Cable early fault parameter extraction method and system fused with temperature rise process data

Abstract

The application relates to a cable early fault parameter extraction method and system fusing temperature rise process data, and relates to the technical field of cable fault monitoring, and comprises the following steps: receiving multi-node temperature rise data uploaded by a temperature measuring optical fiber for a to-be-tested cable in a conduction state, comparing the multi-node temperature rise data with cable rated working conditions, extracting deviation information from a standard temperature rise curve, and setting first and second fault characteristic indexes; combining the deviation information with the first and second fault characteristic indexes to formulate a fault parameter extraction atlas; according to an attenuation coefficient, taking early fault heat generation under different equivalent insulation resistances as core correlation characteristics, evaluating a fault influence range, and using a dynamic threshold adjustment rule to perform fault reminding. The application solves the problems that in traditional cable fault detection, early fault characteristics are not sensitive to capture, are prone to missing detection, cable operation states and environmental factors are not fully combined, and fault judgment accuracy is insufficient.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of cable fault monitoring, in particular to a cable early fault parameter extraction method and system fusing temperature rise process data. BACKGROUND

[0002] With the continuous improvement of the power system's requirement for power supply reliability, as the core component of power transmission, the accurate identification of the early fault of the cable becomes a key technical requirement to ensure the stable operation of the system.

[0003] At present, the traditional cable fault detection method has insufficient ability to capture the weak characteristics of early faults, and is prone to miss detection of faults. Moreover, it does not fully combine the influence of the actual operation state of the cable and the surrounding environmental factors, and it is difficult to fully reflect the evolution law of the cable fault, which not only leads to low accuracy of fault judgment, but also may increase the risk of cable fault expansion and power interruption due to the failure to discover early hidden dangers in time. SUMMARY

[0004] The present application provides a cable early fault parameter extraction method and system fusing temperature rise process data, which improves the current situation that the traditional cable fault detection method is not sensitive to the capture of early fault characteristics, is prone to miss detection, and does not fully combine the operation state of the cable and environmental factors, resulting in insufficient accuracy of fault judgment.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, the embodiments of the present application provide a cable early fault parameter extraction method fusing temperature rise process data, which comprises:

[0007] According to the cable to be tested in the on state, the multi-node temperature rise data uploaded by the temperature measuring optical fiber is received, compared with the rated working condition of the cable, the deviation information from the standard temperature rise curve is extracted, and the first fault characteristic index and the second fault characteristic index are set;

[0008] Through the deviation information, the first fault characteristic index and the second fault characteristic index, a fault parameter extraction atlas conforming to the logical relationship of cable aging is drafted;

[0009] In the fault parameter extraction atlas, according to the attenuation coefficient, the early fault heat generation under different equivalent insulation resistances is taken as the core correlation characteristic, the fault influence range is evaluated in multiple scales, and a dynamic threshold adjustment rule is used for fault reminding.

[0010] In a second aspect, the embodiments of the present application provide a cable early fault parameter extraction system fusing temperature rise process data, which comprises:

[0011] The information extraction and index setting module is used to receive multi-node temperature rise data uploaded by the temperature measuring optical fiber based on the cable under test in the conducting state, compare it with the rated operating conditions of the cable, extract the deviation information from the standard temperature rise curve, and set the first fault characteristic index and the second fault characteristic index.

[0012] The fault parameter extraction map formulation module is used to formulate a fault parameter extraction map that conforms to the cable aging logic relationship by using the deviation information and combining the first fault feature index and the second fault feature index.

[0013] The fault impact range assessment and early warning module is used to assess the fault impact range in the fault parameter extraction map based on the attenuation coefficient and with the early fault heat generation under different equivalent insulation resistances as the core correlation feature, and to provide fault alerts using dynamic threshold adjustment rules.

[0014] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0015] This application proposes a method and system for extracting early fault parameters of cables by integrating temperature rise process data. Through a step-by-step approach, it acquires cable temperature rise deviation and fault characteristic indicators, formulates a fault parameter extraction map, and conducts fault impact assessment and early warning based on the map, achieving accurate extraction of early fault parameters and timely management of safety risks. First, temperature rise data from multiple nodes of the cable is collected using temperature-measuring optical fibers and compared with the cable's rated operating condition data to extract temperature rise deviation information. Simultaneously, a first fault characteristic indicator is set by combining cable load logs and insulation layer node correlation data, and a second fault characteristic indicator is set by combining partial discharge signals and conductor node correlation data. Next, the directed edges of the fault parameter extraction map are determined using a sliding time window, and stage correlation information is labeled. Attenuation coefficients are obtained by fusing stage information based on LSTM time-series perception. Early fault heat generation under different equivalent insulation resistances is used as the core correlation feature, and the fault impact range is evaluated on a multi-scale basis using start / end thresholds. Then, a fault early warning knowledge graph is constructed, and potential fault nodes are identified through a graph convolutional network based on an attention mechanism, performing topological expansion analysis on the impact range. Finally, dynamic threshold adjustment rules are configured based on the evaluation results and node identification information to achieve accurate early warning of cable faults.

[0016] This technical solution integrates temperature rise process data with multi-dimensional fault characteristics, solving problems in traditional cable fault detection such as one-sided parameter extraction due to reliance on a single data point, inability of fixed models to adapt to different operating environments, and lack of foresight in fault impact assessment. It avoids early fault omissions and misjudgments due to insufficient data dimensions or poor model adaptability, as well as fault propagation due to failure to identify potential risk nodes. It improves the accuracy of early fault parameter extraction and the timeliness of fault warning, providing reliable technical support for the safe and stable operation of cables. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the method for extracting early fault parameters of cables by integrating temperature rise process data, as provided in this application embodiment;

[0019] Figure 2 A schematic diagram of the structure of the cable early fault parameter extraction system that integrates temperature rise process data provided in the embodiments of this application.

[0020] The components represented by each number in the attached diagram are explained below:

[0021] Information extraction and indicator setting module 01, fault parameter extraction map formulation module 02, fault impact range assessment and early warning module 03. Detailed Implementation

[0022] This application provides a method and system for extracting early fault parameters of cables by integrating temperature rise process data, which is used to solve the technical problems in the prior art that are not sensitive to the capture of early fault characteristics of cables, are prone to missed detection, and do not fully combine cable operating status and environmental factors, resulting in insufficient accuracy of fault judgment.

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0025] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0026] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for extracting early fault parameters of cables by integrating temperature rise process data. The method includes the following steps:

[0027] S110: Based on the cable under test in the conducting state, receive multi-node temperature rise data transmitted by the temperature measuring optical fiber, compare it with the rated operating conditions of the cable, extract the deviation information from the standard temperature rise curve, and set the first fault characteristic index and the second fault characteristic index.

[0028] In this embodiment of the application, in order to accurately capture abnormal signals during cable operation and establish quantifiable fault judgment criteria, it is necessary to first obtain temperature change data at key locations of the cable and compare it with the cable's rated operating conditions to extract temperature deviation information. Then, in combination with the operating characteristics of the cable's core components, two types of fault characteristic indicators are set to form a fault monitoring basis covering key areas of the cable insulation and conductor, providing data support for the subsequent formulation of fault parameter extraction maps.

[0029] Specifically, for the cable under test that is in a conductive state, the fiber optic temperature measurement system is activated to receive temperature rise data from multiple nodes at different locations on the cable, relying on the temperature-measuring optical fiber that has been laid along the cable in advance. This data covers key parts such as the cable insulation layer, conductor, and joints, and can reflect the dynamic temperature changes of each node in real time during operation. For example, the rising trend of conductor node temperature when the cable load increases, and the slight local temperature anomalies in the early stage of insulation aging, will all be recorded completely.

[0030] Furthermore, the received multi-node temperature rise data is compared node by node with the standard data under the cable's rated operating conditions. The cable's rated operating conditions are a baseline state determined based on cable design parameters, industry standards, and long-term safe operation experience, including the normal temperature range and standard temperature rise curves of each node under different loads.

[0031] By comparison, the deviation information between the actual temperature rise data and the standard temperature rise curve can be accurately extracted. This deviation information is an important initial signal for judging whether there are potential early faults in the cable.

[0032] Furthermore, a first fault characteristic indicator and a second fault characteristic indicator are set to build a more comprehensive basis for early fault monitoring from the operational dimensions of different key components of the cable.

[0033] In the method provided in this application embodiment, "setting the first fault characteristic index and the second fault characteristic index" includes:

[0034] Associate the cable load log with each cable insulation node under the rated operating conditions of the cable, and set the first fault characteristic index.

[0035] The partial discharge signal is recorded and associated with each cable conductor node under the rated operating conditions of the cable, and a second fault characteristic index is set.

[0036] In this embodiment of the application, in order to avoid misjudgment or omission of faults due to reliance on a single data point, it is necessary to combine the core influencing factors and key component characteristics in cable operation to establish fault characteristic indicators for the insulation layer and conductor respectively, so as to form a fault monitoring dimension covering the core operating area of ​​the cable.

[0037] Specifically, when setting the first fault characteristic index, the cable load log generated by the daily operation of the cable is first sorted out. The log records in detail the key data such as the actual current carried by the cable, the continuous operating time, and the frequency of load fluctuations in different time periods. The aging speed of the cable insulation layer and the risk of local overheating are directly related to the load status. That is, the temperature of the insulation layer nodes that are in an overloaded operating state for a long time is likely to exceed the rated range, which will accelerate the aging process and cause faults.

[0038] At this point, the cable load log is associated with each insulation layer node under the cable's rated operating conditions. The cable's rated operating conditions are the baseline state determined according to cable design standards and safe operation specifications, which includes the normal tolerance range of each insulation layer node in different load ranges. Through the above association, the impact of load changes on the insulation layer can be quantified.

[0039] For example, when the cable load log corresponding to a certain insulation layer node shows that the load exceeds the rated value by 15% for 3 consecutive hours and the temperature of the node approaches the rated upper limit, the potential risk can be captured by the first fault characteristic indicator, thereby clarifying the direction of fault warning for the insulation layer node.

[0040] Furthermore, when setting the second fault characteristic index, special attention should be paid to the partial discharge signals recorded during cable operation. Partial discharge is a typical manifestation of defects in the cable conductor, and the discharge intensity gradually increases as the conductor defect worsens. If not monitored in time, it can easily develop into serious faults such as short circuits and breakdowns.

[0041] Specifically, partial discharge signals are recorded and associated with each conductor node under rated cable conditions. First, the normal partial discharge threshold for each conductor node under rated cable conditions is determined, and then the discharge signals recorded during actual operation are compared. When the amplitude of the partial discharge signal of a conductor node continuously exceeds 10% of the rated threshold, or the discharge frequency increases significantly, the abnormality of the node can be quickly identified through the second fault characteristic index, thus achieving early detection of conductor defects.

[0042] At the same time, during the correlation process, it is necessary to ensure that each insulation layer node and conductor node can correspond to specific cable rated operating parameters. Precise correlation can make the first and second fault characteristic indicators more targeted and avoid indicator failure due to ambiguous correlation.

[0043] This step, by setting the first and second fault characteristic indicators respectively, can focus on the two high-incidence fault areas of the cable insulation layer and conductor. Combined with the two core influencing factors of cable load and partial discharge, a more comprehensive fault monitoring system is formed, laying the foundation for subsequent fault assessment.

[0044] S120: Based on the deviation information, combined with the first fault characteristic index and the second fault characteristic index, a fault parameter extraction map that conforms to the cable aging logic relationship is formulated.

[0045] In this embodiment of the application, in order to transform fault monitoring from single data judgment to structured logical relationship analysis, it is necessary to combine the temperature rise deviation information obtained in the early stage with the two types of fault characteristic indicators to build a fault parameter extraction map that fits the actual aging law of the cable, so as to clearly present the complete path of the fault from its inception to its development.

[0046] Specifically, in the process of formulating the fault parameter extraction map, the directed edges corresponding to the map are first determined by using a sliding time window, and the various fault evolution stages of the cable are effectively connected by the directed edges.

[0047] The duration of the sliding time window needs to be determined in conjunction with the typical cycle of cable aging, and the changing trends of each fault characteristic are continuously tracked through this window. When the temperature rise deviation information of a certain node, the associated data of the first fault characteristic indicator, and the associated data of the second fault characteristic indicator simultaneously show phased changes, the connection between different fault evolution stages is completed through directed edges.

[0048] Meanwhile, the preceding causes, time-dependent parameters, and signal correlation information between each stage of fault evolution are marked on the directed edges, making the evolution logic of the fault clearer at different stages and ensuring that the fault parameter extraction map can accurately reflect the inherent relationship between the initial stage and the development of the cable fault.

[0049] In addition, considering that the cable operating environment has a significant impact on the aging rate, the graphs drawn based solely on basic data are insufficient to adapt to the operating conditions under different environments. It is also necessary to integrate current sensing units, voltage monitoring units, and ambient temperature detection units to jointly construct a dynamic operating parameter sensing network. Through this network, operating environment data such as current fluctuations, voltage stability, and ambient temperature during cable operation can be obtained in real time.

[0050] Furthermore, based on the preset fault impact coefficient, the acquired operating environment data is correlated with the aging sub-intervals of the cable. An environmental fault correlation matrix is ​​constructed based on the correlation relationship. The matrix is ​​used to quantify the degree of influence of the operating environment data on each aging sub-interval. The attenuation coefficient of the fault parameter extraction spectrum is dynamically corrected through the matrix, so that the fault parameter extraction spectrum can not only conform to the aging logic of the cable itself, but also be adjusted and adapted according to environmental changes.

[0051] Step S120 in the method provided in this application embodiment includes:

[0052] By using a sliding time window, the directed edges corresponding to the fault parameter extraction graph are determined and connected to each fault evolution stage;

[0053] At the same time, the preceding causes, time-dependent parameters, and signal correlation information between each stage of fault evolution are labeled.

[0054] In this embodiment of the application, in order to ensure that the proposed fault parameter extraction map can accurately fit the actual logic of cable aging, it is necessary to sort out the fault evolution relationship and optimize the map parameters by combining environmental data. This ensures that the map can clearly present the fault development path and dynamically adapt to different operating scenarios, providing reliable support for extracting early fault parameters of cables based on the fault parameter extraction map.

[0055] Specifically, the directed edges corresponding to the fault parameter extraction map are first determined by using a sliding time window. The sliding time window needs to be set in conjunction with the typical process of cable aging to ensure that the characteristic changes of the fault from the initial stage to the development stage can be fully captured. By continuously tracking the linkage changes between temperature rise deviation information and the first and second fault characteristic indicators, when the data shows a stage transition, the corresponding fault evolution stages are connected by directed edges.

[0056] Meanwhile, the preceding causes between each stage are marked on the directed edges to clarify the key factors that trigger the transition of the fault stage; time-dependent parameters are marked to reflect the time range required for the evolution of different stages; signal correlation information is marked to explain the characteristic thresholds of temperature rise deviation, load correlation data, and partial discharge signals corresponding to each stage. These annotations make the internal logic of fault evolution clearer and avoid the fault parameter extraction map only showing stage connections without key correlation information.

[0057] For example, when the fault transitions from the "slight overheating stage" to the "insulation aging stage", the preceding cause relationship is marked as "the cable load exceeds the rated value by 18% for 3 hours or more" next to the directed edge connecting the two stages, clearly indicating that the cable load condition is the key factor triggering the transition stage; the time-dependent parameter is marked as "average evolution time 45-60 hours", reflecting the time range usually required for the development from slight overheating to insulation aging; the signal correlation information is marked as "temperature rise deviation is maintained in the range of 12%-15% of the rated curve, the load correlation value of the insulation layer node in the first fault characteristic index exceeds the threshold by 10%, and the amplitude of the partial discharge signal in the second fault characteristic index is stable at 8-12 pC".

[0058] Ultimately, by using a sliding time window to determine directed edges, marking the relationships between the preceding causes at each stage, and obtaining fault parameter extraction maps based on time-dependent parameters and signal correlation information, the fault parameter extraction maps can accurately match the actual logic of cable aging and clearly present the fault development path.

[0059] The method provided in this application embodiment, which "constructs a fault parameter extraction map that conforms to the cable aging logic relationship", further includes:

[0060] A dynamic operating parameter sensing network is constructed by integrating a current sensing unit, a voltage monitoring unit, and an ambient temperature detection unit to acquire cable operating environment data;

[0061] Based on the joint aging sub-intervals associated with the fault influence coefficient, and combined with the cable operating environment data, an environmental fault correlation matrix is ​​constructed, and the attenuation coefficient of the fault parameter extraction spectrum is dynamically corrected.

[0062] In this embodiment of the application, in order to fully capture the key operating conditions that may affect the aging rate of the cable, it is also necessary to integrate a current sensing unit, a voltage monitoring unit, and an ambient temperature detection unit to build a dynamic operating parameter sensing network, so as to provide data input for subsequent association of aging sub-intervals based on fault influence coefficients and construction of an environmental fault correlation matrix.

[0063] Specifically, the construction of the dynamic operating parameter sensing network is achieved by distributing current sensing units, voltage monitoring units, and ambient temperature detection units according to the cable laying path and high-fault nodes, and connecting them to the data processing terminal through a data transmission device.

[0064] Among them, the current sensing unit is used to capture the current fluctuations in the cable during operation in real time, the voltage monitoring unit records the voltage stability data, and the ambient temperature detection unit collects the temperature changes of the environment in which the cable is located. The three types of units work together to comprehensively acquire cable operating environment data that may affect the cable aging rate, so as to ensure that core influencing dimensions such as current, voltage, and ambient temperature are covered, providing a data foundation for subsequent analysis of the role of the environment in fault evolution.

[0065] Furthermore, after obtaining cable operating environment data, in order to avoid the aging rate of cables in different environments being unable to be accurately reflected by the spectrum due to fixed coefficients, it is also necessary to construct an environmental fault correlation matrix based on the aging sub-intervals associated with the fault influence coefficients and combine it with the cable operating environment data, and dynamically correct the attenuation coefficient of the fault parameter extraction spectrum.

[0066] The method provided in this application embodiment, which "constructs an environmental fault correlation matrix and dynamically corrects the attenuation coefficient of the fault parameter extraction spectrum", further includes:

[0067] The row dimension of the environmental fault correlation matrix is ​​the cable operating environment data, and the column dimension of the environmental fault correlation matrix includes conductor overheating sub-interval, insulation damage sub-interval, and joint aging sub-interval.

[0068] The environmental fault correlation matrix is ​​quantized and dynamically updated.

[0069] When the cable operating environment data does not meet the safe operating environment restrictions, the matrix element values ​​of the environmental fault correlation matrix are set to 0, triggering a forced detection command.

[0070] First, the row and column dimensions of the environmental fault correlation matrix are defined. Specifically, the row dimension is set to cable operating environment data because this data directly affects the cable aging rate, covering key factors such as current fluctuations, voltage stability, and ambient temperature, and can comprehensively reflect the effect of the environment on cable operation.

[0071] In addition, the column dimensions are set as conductor overheating sub-interval, insulation damage sub-interval, and joint aging sub-interval because these three types of sub-intervals are high-incidence areas of early cable faults. Focusing on these sub-intervals can make the matrix more targeted in quantifying the impact of aging, so as to ensure that the fault correlation matrix can accurately correspond to the relationship between environmental data and core aging risk points.

[0072] Furthermore, the environmental fault correlation matrix is ​​quantified and dynamically updated. Element quantification is based on fault impact coefficients, which are determined through extensive cable aging experiments and historical fault cases. Specific numerical values ​​are used to quantify the degree of influence of environmental data on each aging sub-interval.

[0073] In addition, dynamic updates require real-time tracking of changes in cable operating environment data. When environmental data fluctuates, the corresponding matrix element values ​​are adjusted synchronously to ensure that the matrix elements can reflect the impact of the current environment on aging in real time.

[0074] Finally, an environmental safety mechanism is set up. When the cable operating environment data does not meet the limits of safe operating environment, such as the ambient temperature exceeding the cable's tolerance limit or the voltage fluctuation exceeding the safety threshold, the impact of the environment on cable aging has exceeded the normal quantitative range. Continuing to rely on matrix elements to correct the attenuation coefficient may result in a large deviation. Therefore, the corresponding element value of the environmental fault correlation matrix is ​​set to 0 to suspend the normal correction logic.

[0075] At the same time, a mandatory testing command is triggered to initiate a comprehensive cable testing process, such as checking key components like conductors, insulation layers, and joints one by one. This prevents potential faults from being overlooked due to parameter deviations under extreme conditions, ensuring that even under unconventional operating conditions, potential faults can be detected in a timely manner through mandatory testing, providing additional protection for the safe operation of cables.

[0076] This step involves first constructing a dynamic operating parameter sensing network to obtain cable operating environment data, then quantifying the impact of the environment on conductor overheating, insulation damage, and joint aging sub-regions based on the environmental fault correlation matrix, and finally setting up an environmental safety mechanism to cope with extreme environments. This enables dynamic correction of the attenuation coefficient of the fault parameter extraction spectrum, ensuring that the spectrum accurately reflects the actual cable aging under different environments.

[0077] S130: In the fault parameter extraction map, based on the attenuation coefficient, the early fault heat generation under different equivalent insulation resistances is used as the core correlation feature to evaluate the fault influence range on a multi-scale basis, and a dynamic threshold adjustment rule is used to provide fault alerts.

[0078] In this embodiment of the application, in order to make the fault assessment both accurately quantify the evolution rate and fully cover the scope of influence, while ensuring the timeliness and adaptability of the early warning signal, it is necessary to rely on the fault parameter extraction spectrum, combine the attenuation coefficient and the early fault heat generation to carry out the assessment, and configure dynamic threshold adjustment rules to achieve scientific management and control of early cable faults and ensure the stability of cable operation.

[0079] Specifically, the core criteria for evaluation are first determined from the established fault parameter extraction map. The attenuation coefficient is obtained by dynamically weighting the directed edges of the fault parameter extraction map, and then, based on the time-series awareness capability of LSTM, by fusing the pre-existing causes, time-dependent parameters, and signal correlation information between each fault evolution stage, thereby quantifying the development rate of the fault at different stages.

[0080] Furthermore, by using the early fault heat generation under different equivalent insulation resistances as the core correlation feature, the energy release state of the fault and the change in insulation performance are correlated through this feature, allowing the assessment to focus on the fundamental influencing factors of the fault.

[0081] Furthermore, based on the aforementioned core criteria, a multi-scale assessment of the fault's impact range is conducted. During the assessment, temporal characteristic analysis is performed using both start and end thresholds. These two thresholds clearly mark different fault stages, such as the minor overheating stage, insulation aging stage, and partial discharge stage. Then, based on the characteristics of each stage and their correlations in the graph, the assessment dimensions are gradually expanded, extending from a single fault node to surrounding related nodes, achieving comprehensive coverage of the fault's impact range.

[0082] Simultaneously, dynamic threshold adjustment rules are configured based on the temporal relationship between the start and end thresholds. During the configuration process, fault types, historical fault cases, and start and end thresholds need to be associated to build a fault early warning knowledge graph. Potential fault nodes are identified based on this knowledge graph, and topological expansion analysis is performed on the fault impact range. This ensures that the dynamic threshold adjustment rules can adapt to different fault types and evolution stages, guaranteeing that fault alerts neither miss potential hazards nor generate invalid warnings.

[0083] This step provides a complete path for the precise control of early cable faults by combining the attenuation coefficient and early fault heat generation in the fault parameter extraction map and configuring dynamic threshold adjustment rules. This enables fault assessment and early warning to closely match the actual operating status of the cable, laying the foundation for timely fault handling measures.

[0084] Step S130 in the method provided in this application embodiment includes:

[0085] Dynamic weights are assigned to the directed edges of the fault parameter extraction map. Based on LSTM time-aware fusion, the pre-cause relationships, time-dependent parameters and signal correlation information between each fault evolution stage are fused to obtain the attenuation coefficient.

[0086] Time-series feature analysis is performed using start and end thresholds, which are used to mark the mild overheating stage, insulation aging stage, and partial discharge stage.

[0087] The dynamic threshold adjustment rule is configured based on the temporal relationship between the start threshold and the end threshold.

[0088] In this embodiment of the application, in order to accurately quantify the fault evolution rate and scientifically configure early warning rules based on the fault parameter extraction map, it is necessary to first obtain the attenuation coefficient that reflects the fault development rhythm, and then clarify the fault stage through threshold analysis and configure dynamic reminder rules to ensure that the impact range assessment of early cable faults is more accurate and the fault reminders are more timely and adapted to the actual working conditions.

[0089] Specifically, the first step is to obtain the attenuation coefficient. During this process, the directed edges of the fault parameter extraction map need to be dynamically weighted. These directed edges connect to various fault evolution stages of the cable, and the weights must be adjusted based on the degree of correlation between each stage to ensure that the weights reflect the ease or difficulty of transitioning between different stages.

[0090] Furthermore, based on the time-series awareness capability of LSTM, the preceding causes, time-dependent parameters, and signal correlation information between each stage of fault evolution are fused and analyzed to accurately capture the temporal patterns and intrinsic relationships of fault evolution.

[0091] Specifically, the antecedent causal relationship reflects the key conditions that trigger the transition of stages, the time-dependent parameters reflect the duration of stage evolution, and the signal correlation information contains the characteristic data thresholds corresponding to each stage. By deeply fusing these time-related information through the LSTM (Long Short-Term Memory) model, the decay coefficient of the development rate of the quantified fault in different evolution stages is finally calculated, providing core quantitative indicators for the multi-scale assessment of the subsequent fault impact range.

[0092] For example, when analyzing the evolution process of "slight overheating stage - insulation aging stage", the preceding cause relationship is "cable load exceeds rated value by 15% for 2 hours", the time-dependent parameter is "average evolution time of 30-40 hours", and the signal correlation information is "temperature rise deviation is maintained at 10%-13% of the rated curve, and partial discharge signal amplitude is 5-8pC".

[0093] After inputting the above information into the LSTM model, the model will combine the temporal patterns of similar historical faults to perceive the linkage between load, time, and signal during the transition of this stage, and calculate the attenuation coefficient of this evolution path as 0.5.

[0094] However, when analyzing the evolution process of the "insulation aging stage - partial discharge stage", the preceding cause relationship is "insulation layer node temperature exceeds the rated upper limit by 8%", the time-dependent parameter is "average evolution time of 15-25 hours", and the signal correlation information is "temperature rise deviation exceeds the rated curve by 18% and the amplitude of partial discharge signal jumps to 15-20pC". By fusing these more pressing time-series information, the LSTM model calculates the attenuation coefficient of this path to be 0.8, thereby reflecting the difference in the evolution rate of different fault stages.

[0095] After obtaining the attenuation coefficient, further time-series characteristic analysis of the starting and ending thresholds and configuration of dynamic threshold adjustment rules are carried out.

[0096] Specifically, the core function of determining the starting and ending thresholds is to mark the key evolution stages of early cable faults, namely the slight overheating stage, the insulation aging stage, and the partial discharge stage. Through these two types of thresholds, the boundaries of different fault stages can be clearly defined, avoiding deviations in the assessment direction due to ambiguous stage divisions.

[0097] During the time-series feature analysis, it is necessary to continuously track the matching relationship between cable operation data and thresholds. For example, when the data of a certain node reaches the starting threshold of the slight overheating stage, it is determined that the fault has entered the stage; when the data exceeds the termination threshold of the stage and meets the starting threshold of the insulation aging stage, it is determined that the fault has entered the next stage, thereby constructing a complete fault stage time-series evolution process.

[0098] Based on the above temporal characteristic analysis results, dynamic threshold adjustment rules are further configured. During the configuration process, the temporal relationship of the fault stages must be taken as the core basis. Specifically, for the slight overheating stage, the adjustment range of the corresponding warning threshold is set based on the difference between its starting and ending thresholds to ensure timely alerts are issued in the early stages of the fault. For the insulation aging stage and the partial discharge stage, considering their faster evolution rate and greater impact, the threshold adjustment interval is appropriately reduced to improve warning sensitivity.

[0099] At the same time, it is necessary to ensure that the dynamic threshold adjustment rules can adapt to the characteristics of different fault stages, and avoid the situation where fixed thresholds lead to delayed alerts in the rapid development stage of the fault, or excessive warnings in the early stage of the fault. Ultimately, the dynamic threshold adjustment rules should be highly consistent with the temporal pattern of fault evolution, laying the foundation for subsequent multi-scale assessment of the fault impact range and accurate fault alerts based on the attenuation coefficient.

[0100] For example, if the starting threshold for the slight overheating stage is 8% temperature rise deviation and the ending threshold is 15%, with a difference of 7%, then the adjustment range of the warning threshold for this stage is set to be ±0.5% every 2 hours based on the actual temperature rise deviation, so as to ensure that the primary warning can be triggered in time when the temperature rise deviation just exceeds 8%, and that frequent false alarms will not occur due to excessive adjustment range.

[0101] For the insulation aging stage, the starting threshold is a temperature rise deviation of 15% and the ending threshold is 22%. The evolution rate is 30% faster than the slight overheating stage. Therefore, the warning threshold adjustment interval is reduced to every 1 hour, and the adjustment range is set to ±0.8%. When the temperature rise deviation approaches 20%, the warning level is accelerated.

[0102] For the partial discharge stage, the starting threshold is 22% temperature rise deviation and the ending threshold is 30%. The impact of the fault will spread rapidly. Therefore, the adjustment interval of the warning threshold is further shortened to every 30 minutes, and the adjustment range is set to ±1.2%. Once the temperature rise deviation exceeds 25%, an advanced warning is immediately triggered. Through the phased and differentiated threshold configuration of the above steps, the dynamic threshold adjustment rules can accurately match the fault characteristics of each stage.

[0103] The method provided in this application embodiment also includes:

[0104] Associate fault types, historical fault cases, starting thresholds, and ending thresholds to create a fault warning knowledge graph;

[0105] Based on the fault early warning knowledge graph, potential fault nodes are identified, and topological expansion analysis of the fault's impact range is performed.

[0106] In this embodiment of the application, in order to make the multi-scale assessment of the fault impact range more comprehensive and forward-looking, avoid being limited to the current fault node and missing potential risks, and at the same time make the dynamic threshold adjustment rules more accurately adapt to the actual working conditions by combining historical experience, it is necessary to construct a fault early warning knowledge graph by associating core fault information, and rely on the graph to identify potential fault nodes and expand the impact range analysis, so as to improve the depth and accuracy of early fault assessment and early warning of cables.

[0107] Specifically, the first step is to set up a fault early warning knowledge graph. During the setup process, fault types, historical fault cases, and starting and ending thresholds need to be systematically correlated.

[0108] Among them, the fault types cover common early fault categories of cables such as conductor overheating faults, insulation damage faults, and joint aging faults. The historical fault cases include key information such as the occurrence time, evolution path, and handling results of similar cables under similar working conditions in the past. The starting threshold and the termination threshold are the core judgment criteria previously used to mark the slight overheating stage, insulation aging stage, and partial discharge stage.

[0109] Furthermore, after completing the fault early warning knowledge graph setup, potential fault nodes are identified based on this graph, and a topological expansion analysis of the fault impact range is performed to break through the limitations of the current fault nodes, accurately locate surrounding nodes that may be affected by the fault spread, and fully cover the potential fault risk area.

[0110] The method provided in this application embodiment, "identifying potential fault nodes based on the fault early warning knowledge graph," includes:

[0111] Set up a graph convolutional network based on the attention mechanism. The input node features include historical fault frequency and parameter anomaly frequency. Obtain the fault propagation weights between each fault evolution stage.

[0112] Based on the fault propagation weights between each fault evolution stage, the key nodes and fault propagation paths for fault spread are determined.

[0113] Based on the aforementioned fault early warning knowledge graph and key nodes of fault propagation, the fault propagation path is used to simulate a chain reaction process to identify potential fault nodes.

[0114] In this embodiment of the application, in order to avoid focusing only on the current fault node and missing the area that may be affected by the spread, and at the same time to provide a more comprehensive risk basis for the configuration of dynamic threshold adjustment rules, it is necessary to analyze the fault propagation law through graph convolutional networks and locate potential fault nodes by combining fault early warning knowledge graph information, so as to improve the foresight and comprehensiveness of early fault warning for cables.

[0115] Specifically, the first step is to set up a graph convolutional network based on an attention mechanism. During the setup process, it's crucial to understand that the core function of this graph convolutional network is to analyze the propagation patterns between fault evolution stages. Therefore, the input node features should focus on key data that reflects the risk of node failures, specifically including historical fault frequencies and parameter anomaly frequencies.

[0116] Among them, the historical failure frequency reflects how frequently the node has experienced failures in the past, while the parameter anomaly frequency reflects the number of times the node's operating data deviates from the normal range. These two types of features can be directly related to the node's fault susceptibility, providing basic data for obtaining fault propagation weights in the future.

[0117] Meanwhile, the introduction of the attention mechanism allows graph convolutional networks to give more attention to node features that have a greater impact on fault propagation (such as nodes with a high frequency of historical faults) during the computation process, avoiding the bias in propagation pattern analysis caused by equal feature weights, and ensuring that the network can more accurately capture the propagation characteristics of faults at different stages of evolution.

[0118] After setting up the graph convolutional network and inputting node features, the fault propagation weights between each fault evolution stage are further obtained. That is, relying on the computational power of the graph convolutional network, a deep analysis is performed on the input historical fault frequency and parameter anomaly frequency.

[0119] For example, for the evolution path of "slight overheating stage - insulation aging stage", the graph convolutional network will combine the historical fault frequency (e.g., node A has experienced 3 slight overheating faults in the past year) and parameter abnormality frequency (e.g., node A has experienced 5 voltage parameter abnormalities in the past 3 months) of each node in the fault warning knowledge graph to calculate the weight value of node A propagating the fault to surrounding nodes; for nodes with low historical fault frequency and few parameter abnormalities, the corresponding fault propagation weight is relatively low.

[0120] Through the above quantitative calculations, the abstract probability of fault propagation is transformed into specific weight values, providing a clear quantitative basis for subsequently determining key nodes and propagation paths.

[0121] Furthermore, after obtaining the fault propagation weights, the key nodes and fault propagation paths for fault spread are further determined based on the obtained weights.

[0122] Among them, the determination of critical nodes should be based on the propagation weight. The higher the propagation weight of a node, the greater its influence on surrounding nodes during the fault propagation process. For example, if a node has a fault propagation weight of 0.8 (out of 1), which is much higher than the 0.3-0.5 of other nodes, the node will be determined as a critical node for fault propagation.

[0123] Simultaneously, by combining the physical connection relationships of cable lines with the fault evolution logic in the fault early warning knowledge graph, the possible paths for fault propagation are identified starting from key nodes. For example, if the key node is a joint node in the middle of the line, and the "joint aging fault" in the knowledge graph often propagates to conductor nodes at both ends of the line, then "key joint node - left conductor node" and "key joint node - right conductor node" are identified as the main fault propagation paths, ensuring that the path identification conforms to both the quantitative weighting rules and the actual fault evolution logic.

[0124] Finally, based on the fault early warning knowledge graph and the key nodes of fault propagation, the chain reaction process is simulated using the determined fault propagation path to ultimately identify potential fault nodes.

[0125] During the simulation, historical fault cases in the fault early warning knowledge graph should be used as a reference to reconstruct the process of the fault spreading from the key node along the propagation path.

[0126] For example, referring to historical fault cases in the fault warning knowledge graph where "after a critical joint node experiences an insulation aging fault, it usually spreads to the conductor nodes at both ends within 48 hours", and combining the current fault status of the critical node (e.g., it is already in the insulation aging stage), the conductor nodes that the fault may spread to in the next 48 hours can be simulated.

[0127] Meanwhile, if the frequency of abnormal parameters of a node on the propagation path has reached a threshold (e.g., three abnormal partial discharge signals in the past week), the node will be marked as a potential fault node to ensure that the simulation results can be adjusted in conjunction with the real-time node status.

[0128] Through the above steps, not only can potential fault nodes be accurately located, but the identification process of potential nodes can also integrate the quantitative analysis of the network model, the historical experience of the knowledge graph, and the real-time node status, ensuring that the identification results are consistent with the actual fault evolution pattern.

[0129] Ultimately, the information on these potential fault nodes will be synchronized to the multi-scale assessment of the fault impact range, providing a more comprehensive risk area reference for subsequent assessments based on attenuation coefficients and early fault heat generation. This also allows the dynamic threshold adjustment rules to set more precise early warning thresholds for potential fault nodes, further improving the effectiveness of early cable fault warnings.

[0130] Furthermore, after identifying potential fault nodes and sorting out fault propagation paths, a multi-scale assessment of the fault impact range is carried out based on fault parameter extraction maps, combined with the environmentally corrected attenuation coefficient and the early fault heat generation under different equivalent insulation resistances.

[0131] Specifically, the equivalent insulation resistance is first correlated with the heat generated by early faults. Different equivalent insulation resistances correspond to differences in cable insulation performance, while the heat generated directly reflects the intensity of fault energy release. Through the correspondence between the two, the core influence area of ​​a single fault node can be preliminarily delineated.

[0132] Furthermore, by combining the failure evolution rate quantified by the attenuation coefficient and superimposing the distribution of potential failure nodes, the assessment scope is expanded from the core area of ​​a single node to a multi-dimensional range of "core area + diffusion area + potential risk area".

[0133] Furthermore, precise fault alerts are achieved by combining the previously configured dynamic threshold adjustment rules. That is, differentiated early warning strategies are matched for different areas defined by multi-scale assessments.

[0134] Specifically, advanced early warnings are triggered for core affected areas (such as nodes where partial discharge has occurred), and fault location and heat generation data are pushed in real time according to the partial discharge stage threshold rules (such as adjusting ±1.2% every 30 minutes); intermediate early warnings are triggered for diffusion areas (such as the slightly overheated diffusion section corresponding to an attenuation coefficient of 0.5), and diffusion progress is pushed according to the insulation aging stage rules (such as adjusting ±0.8% every 1 hour).

[0135] In addition, a primary warning is triggered for potential risk areas (such as nodes where the frequency of abnormal parameters reaches a threshold), and monitoring suggestions are pushed according to the rules for the slight overheating stage (such as adjusting ±0.5% every 2 hours).

[0136] At the same time, historical handling cases in the fault early warning knowledge graph are synchronously linked to the reminder information to ensure that maintenance personnel can quickly locate the scope of the fault impact and efficiently handle it by referring to past experience, ultimately achieving accurate early fault management of cables.

[0137] The embodiments of this application, through the specific implementation methods described above, achieve the following technical effects:

[0138] This application proposes a method for extracting early fault parameters of cables by integrating temperature rise process data. First, for the cable under test in a conductive state, multi-node temperature rise data transmitted via a temperature-measuring optical fiber is received and compared with the standard temperature rise curve under rated cable conditions to extract deviation information. Then, a first fault characteristic index is set by associating the cable load log with insulation node data, and a second fault characteristic index is set by associating partial discharge signal records with conductor node data, forming the basic data for fault monitoring. Subsequently, directed edges of the fault parameter extraction graph are determined through a sliding time window, and the pre-existing causes, time dependencies, and signal correlation information for each fault evolution stage are labeled. Dynamic weights are then assigned to the directed edges of the fault parameter extraction graph. The system obtains attenuation coefficients based on LSTM time-series sensing fusion stage information, uses early fault heat generation under different equivalent insulation resistances as the core feature, and combines start / end thresholds to conduct multi-scale assessment of the fault impact range. Simultaneously, it constructs a fault early warning knowledge graph by associating fault types, historical cases, and thresholds, sets up an attention-based graph convolutional network, inputs historical fault frequencies and parameter anomaly frequencies to obtain fault propagation weights, determines key nodes and propagation paths, simulates chain reactions to identify potential fault nodes, and expands the impact range analysis. Finally, based on the assessment results and potential node information, it configures dynamic threshold adjustment rules according to the fault stage to achieve accurate early warning of cable faults.

[0139] The method provided in this application, through the technical solution of "data acquisition and feature index setting - fault parameter extraction map construction and optimization - multi-scale assessment of fault impact and potential node identification - dynamic threshold early warning", solves the problems in traditional cable fault detection caused by relying on only a single data, such as one-sided parameter extraction, fixed models that cannot adapt to environmental changes, and lack of foresight in fault impact assessment. It avoids the risks of early fault omission, misjudgment, and fault propagation, improves the accuracy of early fault parameter extraction and the timeliness of early warning, and provides technical support for the safe and stable operation of cables.

[0140] Example 2, as shown in the appendix Figure 2 As shown, based on the inventive concept of the cable early fault parameter extraction method that integrates temperature rise process data provided in Embodiment 1, this application also provides a cable early fault parameter extraction system that integrates temperature rise process data, specifically including:

[0141] The information extraction and index setting module 01 is used to receive multi-node temperature rise data uploaded by the temperature measuring optical fiber based on the cable under test in the conducting state, compare it with the rated operating conditions of the cable, extract the deviation information from the standard temperature rise curve, and set the first fault characteristic index and the second fault characteristic index.

[0142] The fault parameter extraction map formulation module 02 is used to formulate a fault parameter extraction map that conforms to the cable aging logic relationship by using the deviation information and combining the first fault feature index and the second fault feature index.

[0143] The fault impact range assessment and early warning module 03 is used to perform multi-scale assessment of the fault impact range in the fault parameter extraction map based on the attenuation coefficient and with the early fault heat generation under different equivalent insulation resistances as the core correlation feature, and to provide fault reminders using dynamic threshold adjustment rules.

[0144] In one embodiment, the information extraction and indicator setting module 01 further includes:

[0145] Associate the cable load log with each cable insulation node under the rated operating condition of the cable and set a first fault characteristic index; associate the partial discharge signal record with each cable conductor node under the rated operating condition of the cable and set a second fault characteristic index.

[0146] In one embodiment, the fault parameter extraction map formulation module 02 is further configured to:

[0147] An integrated current sensing unit, voltage monitoring unit, and ambient temperature detection unit are used to construct a dynamic operating parameter sensing network to acquire cable operating environment data. Based on the aging sub-intervals associated with the fault influence coefficient, and combined with the cable operating environment data, an environmental fault correlation matrix is ​​constructed to dynamically correct the attenuation coefficient of the fault parameter extraction spectrum.

[0148] Furthermore, the fault parameter extraction map formulation module 02 also includes:

[0149] By using a sliding time window, the directed edges corresponding to the fault parameter extraction map are determined and connected to each fault evolution stage; at the same time, the preceding causes, time-dependent parameters and signal correlation information between each fault evolution stage are labeled.

[0150] Furthermore, the fault parameter extraction map formulation module 02 also includes:

[0151] The row dimension of the environmental fault correlation matrix is ​​the cable operating environment data, and the column dimension of the environmental fault correlation matrix includes conductor overheating sub-interval, insulation damage sub-interval, and joint aging sub-interval; the environmental fault correlation matrix is ​​quantified and dynamically updated; when the cable operating environment data does not meet the safe operating environment restrictions, the matrix element value of the environmental fault correlation matrix is ​​set to 0, triggering a mandatory detection command.

[0152] In one embodiment, the fault impact range assessment and early warning module 03 further includes:

[0153] Dynamic weights are assigned to the directed edges of the fault parameter extraction map. Based on LSTM time-aware fusion, the pre-cause relationships, time-dependent parameters and signal correlation information between each fault evolution stage are fused to obtain the attenuation coefficient.

[0154] Furthermore, the fault impact range assessment and early warning module 03 also includes:

[0155] A time-series feature analysis is performed using a start threshold and a stop threshold, which are used to mark the slight overheating stage, the insulation aging stage, and the partial discharge stage. The dynamic threshold adjustment rule is configured based on the time-series relationship corresponding to the start threshold and the stop threshold.

[0156] Furthermore, the fault impact range assessment and early warning module 03 also includes:

[0157] By associating fault types, historical fault cases, starting thresholds, and ending thresholds, a fault early warning knowledge graph is established; based on the fault early warning knowledge graph, potential fault nodes are identified, and topological expansion analysis of the fault's impact range is performed.

[0158] Furthermore, the fault impact range assessment and early warning module 03 also includes:

[0159] An attention-based graph convolutional network is set up, with input node features including historical fault frequency and parameter anomaly frequency, to obtain fault propagation weights between each fault evolution stage; based on the fault propagation weights between each fault evolution stage, key nodes and fault propagation paths for fault spread are determined; based on the fault early warning knowledge graph and key nodes for fault spread, the fault propagation paths are used to simulate a chain reaction process to determine the potential fault nodes.

[0160] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0161] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0162] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for extracting early fault parameters of cables by integrating temperature rise process data, characterized in that, The method includes: Based on the cable under test in the conducting state, receive multi-node temperature rise data transmitted by the temperature measuring optical fiber, compare it with the rated operating conditions of the cable, extract the deviation information from the standard temperature rise curve, and set the first fault characteristic index and the second fault characteristic index. Based on the deviation information, combined with the first fault characteristic index and the second fault characteristic index, a fault parameter extraction map that conforms to the logic relationship of cable aging is formulated. In the fault parameter extraction map, based on the attenuation coefficient, the early fault heat generation under different equivalent insulation resistances is used as the core correlation feature to evaluate the fault impact range on a multi-scale basis, and a dynamic threshold adjustment rule is used to provide fault alerts. The first fault characteristic indicator and the second fault characteristic indicator are set, including: Associate the cable load log with each cable insulation node under the rated operating conditions of the cable, and set the first fault characteristic index. The partial discharge signal is recorded and associated with each cable conductor node under the rated operating condition of the cable, and a second fault characteristic index is set. Specifically, a sliding time window is used to determine the directed edges corresponding to the fault parameter extraction map and connect each fault evolution stage; At the same time, the preceding causes, time-dependent parameters, and signal correlation information between each stage of the fault evolution are labeled; This includes multi-scale assessment of the fault's impact range and the use of dynamic threshold adjustment rules for fault alerts, including: Time-series feature analysis is performed using start and end thresholds, which are used to mark the mild overheating stage, insulation aging stage, and partial discharge stage. The dynamic threshold adjustment rule is configured based on the temporal relationship between the start threshold and the end threshold.

2. The method for extracting early fault parameters of cables by integrating temperature rise process data as described in claim 1, characterized in that, Based on the attenuation coefficient, and taking the early fault heat generation under different equivalent insulation resistances as the core correlation feature, the method further includes: Dynamic weights are assigned to the directed edges of the fault parameter extraction map. Based on LSTM time-aware fusion, the pre-cause relationships, time-dependent parameters and signal correlation information between each fault evolution stage are fused to obtain the attenuation coefficient.

3. The method for extracting early fault parameters of cables by integrating temperature rise process data as described in claim 1, characterized in that, The method includes: Associate fault types, historical fault cases, starting thresholds, and ending thresholds to create a fault warning knowledge graph; Based on the fault early warning knowledge graph, potential fault nodes are identified, and topological expansion analysis of the fault's impact range is performed.

4. The method for extracting early fault parameters of cables by integrating temperature rise process data as described in claim 3, characterized in that, Based on the aforementioned fault early warning knowledge graph, potential fault nodes are identified, the method comprising: Set up a graph convolutional network based on the attention mechanism. The input node features include historical fault frequency and parameter anomaly frequency. Obtain the fault propagation weights between each fault evolution stage. Based on the fault propagation weights between each fault evolution stage, the key nodes and fault propagation paths for fault spread are determined. Based on the aforementioned fault early warning knowledge graph and key nodes of fault propagation, the fault propagation path is used to simulate a chain reaction process to identify potential fault nodes.

5. The method for extracting early fault parameters of cables by integrating temperature rise process data as described in claim 1, characterized in that, The method further includes: developing a fault parameter extraction map that conforms to the logic relationship of cable aging. A dynamic operating parameter sensing network is constructed by integrating a current sensing unit, a voltage monitoring unit, and an ambient temperature detection unit to acquire cable operating environment data; Based on the joint aging sub-intervals associated with the fault influence coefficient, and combined with the cable operating environment data, an environmental fault correlation matrix is ​​constructed, and the attenuation coefficient of the fault parameter extraction spectrum is dynamically corrected.

6. The method for extracting early fault parameters of cables by integrating temperature rise process data as described in claim 5, characterized in that, The method further includes constructing an environmental fault correlation matrix and dynamically correcting the attenuation coefficient of the fault parameter extraction spectrum. The row dimension of the environmental fault correlation matrix is ​​the cable operating environment data, and the column dimension of the environmental fault correlation matrix includes conductor overheating sub-interval, insulation damage sub-interval, and joint aging sub-interval. The environmental fault correlation matrix is ​​quantized and dynamically updated. When the cable operating environment data does not meet the safe operating environment restrictions, the matrix element values ​​of the environmental fault correlation matrix are set to 0, triggering a forced detection command.

7. A cable early fault parameter extraction system integrating temperature rise process data, characterized in that, The system is used to execute the cable early fault parameter extraction method based on the fused temperature rise process data according to any one of claims 1-6, the system comprising: The information extraction and index setting module is used to receive multi-node temperature rise data uploaded by the temperature measuring optical fiber based on the cable under test in the conducting state, compare it with the rated operating conditions of the cable, extract the deviation information from the standard temperature rise curve, and set the first fault characteristic index and the second fault characteristic index. The fault parameter extraction map formulation module is used to formulate a fault parameter extraction map that conforms to the cable aging logic relationship by using the deviation information and combining the first fault feature index and the second fault feature index. The fault impact range assessment and early warning module is used to assess the fault impact range in the fault parameter extraction map based on the attenuation coefficient and with the early fault heat generation under different equivalent insulation resistances as the core correlation feature, and to provide fault alerts using dynamic threshold adjustment rules.

Images