Method and system for predicting and analyzing health state of cable in full life cycle

Abstract

The invention discloses a cable full life cycle health state prediction analysis method and system, and relates to the technical field of cable state monitoring, and the system comprises a power supply management module, a health comprehensive analysis module, and an adaptive sampling communication module. The power supply management module collects an electric parameter flow, generates an instantaneous power density sequence, and detects parameters such as a charge-discharge state, an output power failure type, energy supply reliability and the like in combination with historical and real-time data; the health comprehensive analysis module dynamically calculates a sampling power consumption budget, predicts an output energy curve and a risk level through graph time sequence fusion, and determines a recommended sampling level after correction; the adaptive sampling communication module maps a layering strategy and dynamically executes sampling and communication scheduling; the system solves the problem of adaptive sampling strategy matching in multi-node high-voltage distribution network monitoring, realizes accurate monitoring and strategy adaptation under differentiated aging, and improves the cable health management efficiency and reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable state monitoring, and more particularly to a cable full life cycle health state prediction analysis method and system. BACKGROUND

[0002] In the prior art, the monitoring of the health state of the cable mainly relies on manual inspection or periodic detection by handheld devices. This approach has many shortcomings. Since it cannot be continuously monitored online, the operation and maintenance personnel cannot obtain complete state data of the cable throughout its life cycle in a timely manner, which makes it impossible to accurately predict the aging trend and remaining service life of the cable. Often, hidden dangers are not discovered until a serious fault or power outage occurs, which affects the reliability of power supply of the power grid. At the same time, some existing online monitoring devices have single functions, such as monitoring only a single parameter such as partial discharge or temperature of the cable, and cannot fuse and analyze multiple state parameters, making it difficult to assess the health condition of the cable in a timely and comprehensive manner.

[0003] In addition, online monitoring devices are limited by the power supply mode. Usually, such devices are installed at outdoor cable joints and use battery power supply or current transformer induction power supply to obtain energy. Existing monitoring systems mostly use fixed frequency sampling and data uploading strategies, and do not dynamically adjust the sampling frequency and power consumption distribution according to the aging degree of the cable or the load change. On the one hand, too low sampling frequency may miss fault precursor data, reducing the accuracy of diagnosis. On the other hand, too high sampling frequency will accelerate the consumption of battery energy, especially when the load current is low, the mutual inductance power supply is insufficient, leading to power supply shortage or even power failure of the monitoring device.

[0004] In view of the above problems, the present application provides a solution. SUMMARY

[0005] The purpose of the present application is to provide a cable full life cycle health state prediction analysis method and system to solve the problem of insufficient adaptive sampling strategy matching under cable differential aging in the real-time monitoring scene of multi-node high-voltage distribution network.

[0006] The purpose of the present application can be achieved by the following technical solutions:

[0007] The cable full life cycle health state prediction analysis system comprises the following modules:

[0008] The power supply management module is used to collect cable monitoring device node cable electrical parameter flow and aging basic data, and to perform derivative operation on the cable electrical parameter flow to generate an instantaneous power density sequence. The charging and discharging state of the cable monitoring device node is detected based on the instantaneous power density sequence and the cable electrical parameter flow, and the power supply reliability parameters including the mutual inductance power supply remaining ratio, the battery state of charge percentage, the cable aging grade coefficient and the actual supportable sampling power consumption upper limit are outputted.

[0009] a health comprehensive analysis module configured to dynamically calculate an upper limit of a sampling power consumption budget according to the power-off type identifier, the cable aging level coefficient, and the power supply reliability parameter, perform time synchronization based on cable state data, generate a health initial vector, and based on the health initial vector, take the cable monitoring device node topology as a graph structure, perform graph time sequence fusion to predict an output residual available energy curve, an uncertainty curve, and a preliminary risk level, correct the preliminary risk level based on the uncertainty curve to perform confidence correction to obtain a final risk level, and determine the upper limit of the sampling power consumption budget and a recommended sampling level;

[0010] an adaptive sampling communication module configured to receive the final risk level and the recommended sampling level, map the final risk level and the recommended sampling level into a hierarchical sampling strategy, and dynamically execute a sampling strategy and a communication strategy of each cable monitoring device node.

[0011] As a further scheme of the present application, the power supply management module is configured to perform the following steps: collecting bus current, bus voltage, backup battery current, and battery terminal voltage according to a fixed sampling frequency to generate a cable electrical parameter stream; multiplying the bus voltage and the current corresponding to each sampling point to obtain bus instantaneous power; performing first derivative operation on the bus instantaneous power to generate an instantaneous power density sequence; and determining the power-off type identifier as a deep discharge, normal discharge, or charging operation state based on the instantaneous power density sequence, the cable electrical parameter stream, and preset threshold values corresponding thereto.

[0012] As a further scheme of the present application, the determination conditions of the power-off type identifier include: detecting the charging and discharging state of the cable monitoring device node based on the obtained instantaneous power density sequence, the cable electrical parameter stream, and preset battery continuous discharge threshold values and preset battery static current threshold values, preset power density positive fluctuation upper threshold values, and preset power density negative fluctuation lower threshold values, including:

[0013] when the instantaneous power density sequence of the corresponding monitoring device node at consecutive N sampling points is less than the preset power density negative fluctuation lower threshold value and the backup battery discharge current is less than the preset battery continuous discharge threshold value, determining that the corresponding monitoring device node is in a deep discharge power-off state, and outputting a deep discharge power-off identifier;

[0014] when the instantaneous power density sequence is between the power density negative fluctuation lower threshold value and the preset power density positive fluctuation upper threshold value, and the absolute value of the backup battery discharge current is less than or equal to the preset battery static current threshold value, the corresponding monitoring device node is in a normal discharge state, and a normal discharge power-off identifier is outputted;

[0015] when the instantaneous power density sequence is greater than the preset power density positive fluctuation upper threshold value and the backup battery discharge current is greater than the preset battery continuous discharge threshold value, the corresponding cable monitoring device node is significantly charged, and a charging operation identifier is outputted.

[0016] As a further scheme of the present application: the energy reliability parameter, the cable aging grade coefficient and the actual supportable sampling power upper limit are generated by the following way: the instantaneous power density sequence is integrated to obtain the energy integral value in a certain period, the energy integral value is compared with the rated mutual inductance power capacity to calculate the mutual inductance power remaining ratio; the battery state of charge percentage is derived based on the ampere-hour integration method according to the backup battery current integration and the battery rated capacity; the formula is as follows: ; in the formula, represents the battery state of charge at time t, is the battery state of charge at the initial time , represents the battery rated capacity unit, ampere-hour; is the battery current at time t, unit ampere, positive value when charging, negative value when discharging, represents the integral of the battery current from the initial time to the current time t, that is, the change amount of the battery charge in this period of time;

[0017] Based on the pre-stored and provided cable service life and historical failure data in the system, the ratio of the cable service life to the preset life reference value is calculated to obtain the basic aging proportion;

[0018] Based on the historical failure times, the offset is set, the basic aging proportion is added to the offset to generate a quantitative value, which is recorded as the cable aging grade coefficient; the product of the mutual inductance power remaining ratio, the battery state of charge percentage and the cable aging grade coefficient is multiplied by the device preset power consumption to calculate the actual supportable sampling power upper limit.

[0019] As a further scheme of the present application: the dynamic calculation of the sampling power budget upper limit in the health comprehensive analysis module includes: when the cable aging grade coefficient is less than or equal to the low aging risk threshold in the deep discharge power-off identification, it is marked as a low aging label, and the corresponding first compression ratio is configured;

[0020] When the cable aging grade coefficient is greater than the low aging risk threshold and less than the high aging threshold, it is marked as a medium aging label, and the corresponding second compression ratio is configured;

[0021] When the cable aging grade coefficient is greater than or equal to the high aging threshold, it is marked as a high aging label, and the corresponding third compression ratio is configured; and the first compression ratio is greater than the second compression ratio, which is greater than the third compression ratio; and based on the above aging label grading, the different compression ratios are multiplied by the actual supportable sampling power upper limit to obtain the maximum sampling power allowed in the current node deep discharge mode, that is, the sampling power budget upper limit.

[0022] As a further scheme of the present application: the health comprehensive analysis module adopts a graph time sequence fusion prediction algorithm, including: a graph convolution network processing, taking the cable monitoring device node as a graph node and the cable connection topology as an edge to construct a relationship graph; a long short-term memory network processing cable multi-source sensing data of each cable monitoring device node; a fusion of a recurrent neural network and a graph neural network, outputting a residual available energy curve and a preliminary risk level; based on an uncertainty curve, calculating the confidence weight of each time step; using a weighted average formula to bias correct the preliminary risk level to generate a final risk level; wherein the residual available energy curve is used to reflect the possible change of the energy remaining in the power supply of the cable monitoring device node in the future period of time.

[0023] As a further scheme of the present application: the determination rule of the recommended sampling level in the health comprehensive analysis module includes: triggering high-priority monitoring when the preliminary risk level is greater than or equal to the high-risk level threshold or the aging label is a high aging label, that is, sampling the core parameters with high frequency according to the upper limit of the sampling power consumption budget;

[0024] triggering medium-priority monitoring when the preliminary risk level is greater than or equal to the medium-risk threshold and less than the high-risk level threshold or the aging label is a medium aging label, that is, balancing the sampling of core and non-core parameters according to the upper limit of the sampling power consumption budget;

[0025] triggering low-priority monitoring when the preliminary risk level is less than the low-risk threshold and the aging label is a low aging label, that is, reducing the sampling frequency of non-core parameters according to the upper limit of the sampling power consumption budget;

[0026] triggering normal energy supply scene adaptation when the aging label is a normal aging label, that is, directly determining according to the preliminary risk level, and the sampling frequency is still according to the fixed sampling frequency.

[0027] As a further scheme of the present application: it also includes a life attenuation evaluation function: based on the residual available energy curve and the cable aging level coefficient, generating a high life attenuation risk, energy supply warning or aging warning determination result; when the residual available energy curve shows that the energy will drop below the energy warning threshold for maintaining basic monitoring in the future period of time, and the cable aging level coefficient is still greater than or equal to the high aging threshold, it is determined that the cable monitoring device node is a high life attenuation risk;

[0028] if only the residual available energy drops below the energy warning threshold, and the cable aging level coefficient is still less than or equal to the low aging threshold, it is determined as an energy supply warning;

[0029] if only the aging level coefficient is higher than the energy warning threshold, and the cable aging level coefficient is still greater than or equal to the high aging threshold, it is determined as an aging warning;

[0030] for the cable monitoring device node with high life attenuation risk, an operation and maintenance work order containing joint maintenance suggestions and sampling strategy upgrade is automatically triggered.

[0031] As a further scheme of the application: when the adaptive sampling communication module executes the sampling strategy of each cable monitoring device node, it includes: when the recommended sampling level corresponding to the cable monitoring device node is high-priority monitoring, entering a first-level strategy mode, and using an event-triggered mechanism to upload key data in real time; when the recommended sampling level corresponding to the cable monitoring device node is medium-priority monitoring, entering a second-level strategy mode, and using a data compression and batch uploading strategy to reduce communication energy consumption;

[0032] When the recommended sampling level corresponding to the cable monitoring device node is low-priority monitoring, entering a third-level strategy mode oriented to energy saving.

[0033] The cable full-life-cycle health state prediction analysis method includes the following steps: including the following steps: step one: collecting bus current, bus voltage, backup battery current and battery terminal voltage of the cable monitoring device node, generating cable electric parameter flow; calculating the first derivative of bus instantaneous power, generating instantaneous power density sequence; based on the instantaneous power density sequence and the preset threshold, determining the power failure type identifier as deep discharge, normal discharge or charging operation state;

[0034] Step two: obtaining cable service life and historical failure data, calculating the ratio of service life to preset life reference value to obtain a basic aging proportion; based on the historical failure times, setting an offset, adding the basic aging proportion and the offset to generate a cable aging level coefficient;

[0035] Step three: taking the cable monitoring device node as a graph node and the cable connection topology as an edge to construct a relationship graph; using a graph time sequence fusion prediction algorithm to process cable state data, outputting a residual available energy curve, an uncertainty curve and a preliminary risk level; based on the uncertainty curve, calculating a confidence weight to bias correct the preliminary risk level to generate a final risk level;

[0036] Step four: when the power failure type identifier is deep discharge, according to the cable aging level coefficient, compressing the sampling power consumption upper limit by grades; and combining the final risk level and the aging level coefficient, distributing a recommended sampling level, mapping the recommended sampling level to a hierarchical sampling strategy, and based on the hierarchical sampling strategy, dynamically adjusting the sampling and communication operations of each cable monitoring device node according to the sampling power consumption upper limit.

[0037] The beneficial effects of the application are:

[0038] (1) The present application realizes dynamic monitoring and adaptive sampling of the health status of the cable full life cycle by constructing a cooperative architecture of a power supply management module, a health comprehensive analysis module and a self-adaptive sampling communication module. The power supply management module collects cable electrical parameter flow in real time and generates instantaneous power density sequence, accurately identifies power-off type and power supply state; the health comprehensive analysis module generates risk level based on graph time sequence fusion prediction algorithm, combines cable topological relationship and multi-source data to dynamically match sampling power consumption budget; the self-adaptive sampling communication module flexibly adjusts sampling frequency and communication strategy according to risk level and power supply state. The three cooperatively solve the contradiction between high energy consumption and missing hidden dangers caused by insufficient sampling in the traditional fixed sampling mode, while ensuring the continuity of monitoring and improving the fault warning accuracy;

[0039] (2) The present application realizes full-chain management from state monitoring to life prediction by introducing the correlation analysis of cable aging grade coefficient and residual available energy curve. On the one hand, based on the aging grade coefficient and power supply state, the sampling power consumption is dynamically allocated, such as using differentiated compression ratio according to the aging label when deep discharging, and preferentially guaranteeing core parameter monitoring when power supply is tight, prolonging the device endurance; on the other hand, through the coupling judgment of residual available energy curve and aging grade, results such as high life attenuation risk and power supply warning are generated to link operation and maintenance work order, promoting the operation and maintenance mode from post-fault maintenance to predictive maintenance, significantly improving the reliability and operation and maintenance efficiency of power grid cable operation. BRIEF DESCRIPTION OF DRAWINGS

[0040] The present application will be further described below with reference to the accompanying drawings.

[0041] Figure 1 is a system framework structure diagram of the cable full life cycle health status prediction analysis system of the present application;

[0042] Figure 2 is a principle diagram of the power supply management module in the cable full life cycle health status prediction analysis system of the present application;

[0043] Figure 3 is an implementation process diagram of the cable full life cycle health status prediction analysis method of the present application. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0045] Embodiment one

[0046] Please refer to Figure 1 The application is a cable full life cycle health state prediction analysis system, comprising the following modules: a power supply management module, a health comprehensive analysis module, and a self-adaptive sampling communication module.

[0047] The power supply management module is used to obtain cable electrical parameter flow and aging basic data, and perform first derivative operation on the cable electrical parameter flow to generate a transient power density sequence; based on the transient power density sequence and the cable electrical parameter flow, the charging and discharging state of the cable monitoring device node is detected, the cable power-off type identifier is obtained, and the mutual inductance power supply remaining ratio, the power supply reliability parameter including the battery state of charge percentage, and the actual supportable sampling power consumption upper limit are output in real time.

[0048] In this embodiment, the power supply management module is mainly responsible for power supply and electrical parameter calculation of the cable monitoring device node, and the circuit principle is as shown in Figure 2 The power supply management module is as shown in Figure 2 CT1 is a buckle type current transformer, which senses the alternating current signal in the cable bus, filters through the inductor L2 (I-shaped inductor), rectifies through the rectifier bridge DB1 to a direct current signal, and then amplifies and stabilizes the direct current signal through the charging chip U1, and outputs about 12V voltage to charge the battery BT1. The above power supply management module as a whole is installed at the cable joint as a cable monitoring device node. Multiple cable monitoring device nodes can be arranged along the line, and each cable monitoring device node forms a distributed monitoring network with the background platform through a wireless network, to realize continuous state perception and risk prediction of the cable full life cycle.

[0049] While realizing the power supply function, this module judges the power supply state and the basic working condition of the cable by collecting and analyzing key electrical parameters and historical data, to provide a basis for subsequent analysis; the implementation steps are as follows:

[0050] The electrical parameters reflecting the cable power supply and battery state are collected to form an electrical parameter time sequence, i.e., cable electrical parameter flow; specifically including the bus current, i.e., the cable load current size sensed by CT1, reflecting the load current intensity passing through the node, the bus voltage, i.e., the direct current bus voltage after rectification and filtering of the mutual inductance power supply circuit, indicating the current mutual inductor power supply capability output level, the standby battery current, i.e., the charging and discharging current of the battery BT1, the positive value usually indicating the charging current, the negative value indicating the discharging current, i.e., the battery discharging to supply power to the system, and the battery terminal voltage, i.e., the voltage between the two terminals of the battery BT1, which can reflect the battery charge level.

[0051] It should be noted that these electrical parameters can be obtained through the ADC sampling circuit and the battery management unit in the power supply management module, and can be recorded continuously according to the fixed sampling frequency of 1s-10s, which can be adjusted according to the actual response demand, to obtain the cable electrical parameter flow changing with time;

[0052] Based on the collected cable electrical parameters, the instantaneous power of the bus is approximately calculated by multiplying the bus voltage and current at each sampling point. The first derivative of the instantaneous power is then calculated to obtain the instantaneous power density sequence. By taking the first derivative of relevant parameters, the transient characteristics of parameter abrupt changes can be highlighted. For example, when the cable bus current suddenly drops, the first derivative of the instantaneous power of the bus shows a significant negative peak; when the backup battery switches from charging to discharging, the rate of change of battery current changes drastically. This instantaneous power change sequence can sensitively capture transient events such as power outages and sudden load changes, providing a basis for identifying power supply anomalies.

[0053] Integrating the instantaneous power density sequence yields the energy integral value over a certain period. This energy integral value is then compared to the rated capacity of the mutual inductance power supply to calculate the remaining capacity ratio. Simultaneously, the state of charge (SCC) is obtained from the battery terminal voltage and standby battery current collected by the battery management unit. Based on the ampere-hour integration method, the SCC is derived by integrating the standby battery current and combining it with the battery's rated capacity, and the output is... The battery's state of charge percentage reflects the remaining battery capacity; specifically, the calculation of the battery's state of charge percentage is based on the following formula: In the formula, This represents the state of charge of the battery at time t. It is the initial moment The state of charge (SOC) of a battery is typically determined after battery initialization or calibration, such as when a new battery is fully charged. Can be set to , The rated capacity of a battery is expressed in ampere-hours (AH). This value can be obtained from the battery's specifications and reflects the total amount of charge the battery can store. It is the battery current at time t, in amperes, during charging. It is a positive value during discharge and a negative value during discharge. Indicates from the initial time The integral of the battery current up to the current time t, that is, the change in the amount of charge in the battery during this time.

[0054] Furthermore, the system pre-stores and provides cable service life and historical fault data as auxiliary inputs for assessing the current state. The cable service life is divided by a preset cable life reference value to obtain a basic aging ratio. An offset is then set based on the historical fault data, and the basic aging ratio is added to the set offset to obtain the cable aging level coefficient. For example, if a cable has been in operation for 15 years and has experienced two faults in the past 3 years, the basic aging ratio based on a 30-year lifespan is 15 divided by 30, which equals 0.5. The two faults may correspond to an additional 0.1, resulting in a cable aging level coefficient of 0.6. The value range can be 0 to 1 or 0 to 100%.

[0055] Based on the obtained instantaneous power density sequence, the cable electrical parameter flow and its corresponding preset battery continuous discharge threshold and preset battery static current threshold, the preset power density positive fluctuation upper threshold and the preset power density negative fluctuation lower threshold, the charge and discharge state of the cable monitoring device node is detected, including:

[0056] When the following conditions of the continuous N sampling points (N≥3) are simultaneously satisfied and continuously established, the corresponding monitoring device node is in a deep discharge power-off state, and a deep discharge power-off identifier is output; the conditions are that when the instantaneous power density sequence is less than the preset power density negative fluctuation lower threshold and the backup battery discharge current is less than the preset battery continuous discharge threshold;

[0057] When the instantaneous power density sequence is between the power density negative fluctuation lower threshold and the preset power density positive fluctuation upper threshold, and the absolute value of the backup battery discharge current is less than or equal to the preset battery static current threshold, the corresponding monitoring device node is in a normal discharge state, i.e. no significant discharge or charge, and a normal discharge power-off identifier is output;

[0058] When the instantaneous power density sequence is greater than the preset power density positive fluctuation upper threshold and the backup battery discharge current is greater than the preset battery continuous discharge threshold, the corresponding cable monitoring device node is significantly charging, and a charging operation identifier is output;

[0059] It should be noted that the above-mentioned preset battery continuous discharge threshold is set to a negative value such as continuous discharge current = 5A, and the preset battery continuous discharge threshold is -5A, which is used to identify the battery deep discharge state; the preset power density positive fluctuation upper threshold is the maximum positive fluctuation value of the instantaneous power density sequence under normal working conditions, such as the power rise caused by short-time load fluctuation, which is used to define that the power fluctuation is still within the normal range; the preset power density negative fluctuation lower threshold is based on historical normal working condition data, and the minimum negative mutation value of the instantaneous power density sequence is calculated, such as the negative extreme value of the instantaneous power density sequence in 1000 hours of normal operation, and the absolute value thereof is taken as the preset power density negative fluctuation lower threshold, which is used to identify abnormal power supply power drop;

[0060] The mutual inductance power source remaining ratio and the battery state of charge percentage calculated above are taken as the energy supply reliability parameters, and are output together with the power-off type identifier and the cable aging grade coefficient; at the same time, the product of the mutual inductance power source remaining ratio, the battery state of charge percentage and the cable aging grade coefficient is multiplied by the device preset power consumption to calculate the actual supportable sampling power consumption upper limit, which provides a basis for subsequent module control sampling strategy; wherein the device preset power consumption can be obtained through the basic configuration information of the corresponding cable monitoring device node;

[0061] The health comprehensive analysis module is used for dynamically adjusting the sampling power consumption according to the energy supply reliability parameter and the cable aging grade coefficient after receiving the discharge type identification, performing multi-path synchronous collection on the cable state data, obtaining a health initial vector, performing time sequence fusion prediction on a cable joint branch topology execution graph, modeling the health initial vector, outputting a cable joint monitoring node residual available energy curve, an uncertainty curve and a preliminary risk grade within a variable prediction window, and calculating a confidence weight according to the uncertainty curve to bias correct the preliminary risk grade and generate a final risk grade;

[0062] The core data output by the power supply management module includes: a power-off type identification, i.e., deep discharge, normal discharge and charging operation, an energy supply reliability parameter, i.e., a mutual inductance power supply residual ratio and a battery state of charge percentage, and actual supportable sampling power consumption upper limit, cable aging grade coefficient and cable electrical parameter flow containing real-time load current;

[0063] Further, the multi-source sensing data of each cable monitoring device node is synchronously acquired and preprocessed to form a health initial vector for analysis; since each cable monitoring device node is equipped with multiple sensors such as the aforementioned partial discharge, temperature, current, humidity and the like and regularly uploads data through a 4G module, first, the data of different sources are synchronized according to time stamps, and the sensor readings are aligned to ensure that the parameters collected at the same monitoring time point jointly constitute the health state vector of the cable monitoring device node; for the cable monitoring device nodes distributed in multiple places, the data of each cable monitoring device node is processed in parallel to obtain a respective health initial vector set;

[0064] Based on the above core data and health initial vector set, a graph time sequence fusion prediction is performed to generate a preliminary risk grade, including: taking the cable monitoring device node as a graph node, and the edge representing the connection relationship or mutual influence relationship between the cable monitoring device nodes; usually, the high-voltage cable line is connected in sections, and each two adjacent joints are connected through a cable conductor, and there is coupling in electrical and thermal behavior;

[0065] The graph time sequence fusion prediction algorithm takes the core data and health initial vector set of each cable monitoring device node as the prediction input: the algorithm can be understood as a fusion model combining a time sequence model such as a recurrent neural network (RNN), LSTM, etc. to capture the time dependence of a single node and a graph neural network (GNN) or a spatial weight matrix to capture the spatial correlation between nodes; in implementation, a prediction model can be established for each cable monitoring device node, and the information of neighboring nodes can be introduced through graph convolution, etc. to realize the propagation of information on the topology; in order to flexibly adapt to different scene requirements, the prediction algorithm adopts a variable prediction window, that is, the time span of prediction is dynamically adjusted according to the load fluctuation and the required prediction accuracy; for example, when the load and state are stable, a longer prediction window such as 24 hours in the future can be selected to plan long-term trends; when the state changes dramatically or uncertainty is large, a shorter window such as 2 hours in the future can be used to improve prediction accuracy; through the above model operation, the residual energy curve of each cable monitoring device node within the prediction window is output, representing the possible energy change of the monitoring device's own power supply in the future period of time, which is obtained from the current energy state provided by the power supply management module, that is, the residual ratio of mutual inductance power supply and the battery state of charge percentage, the uncertainty curve reflects the prediction error range, such as the standard deviation of each time step and the preliminary risk level including 0-5 levels; specifically, the graph time sequence fusion prediction algorithm fuses LSTM and GCN: the input layer receives a 6-dimensional health initial vector; 2 layers of LSTM (32 neurons per layer) process the cable multi-source sensing data of each cable monitoring device node for 24 hours, and output the energy trend of a single node; 1 layer of GCN (3x3 convolution kernel) fuses the features of adjacent nodes with the cable topology as the adjacency matrix; the fusion layer fuses the space-time features through an attention mechanism; the output layer outputs the energy curve for 48 hours in the future, the preliminary risk level of 0-5 levels, and the uncertainty curve;

[0066] Further, the preliminary risk level is corrected based on the uncertainty curve to obtain the final risk level, including:

[0067] According to the standard deviation σ(t) of each time step in the uncertainty curve, the confidence weight is calculated as The weighted average formula is used to correct the preliminary risk level, that is, the result of subtracting the confidence weight from 1 is multiplied by the reference low risk value, and the product of the confidence weight and the preliminary risk level is added to the multiplication result, and the sum is the final risk level; wherein the reference low risk value is generally a lower risk level such as level 0, which is used to make the risk assessment more biased towards the safe side when the prediction is uncertain;

[0068] The final risk level after correction is smoothed to avoid unfounded level jumps, and the continuous high-risk section is marked as the basis for subsequent sampling level determination;

[0069] For each cable monitoring device node, the upper limit of the sampling power consumption budget is dynamically calculated according to the corresponding power-off type identifier, to ensure that the power consumption is adapted to the power supply state, including:

[0070] When it is detected that the power-off type identifier of a certain cable monitoring device node is deep discharge, that is, the cable monitoring device node mainly relies on battery discharge to maintain, and the power supply is tight; and the deep discharge power consumption compression coefficient is dynamically adjusted based on the corresponding cable aging grade coefficient and the corresponding aging risk threshold value; the aging risk threshold value includes a low aging risk threshold value and a high aging threshold value; the setting is based on the comprehensive consideration of the cable aging law, historical failure data and operating environment. In actual operation, cable aging is affected by many factors, such as operating life, load condition, environmental temperature and humidity, etc.; by collecting and analyzing a large amount of historical data of cables, the correlation between different aging degrees and failure probability can be found out; for cables with longer operating life and more historical failure times, the aging risk is relatively high;

[0071] When the cable aging grade coefficient is less than or equal to the low aging risk threshold value, it is marked as a low aging label, and since the cable state is relatively healthy, the power consumption can be moderately compressed, and the corresponding first compression ratio can be 50%;

[0072] When the cable aging grade coefficient is greater than the low aging risk threshold value and less than the high aging threshold value, it is marked as a medium aging label, and the corresponding second compression ratio can be 40%;

[0073] When the cable aging grade coefficient is greater than or equal to the high aging threshold value, it is marked as a high aging label, and the corresponding third compression ratio can be 30%; the first compression ratio, the second compression ratio and the third compression ratio are system-pre-set deep discharge power consumption ratio factors and satisfy the first compression ratio > the second compression ratio > the third compression ratio;

[0074] And the maximum sampling power consumption allowed in the deep discharge mode of the current node is obtained by multiplying the deep discharge power consumption compression coefficient obtained above and the actual supportable sampling power consumption upper limit, that is, the sampling power consumption budget upper limit;

[0075] When it is detected that the power-off type identifier of a certain cable monitoring device node is normal discharge, it is marked as a normal aging label, and the power supply system is in a steady state, and the aging load sampling power consumption upper limit output by the power supply management module is directly used, that is, the sampling power consumption budget upper limit is equal to the actual supportable sampling power consumption upper limit;

[0076] Further, in combination with the final risk level and aging label, a recommended sampling level is assigned according to a preset risk level threshold, wherein the preset risk level threshold includes a low risk threshold, a medium risk threshold, and a high risk threshold, such as 0-5 level risks, the low risk threshold = 1, the medium risk threshold = 2, 3, the high risk threshold = 4, 5, the values of which are obtained according to historical failure data and the failure probability under different risk levels; including:

[0077] When the preliminary risk level is greater than or equal to the high risk level threshold or the aging label is a high aging label, a high-priority monitoring is triggered, that is, the core parameters such as partial discharge and bus current high-frequency sampling such as 1s are preferentially guaranteed according to the upper limit of the sampling power consumption budget;

[0078] When the preliminary risk level is greater than or equal to the medium risk threshold and less than the high risk level threshold or the aging label is a medium aging label, a medium-priority monitoring is triggered, that is, the core and non-core parameter sampling is balanced according to the upper limit of the sampling power consumption budget, such as 5s for core parameters and 10s for non-core parameters;

[0079] When the preliminary risk level is less than the low risk threshold and the aging label is a low aging label, a low-priority monitoring is triggered, that is, the non-core parameter sampling frequency is reduced according to the upper limit of the sampling power consumption budget, such as 10s for core parameters and 30s for non-core parameters;

[0080] When the aging label is a normal aging label, a normal power supply scene is triggered to adapt, and the sampling frequency is directly determined according to the preliminary risk level, such as high-priority monitoring for high risk level and medium-priority monitoring for medium risk level, and the sampling frequency is referred to a preset strategy, such as 5s for core parameters.

[0081] The adaptive sampling communication module utilizes the final risk level and the recommended sampling level provided by the health comprehensive analysis module to perform dynamic adjustment of the data acquisition and communication strategy for each cable monitoring device node, so as to realize the unification of fine monitoring and energy efficient utilization;

[0082] The core data output by the health comprehensive analysis module is received in real time as the basis for strategy adjustment, including decision instructions, that is, the recommended sampling level and the upper limit of the sampling power consumption budget; state parameters include the final risk level, the aging label, the residual available energy curve, and the cable electrical parameter flow;

[0083] The recommended sampling level output by the health comprehensive analysis module is directly mapped to the preset hierarchical sampling strategy; when the recommended sampling level is high-priority monitoring, it is mapped to a first-level strategy, the sampling frequency is the highest, and the collected parameters are the most comprehensive;

[0084] When the recommended sampling level is medium-priority monitoring, it is mapped to a second-level strategy, the sampling frequency is medium, and the collected parameters are core plus part of non-core;

[0085] When the recommended sampling level is low, the low-priority monitoring is mapped to a three-level strategy, with the lowest sampling frequency, and only key indicators are collected;

[0086] When the recommended sampling level is normal, the normal power supply scenario is mapped to a normal strategy, and the sampling frequency is based on the preset reference;

[0087] In addition, the module introduces new discriminators such as risk change trend in strategy judgment: for example, when it is detected that the risk level has not been upgraded but is showing a rapid upward trend, it can be upgraded to a higher level of monitoring strategy in advance to avoid lag response; Similarly, if the key indicators such as partial discharge are continuously abnormal but the comprehensive risk assessment has not yet reached the threshold, the module can temporarily increase the sampling density of the corresponding parameters to ensure that potential hidden dangers are not ignored. Through the above fine-grained rule extension, the differentiated sampling strategy matching for different risk levels and aging degree nodes is realized, and the coverage accuracy of adaptive sampling for diversified cable aging conditions is improved;

[0088] According to the determined hierarchical sampling strategy and sampling power consumption budget, the module issues collection and communication scheduling instructions in real time to coordinate and control each sensing unit and communication unit, dynamically executes the optimal sampling communication scheme, and realizes the unification of fine monitoring and energy-efficient utilization. Under high-priority monitoring, the module drives the node into a first-level strategy mode: as much as possible to collect core parameters such as partial discharge and current at the highest frequency, while appropriately increasing the collection frequency of auxiliary parameters such as temperature and humidity; and shortens the data upload period, and if necessary, uses event-triggered real-time reporting of key data to ensure that the background can obtain detailed front-line status in a timely manner;

[0089] Under medium-priority monitoring, the module drives the node into a second-level strategy mode to balance the sampling task: continues to ensure a high sampling frequency of core parameters, while periodically collecting some non-core parameters, such as collecting core parameters every 5 seconds and general parameters every 10 seconds, and using batch upload or data compression upload to reduce communication energy consumption; Under low-priority monitoring, the module executes into a third-level strategy mode with energy saving as the guide: significantly reduces most sampling activities except key indicators, such as relaxing the core indicator sampling period to 10 seconds or more, and extending the non-core indicator sampling interval to 30 seconds or even collecting on demand, and periodically reporting data after storing locally in the buffer, thereby reducing the working duty cycle of the wireless communication module to the minimum. Throughout the process, the module strictly constrains each sampling and transmission operation according to the upper limit of the sampling power consumption budget provided by the power supply management module, ensuring that the actual executed sampling frequency and communication frequency are within the current energy supply capability range; When monitoring that the power supply state of a cable monitoring device node is switched, such as from normal power supply to battery discharge or the health status evaluation result is updated, the corresponding sampling communication strategy is immediately switched according to the new recommended sampling level, ensuring the real-time and continuity of strategy adjustment;

[0090] For the case of limited sampling energy or communication resources, the module enhances the flexibility of power consumption strategy allocation, and intelligently optimizes the allocation of sampling resources to adapt to the actual deployment conditions. First, according to the current energy supply margin provided by the power supply management module, such as the residual ratio of mutual inductor power supply and the battery state of charge percentage, the power consumption demand of the current sampling scheme is evaluated to assess the pressure of the sampling task on the power supply. When it is detected that the node is in a deep discharge state and only relies on battery power supply and the energy supply is tight, the module will compress the sampling power consumption overhead under the premise of maintaining basic monitoring: combined with the aging level of the node cable monitoring device node, the power consumption allocation ratio is dynamically adjusted, for example, for a low-aging node in a healthy state, the current actual supportable sampling power consumption upper limit can be allowed to be used for sampling, while for a high-aging risk node, the sampling power consumption is strictly controlled at about 30% of the supportable upper limit, to prolong the device endurance; Finally, the corresponding proportion is calculated to obtain the sampling power consumption budget upper limit in the deep discharge mode, and according to this, part of the collection of non-critical parameters is reduced or suspended; secondly, the module realizes the on-demand allocation of sampling resources through task scheduling optimization: for example, in the case of simultaneous high-frequency monitoring of multiple cable monitoring device nodes, the sampling and sending timing of each node is staggered to prevent communication conflicts or instantaneous power consumption from affecting the overall monitoring effect; During the entire adaptation process, the module continuously balances the monitoring accuracy and energy consumption constraints, and uses intelligent strategies to ensure that key health status parameters can be effectively monitored with the lowest necessary power consumption under harsh conditions such as low battery power and insufficient electromagnetic energy, which reflects high environmental adaptability and deployment practicality;

[0091] Further, based on the life analysis and operation linkage of the remaining available energy curve and the aging level coefficient, including: the module executes the sampling communication strategy while using the remaining available energy curve and the cable aging level coefficient output by the health comprehensive analysis module for life decay evaluation; wherein, the remaining available energy curve is used to reflect the energy change that the cable monitoring device node itself power supply may remain in the future period of time, such as the energy change trend in the next 6 months, and the cable aging level coefficient reflects the aging degree of the cable body, such as a quantitative value in the range of 0-1; Specifically, when the remaining available energy curve shows that the energy will drop below the energy warning threshold for maintaining basic monitoring in the future period of time, such as in the next 3 months, and the cable aging level coefficient is still greater than or equal to the high aging threshold, it is determined that the cable monitoring device node is at high risk of life decay; If only the remaining available energy drops below the energy warning threshold, and the cable aging level coefficient is still less than or equal to the low aging threshold, i.e. the cable is in a healthy state, it is determined as energy warning; If only the aging level coefficient is higher than the energy warning threshold, and the cable aging level coefficient is still greater than or equal to the high aging threshold, it is determined as aging warning; wherein the energy warning threshold is configured by statistical historical monitoring device power interruption data and cable fault data;

[0092] According to the above determination result, the corresponding operation and maintenance work order is generated: for the high service life attenuation risk node, the work order contains joint maintenance suggestions such as insulation test, joint replacement and sampling strategy temporary adjustment instructions such as upgrading to the first level strategy intensive monitoring, and synchronously increasing power supply guarantee priority; for the power supply warning node, the work order suggests replacing the monitoring device battery or optimizing the mutual inductance energy taking structure, and temporarily reducing the non-core parameter sampling frequency to prolong the endurance; for the aging warning node, the work order suggests strengthening the sampling accuracy of aging related parameters such as partial discharge and temperature, such as increasing the partial discharge pulse collection time, without adjusting the power supply strategy;

[0093] The module cyclically executes the above sampling communication strategy adjustment process in a preset period or event triggered manner, forming an online closed loop optimization mechanism; that is, the new power supply state and health state information are monitored in real time, the latest data is continuously fed back to the decision link, and the final risk level and recommended sampling level are updated in cooperation with the power supply management module and the health comprehensive analysis module, thereby driving the adaptive adjustment of the next round of sampling strategy;

[0094] In the embodiment, the power supply management, health comprehensive analysis and adaptive sampling communication modules are included; the power supply management module collects the electric parameter flow, generates the instantaneous power density sequence, detects the charging and discharging state in combination with the historical and real-time data, and outputs parameters such as power-off type and power supply reliability; the health comprehensive analysis module dynamically calculates the sampling power consumption budget, predicts the energy curve and risk level through graph time sequence fusion, and determines the recommended sampling level after correction; the adaptive sampling communication module maps the hierarchical strategy and dynamically executes the sampling and communication scheduling; the system solves the adaptive sampling strategy matching problem in the multi-node high-voltage distribution network monitoring, realizes the precise monitoring and strategy adaptation under the differentiation of aging, and improves the efficiency and reliability of cable health management

[0095] As shown in Figure 3 The cable full life cycle health state prediction analysis method includes the following steps: step one: collecting the bus current, bus voltage, standby battery current and battery terminal voltage of the cable monitoring device node to generate the cable electric parameter flow; calculating the first derivative of the bus instantaneous power to generate the instantaneous power density sequence; determining the power-off type identifier as deep discharge, normal discharge or charging operation state based on the instantaneous power density sequence and the preset threshold;

[0096] Step two: obtaining the cable service life and historical failure data, calculating the ratio of the service life to the preset life reference value to obtain the basic aging proportion; setting an offset based on the historical failure times, adding the basic aging proportion and the offset to generate the cable aging level coefficient;

[0097] Step three: the relationship graph is constructed with the cable monitoring device nodes as graph nodes and the cable connection topology as edges; the graph time sequence fusion prediction algorithm is used to process the cable state data, and the residual available energy curve, the uncertainty curve and the preliminary risk level are output; the confidence weight is calculated based on the uncertainty curve, and the preliminary risk level is biased and corrected to generate the final risk level;

[0098] Step four: when the power-off type identifier is deep discharge, the upper limit of the sampling power consumption budget is compressed according to the cable aging grade coefficient; and combined with the final risk level and the aging grade coefficient, the recommended sampling level is allocated, the recommended sampling level is mapped to the hierarchical sampling strategy, and the sampling and communication operations of each cable monitoring device node are dynamically adjusted based on the hierarchical sampling strategy according to the upper limit of the sampling power consumption budget.

[0099] The above formulas are dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0100] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0101] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application of the technical solution and the constraints of the invention. A person skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0102] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically, or two or more modules can be integrated in one module.

[0103] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0104] Finally, the above is only the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A cable full life cycle health status prediction and analysis system, characterized in that, Includes the following modules: The power supply management module is used to collect cable electrical parameter flow and aging basic data of the cable monitoring device nodes, and to generate an instantaneous power density sequence by performing derivative calculations on the cable electrical parameter flow. By comparing the changing trend of the instantaneous power density sequence and the bus voltage and current relationship in the cable electrical parameter flow, the module determines the charging and discharging status of the cable monitoring device nodes and outputs a power outage type identifier. It combines historical data fitting and real-time values ​​to calculate the mutual inductance power remaining ratio and battery state of charge percentage as power supply reliability parameters, and implements the output cable aging level coefficient and the upper limit of actual supportable sampling power consumption. The health comprehensive analysis module is used to dynamically calculate the upper limit of the sampling power consumption budget based on the power outage type identifier, cable aging level coefficient, and power supply reliability parameters. It also performs time synchronization based on cable status data to generate a health initial vector. Based on the health initial vector and the node topology of the cable monitoring device as a graph structure, it predicts and outputs the remaining available energy curve, uncertainty curve, and preliminary risk level through graph time series fusion. Based on the uncertainty curve correction, the preliminary risk level is confidence-corrected to obtain the final risk level, and the upper limit of the sampling power consumption budget and the recommended sampling level are determined. The adaptive sampling communication module receives the final risk level and recommended sampling level, maps them to a hierarchical sampling strategy, and dynamically executes the sampling and communication strategies of each cable monitoring device node.

2. The cable lifecycle health status prediction and analysis system according to claim 1, characterized in that, The power supply management module performs the following steps: It collects bus current, bus voltage, backup battery current, and battery terminal voltage at a fixed sampling frequency to generate cable electrical parameter flow; it calculates the instantaneous power of the bus by multiplying the bus voltage and current corresponding to each sampling point; it performs a first derivative operation on the instantaneous power of the bus to generate an instantaneous power density sequence; and based on the instantaneous power density sequence and the cable electrical parameter flow and their corresponding preset thresholds, it determines the power outage type as deep discharge, normal discharge, or charging operation.

3. The cable lifecycle health status prediction and analysis system according to claim 2, characterized in that, The criteria for determining the power outage type include: detecting the charging and discharging status of the cable monitoring device nodes based on the obtained instantaneous power density sequence, cable electrical parameters and their corresponding preset battery continuous discharge threshold, preset battery static current threshold, preset upper limit threshold for positive power density fluctuation, and preset lower limit threshold for negative power density fluctuation, including: When the instantaneous power density sequence of the monitoring device node corresponding to N consecutive sampling points is less than the preset negative power density fluctuation lower limit threshold and the backup battery discharge current is less than the preset battery continuous discharge threshold, it is determined to be in a deep discharge power failure state and a deep discharge power failure flag is output. When the instantaneous power density sequence is between the lower threshold of negative power density fluctuation and the upper threshold of preset positive power density fluctuation, and the absolute value of the backup battery discharge current is less than or equal to the preset battery static current threshold, the corresponding monitoring device node is in normal discharge and outputs a normal discharge power failure flag. When the instantaneous power density sequence is greater than the preset upper limit threshold for positive power density fluctuation and the backup battery discharge current is greater than the preset battery continuous discharge threshold, the corresponding cable monitoring device node is significantly charging and outputs a charging operation indicator.

4. The cable lifecycle health status prediction and analysis system according to claim 1, characterized in that, The power supply reliability parameters, cable aging level coefficient, and actual supportable sampling power consumption upper limit are generated as follows: The instantaneous power density sequence is integrated to obtain the energy integral value over a certain period; this energy integral value is compared with the rated mutual inductance power supply capacity to calculate the mutual inductance power supply remaining ratio; the battery state-of-charge percentage is derived based on the ampere-hour integration method, combining the backup battery current integration with the battery's rated capacity; according to the following formula: In the formula, This represents the state of charge of the battery at time t. It is the initial moment The state of charge of the battery. The rated capacity of a battery is represented by an ampere-hour (AH). It is the battery current at time t, in amperes, during charging. It is a positive value during discharge and a negative value during discharge. Indicates from the initial time The integral of the battery current up to the current time t, that is, the change in the amount of charge in the battery during this time. Based on the pre-stored and provided cable service life and historical fault data within the system, the ratio of the cable service life to the preset life reference value is calculated to obtain the basic aging ratio; The offset is set based on the number of historical faults, and the basic aging ratio is added to the offset to generate a quantitative value, which is recorded as the cable aging level coefficient. The upper limit of the actual sampling power consumption that can be supported is calculated by multiplying the product of the mutual inductance power remaining ratio, the battery state of charge percentage, and the cable aging level coefficient by the preset power consumption of the device.

5. The cable lifecycle health status prediction and analysis system according to claim 1, characterized in that, The health comprehensive analysis module dynamically calculates the upper limit of the sampling power consumption budget, including: under the deep discharge power failure flag, when the cable aging level coefficient is less than or equal to the low aging risk threshold, it is marked as a low aging tag and its corresponding first compression ratio is configured; When the cable aging level coefficient is greater than the low aging risk threshold but less than the high aging threshold, it is marked as a medium aging label and its corresponding second compression ratio is configured. When the cable aging level coefficient is greater than or equal to the high aging threshold, it is marked as a high aging label and its corresponding third compression ratio is configured; and the first compression ratio is greater than the second compression ratio is greater than the third compression ratio; and based on the above classification by aging label, the different configured compression ratios are multiplied by the actual supportable upper limit of sampling power consumption to obtain the maximum allowable sampling power consumption, i.e., the upper limit of sampling power consumption budget, under the current node deep discharge mode.

6. The cable full life cycle health status prediction and analysis system according to claim 1, characterized in that, The comprehensive health analysis module employs a graph-time series fusion prediction algorithm, which includes: a fusion architecture of graph convolutional networks and long short-term memory networks. The graph convolutional network processes the relationship graph constructed with cable monitoring device nodes as nodes and cable connection topology as edges; the long short-term memory network processes the multi-source sensor data of each cable monitoring device node; the recurrent neural network and graph neural network are fused to output the remaining available energy curve and preliminary risk level; the confidence weight of each time step is calculated based on the uncertainty curve; and a weighted average formula is used to bias-correct the preliminary risk level to generate the final risk level. The remaining available energy curve reflects the potential changes in the remaining energy of the cable monitoring device node's own power supply over a future period.

7. The cable full life cycle health status prediction and analysis system according to claim 1, characterized in that, The rules for determining the recommended sampling level in the comprehensive health analysis module include: when the initial risk level is greater than or equal to the high risk level threshold or the aging label is a high aging label, high-priority monitoring is triggered, that is, the core parameters should be sampled at high frequency according to the upper limit of the sampling power consumption budget. When the initial risk level is greater than or equal to the medium risk threshold and less than the high risk level threshold, or when the aging label is a medium aging label, medium priority monitoring is triggered, i.e., the sampling of core and non-core parameters is balanced according to the upper limit of the sampling power consumption budget. When the initial risk level is less than the low risk threshold and the aging label is a low aging label, low priority monitoring is triggered, that is, the sampling frequency of non-core parameters is reduced according to the upper limit of the sampling power consumption budget. When the aging label is a normal aging label, the normal power supply scenario adaptation is triggered and the initial risk level is determined directly, and the sampling frequency is still the fixed sampling frequency.

8. The cable lifecycle health status prediction and analysis system according to claim 6, characterized in that, It also includes a lifespan degradation assessment function: based on the remaining available energy curve and the cable aging level coefficient, it generates a judgment result for high lifespan degradation risk, power supply warning or aging warning; When the remaining available energy curve shows that the energy will drop below the energy warning threshold for maintaining basic monitoring in the future, and the cable aging level coefficient is still greater than or equal to the high aging threshold, the cable monitoring device node is determined to be at high risk of life degradation. If the remaining available energy drops below the energy warning threshold and the cable aging level coefficient is still less than or equal to the low aging threshold, then it is determined to be an energy supply warning. If only the aging level coefficient is higher than the energy warning threshold, and the cable aging level coefficient is still greater than or equal to the high aging threshold, then it is judged as an aging warning. For cable monitoring device nodes with high lifespan degradation risk, maintenance work orders that include joint inspection suggestions and sampling strategy upgrades are automatically triggered.

9. The cable full life cycle health status prediction and analysis system according to claim 1, characterized in that, When the adaptive sampling communication module executes the sampling strategy of each cable monitoring device node, it includes: when the recommended sampling level of the cable monitoring device node is high-priority monitoring, it enters the first-level strategy mode and uses an event-triggered mechanism to upload key data in real time; when the recommended sampling level of the cable monitoring device node is medium-priority monitoring, it enters the second-level strategy mode and uses data compression and batch upload strategies to reduce communication energy consumption. When the recommended sampling level for the cable monitoring device node is low-priority monitoring, it enters a three-level strategy mode oriented towards energy saving.

10. A method for predicting and analyzing the health status of cables throughout their entire life cycle, used to implement the cable life cycle health status prediction and analysis system according to any one of claims 1-9, characterized in that, Includes the following steps: Step 1: Collect bus current, bus voltage, backup battery current, and battery terminal voltage from the cable monitoring device nodes to generate cable electrical parameter flow; calculate the first derivative of the instantaneous power of the bus to generate an instantaneous power density sequence; based on the instantaneous power density sequence and a preset threshold, determine the power outage type as deep discharge, normal discharge, or charging operation. Step 2: Obtain cable service life and historical fault data, and calculate the ratio of service life to preset life reference value to obtain the basic aging ratio; The offset is set based on the number of historical failures, and the basic aging ratio is added to the offset to generate the cable aging level coefficient. Step 3: Construct a relationship graph using the cable monitoring device nodes as graph nodes and the cable connection topology as edges; A graph-time series fusion prediction algorithm is used to process cable status data and output the remaining available energy curve, uncertainty curve, and preliminary risk level. Confidence weights are calculated based on the uncertainty curve, and bias corrections are applied to the preliminary risk level to generate the final risk level. Step 4: When the power outage type is identified as deep discharge, compress the upper limit of the sampling power consumption budget according to the cable aging level coefficient. Combined with the final risk level and aging level coefficient, a recommended sampling level is assigned, and the recommended sampling level is mapped to a hierarchical sampling strategy. Based on the hierarchical sampling strategy, the sampling and communication operations of each cable monitoring device node are dynamically adjusted according to the upper limit of the sampling power consumption budget.

Images