Multi-mode cable defect identification method and system

Abstract

The embodiment of the invention discloses a multi-mode cable defect identification method and system. According to the technical scheme provided by the embodiment of the invention, under the condition that the cable parameter of any mode is monitored to be abnormal, at least one other available mode associated with the current abnormal parameter is determined; acquiring real-time parameter data corresponding to other available modes; and performing independent evaluation on the real-time parameter data of each mode to obtain a mode evaluation result, performing fusion decision on all the mode evaluation results, and generating a final monitoring result of the cable defect. By adopting the technical means, the real-time parameter data of the available modes associated with the abnormal parameters are subjected to fusion evaluation, so that an accurate and comprehensive cable defect monitoring result is obtained, the situation of misjudgment of single-mode cable defect detection is avoided, the cable defect detection precision is improved, the cable operation stability and reliability are further improved, and the cable defect detection efficiency is improved. And the cable operation and maintenance effect is improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of cable, in particular to a multi-modal cable defect identification method and system. BACKGROUND

[0002] At present, as the core transmission carrier of key infrastructure such as power and communication, the operation state of the cable is directly related to the power supply reliability and communication stability. The traditional cable defect detection technology mainly relies on single modal sensing data for fault diagnosis, such as monitoring the cable temperature field change through distributed temperature sensors, or collecting electrical signal changes through partial discharge sensors. Whether the cable has defects or potential abnormal conditions is analyzed through the change of single modal sensor data.

[0003] However, the existing cable defect detection scheme is mostly based on single modal sensor data. Even if multiple types of sensors are added, each sensor is analyzed independently in parallel mode (such as monitoring temperature, current, and partial discharge at the same time, but determining the defect detection result independently). Single modal sensor data is one-sided and difficult to accurately and comprehensively reflect the cable defect situation, which can easily lead to defect detection misjudgment and low defect detection accuracy. SUMMARY

[0004] Embodiments of the present application provide a multi-modal cable defect identification method and system, which can dynamically associate multi-modal information of the cable, improve the cable defect detection accuracy through cross-modal collaborative verification, and solve the technical problem of single modal cable defect detection misjudgment.

[0005] In a first aspect, embodiments of the present application provide a multi-modal cable defect identification method, comprising: In the case that any modal cable parameter is monitored to have an abnormality, at least one other available modal associated with the current abnormal parameter is determined; Real-time parameter data corresponding to the other available modal is obtained; Each modal real-time parameter data is separately evaluated to obtain a modal evaluation result, all modal evaluation results are fused and decided to generate a final monitoring result of the cable defect.

[0006] Further, the determination of at least one other available modal associated with the current abnormal parameter comprises: According to the abnormal modal to which the current abnormal parameter belongs, a pre-defined modal association mapping table is queried to determine at least one other available modal associated with the abnormal modal.

[0007] Further, before the pre-defined modal association mapping table is queried according to the abnormal modal to which the current abnormal parameter belongs, it further comprises: According to the cable topology structure and a historical defect data set, an association weight matrix between different modal parameters is established, and a modal association mapping table is constructed based on the association weight matrix, and the association weight matrix is dynamically corrected by a Pearson correlation coefficient between modes.

[0008] Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude. Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude.

[0009] Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude. Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude.

[0010] Further, before the weighted sum based on the fusion weight is generated to generate the final monitoring result of the cable defect, the method further includes: Further, before the weighted sum based on the fusion weight is generated to generate the final monitoring result of the cable defect, the method further includes:

[0011] Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude. Further, the real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine a data deviation amplitude, and a modal evaluation result is generated based on the data deviation amplitude.

[0012] In a second aspect, an embodiment of the present application provides a multi-modal cable defect identification system, comprising: A detection module is configured to determine at least one other available mode associated with a current abnormal parameter when it is monitored that the cable parameter of any mode is abnormal. An acquisition module is configured to acquire real-time parameter data corresponding to the other available mode. A fusion module is configured to perform individual evaluation on the real-time parameter data of each mode to obtain a modal evaluation result, and perform fusion decision on all modal evaluation results to generate a final monitoring result of a cable defect.

[0013] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory and one or more processors; the memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-modal cable defect identification method as described in the first aspect.

[0014] In a fourth aspect, the embodiments of the present application provide a storage medium containing computer executable instructions for executing the multi-modal cable defect identification method as described in the first aspect when executed by a computer processor.

[0015] The embodiments of the present application determine at least one other available modality associated with the current abnormal parameter when any modality of the cable parameter is monitored to be abnormal, obtain real-time parameter data corresponding to the other available modality, perform individual evaluation on the real-time parameter data of each modality to obtain modality evaluation results, fuse all modality evaluation results to generate a final monitoring result of the cable defect. By using the above technical means, the real-time parameter data of the available modality associated with the abnormal parameter is fused and evaluated, so as to obtain an accurate and comprehensive cable defect monitoring result, avoid the situation of misjudgment of single modality cable defect detection, improve the cable defect detection precision, and further improve the cable operation stability and reliability, and improve the cable operation and maintenance effect. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a flowchart of a multi-modal cable defect identification method provided by the first embodiment of the present application; Figure 2 is a structural schematic diagram of a cable defect identification architecture in the first embodiment of the present application; Figure 3 is a flowchart of generation of a modality evaluation result in the first embodiment of the present application; Figure 4 is a flowchart of generation of another modality evaluation result in the first embodiment of the present application; Figure 5 is a structural schematic diagram of a multi-modal cable defect identification system provided by the second embodiment of the present application; Figure 6 is a structural schematic diagram of an electronic device provided by the third embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to make the purposes, technical solutions and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below with reference to the drawings. It can be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, for the convenience of description, only parts related to the present application are shown in the drawings, but not all. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowchart describes each operation (or step) as a sequential process, many of the operations can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operations are completed, but can also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, etc.

[0018] Embodiment one: Figure 1 A flowchart of a multi-modal cable defect identification method provided by the embodiment one of the present application is given. The multi-modal cable defect identification method provided in the embodiment can be executed by a multi-modal cable defect identification device. The multi-modal cable defect identification device can be realized by software and / or hardware. The multi-modal cable defect identification device can be composed of two or more physical entities, or can be composed of one physical entity. Generally, the multi-modal cable defect identification device can be a server host or other computing device.

[0019] The following describes a server host as an example of a main body for executing the multi-modal cable defect identification method. Referring to Figure 1 The multi-modal cable defect identification method specifically includes: S110, in the case that any modal cable parameter is monitored to exist an abnormality, determining at least one other available modal associated with the current abnormal parameter.

[0020] When any modal cable parameter anomaly is monitored, the application first automatically identifies other available sensor modalities that have a physical correlation or causal relationship with the current abnormal parameter through a pre-set modal correlation rule base or dynamic learning model. For example, when a distributed temperature sensor detects a local temperature rise anomaly, the system will associate and call the partial discharge detection data, current load data, and mechanical stress monitoring data of the same section of the cable. This dynamic association mechanism can be achieved by pre-constructing a logical association network between multi-modal data, and its core is to use the multi-physical field coupling characteristics of cable defects to map isolated abnormal signals to a multi-dimensional cause space that may trigger the anomaly, thereby establishing a data basis for subsequent collaborative analysis, breaking the limitations of traditional single-modal analysis, and providing a more comprehensive observation dimension for defect diagnosis by establishing a mapping relationship between abnormal parameters and potentially associated modalities, effectively avoiding misjudgment caused by false positives or local interference of a single sensor.

[0021] Exemplarily, with reference to Figure 2 The sensor module 111 is distributedly arranged corresponding to the cable 11, and can be a sensor group for monitoring temperature, humidity, partial discharge, current load, mechanical stress, and the like. The data collected by the sensor module 111 is reported to the background server 112. The background server 112 determines whether there is an abnormality in each type of data through a set judgment rule, and when there is a cable parameter anomaly, the parameter collection of the associated available modalities is triggered.

[0022] Optionally, determining at least one other available modality associated with the current abnormal parameter comprises: According to the abnormal modality to which the current abnormal parameter belongs, a pre-defined modal correlation mapping table is queried to determine at least one other available modality associated with the abnormal modality.

[0023] When the other available modalities associated with the current abnormal parameter are determined, the association path of the abnormal modality and other potentially related modalities is first established based on the pre-defined modal correlation mapping table. The mapping table is constructed by integrating cable physical characteristics, historical defect cases, and expert experience knowledge, for example, the temperature abnormality modality is associated with the partial discharge, mechanical stress, current harmonic, and the like, and the sheath grounding current abnormality is associated with the insulation aging, environmental humidity, and the like. When a parameter anomaly of a certain modality (such as temperature) is monitored, the system quickly queries the mapping table through a modal coding matching mechanism to extract a list of other modalities that have a causal relationship or coupling effect with the current abnormal modality.

[0024] Optionally, to achieve dynamic adaptability, the mapping table supports real-time running condition-based associated weight adjustment, for example, enhancing the associated strength of temperature-humidity-sheath ground current in a humid environment. Mapping isolated abnormal signals to a multi-dimensional defect cause network through a structured knowledge base not only avoids the computational overhead of full-modal data acquisition, but also ensures the physical reasonableness of associated modalities, providing an interpretable data association framework for subsequent cross-modal collaborative analysis, while the self-adaptive mechanism of working conditions improves the association accuracy in complex scenarios.

[0025] Optionally, before querying the pre-defined modal association mapping table according to the abnormal modality to which the current abnormal parameter belongs, the following steps are further included: An associated weight matrix between different modal parameters is established according to the cable topology structure and historical defect data set, and a modal association mapping table is constructed based on the associated weight matrix, and the associated weight matrix is dynamically modified through the Pearson correlation coefficient between modalities.

[0026] Before constructing the modal association mapping table, the present application constructs a dynamic associated weight matrix between multi-modal parameters based on the physical topology structure of the cable and the historical defect data set. This process performs three-dimensional modeling of the cable line through digital twinning technology, analyzes the topological characteristics of the conductor structure, insulating material, laying environment, etc. of each monitoring section, and quantifies the associated strength of different modal parameters in the physical space in combination with the defect modality co-occurrence rules recorded in the historical defect case library. For example, in the cable joint area, the associated weight of temperature abnormality and partial discharge will be significantly enhanced due to local electric field distortion. The system continuously collects multi-modal historical data, calculates the Pearson correlation coefficient between modal parameters, and realizes dynamic updating of the weight matrix through a sliding time window mechanism. For example, when it is detected that the humidity in a certain area is long-term high, the associated weight of the sheath ground current and the insulation resistance modality is automatically increased. When constructing the modal association mapping table based on the associated weight matrix, the first N modalities with higher weights are selected to construct the association mapping table. This adaptive adjustment mechanism based on actual running data enables the modal association mapping table to accurately reflect the time-varying characteristics of the cable state. By upgrading the static knowledge base to a dynamic decision engine with environmental perception capability, the prior knowledge of physical topology constraints is retained, and the evolution law of modal association is captured through data-driven methods, providing a more realistic working condition-based decision basis for subsequent abnormal modality association analysis, significantly improving the confidence of cross-modal diagnosis in complex scenarios.

[0027] S120, acquiring real-time parameter data corresponding to other available modalities.

[0028] After determining the other available modalities associated, real-time parameter data of all relevant modalities is synchronously collected. Considering the differences in spatial and temporal resolution, data format and sampling frequency of different sensors, time alignment algorithms are needed to ensure the synchronization of modal data in the time dimension, and feature standardization processing is used to eliminate dimensional differences. For example, high-frequency partial discharge pulse data and low-frequency temperature patrol data are time-axis interpolated and aligned, and current harmonic components are converted in the frequency domain. This multi-modal data space-time alignment mechanism can ensure the comparability of features of each modality in subsequent analysis, provide a high-quality data basis for cross-modal collaborative verification, avoid analysis bias caused by data mismatch, and ensure the timeliness of defect warning through real-time data stream processing.

[0029] S130, performing individual evaluation on real-time parameter data of each modality to obtain modality evaluation results, fusing all modality evaluation results to generate the final monitoring result of the cable defect.

[0030] Subsequently, when performing independent evaluation on each modality data, for real-time parameter data of different modalities, intelligent diagnostic algorithms specific to the corresponding modality are used for preliminary analysis. For example, a space-time gradient analysis algorithm is applied to temperature data to detect thermal field diffusion characteristics, a pulse sequence pattern recognition technique is used for partial discharge data, and a frequency spectrum energy analysis is performed on mechanical vibration signals. Each modality evaluation result is then input into a fusion decision engine, which dynamically allocates confidence weight of each modality diagnostic result through weighted evidence fusion, Bayesian inference or deep learning attention mechanism. The final fusion decision combines the cable operating condition database and historical defect case library to generate a comprehensive diagnostic report containing defect type, location and severity, i.e. the final monitoring result of the cable defect. Thus, both the professional diagnostic advantages of each modality and the complementarity of cross-modal information are retained to eliminate the uncertainty of a single modality, significantly improving the diagnostic accuracy in complex defect scenarios, while providing an interpretable defect traceability path to provide quantitative basis for operation and maintenance decisions.

[0031] Optionally, with reference to Figure 3 performing individual evaluation on real-time parameter data of each modality to obtain modality evaluation results, comprising: S1301, comparing real-time parameter data of each modality with historical baseline data of the corresponding modality to determine data deviation amplitude; S1302, generating modality evaluation results based on the data deviation amplitude.

[0032] In the modal evaluation process, the real-time parameter data of each monitoring mode is standardized and dynamically compared with the historical baseline model of the corresponding mode. The baseline model is constructed based on massive historical data collected under normal operation of the cable, and uses a time series decomposition algorithm to extract periodic variation rules and combines environmental parameters (such as load level, environmental temperature) to establish a multi-dimensional baseline surface. For newly collected real-time data, the system uses a sliding window statistical method to calculate the data deviation amplitude from the baseline model, which includes amplitude deviation, change rate, and trend consistency. For example, temperature modal evaluation will detect whether the current temperature rise exceeds the seasonal baseline threshold, whether the temperature rise rate exceeds the short-term fluctuation limit, and whether the spatial distribution conforms to the conductor heating conduction rule. This multi-dimensional deviation analysis quantifies the difference between real-time data and historical normal patterns, effectively filtering out false anomalies caused by measurement noise or short-term disturbances, and providing interpretable deviation quantification indicators for subsequent decision-making.

[0033] Further, when generating the modal evaluation results based on the data deviation amplitude, a fuzzy logic reasoning engine is used to convert the numerical deviation indicators into semanticized anomaly confidence. The engine integrates cable expert experience rules and machine learning classification models, and performs pattern matching on the deviation patterns through a pre-set anomaly feature library. For example, partial discharge modal evaluation will combine pulse amplitude, frequency, and phase distribution characteristics to comprehensively determine whether it is surface discharge, air gap discharge, or corona discharge type. The evaluation results are output in the form of a structured diagnostic report, including anomaly type, severity level, and recommended response measures. This evaluation mechanism based on quantitative deviation converts raw data into diagnostic conclusions with engineering significance, retaining the objectivity of data-driven while improving the accuracy of anomaly interpretation through expert knowledge injection, and providing a standardized evaluation interface for cross-modal fusion decision-making, ensuring that different modal diagnostic results have semantic comparability.

[0034] Optionally, all modal evaluation results are fused for decision-making to generate the final monitoring result of the cable defect, including: Obtaining the preconfigured fusion weight of each modal evaluation result, and performing weighted summation based on the fusion weight to generate the final monitoring result of the cable defect.

[0035] In the fusion decision stage, the application dynamically loads the pre-configured weight parameters of the evaluation results of each modality from the configuration management center. These weights are determined by combining offline training and online adaptive adjustment. In the offline training stage, based on the historical defect case library, the decision tree algorithm is used to analyze the contribution of different modalities in various typical defect scenarios. For example, in the cable joint defect, higher weight is given to the partial discharge modality, and in the overload heating scenario, the weight of the temperature modality is strengthened. In online operation, the system combines the real-time operating conditions of the cable (such as load level, environmental temperature and humidity) to dynamically adjust the weight distribution through the fuzzy reasoning engine, ensuring that the weight parameters are strongly related to the current operating state. After obtaining the weights, the system normalizes the evaluation results of each modality, maps the abnormal confidence of different dimensions to the [0, 1] interval, and then performs a weighted sum operation to generate a comprehensive defect index. The final monitoring result is analyzed by the defect type classifier, combined with the spatial positioning algorithm to mark the defect occurrence section, and supplemented by the operation and maintenance suggestions based on the rule engine. This fusion mechanism breaks through the limitations of average weighting by differentiating the weights, allowing key modalities to play a dominant role in corresponding defect types. The dynamic weight adjustment mechanism enables the system to be environmentally adaptive, increasing the weight of the sheath ground current modality in the rainy season and the weight of the temperature modality in the hot season. The final result is presented in the form of a quantitative index, preserving the completeness of the multi-modal evidence chain while suppressing single modality noise interference through weighted fusion, significantly improving the diagnostic robustness in complex defect scenarios and providing traceable quantitative evidence for operation and maintenance decisions.

[0036] Optionally, before generating the final monitoring result of the cable defect based on the weighted sum of the fusion weights, it further includes: Generating a correction coefficient based on the data deviation amplitude corresponding to the modality evaluation result, and correcting the fusion weight pre-configured by the modality evaluation result based on the correction coefficient.

[0037] The application also introduces a correction coefficient adjustment mechanism in the fusion decision stage, and generates a dynamic correction factor according to the data deviation range in the real-time evaluation result of each modality. The correction coefficient converts the original deviation amount into a weight adjustment amount through a nonlinear mapping function, for example, an exponential correction function is used for the temperature modality to strengthen the weight contribution of serious temperature rise, and a segmented linear function is used for the partial discharge modality to balance the influence of pulse amplitude and frequency. A time-space attenuation factor is introduced in the correction coefficient generation process, so that the influence of recent data deviation on weight adjustment is greater than that of historical data. At the same time, the correction amplitude is normalized and constrained in combination with the cable health index (such as insulation aging degree and historical defect frequency), to prevent excessive amplification of the weight of a single modality. The corrected weight keeps the sum of the weights of each modality constant through an adaptive normalization algorithm, avoiding distortion of the overall confidence due to dynamic adjustment. Through the weight correction mechanism driven by the deviation range, the key abnormal modality plays a role matching its abnormality degree in the fusion decision, for example, in the case of serious partial discharge, even if the temperature modality does not reach the threshold, its weight will be dynamically increased due to the heat effect associated with discharge; The time-space attenuation factor ensures that the system has a rapid response capability to sudden abnormalities while maintaining the stability of the long-term operating state; The normalization constraint mechanism prevents misjudgment risk caused by weight imbalance, so that the final monitoring result has dynamic sensitivity and decision robustness, significantly improving the diagnosis confidence in complex coupled defect scenarios.

[0038] Optionally, with reference to Figure 4 A separate evaluation is performed on the real-time parameter data of each modality to obtain a modality evaluation result, including: S1303, inputting the real-time parameter data of each modality into a preset defect probability evaluation model to output a defect probability value of the corresponding modality; S1304, constructing a modality evaluation result based on the defect probability value, and the defect probability evaluation model is generated by training historical normal data and defect data of the corresponding modality.

[0039] Unlike the above-mentioned method of comparing the deviation of the data from the historical baseline model, the application can also deploy a dedicated defect probability evaluation model for each monitoring mode during the modal evaluation stage. These models use deep learning architecture designs, such as long short-term memory networks (LSTM) for time series data or convolutional neural networks (CNN) for high-dimensional features, the core of which is to capture abnormal patterns of this type of sensor data through modal-specific feature extraction layers. Taking the temperature mode as an example, the model input includes current temperature value, historical temperature rise curve, spatial gradient distribution and other features, and the output is a defect probability value between 0 and 1. During model training, a transfer learning strategy is used. First, pre-training is performed on a public cable defect dataset to learn general abnormal feature representations, and then fine-tuning is performed on the historical data of the target cable line. The historical data set needs to include normal operating condition data, confirmed defect case data, and simulated defect data generated by the digital twin system to ensure that the model can cover the full life cycle features from early degradation to severe defects. By deep feature learning, implicit correlation patterns that are difficult to explicitly encode in each modal data are captured, such as the mapping relationship between the phase distribution characteristics of partial discharge pulses and the insulation defect type, significantly improving the sensitivity of single modal evaluation.

[0040] Further, when constructing the evaluation results based on the defect probability value, a probability calibration technique is used to convert the original model output into a confidence index with actual engineering significance. Among them, the probability value output by the model is mapped and correlated with the actual severity in historical defect cases, for example, a probability value of 0.7 or higher corresponds to an urgent defect level that needs immediate repair. The evaluation results are presented in the form of a structured probability atlas, including defect type probability distribution, spatial positioning heat map, and development trend prediction curve. For multi-modal fusion scenarios, the system inputs each modal probability atlas into a graph neural network for correlation analysis, and identifies cross-modal coupling abnormal patterns through node feature aggregation.

[0041] Through probability quantization, a graded response for defects of different severity levels is achieved, avoiding false judgments of black or white; probability calibration ensures that the model output aligns with actual operation and maintenance needs, so that a probability value of 0.3 corresponds to an observable temperature rise threshold; the structured atlas provides a traceable decision basis for operation and maintenance personnel, retaining model interpretability while displaying the risk path of defect evolution through probability gradient.

[0042] The above, by monitoring any modal cable parameter exists in the case of abnormal, determine at least one other available modal associated with the current abnormal parameter; obtain the real-time parameter data corresponding to the other available modal; each modal real-time parameter data is executed separately to obtain the modal evaluation result, all modal evaluation results are fused to generate the final monitoring result of the cable defect. Adopt the above technical means, through the real-time parameter data of the available modal associated with the abnormal parameter is fused and evaluated, so as to obtain accurate and comprehensive cable defect monitoring result, avoid the single modal cable defect detection misjudgment, improve the cable defect detection precision, and then improve the cable operation stability and reliability, improve the cable operation effect.

[0043] Embodiment two: On the basis of the above embodiment, Figure 5 A structure diagram of a multi-modal cable defect identification system provided by the second embodiment of the application is provided. Referring to Figure 5 The multi-modal cable defect identification system provided by the embodiment specifically includes: The detection module 21 is configured to determine at least one other available modal associated with the current abnormal parameter in the case of monitoring any modal cable parameter exists in the case of abnormal; The acquisition module 22 is configured to acquire real-time parameter data corresponding to the other available modal; The fusion module 23 is configured to execute separate evaluation on the real-time parameter data of each modal to obtain the modal evaluation result, and fuse all modal evaluation results to generate the final monitoring result of the cable defect.

[0044] Specifically, determining at least one other available modal associated with the current abnormal parameter includes: According to the abnormal modal to which the current abnormal parameter belongs, the pre-defined modal correlation mapping table is queried to determine at least one other available modal associated with the abnormal modal.

[0045] Before querying the pre-defined modal correlation mapping table according to the abnormal modal to which the current abnormal parameter belongs, further includes: According to the cable topology structure and the historical defect data set, the correlation weight matrix between different modal parameters is established, and the modal correlation mapping table is constructed based on the correlation weight matrix, and the correlation weight matrix is dynamically corrected by the Pearson correlation coefficient between modes.

[0046] Specifically, executing separate evaluation on the real-time parameter data of each modal to obtain the modal evaluation result includes: The real-time parameter data of each modal is compared with the historical baseline data of the corresponding modal to determine the data deviation amplitude, and the modal evaluation result is generated based on the data deviation amplitude.

[0047] The fusion decision is made on all modal evaluation results to generate the final monitoring result of the cable defect, including: The fusion weight preconfigured for each modal evaluation result is acquired, and the fusion weight is used for weighted summation to generate the final monitoring result of the cable defect.

[0048] Before the weighted summation based on the fusion weight is performed to generate the final monitoring result of the cable defect, the method further includes: The correction coefficient is generated based on the data deviation amplitude corresponding to the modal evaluation result, and the fusion weight preconfigured for the modal evaluation result is corrected based on the correction coefficient.

[0049] The real-time parameter data of each mode is evaluated to obtain the modal evaluation result, including: The real-time parameter data of each mode is input into a preset defect probability evaluation model, a defect probability value of the corresponding mode is output, and the modal evaluation result is constructed based on the defect probability value. The defect probability evaluation model is generated by training historical normal data and defect data of the corresponding mode.

[0050] The above, by monitoring any mode of cable parameter exists abnormal situation, determine at least one other available mode associated with the current abnormal parameter; acquire the real-time parameter data corresponding to the other available mode; the real-time parameter data of each mode is evaluated to obtain the modal evaluation result, all modal evaluation results are fused to make decisions to generate the final monitoring result of the cable defect. By using the above technical means, the real-time parameter data of the available mode associated with the abnormal parameter is fused and evaluated, so as to obtain accurate and comprehensive cable defect monitoring result, avoid single mode cable defect detection misjudgment, improve cable defect detection precision, and improve cable operation stability and reliability, and improve cable operation effect.

[0051] The multi-modal cable defect identification system provided by the second embodiment of the application can be used to execute the multi-modal cable defect identification method provided by the first embodiment of the application, and has the corresponding functions and beneficial effects.

[0052] Embodiment three: The electronic device provided by the third embodiment of the application is described with reference to Figure 6 The electronic device includes a processor 31, a memory 32, a communication module 33, an input device 34, and an output device 35. The number of processors in the electronic device can be one or more, and the number of memories in the electronic device can be one or more. The processor, memory, communication module, input device, and output device of the electronic device can be connected through a bus or other means.

[0053] The memory, as a computer readable storage medium, can be used to store software programs, computer executable programs and modules, such as the program instructions / modules of the multi-modal cable defect identification method according to any embodiment of the present application (for example, each module in the multi-modal cable defect identification system). The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and application programs required by at least one function; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some examples, the memory can further include a memory remotely arranged with respect to the processor, which can be connected to the device through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0054] The communication module is used for data transmission.

[0055] The processor executes the software programs, instructions and modules stored in the memory, thereby performing various functional applications and data processing of the device, that is, implementing the multi-modal cable defect identification method described above.

[0056] The input device can be used to receive input digital or character information, and generate key signal input related to user settings and function control of the device. The output device can include a display device such as a display screen.

[0057] The electronic device provided above can be used to perform the multi-modal cable defect identification method provided in Embodiment One, and has corresponding functions and advantages.

[0058] Embodiment Four: The embodiments of the present application also provide a storage medium containing computer executable instructions, which, when executed by a computer processor, are used to perform a multi-modal cable defect identification method, the multi-modal cable defect identification method comprising: in the case that any modal cable parameter is monitored to have an abnormality, determining at least one other available mode associated with the current abnormal parameter; obtaining real-time parameter data corresponding to the other available mode; performing individual evaluation on the real-time parameter data of each mode to obtain a modal evaluation result, fusing all modal evaluation results to generate a final monitoring result of the cable defect.

[0059] Storage medium - any type of memory device or storage device. The term "storage medium" is intended to include an installation medium, e.g., a CD-ROM, floppy disks, or tape apparatus; computer system memory or random access memory such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; or a non-volatile memory such as a magnetic medium (e.g., a hard drive or optical storage); registers or other similar types of memory elements, etc. The memory medium can also include other types of storage medium or combinations thereof. In addition, the memory medium can reside in a first computer system's main memory, or in a second different computer system's memory, which second computer system can provide the first computer system with an instruction for execution. The term "storage medium" can also be

[0060] Of course, the storage medium provided by the embodiments of the present application includes computer executable instructions, and the computer executable instructions are not limited to the multi-modal cable defect identification method as described above, but can also perform the related operations in the multi-modal cable defect identification method provided by any of the embodiments of the present application.

[0061] The multi-modal cable defect identification system, storage medium and electronic device provided in the above embodiments can execute the multi-modal cable defect identification method provided by any of the embodiments of the present application, and the technical details not described in detail in the above embodiments can be referred to the multi-modal cable defect identification method provided by any of the embodiments of the present application.

[0062] The above are only the preferred embodiments of the present application and the technical principles applied. The present application is not limited to the specific embodiments described herein, and various obvious changes, readjustments and substitutions made by those skilled in the art will not deviate from the protection scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can also include more other equivalent embodiments without deviating from the concept of the present application, and the scope of the present application is determined by the scope of the claims.

Claims

1. A multimodal cable defect identification method, characterized in that, include: If an anomaly is detected in the cable parameters of any mode, determine at least one other available mode associated with the current anomaly parameter; Obtain the real-time parameter data corresponding to the other available modes; Each modality's real-time parameter data is evaluated individually to obtain modal evaluation results. All modal evaluation results are then fused to make a decision, generating the final monitoring results for cable defects.

2. The multimodal cable defect identification method according to claim 1, characterized in that, The determination of at least one other available modality associated with the current abnormal parameter includes: Based on the abnormal mode to which the current abnormal parameter belongs, query the predefined mode association mapping table to determine at least one other available mode associated with the abnormal mode.

3. The multimodal cable defect identification method according to claim 2, characterized in that, Before querying the predefined modality association mapping table based on the abnormal modality to which the current abnormal parameter belongs, the following is also included: Based on the cable topology and historical defect dataset, an association weight matrix is ​​established between different modal parameters. A modal association mapping table is constructed based on the association weight matrix. The association weight matrix is ​​dynamically corrected using the intermodal Pearson correlation coefficient.

4. The multimodal cable defect identification method according to claim 1, characterized in that, The modal evaluation results are obtained by performing individual evaluations on the real-time parameter data for each modality, including: The real-time parameter data of each mode is compared with the historical baseline data of the corresponding mode to determine the data deviation magnitude, and the mode evaluation result is generated based on the data deviation magnitude.

5. The multimodal cable defect identification method according to claim 4, characterized in that, The process of fusing all modal evaluation results to generate the final monitoring results for cable defects includes: Obtain the pre-configured fusion weights of each modal evaluation result, and perform weighted summation based on the fusion weights to generate the final monitoring result of cable defects.

6. The multimodal cable defect identification method according to claim 5, characterized in that, Before generating the final monitoring result of the cable defect by weighted summation based on the fusion weights, the method further includes: A correction coefficient is generated based on the data deviation magnitude corresponding to the modal evaluation result, and the pre-configured fusion weights of the modal evaluation result are corrected based on the correction coefficient.

7. The multimodal cable defect identification method according to claim 1, characterized in that, The modal evaluation results are obtained by performing individual evaluations on the real-time parameter data for each modality, including: Real-time parameter data for each modality is input into a preset defect probability assessment model, which outputs the defect occurrence probability value for the corresponding modality. Modality assessment results are constructed based on the defect occurrence probability value. The defect probability assessment model is generated by training with historical normal data and defect data for the corresponding modality.

8. A multimodal cable defect identification system, characterized in that, include: The detection module is used to determine at least one other available mode associated with the current abnormal parameter when an abnormality is detected in the cable parameters of any mode. The acquisition module is used to acquire real-time parameter data corresponding to the other available modes; The fusion module is used to perform separate evaluations on the real-time parameter data of each mode to obtain modal evaluation results, and then fuse all modal evaluation results to generate the final monitoring results of cable defects.

9. An electronic device, characterized in that, include: Memory and one or more processors; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the multimodal cable defect identification method as described in any one of claims 1-7.

10. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the multimodal cable defect identification method as described in any one of claims 1-7.

Images