A compensation branch health identification and life evaluation method of an intelligent metal comprehensive distribution box

Abstract

The present application belongs to the technical field of low-voltage reactive power compensation device state monitoring, and discloses a compensation branch health identification and life evaluation method of an intelligent metal comprehensive distribution box. The present application aims at the failure problem caused by the coupling of the wear of the composite switch action, the attenuation of the compensation capacity and the harmonic stress in the compensation branch, collects the response residual, action delay, temperature rise, harmonic stress and action frequency of each compensation branch, constructs a light gradient boosting classification model to identify the health grade and calculate the health score, combines the health score, action life consumption and temperature rise stress damage to estimate the remaining life, and finally performs risk sorting and outputs maintenance suggestions such as reexamination, replacement or reduced capacity operation. The present application can comprehensively compensate the multi-dimensional characteristics such as response, switch action, thermal stress and harmonic stress, quantifies the health degree and remaining life before the branch completely fails, and effectively improves the accuracy and interpretability of the predictive maintenance of the compensation branch of the distribution box.

Description

Technical Field

[0001] This invention belongs to the field of condition monitoring and life management technology of low-voltage complete reactive power compensation devices, and particularly relates to a method for health identification and life assessment of compensation branches of an intelligent metal integrated distribution box. Background Technology

[0002] JP intelligent metal integrated distribution boxes typically include components such as disconnect switches, circuit breakers, intelligent harmonic suppression reactive power compensation modules, composite switches, distribution monitoring and compensation controllers, and copper busbars. During long-term operation, the compensation branches are affected by multiple factors, including the cumulative number of operations, harmonic current surges, ambient temperature changes, capacitor parameter decay, and wear of composite switch contacts. These factors can lead to decreased compensation capacity, increased operating delays, increased localized temperature rise, and shortened lifespan.

[0003] In existing projects, the identification of the health status of compensation branches mostly relies on manual inspections, single temperature rise judgments, or replacement only after the branch has obviously failed. This type of method has at least the following shortcomings:

[0004] First, judging solely by temperature levels makes it difficult to distinguish between normal temperature rise caused by short-term high loads and abnormal deterioration caused by the aging of the branch circuit itself.

[0005] Second, even if a decline in the compensation effect is found, it is often only possible to determine that the branch is "possibly abnormal", and it is difficult to quantify its health status and remaining lifespan.

[0006] Third, the compensation branch inside the JP distribution box is composed of composite switches, capacitor modules, connecting conductors and control logic. The fault mechanism is coupled. If the compensation response residual, action delay, temperature rise and harmonic stress are not considered at the same time, the health assessment results will be unstable and difficult to directly guide operation and maintenance decisions.

[0007] Therefore, there is an urgent need to provide a method that can combine compensation response characteristics, switching action characteristics, temperature rise characteristics, and harmonic stress to identify the health status of compensation branches online and further assess their remaining lifespan. Summary of the Invention

[0008] To overcome the shortcomings of existing methods for identifying the condition of compensation branches, which rely on a single temperature index, struggle to detect branch aging caused by the coupling of wear and tear of composite switch actions and compensation capacity decay, and fail to provide a remaining lifespan assessment, this invention proposes a method for health identification and lifespan assessment of compensation branches in intelligent metal integrated distribution boxes. This method jointly analyzes compensation response residuals, action delay increments, critical temperature rise, harmonic stress quantification values, and cumulative action counts. First, a lightweight gradient boosting classification model is used to identify the health level of the compensation branch. Then, based on the health score, action lifespan consumption, and temperature rise stress damage, the remaining lifespan is estimated, and risk ranking and maintenance recommendations are output.

[0009] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0010] A method for health identification and life assessment of compensation branches in an intelligent metal integrated distribution box, applied to an intelligent metal integrated distribution box with current, temperature, harmonic acquisition, and branch action recording functions, includes the following steps: S1: Collect the operating status information of each compensation branch, including at least the commanded compensation capacity, actual response compensation capacity, action duration, temperature of key measuring points, total harmonic distortion rate, cumulative number of actions, ambient temperature and rated capacity; S2: Based on the collected information, construct a branch health feature vector containing five dimensions: compensation response residual, action delay increment, branch temperature rise, harmonic stress value, and action consumption ratio. S3: Input the health feature vector into a lightweight gradient boosting classification model. The model outputs the probability that the branch belongs to one of four levels: normal, mildly deteriorated, moderately deteriorated, or severely deteriorated. A continuous health score is then calculated based on these probabilities and weighted accordingly. S4: Combining the health score, the ratio of action frequency consumption, and a temperature rise stress coefficient that reflects the degree of recent temperature rise exceeding the limit, the remaining life of the compensation branch is estimated according to the preset life damage formula. S5: Based on the health score, classify the health level and calculate the risk index in combination with the remaining life expectancy, and rank all compensation branches by risk. S6: Based on the health level and remaining life assessment results, output maintenance recommendations including re-inspection, replacement, de-rated operation or decommissioning.

[0011] Furthermore, the compensation response residual is obtained by dividing the absolute value of the difference between the commanded compensation capacity and the actual response compensation capacity by the rated compensation capacity of the compensation branch, and is used to quantify the degree of attenuation of the branch compensation capacity.

[0012] Furthermore, the action delay increment is taken as the larger value between "the difference between the current action duration and the reference action duration" and "0". When the action delay increment increases significantly, it is considered that the composite switch or related branch has an aging trend.

[0013] Furthermore, the branch temperature rise is obtained by subtracting the current ambient temperature from the temperature of the key measuring point of the branch, which is used to quantify the degree of heat generation during the operation of the branch.

[0014] Furthermore, the harmonic stress value comprehensively considers the square of the current total harmonic distortion rate, the ratio of the rated capacity of the compensation branch to a reference capacity, and a sensitivity coefficient that is dynamically adjusted according to the degree to which the compensation response residual exceeds the reference value, in order to quantify the additional stress of the harmonic environment on the branch.

[0015] Furthermore, the reference base capacity is set to 40 kvar, which is used to achieve normalized comparison between branches with different capacities.

[0016] Furthermore, the health score is obtained by multiplying the probabilities of the four levels output by the gradient boosting classification model by the central values ​​of the levels corresponding to normal, mild deterioration, moderate deterioration, and severe deterioration, and then summing them up. The higher the score, the worse the health status.

[0017] Furthermore, the temperature rise stress coefficient is obtained by averaging the ratio of the portion of the measured temperature rise exceeding the rated temperature rise benchmark in each sampling to the benchmark within a statistical window. If the temperature rise does not exceed the benchmark, it is taken as zero. This is used to quantify the cumulative damage caused to the branch by long-term over-temperature operation.

[0018] Furthermore, the rules for classifying health levels are based on health scores:

[0019] When the health score is below the first threshold, it is considered normal;

[0020] When the health score is between the first threshold and the second threshold, it is considered to be slightly deteriorated;

[0021] When the health score is between the second and third thresholds, it is considered to be moderately deteriorated;

[0022] When the health score is greater than or equal to the third threshold, it is judged as severely deteriorated.

[0023] Furthermore, the maintenance recommendations are as follows: for slightly deteriorated branches, it is recommended to re-inspect and shorten the inspection cycle; for moderately deteriorated branches, it is recommended to arrange power outage windows for re-inspection or reduce the operating rate; for severely deteriorated branches or branches with remaining life below the preset threshold, it is recommended to replace or decommission them as soon as possible and list them as priority maintenance targets.

[0024] The present invention provides a method for health identification and lifespan assessment of compensation branches in intelligent metal integrated distribution boxes, which has the following advantages: The method does not merely compare single-point thresholds for temperature rise, but rather addresses the failure mechanisms of composite switch wear, compensation capacity decay, and harmonic stress coupling in the compensation branches of JP intelligent metal integrated distribution boxes. It combines compensation response residuals, action delay increments, critical temperature rises, harmonic stress quantification values, and action lifespan consumption ratios for branch health identification, and estimates the remaining lifespan using a clear lifespan damage formula. Compared to methods relying solely on a single temperature rise baseline or solely on empirical inspections, the present invention can detect compensation branch degradation earlier and provide interpretable health levels, risk rankings, and remaining lifespan results. Attached Figure Description

[0025] Figure 1 Flowchart for health identification and life assessment of compensation branch circuits in intelligent metal integrated distribution boxes.

[0026] Figure 2 The graph shows the variation of the compensation response residual and action delay of the aging branch.

[0027] Figure 3 A comparison chart of health scores and degradation thresholds for aging branches.

[0028] Figure 4 A comparison chart showing the accuracy of degradation level identification.

[0029] Figure 5 This is a comparison chart of remaining lifetime error.

[0030] Figure 6 This is a comparison chart of the remaining life assessment for aging branches. Detailed Implementation

[0031] To better understand the purpose, structure, and function of this invention, the following description, in conjunction with the accompanying drawings, provides a more detailed account of a method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box.

[0032] This invention discloses a method for health identification and lifespan assessment of compensating branches in an intelligent metal integrated distribution box. The method is applied to an intelligent metal integrated distribution box, which at least includes a current acquisition unit, a temperature acquisition unit, a harmonic acquisition unit, a branch action recording unit, a distribution monitoring and compensation control unit, a storage unit, and a maintenance suggestion output unit. Figure 1 As shown, the method includes the following steps:

[0033] S1. Collect the operating status information of each compensation branch;

[0034] The operational status information includes at least the branch command compensation capacity. Actual response compensation capacity of branch lines Action time from receiving the switching command to the branch state stabilizing Temperature at key measuring points of branch circuits Total harmonic current distortion rate Cumulative number of actions Ambient temperature and branch line rated compensation capacity ; of which the actual response compensation capacity It is obtained by converting branch current or branch reactive power. Indicates the first One sampling point.

[0035] S2, Characteristics of healthy branch structures;

[0036] Based on the compensation response residual Action delay increment Branch circuit temperature rise Harmonic stress quantification value The ratio of the number of actions consumed Construct the health feature vector of the compensation branch ,in, This refers to the rated operating life.

[0037] Preferably, the compensated response residual can be constructed as follows: .

[0038] in, For the first Each compensation branch is at any time Instruction compensation capacity, For its corresponding actual response compensation capacity, For the first The rated compensation capacity of each compensation branch.

[0039] Preferably, the action delay increment can be calculated according to... Obtain, among which For the first The baseline operating time of each branch under the calibrated operating conditions; when the increase in operating delay is significant, it is considered that the composite switch or related branch is aging.

[0040] Preferably, the branch temperature rise term can be determined by the following formula: .

[0041] in, For the first Each compensation branch corresponds to the temperature of key measuring points. This represents the current ambient temperature.

[0042] Preferably, the harmonic stress quantization value It can be calculated as follows:

[0043] .

[0044] in, For reference rated compensation capacity used in normalized comparisons between different compensation branches, the preferred value is... ; This represents the current total harmonic current distortion rate. The harmonic stress sensitivity coefficient, For the first The baseline response residuals of each branch under the calibrated operating conditions.

[0045] S3. Estimate the health level of the compensation branch;

[0046] Input the health feature vector obtained in step S2 into the lightweight gradient boosting classification model to obtain the probability that the corresponding compensation branch belongs to one of four levels: normal, mildly deteriorated, moderately deteriorated, and severely deteriorated. and , and according to Calculate the health score; the lightweight gradient boosting classification model consists of 60 to 100 CART classification trees, each with a maximum depth of 2 to 4, a learning rate of 0.02 to 0.10, a minimum number of leaf node samples of 3 to 10, and a model storage footprint of no more than 256kB.

[0047] Preferably, the lightweight health assessment model is a lightweight gradient boosting classification model, consisting of 60 to 100 CART classification trees, with a maximum depth of 2 to 4, a minimum number of leaf node samples of 3 to 10, and a total model storage footprint of no more than 256kB.

[0048] Preferably, the health score The weighted summation of the probabilities of each level output by the lightweight gradient boosting classification model can be used to obtain the following:

[0049] .

[0050] in, , , and They represent the first Each branch road at any time The probability of belonging to four levels: normal, mild, moderate, and severe degradation; 1.00, 1.24, 1.46, and 1.94 are the level center values ​​obtained by fitting the damage center values ​​of the four health states in the labeled aging samples.

[0051] S4. Estimate the remaining lifespan of the compensation branch;

[0052] Based on the health score output in step S3 And combined with the ratio of branch action times consumption and temperature rise stress coefficient ,according to Estimate the remaining lifespan of the compensation branch, where , This represents the number of sampling points within the most recent temperature rise stress statistical window. For the first The stable temperature rise benchmark of each branch under rated operating conditions This is the nominal lifespan of the branch under rated operating conditions. To meet Damage weighting coefficient.

[0053] Preferably, The preferred value is 15. The life loss is determined by least squares fitting of the prototype aging sample, and the remaining life assessment result can be output in the form of days, number of operation cycles or the length of time that can continue to run.

[0054] S5. Implement branch classification and risk ranking;

[0055] The compensation branches are classified into normal, slightly deteriorated, moderately deteriorated, and severely deteriorated levels, and further classified according to risk index. The risks are ranked from highest to lowest, among which... and To meet The ranking weight.

[0056] Preferably, the health level classification can be performed according to the following rules, where 1.12, 1.35, and 1.70 are the midpoints of the adjacent level center values ​​1.00 and 1.24, 1.24 and 1.46, and 1.46 and 1.94, respectively:

[0057] when A value less than 1.12 is considered normal.

[0058] When 1.12 is less than or equal to and A value less than 1.35 is considered a slight degradation.

[0059] When 1.35 is less than or equal to and A value less than 1.70 is considered moderate degradation.

[0060] when A value greater than or equal to 1.70 is considered severely degraded.

[0061] S6. Output health assessment results and maintenance recommendations;

[0062] Based on the results of step S5, output suggestions for re-inspection, replacement, derating operation, or decommissioning of the compensation branch; when When the health level falls below the preset threshold or reaches a severely degraded state, the corresponding branch will be listed as a priority maintenance target.

[0063] The following description is based on specific embodiments.

[0064] Step 1: Determine the target audience and evaluation criteria

[0065] The JP intelligent metal integrated distribution box was selected as the implementation object. The distribution box is equipped with an intelligent harmonic suppression reactive power compensation module, a composite switch, and a power distribution monitoring and compensation controller. The compensation branches include single-phase and three-phase compensation branches. Based on the product type test report, composite switch specifications, and on-site operation and maintenance records, the typical compensation branch capacity includes a 20kvar single-phase branch and a 40kvar three-phase branch, where the 40kvar three-phase branch is used as the reference rated compensation capacity Qbase in the harmonic stress quantification of this embodiment.

[0066] In this embodiment, the branch health assessment targets include the C4 single-phase compensated branch and the C5, C6, C7, and C8 three-phase compensated branches. The sampling period is preferably 1 minute; the training and validation datasets mainly consist of accelerated aging test data from physical prototypes, type test data, and historical field operation data, supplemented by boundary condition augmentation data generated by digital prototypes.

[0067] Step 2: Constructing the health characteristics of the compensating branch

[0068] At each evaluation time, the commanded compensation capacity and actual response capacity of each compensation branch are collected to obtain the compensation response residual. Simultaneously, the branch operation duration, branch temperature rise, harmonic stress quantification value, and cumulative number of operations are statistically analyzed. The actual response capacity Qact,i(k) is obtained by converting the branch current or branch reactive power. The above features are collectively used as the branch health feature vector and input into the subsequent model.

[0069] In this embodiment, for the i-th branch, the following features are constructed: ,in , , , .

[0070] in, Let i be the actual action time of the i-th branch from receiving the switching command to the state stabilizing. To calibrate the baseline motion time under the operating conditions, To accumulate the number of actions, The rated operating life of the composite switch.

[0071] Step 3: Health Level Assessment

[0072] The health feature vector is input into a lightweight gradient boosting classification model to obtain the probabilities of each level. , , and Press again Calculate the health score. In this embodiment, each model consists of 80 CART classification trees with a maximum depth of 3, a learning rate of 0.05, a minimum number of leaf node samples of 5, and a total model storage footprint of no more than 256kB.

[0073] The health level is then given according to the following rules: based on the four damage center values ​​of 1.00, 1.24, 1.46 and 1.94 in the labeled aging samples, the midpoints of adjacent center values ​​of 1.12, 1.35 and 1.70 are taken as the grading thresholds.

[0074] when A value less than 1.12 is considered normal.

[0075] When 1.12 is less than or equal to and A value less than 1.35 is considered a slight degradation.

[0076] When 1.35 is less than or equal to and A value less than 1.70 is considered moderate degradation.

[0077] when A value greater than or equal to 1.70 is considered severely degraded.

[0078] The above grading rules mean that the health level is no longer determined by a single temperature threshold, but is jointly determined by compensation response performance, dynamic characteristics of motion, thermal stress, and harmonic stress; the level center value and grading threshold are both determined by the statistical results of labeled aging samples and the optimal classification effect of the validation set.

[0079] Step 4: Remaining life assessment

[0080] In this embodiment, the health score output in step S3 is used as the basis for... Action frequency consumption ratio and temperature rise stress coefficient ,according to Calculate the remaining lifespan in days for the branch. Among these, Determined by the composite switch specifications and type test life data. The lifespan loss is obtained by least-squares fitting of the prototype aging sample; when the remaining lifespan of a certain branch is lower than the preset threshold, the branch is listed as a priority maintenance target.

[0081] Step 5: Output Maintenance Suggestions

[0082] For lightly degraded branches, recommendations are made to re-inspect and shorten the inspection cycle; for moderately degraded branches, recommendations are made to schedule power outages for re-inspection or reduce operating limits; for severely degraded branches, recommendations are made to replace or decommission them as soon as possible. Furthermore, branches can be prioritized by remaining lifespan from low to high risk, allowing maintenance personnel to address high-risk branches first.

[0083] Step Six: Joint Verification and Result Analysis of Physical and Digital Prototypes

[0084] In the development phase of this embodiment, training and validation datasets were constructed using accelerated aging tests, type tests, and historical field operation data from physical prototypes. A system-level digital prototype built using Python was used to supplement boundary conditions related to branch aging, harmonic stress, and motion delay variations. The digital prototype data is only used to expand upon conditions that the physical prototype cannot cover long-term and is not used as the sole basis for the final results.

[0085] In this embodiment, a branch health dataset is constructed by sampling the 24-hour operating sequence over time. The true health level label is determined based on a comprehensive analysis of capacity retention rate, action delay increment, steady-state temperature rise, and contact wear inspection results: preferably, a capacity retention rate of not less than 95%, action delay increment of not more than 5ms, and steady-state temperature rise of not more than 15 degrees Celsius is marked as normal; a capacity retention rate of 90% to 95%, action delay increment of 5ms to 10ms, or steady-state temperature rise of 15 degrees Celsius to 20 degrees Celsius is marked as slightly degraded; a capacity retention rate of 80% to 90%, action delay increment of 10ms to 20ms, or steady-state temperature rise of 20 degrees Celsius to 30 degrees Celsius is marked as moderately degraded; and a capacity retention rate of less than 80%, action delay increment of more than 20ms, or steady-state temperature rise of more than 30 degrees Celsius is marked as severely degraded. If multiple indicators correspond to different levels, the label is determined according to the most severe level.

[0086] The joint test results show that, within the test set comprised of physical prototype test data, field operation history data, and digital prototype boundary expansion data, the branch health level identification accuracy of the method of this invention is approximately 98.06%, the macro-average F1 value is approximately 0.9097, and the average absolute error of remaining lifetime assessment is approximately 2.31 days; the level identification accuracy of the temperature rise baseline method is approximately 76.46%, the macro-average F1 value is approximately 0.7375, and the average absolute error of remaining lifetime assessment is approximately 73.80 days.

[0087] like Figure 2 As shown, as the health status of aging branches deteriorates, their compensation response residuals and action delays gradually increase; for example... Figure 3As shown, the health score output by this invention can effectively follow changes in labeled health status and corresponds to the thresholds for mild, moderate, and severe degradation; for example... Figure 4 Figure 5 As shown in Figure 6, the method of the present invention is superior to the temperature rise baseline method in both grade identification and life assessment.

[0088] Therefore, the method of the present invention can perform a more targeted assessment of the health status and remaining lifespan of the compensation branch before it completely fails, making it suitable for predictive maintenance and compensation branch lifespan management in JP smart metal integrated distribution boxes.

[0089] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box, characterized in that, The application in intelligent metal integrated distribution boxes with current, temperature, harmonic acquisition, and branch operation recording functions includes the following steps: S1: Collect the operating status information of each compensation branch, including at least the commanded compensation capacity, actual response compensation capacity, action duration, temperature of key measuring points, total harmonic distortion rate, cumulative number of actions, ambient temperature and rated capacity; S2: Based on the collected information, construct a branch health feature vector containing five dimensions: compensation response residual, action delay increment, branch temperature rise, harmonic stress value, and action consumption ratio. S3: Input the health feature vector into a lightweight gradient boosting classification model. The model outputs the probability that the branch belongs to one of four levels: normal, mildly deteriorated, moderately deteriorated, or severely deteriorated. A continuous health score is then calculated based on these probabilities and weighted accordingly. S4: Combining the health score, the ratio of action frequency consumption, and a temperature rise stress coefficient that reflects the degree of recent temperature rise exceeding the limit, the remaining life of the compensation branch is estimated according to the preset life damage formula. S5: Based on the health score, classify the health level and calculate the risk index in combination with the remaining life expectancy, and rank all compensation branches by risk. S6: Based on the health level and remaining life assessment results, output maintenance recommendations including re-inspection, replacement, de-rated operation or decommissioning.

2. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The compensation response residual is obtained by dividing the absolute value of the difference between the commanded compensation capacity and the actual response compensation capacity by the rated compensation capacity of the compensation branch, and is used to quantify the degree of attenuation of the branch compensation capacity.

3. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The action delay increment is taken as the larger value between "the difference between the current action duration and the reference action duration" and "0". When the action delay increment increases significantly, it is considered that the composite switch or related branch has an aging trend.

4. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The branch temperature rise is obtained by subtracting the current ambient temperature from the temperature at the key measuring point of the branch, and is used to quantify the degree of heat generation during branch operation.

5. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The harmonic stress value comprehensively considers the square of the current total harmonic distortion rate, the ratio of the rated capacity of the compensated branch to a reference capacity, and a sensitivity coefficient that is dynamically adjusted according to the degree to which the compensation response residual exceeds the reference value, and is used to quantify the additional stress of the harmonic environment on the branch.

6. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 5, characterized in that, The reference base capacity is set to 40 kvar, which is used to achieve normalized comparison between branches with different capacities.

7. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The health score is obtained by multiplying the probabilities of the four levels output by the gradient boosting classification model by the central values ​​of the levels corresponding to normal, mild, moderate and severe deterioration, respectively, and then summing them up. The higher the score, the worse the health status.

8. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The temperature rise stress coefficient is obtained by averaging the ratio of the portion of the measured temperature rise exceeding the rated temperature rise benchmark in each sampling to the benchmark within a statistical window. If the temperature rise does not exceed the benchmark, it is taken as zero. This is used to quantify the cumulative damage caused to the branch circuit by long-term over-temperature operation.

9. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The rules for classifying health levels are based on health scores: When the health score is below the first threshold, it is considered normal; When the health score is between the first threshold and the second threshold, it is considered to be slightly deteriorated; When the health score is between the second and third thresholds, it is considered to be moderately deteriorated; When the health score is greater than or equal to the third threshold, it is judged as severely deteriorated.

10. The method for health identification and lifespan assessment of compensation branches in an intelligent metal integrated distribution box according to claim 1, characterized in that, The specific maintenance recommendations are as follows: For slightly deteriorated branches, it is recommended to re-inspect and shorten the inspection cycle; for moderately deteriorated branches, it is recommended to arrange a power outage window for re-inspection or reduce the operating rate; for severely deteriorated branches or branches with remaining life below the preset threshold, it is recommended to replace or decommission them as soon as possible and list them as priority maintenance targets.

Images