features, including: Parse the cable type identifiers in the synchronous test dataset, retrieve the pre-configured electrical feature rule base, and apply the rules to extract the time-domain and frequency-domain features of the electrical performance test signals; Parse the cable type identifier in the synchronous test dataset, retrieve the pre-configured structural feature rule base, and apply the rules to extract the geometric and texture features of the structural integrity image data; Based on the cable type identifier, a predefined association mapping table is loaded to match electrical and structural features and verify feature consistency. Integrate and verify consistent electrical and structural features to generate a multimodal feature set.

[0030] Specifically, the cable type identifier is read from the synchronous test dataset, and the corresponding entry is retrieved from the pre-configured electrical feature rule base. According to the extraction requirements specified in the entry, the waveform changes of the electrical performance test signal are observed to obtain the time domain features, the frequency domain features are obtained by observing the frequency distribution of the signal, and the verification is completed by checking that the extraction results are consistent with the feature types described in the rule base entries.

[0031] The cable type identifier is read from the synchronous test dataset. Matching entries are retrieved from the pre-configured structural feature rule base. Based on the extraction criteria specified in the entries, the geometric features of the cable are obtained by observing its external dimensions and bending angles in the structural integrity image. Texture features are obtained by observing the light and dark distribution and texture direction on the image surface. Verification is completed by checking that the extracted features completely match the requirements of the rule base entries.

[0032] Based on the parsed cable type identifier, the predefined association mapping table is retrieved, and the correspondence between electrical features and structural features in the table is searched one by one. The extracted electrical features are compared with the structural features to confirm that the performance status reflected by the electrical features matches the morphological status presented by the structural features. The verification is completed by confirming that the two match without deviation through the mapping table entries.

[0033] The time-domain and frequency-domain characteristics of the electrical performance test signals that have been verified to be consistent, along with the geometric and textural characteristics of the structural integrity image data, are organized into the same data file in a unified format to form a multimodal feature set. Confirmation is completed by checking that the file contains all target features and that there are no conflicts between the features.

[0034] In this embodiment, establishing the association mapping relationship between electrical parameter features and structural morphology features includes: loading a predefined association rule table based on the cable type identifier, and verifying the consistency between the electrical parameter feature set and the structural morphology feature set; Based on the verification results, a feature mapping index is generated to establish a dynamic correlation between electrical parameter features and structural morphology features. Based on dynamic correlation, electrical parameter features and structural morphology features are integrated to generate a multimodal feature set.

[0035] Specifically, the cable type identifier in the synchronous test dataset is read, the predefined association rule table is retrieved, and each feature in the electrical parameter feature set is compared with the corresponding feature in the structural morphology feature set. For example, the electrical signal transmission stability feature corresponds to the structural outer sheath without damage feature. Once all features are confirmed to meet the requirements of the rule table, the consistency verification is completed.

[0036] Based on the results of the consistency verification, a unique matching number is assigned to each electrical parameter feature and its corresponding structural morphology feature, and a feature mapping index table is generated. When a feature is updated, the index number is adjusted synchronously. Confirmation is completed by verifying that the correspondence between the number and the feature in the index table is accurate.

[0037] Based on the dynamic association relationship in the feature mapping index table, the electrical parameter feature set and the structural morphology feature set are integrated into a unified data document according to their numbers, ensuring that all features are completely associated without omission. By checking that the document contains all target features and that the association relationship is clear and accurate, a multimodal feature set is generated.

[0038] In summary, this solution extracts features by calling a pre-configured rule library based on cable type, accurately acquiring electrical time-domain / frequency-domain features and structural geometric / texture features, avoiding the bias of general extraction, and ensuring the specificity and accuracy of features; it establishes an association mapping and verifies consistency, breaking down the independent barriers between electrical and structural features, allowing the two types of features to form a complementary relationship, and solving the problem that a single feature cannot fully reflect the cable status.

[0039] The integrated multimodal feature set can comprehensively characterize the cable performance and structural status, providing a comprehensive and reliable feature foundation for subsequent construction of dynamic correlation networks and identification of degradation modes, thus supporting the accuracy of subsequent analysis.

[0040] As a preferred embodiment of the present invention, a dynamic correlation network between features is constructed based on historical fault data. Degradation feature patterns characterizing progressive performance degradation are identified from a multimodal feature set, and feature degradation trajectories describing the performance degradation rate and path are generated, including: Based on the cable type identifier in the multimodal feature set, the pre-stored historical fault feature database is activated, and a hierarchical retrieval strategy is used to locate the fault feature group that matches the current test conditions. The multimodal feature set is segmented into time windows, and the segmented feature sequences are compared with the matched fault feature groups to mark the feature change points that conform to the gradual degradation law. A timeline sequence is established based on the labeled feature change points to determine the direction of feature evolution. The feature change points are then connected to form a degradation path, generating a feature degradation trajectory.

[0041] Specifically, the cable type identifier in the multimodal feature set is read, the pre-stored historical fault feature database is activated, and a hierarchical retrieval strategy of first classifying by cable type and then filtering by test environment conditions is adopted to locate the fault feature group that is completely consistent with the current test conditions. The verification is completed by checking that the feature group label is consistent with the current information.

[0042] The multimodal feature set is divided into time windows according to a fixed collection time interval. The feature sequence of each window is compared with the matching fault feature group one by one. It is observed whether the feature shows a gradual change from normal to abnormal. Feature change points that conform to the gradual deterioration law are marked. Confirmation is completed by confirming that the marked points are consistent with the deterioration trend of the fault feature group.

[0043] The marked feature change points are sorted according to a unified time identifier to establish a time axis sequence. The trend of feature change before and after is observed to determine the direction of evolution. The feature change points on the time axis are connected sequentially with straight lines to form a degradation path. The path information is integrated to generate a feature degradation trajectory. The performance degradation rate and path are clearly presented by viewing the trajectory to complete the verification.

[0044] In this embodiment, establishing a timeline sequence based on marked feature change points includes: Identify the spatiotemporal correspondence between electrical feature mutation points and structural feature deformation points, and mark electrical-structural co-occurrence degradation points; By connecting the electrical-structural co-occurrence degradation points in chronological order, a spatiotemporal evolution path is constructed. Based on the spatiotemporal evolution path, the electrical performance degradation rate and structural deformation expansion trend are integrated to generate a feature degradation trajectory with multidimensional degradation characteristics.

[0045] Specifically, examine the unified time identifiers of electrical feature mutation points and structural feature deformation points in the multimodal feature set to confirm whether abnormal electrical signal mutations and obvious structural deformations occur simultaneously at the same time point. For example, if there is a sudden change in electrical transmission stability at a certain time point and the cable in the corresponding structural image shows bending deformation, mark this time point as an electrical-structural co-occurrence degradation point. Verification is completed by checking that the time identifiers are consistent and that the two types of feature changes occur simultaneously.

[0046] All electrical-structural co-occurrence degradation points are arranged sequentially according to their unified time identifiers. Co-occurrence degradation points at adjacent time points are connected by line segments to form a coherent spatiotemporal evolution path. Confirmation is completed by checking that the points on the path are arranged in chronological order without reversal or breakage.

[0047] By observing the changes in electrical performance between adjacent co-occurring degradation points in the spatiotemporal evolution path, the degradation rate can be determined. The spread of structural deformation can be examined to clarify the expansion trend. The changes in both are synchronously integrated into the spatiotemporal evolution path to form a characteristic degradation trajectory that includes multi-dimensional degradation features of electrical and structural components. The verification is completed by observing the trajectory to simultaneously present the performance degradation rate and the expansion trend of structural deformation.

[0048] In a preferred embodiment, the construction of the dynamic association network includes: using each feature vector in the historical fault feature group as a network node, and calculating association weights based on the similarity between features to form edges. The association weights are calculated using the cosine similarity formula to quantify the synchronicity of feature changes.

[0049] In the formula, Represents the feature vector in the current multimodal feature set. Feature vectors in historical fault feature groups The correlation weight between them; Representing the eigenvector The Each dimension value Representing the eigenvector The Each dimension value; Indicates the dimension of the feature vector.

[0050] Based on the calculated association weights, the dynamic association network updates the edges between nodes. Edges with weight values ​​higher than a preset threshold are retained for subsequent identification of degradation feature patterns. This network dynamically adjusts as historical fault data is updated, ensuring accurate capture of progressive performance degradation.

[0051] Specifically, each dimension value of the feature vector in the current multimodal feature set comes from electrical parameter features and structural morphology features extracted from the synchronous test dataset and verified to be consistent through association mapping. These features are formed by parsing the cable type identifier in the synchronous test dataset, retrieving the pre-configured electrical feature rule base to extract electrical time-domain and frequency-domain features, retrieving the pre-configured structural feature rule base to extract structural geometry and texture features, and then integrating them after verification to form a multimodal feature set.

[0052] Each dimension value of the feature vector in the historical fault feature group comes from the pre-stored historical fault feature database. This database stores historical fault feature groups obtained by collecting electrical performance test signals and structural integrity image data of faulty cables during past cable fault tests, and then extracting them through the same feature extraction process.

[0053] The association weight is calculated using cosine similarity. The calculation first multiplies the feature vectors in the current multimodal feature set with the corresponding dimension values ​​of the feature vectors in the historical fault feature set, and then sums all the products. Next, the squares of each dimension value of the two feature vectors are calculated and summed. The square roots of the two results are then multiplied. Finally, the sum is divided by the product to obtain the association weight, which quantifies the synchronicity of the feature changes of the two. Edges with weights higher than a preset threshold are retained for subsequent identification of degenerate feature patterns.

[0054] When the numerical changes of two feature vector dimensions are in the same direction and have similar magnitudes, the sum of their products is large, the product of the square and the square root changes little, the correlation weight increases, and the synchronization is strong; when the changes are in different directions and have large differences in magnitude, the sum of their products is small, the correlation weight decreases, and the synchronization is weak.

[0055] The dynamic association network is updated and adjusted according to historical fault data. After new historical fault data is added to the database, the new historical fault feature group is added to the network as a node. The association weight between the new node and the original node is recalculated, and edges above the threshold are retained to ensure accurate capture of progressive performance degradation.

[0056] In summary, building a dynamic correlation network based on historical fault data allows feature associations to be supported by actual fault cases, avoiding abstract associations detached from real-world scenarios, improving the reliability of electrical and structural feature associations, and providing accurate references for subsequent degradation identification.

[0057] It can accurately identify the patterns of progressive performance degradation in multimodal features, making up for the shortcomings of existing methods in identifying cables that are currently qualified but have the potential for rapid degradation. It can capture hidden degradation risks in advance and avoid missing potential products by relying solely on fixed thresholds.

[0058] The system generates characteristic degradation trajectories containing decay rates and paths, clearly presenting the dynamic process of cable performance deterioration. This provides intuitive and accurate characteristic basis for subsequent analysis, enabling the deduction of fault development paths and prediction of remaining service life under preset aging conditions, thus supporting the accuracy of subsequent analysis.

[0059] As a preferred embodiment of the present invention, the cable is operated under preset aging conditions, and the co-evolution of electrical parameter characteristics and structural morphology characteristics is tracked simultaneously. Based on the feature degradation trajectory, the fault development path is deduced, and the remaining service life prediction of the cable is generated. Apply a rated current load to the cable in a constant temperature environment to trigger a stable temperature rise in the cable, and simultaneously collect electrical parameter drift data and structural morphology change data. By comparing the spatiotemporal correspondence between the drift trajectories of electrical parameters and the trajectories of structural morphological changes, key turning points of coordinated degradation can be identified. Based on key inflection points, cable aging stages are divided, the duration of each stage is determined, and a prediction of remaining service life is generated.

[0060] Specifically, the cable is placed in a sealed constant temperature environment chamber, and the rated current load is applied through the automatic power control unit to make the cable generate a stable temperature rise. At the same time, the electrical performance testing equipment and the structural integrity image acquisition equipment are started to track the coordinated evolution of electrical parameter characteristics and structural morphological characteristics. The fault development direction is analyzed by comparing with the generated feature degradation trajectory. The remaining service life prediction is generated by combining the current deterioration state of the cable. The verification is completed by confirming that the equipment continues to operate stably and the data acquisition is uninterrupted.

[0061] Close the constant temperature chamber door and set a fixed temperature value to maintain environmental stability. Connect the cable to the circuit and apply the rated current. After the equipment displays a stable cable temperature rise value, the electrical performance testing equipment records the electrical parameter drift data in real time. The structural integrity image acquisition equipment takes a picture of the cable every hour to obtain data on structural morphological changes. Confirmation is completed by observing that the temperature rise value is constant and the data recording is continuous and without gaps.

[0062] The electrical parameter drift trajectory and structural morphology change trajectory are organized into a time series using a unified time identifier. The parameter changes and morphological changes at each time point are compared one by one. When a significant drift of electrical parameters occurs at a certain time point and morphological changes such as skin cracking appear in the corresponding structural image, the time point is marked as a critical turning point of co-deterioration. Verification is completed by checking that the time identifiers are consistent and that the two types of characteristic changes occur synchronously.

[0063] Using the key inflection points as the dividing points, the cable aging process is divided into three stages: initial stabilization, slow degradation, and rapid degradation. The start and end times of each stage are recorded to determine the duration. Combined with the standard duration from the same stage to complete cable failure in historical data, and deducting the current running time, the remaining service life of the cable is predicted. Confirmation is completed by checking the stage division and matching the characteristic degradation degree.

[0064] In this embodiment, the spatiotemporal correspondence between the electrical parameter drift trajectory and the structural morphology change trajectory is compared to identify key inflection points of collaborative degradation, including: Compare the degree of matching between the drift rate of electrical parameters and the rate of change of structural morphology, and mark the periods of abrupt rate changes; Feature correlation patterns were extracted during periods of rapid rate change to verify the causal relationship between electrical degradation and structural degradation. Based on the verified causal relationships, the key inflection points of collaborative degradation were identified, and the boundaries of the cable aging stages were delineated.

[0065] Specifically, the electrical parameter drift data and structural morphology change data are divided into fixed time periods according to a unified time identifier. The rate of change of electrical parameters and the rate of change of structural morphology are observed in each time period to see if they are synchronized. If the electrical parameter drift suddenly accelerates in a certain time period and the crack propagation rate in the structural morphology also increases significantly at the same time, the time period is marked as the rate change period. The verification is completed by checking that the rate change trends of the two are completely consistent within the time period.

[0066] From the marked rate abrupt change periods, the correspondence between electrical parameter drift characteristics and structural morphology change characteristics is extracted. For example, after the appearance of outer skin cracking characteristics in the structural morphology, the insulation resistance drift characteristics in the electrical parameters occur immediately, confirming that structural morphology deterioration is the direct cause of electrical parameter drift. The causal relationship is verified by excluding other irrelevant factors and confirming the logical sequence of characteristic changes.

[0067] Based on the start time of the rate mutation period corresponding to the verified causal relationship, this time point is determined as the key turning point of synergistic degradation. This turning point is used as the dividing point of the cable aging stage to distinguish the different aging states before and after. By checking the obvious differences in the manifestation of electrical and structural degradation characteristics before and after the turning point, the boundary of the aging stage is completed.

[0068] As a preferred embodiment, a degradation state-space model is used to quantitatively extrapolate the prediction of the cable's remaining service life, as shown in the following formula:

[0069] Indicates time Predicted remaining cable lifespan Indicates time The comprehensive degradation index is calculated by fusing electrical performance and structural morphology data from characteristic degradation trajectories. Indicates the overall degradation index over time The instantaneous rate of change, i.e. the performance degradation rate, is obtained by temporal difference of the feature degradation trajectory; This indicates the preset failure threshold, which is derived from failure data of similar cables in the historical fault feature database.

[0070] Specifically, time is first extracted from the feature degradation trajectory. The corresponding electrical performance data and structural morphology data are used. The electrical performance data represents the quantified values ​​of the electrical parameters at that point in time, and the structural morphology data represents the quantified values ​​of the structural morphology at that point in time. Then, based on the cable type, the weights of the two types of data are determined from predefined fusion rules. The quantified values ​​of the electrical performance data and structural morphology data are multiplied by their respective weights. Finally, the two weighted results are added together, and the resulting value is the time value. The comprehensive degradation index, while the feature degradation trajectory is generated by constructing a dynamic correlation network between features based on historical fault data, identifying degradation feature patterns that characterize progressive performance degradation from a multimodal feature set, and describing the performance degradation rate and path.

[0071] Finding time from feature degradation trajectory Previous adjacent time points and the adjacent time points afterwards ,extract Corresponding comprehensive degradation index and Corresponding comprehensive degradation index ,calculate minus The numerical difference is then calculated. minus The time interval is calculated by dividing the obtained numerical difference by the time interval, and the result is the comprehensive degradation index over time. The instantaneous rate of change.

[0072] The system filters out all failure data of the same type of cable that matches the current test cable type identifier from the historical fault feature database. It extracts the comprehensive degradation index of the cable when it reaches the failure state from each failure data. It summarizes all the extracted comprehensive degradation index values. If there are extreme values ​​that exceed the normal range, the extreme values ​​are removed first. Then, the arithmetic mean of all the remaining comprehensive degradation index values ​​is calculated. The arithmetic mean is the preset failure threshold.

[0073] First, obtain the preset failure threshold and time. The comprehensive degradation index is calculated by subtracting time from a preset failure threshold. The comprehensive degradation index is used to obtain the difference between the current comprehensive degradation index and the failure state, and then the comprehensive degradation index over time is obtained. The instantaneous rate of change is calculated by dividing the obtained difference value by the instantaneous rate of change; the result is the value over time. The predicted remaining lifespan of the cable, which quantifies the cable's lifespan from the current time. The duration required for a state to progress from a state of failure to a state of degradation.

[0074] When time As the overall degradation index gradually approaches the preset failure threshold, the difference between the preset failure threshold and the overall degradation index will decrease. If the instantaneous rate of change remains constant, over time... The predicted remaining cable lifespan will decrease; as the overall degradation index over time... When the instantaneous rate of change increases, even if the difference between the preset failure threshold and the comprehensive degradation index remains unchanged, over time... The predicted remaining lifespan of the cable will also be reduced.

[0075] When time As the overall degradation index gradually deviates from the preset failure threshold, the gap between the preset failure threshold and the overall degradation index will increase. If the instantaneous rate of change remains constant, over time... The predicted remaining cable lifespan will increase; as the overall degradation index over time... When the instantaneous rate of change decreases, even if the difference between the preset failure threshold and the comprehensive degradation index remains unchanged, over time... The predicted remaining lifespan of the cables will also increase.

[0076] In summary, this solution presets aging conditions, such as a constant temperature environment with a rated current load, which can simulate the aging scenario of cables in actual operation, triggering a stable and realistic degradation process. This avoids prediction deviations caused by the disconnect between the test environment and reality, and provides a real and reliable degradation data basis for subsequent analysis.

[0077] Synchronous tracking of the co-evolution of electrical and structural characteristics can accurately capture the correlation and degradation patterns between the two, such as the drift of electrical parameters caused by structural deformation. This avoids missing key turning points of co-degradation when tracking a single feature, and ensures comprehensive control over the overall degradation status of the cable.

[0078] Based on the previously generated feature degradation trajectories, fault paths can be deduced, allowing for the anchoring of the current fault development direction by relying on historical degradation patterns, rather than blindly predicting, thus improving the accuracy of fault path deduction. The final generated remaining service life prediction provides a quantitative basis for cable maintenance and replacement, solving the problem of the disconnect between traditional life prediction methods and real-time detection, and avoiding the risk of over-maintenance or sudden faults.

[0079] As a preferred embodiment of the present invention, the remaining service life prediction and characteristic anomalies in real-time testing are integrated to perform risk assessment and level classification, and output cable quality assessment and early warning reports, including: Establish a lifetime-anomaly correlation mapping by associating remaining useful life predictions with characteristic anomalies in real-time test data; Risk levels are determined and warning levels are classified based on lifetime-anomaly correlation mapping. Integrate risk levels and warning levels to generate cable quality assessment and warning reports.

[0080] Specifically, time nodes in the remaining service life prediction are extracted and characteristic anomalies in real-time test data are matched, such as sudden drops in electrical parameters and damage to the structural skin. Each anomaly is matched with its corresponding remaining service life using a unified time identifier. The impact of the anomaly on the remaining service life is clarified, and a service life-anomaly correlation mapping is established. Verification is completed by checking that the correspondence between anomalies and service life nodes in the mapping table is complete.

[0081] Referring to the predefined risk level standards, if the lifetime-anomaly correlation mapping shows a high density of anomalies and an extremely short remaining lifetime, it is judged as high risk; if there are few anomalies and the remaining lifetime is sufficient, it is judged as low risk, and correspondingly classified into three warning levels. Confirmation is completed by checking that the risk level and warning level match the rules perfectly.

[0082] The identified risk level and corresponding warning level, combined with details of characteristic anomalies and the prediction results of remaining service life, are compiled into a cable quality assessment and warning report in a fixed format. The report must clearly indicate the risk level, warning level, and key anomaly information. Verification can be completed by reviewing the report to ensure that it contains all core content and is logically coherent.

[0083] In summary, by integrating remaining useful life prediction with real-time anomaly detection, the limitations of traditional testing, which separates "life prediction from real-time anomaly detection," are overcome. This allows risk assessment to rely on both "quantified life duration" and "real-time hazard signals," avoiding situations where only lifespan is considered while ignoring sudden anomalies or where anomalies are considered before lifespan is deemed sufficient. This improves the comprehensiveness and accuracy of risk assessment.

[0084] Risk levels are categorized based on the integration results, transforming the abstract concept of "risk" into a concrete classification, such as high / medium / low risk. This avoids vague assessments and facilitates the development of targeted control strategies, such as prioritizing replacement for high-risk risks and conducting regular monitoring for low-risk risks. This reduces the waste of resources or the omission of risks associated with "one-size-fits-all" maintenance.

[0085] The output quality assessment and early warning reports clearly present the risk level, early warning level, and core anomaly information, providing an intuitive basis for decision-making in the full life cycle management of cables. This can help avoid equipment downtime or safety accidents caused by cable failures in advance, while supporting the stable operation of downstream systems and solving the problems of "untimely risk feedback and lack of direction in control" in traditional testing.

[0086] Example 2, as Figure 2 The diagram shown is a functional block diagram of an automated cable testing system based on multimodal data fusion provided in an embodiment of the present invention.

[0087] This invention discloses an automated cable testing system 100 based on multimodal data fusion, which can be installed in an electronic device. Depending on the functions implemented, the automated cable testing system 100 based on multimodal data fusion may include a data acquisition module 101, a feature extraction module 102, a degradation trajectory generation module 103, a fault evolution module 104, and a quality assessment module 105. The modules of this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0088] In this embodiment, the functions of each module / unit are as follows: The data acquisition module 101 is used to synchronously acquire electrical performance test signals and structural integrity image data of the cable, and generate a time-aligned synchronous test dataset. The feature extraction module 102 is used to extract electrical parameter features and structural morphology features from the synchronous test dataset, and generate a multimodal feature set by establishing the correlation mapping relationship between electrical parameter features and structural morphology features. The degradation trajectory generation module 103 is used to construct a dynamic correlation network between features based on historical fault data, identify degradation feature patterns that characterize progressive performance degradation from a multimodal feature set, and generate feature degradation trajectories that describe the performance degradation rate and path. The fault evolution module 104 is used to operate the cable under preset aging conditions, synchronously track the coordinated evolution of electrical parameter characteristics and structural morphological characteristics, infer the fault development path based on the feature degradation trajectory, and generate a prediction of the cable's remaining service life. The quality assessment module 105 is used to integrate the remaining service life prediction with the characteristic anomalies in real-time testing, to conduct risk assessment and classification, and to output cable quality assessment and early warning reports.

[0089] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0090] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0092] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0093] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0094] Finally, it should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An automated cable testing method based on multimodal data fusion, characterized in that, The method includes: S1. Synchronously acquire electrical performance test signals and structural integrity image data of the cable to generate a time-aligned synchronous test dataset; S2. Extract electrical parameter features and structural morphology features from the synchronous test dataset, and generate a multimodal feature set by establishing the correlation mapping relationship between electrical parameter features and structural morphology features. S3. Based on historical fault data, construct a dynamic correlation network between features, identify degradation feature patterns that characterize progressive performance degradation from the multimodal feature set, and generate feature degradation trajectories that describe the performance degradation rate and path. S4. Under preset aging conditions, the cable is operated and the co-evolution of electrical parameter characteristics and structural morphology characteristics is tracked synchronously. Based on the characteristic degradation trajectory, the fault development path is deduced and the remaining service life prediction of the cable is generated. S5 integrates the remaining service life prediction with the characteristic anomalies in real-time testing, performs risk assessment and classification, and outputs cable quality assessment and early warning reports.

2. The automated cable testing method based on multimodal data fusion as described in claim 1, characterized in that, Synchronously acquire electrical performance test signals and structural integrity image data of the cable to generate a time-aligned synchronous test dataset, including: Identify the cable type identifier and determine the corresponding electrical test parameter set and image acquisition parameter set from a predefined parameter set based on the type identifier; Use the electrical test parameter set to configure the acquisition mode of the electrical performance test equipment, and use the image acquisition parameter set to configure the acquisition area and resolution of the structural integrity image acquisition equipment. A hardware synchronization trigger signal is generated and simultaneously sent to the electrical performance testing equipment and the structural integrity image acquisition equipment to initiate synchronous data acquisition; Electrical performance test signals and structural integrity image data are acquired, and a unified time identifier is attached to each data sample based on the hardware synchronization trigger signal to generate a time-aligned synchronization test dataset.

3. The automated cable testing method based on multimodal data fusion as described in claim 1, characterized in that, Electrical parameter features and structural morphology features are extracted from the synchronous test dataset, and a multimodal feature set is generated by establishing a correlation mapping relationship between electrical parameter features and structural morphology features, including: Parse the cable type identifiers in the synchronous test dataset, retrieve the pre-configured electrical feature rule base, and apply the rules to extract the time-domain and frequency-domain features of the electrical performance test signals; Parse the cable type identifier in the synchronous test dataset, retrieve the pre-configured structural feature rule base, and apply the rules to extract the geometric and texture features of the structural integrity image data; Based on the cable type identifier, a predefined association mapping table is loaded to match electrical and structural features and verify feature consistency. Integrate and verify consistent electrical and structural features to generate a multimodal feature set.

4. The automated cable testing method based on multimodal data fusion as described in claim 3, characterized in that, Establishing the correlation mapping relationship between electrical parameter characteristics and structural morphological characteristics includes: Based on the cable type identifier, a predefined association rule table is loaded to verify the consistency between the electrical parameter feature set and the structural morphology feature set; Based on the verification results, a feature mapping index is generated to establish a dynamic correlation between electrical parameter features and structural morphology features. Based on dynamic correlation, electrical parameter features and structural morphology features are integrated to generate a multimodal feature set.

5. The automated cable testing method based on multimodal data fusion as described in claim 1, characterized in that, A dynamic correlation network between features is constructed based on historical fault data. Degradation feature patterns characterizing progressive performance degradation are identified from the multimodal feature set, generating feature degradation trajectories describing the performance degradation rate and path, including: Based on the cable type identifier in the multimodal feature set, the pre-stored historical fault feature database is activated, and a hierarchical retrieval strategy is used to locate the fault feature group that matches the current test conditions. The multimodal feature set is segmented into time windows, and the segmented feature sequences are compared with the matched fault feature groups to mark the feature change points that conform to the gradual degradation law. A timeline sequence is established based on the labeled feature change points to determine the direction of feature evolution. The feature change points are then connected to form a degradation path, generating a feature degradation trajectory.

6. The automated cable testing method based on multimodal data fusion as described in claim 5, characterized in that, Building a timeline sequence based on labeled feature change points includes: Identify the spatiotemporal correspondence between electrical feature mutation points and structural feature deformation points, and mark electrical-structural co-occurrence degradation points; By connecting the electrical-structural co-occurrence degradation points in chronological order, a spatiotemporal evolution path is constructed. Based on the spatiotemporal evolution path, the electrical performance degradation rate and structural deformation expansion trend are integrated to generate a feature degradation trajectory with multidimensional degradation characteristics.

7. The automated cable testing method based on multimodal data fusion as described in claim 1, characterized in that, The cable is operated under preset aging conditions. The co-evolution of electrical parameter characteristics and structural morphology characteristics is tracked simultaneously. Based on the characteristic degradation trajectory, the fault development path is deduced, and the remaining service life prediction of the cable is generated. Apply a rated current load to the cable in a constant temperature environment to trigger a stable temperature rise in the cable, and simultaneously collect electrical parameter drift data and structural morphology change data. By comparing the spatiotemporal correspondence between the drift trajectories of electrical parameters and the trajectories of structural morphological changes, key turning points of coordinated degradation can be identified. Based on key inflection points, cable aging stages are divided, the duration of each stage is determined, and a prediction of remaining service life is generated.

8. The automated cable testing method based on multimodal data fusion as described in claim 7, characterized in that, By comparing the spatiotemporal correspondence between the drift trajectories of electrical parameters and the trajectories of structural morphological changes, key inflection points of coordinated degradation are identified, including: Compare the degree of matching between the drift rate of electrical parameters and the rate of change of structural morphology, and mark the periods of abrupt rate changes; Feature correlation patterns were extracted during periods of rapid rate change to verify the causal relationship between electrical degradation and structural degradation. Based on the verified causal relationships, the key inflection points of collaborative degradation were identified, and the boundaries of the cable aging stages were delineated.

9. The automated cable testing method based on multimodal data fusion as described in claim 1, characterized in that, Integrating remaining service life predictions with characteristic anomalies from real-time testing, risk assessment and classification are performed, outputting cable quality assessment and early warning reports, including: Establish a lifetime-anomaly correlation mapping by associating remaining useful life predictions with characteristic anomalies in real-time test data; Risk levels are determined and warning levels are classified based on lifetime-anomaly correlation mapping. Integrate risk levels and warning levels to generate cable quality assessment and warning reports.

10. An automated cable testing system based on multimodal data fusion, characterized in that, An automated cable testing method based on multimodal data fusion as described in any one of claims 1-9, the system comprising: The data acquisition module is used to synchronously acquire electrical performance test signals and structural integrity image data of the cable, and generate a time-aligned synchronous test dataset. The feature extraction module is used to extract electrical parameter features and structural morphology features from the synchronous test dataset, and generate a multimodal feature set by establishing the correlation mapping relationship between electrical parameter features and structural morphology features. The degradation trajectory generation module is used to construct a dynamic correlation network between features based on historical fault data, identify degradation feature patterns that characterize progressive performance degradation from a multimodal feature set, and generate feature degradation trajectories that describe the rate and path of performance degradation. The fault evolution module is used to operate cables under preset aging conditions, synchronously track the coordinated evolution of electrical parameter characteristics and structural morphological characteristics, deduce the fault development path based on the feature degradation trajectory, and generate a prediction of the remaining service life of the cable. The quality assessment module integrates remaining service life prediction with characteristic anomalies in real-time testing to perform risk assessment and classification, and output cable quality assessment and early warning reports.

Images