GIS basin-type insulator service life estimation method and system based on combined empowerment

Abstract

The invention discloses a GIS basin-type insulator life estimation method and system based on combined weighting, and the method comprises the steps: determining and collecting a plurality of feature indexes affecting the life of a basin-type insulator, and constructing an initial sample evaluation matrix; based on the initial sample evaluation matrix, an entropy weight method is adopted to calculate objective weights of the feature indexes, the objective weights are recorded as entropy weight factors, and a life contribution value of each feature index based on the entropy weight method is calculated; based on the initial sample evaluation matrix, a random forest algorithm is adopted to calculate the feature importance degree of each feature index for the life evaluation result, and the life contribution value of each feature index based on the random forest algorithm is calculated; according to the entropy weight factor of each characteristic index and the importance degree of each characteristic index, constructing a comprehensive life estimation model; and inputting the characteristic index data of the basin-type insulator to be evaluated into the comprehensive life estimation model, and outputting the estimated life value of the insulator. The method has the advantages of remarkably improving the accuracy and reliability of life evaluation and the like.

Description

Technical Field

[0001] This invention mainly relates to the field of power equipment technology, specifically to a method and system for estimating the lifespan of GIS basin insulators based on combined weighting. Background Technology

[0002] With the rapid development of ultra-high voltage and extra-high voltage power grids, gas-insulated switchgear (GIS) has become a key piece of equipment in power transmission and transformation systems due to its advantages such as small footprint, high reliability, and low environmental impact. The core insulating component of GIS equipment—the basin insulator—operates for extended periods in environments with strong electric fields, high gas pressure, and complex conditions; its insulation performance directly affects the safety and stability of the entire GIS system. In recent years, operational statistics have shown that failures caused by basin insulator problems account for a consistently high proportion of GIS equipment failures, with surface charge accumulation leading to insulation performance degradation being a significant contributing factor.

[0003] During long-term operation, basin-type insulators are affected by various factors such as operating voltage, operating temperature, gas pressure, SF6 content, and local defects, causing a gradual accumulation of surface charge. This continuous accumulation of surface charge leads to localized distortion of the electric field, accelerates the electro-aging process of the insulation material, significantly reduces the surface flashover voltage of the insulator, and ultimately triggers surface discharge or even breakdown faults. Determining the contribution of surface charge accumulation factors to the aging of basin-type insulators allows for the prediction of insulator lifespan, providing important insights for the safe and reliable operation of GIS systems.

[0004] Most existing studies on basin insulator life prediction methods employ the entropy weight method. This method is difficult to comprehensively and objectively predict the lifespan of basin insulators under surface charge accumulation. Furthermore, the entropy weight method mechanically considers the weight of each quantitative indicator on the output result, while neglecting the interaction between the output result and each indicator. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method and system for predicting the life of GIS basin insulators based on combined weighting, which significantly improves the accuracy and reliability of life assessment.

[0006] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows: A method for predicting the life of GIS basin insulators based on combined weighting includes the following steps: Multiple characteristic indicators affecting the lifespan of basin insulators were identified and collected, and an initial sample evaluation matrix was constructed. Based on the initial sample evaluation matrix, the objective weight of each feature index is calculated using the entropy weight method, denoted as the entropy weight factor. And calculate the lifetime contribution value of each feature index based on the entropy weight method. ; Based on the initial sample evaluation matrix, the random forest algorithm is used to calculate the feature importance of each feature index to the life assessment result, denoted as […]. And calculate the lifetime contribution value of each feature index based on the random forest algorithm. ; Based on the entropy weight factor of each feature index Importance of each feature indicator A comprehensive lifespan prediction model is constructed; the model is fused using a weighted method. and And introduce the interaction function between feature indices. The estimated lifespan of the basin-type insulator under the combined effect of multiple characteristic indicators was calculated. T ; The characteristic index data of the pot insulator to be evaluated are input into the comprehensive life prediction model, and the predicted life value of the insulator is output to achieve a quantitative assessment of the remaining service life.

[0007] Preferably, based on the initial sample evaluation matrix, the objective weight of each feature index is calculated using the entropy weight method, and denoted as the entropy weight factor. And calculate the lifetime contribution value of each feature index based on the entropy weight method. The specific process is as follows: The initial sample evaluation matrix is ​​normalized, and different standardization algorithms are used to distinguish between positive and negative indicators; Calculate the proportion of the i-th sample value under the j-th indicator to that indicator. ; Calculate the information entropy value of the j-th indicator. and information entropy redundancy ; Calculate the entropy weight factor weights of each indicator. Then, the lifetime contribution value of each characteristic indicator is calculated. .

[0008] Preferably, the lifetime contribution value of each characteristic index The calculation formula is:

[0009] in This represents the average insulator lifetime from historical databases. ; ; ;

[0010] in Let be the value of the j-th indicator for the i-th sample.

[0011] Preferably, based on the initial sample evaluation matrix, the random forest algorithm is used to calculate the feature importance of each feature index to the life assessment result, denoted as . And calculate the lifetime contribution value of each feature index based on the random forest algorithm. The specific process is as follows: A random forest model consisting of multiple decision trees is constructed using sample data, with each tree trained based on a random subset and random feature selection; During the construction of each decision tree, the contribution of each feature to the reduction of impurity is recorded when a node splits; Summarize the reduction in impurity for each feature across all decision trees, calculate the average, and obtain the overall importance of each feature indicator. ; Based on the overall importance of each feature indicator The lifespan contribution value of each characteristic indicator was calculated. .

[0012] Preferably, the specific calculation formula of the integrated life prediction model is as follows:

[0013] in T Lifetime under the combined effect of various characteristics a , b To adjust the parameters; n The number of characteristic indicators; To describe features i and characteristics j Functions that interact with each other.

[0014] Preferably, Interaction factors are assigned values ​​based on general physical principles: positive correlation is assigned a value of 1, negative correlation is assigned a value of -1, and no significant interaction is assigned a value of 0.

[0015] Preferably, the characteristic indicators include the operating voltage amplitude, operating ambient temperature, gas pressure, and SF6 gas content.

[0016] The present invention also discloses a computer program product, comprising a computer program that, when executed by a processor, performs the steps of the method described above.

[0017] The present invention further discloses a computer-readable storage medium having a computer program stored thereon, the computer program executing the steps of the method described above when run by a processor.

[0018] The present invention also discloses a computer system including a memory and a processor interconnected thereon, wherein the memory stores a computer program that, when run by the processor, performs the steps of the method described above.

[0019] Compared with the prior art, the advantages of the present invention are as follows: This invention constructs a basin insulator life prediction model that comprehensively reflects the influence of multiple factors and the interaction between indicators by integrating the objective weight evaluation of the entropy weight method with the feature importance analysis of the random forest algorithm. This method overcomes the limitations of single weighting methods that neglect the correlation between indicators or are sensitive to outliers, achieving in-depth utilization and collaborative interpretation of multi-source monitoring data such as operating voltage, temperature, air pressure, and SF6 content. Practical verification shows that the model's prediction error does not exceed 6%, significantly improving the accuracy and reliability of life assessment. It can provide quantitative and refined decision support for insulation condition assessment, maintenance strategy formulation, and life cycle management of ultra-high voltage and extra-high voltage GIS equipment. Attached Figure Description

[0020] Figure 1 The flowchart shows an embodiment of the GIS basin insulator life prediction method of the present invention.

[0021] Figure 2 This is a schematic diagram of the objective weight distribution of each feature index calculated using the entropy weight method in this invention.

[0022] Figure 3 This is a schematic diagram of the importance results of each feature index obtained by training with the random forest algorithm in this invention.

[0023] Figure 4 This is a graph showing the interactions between the various feature indices defined in this invention.

[0024] Figure 5 This figure shows the comparison results of the actual lifespan and the model-predicted lifespan of insulators for different samples in this invention. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0026] like Figure 1 As shown, the method for estimating the life of GIS basin insulators based on combined weighting provided in this embodiment of the invention includes the following steps: S1. Based on voltage monitoring host computer, infrared temperature measurement equipment, pressure sensing device and SF 6. Gas micro-moisture online monitoring system: Collects key characteristic indicators affecting surface charge accumulation and insulation performance of basin insulators, including operating voltage amplitude, operating ambient temperature, gas insulating medium pressure, etc. SF6. Gas content, constructing an initial sample evaluation matrix; The contribution of characteristic indicators to the service life of insulators is scored based on expert experience, and the scoring results can be found in Table 1.

[0027] In information theory, entropy is an indicator of the degree of disorder in a system. The greater the disorder and uncertainty of a system's information, the higher the entropy value and the greater the amount of information it contains; conversely, the lower the disorder and uncertainty, the lower the entropy value and the smaller the amount of information it contains. Based on the characteristics of entropy, calculating its value can determine the degree of disorder and randomness of an event. It can also be used to determine the dispersion of a certain indicator; the greater the dispersion of an indicator, the greater its influence on the result. If the sample data all take equal values ​​under a certain indicator, then the indicator's influence on the target result is 0, and therefore its weight is also 0. The entropy weighting method is an objective weighting method because it relies only on the dispersion of the data itself and is unrelated to the target result. Its specific implementation steps are as follows: S201. Let the sample size be... m The number of indicators is n ,but For the first i The first sample j The values ​​of each indicator ( i =1, 2, ..., m ; j =1, 2, ..., n ), initial sample evaluation matrix As shown below:

[0028] S202. Normalize the various characteristic indicators to solve the problem of homogenization of values ​​of different indicators; Specifically, because the units of measurement for various characteristic indicators are not uniform, standardization processing is required before calculating the indicators. This involves converting the absolute values ​​of the indicators into relative values ​​to solve the problem of homogenization among different indicator values. Furthermore, positive and negative indicators represent different meanings (higher values ​​for positive indicators are better, while lower values ​​for negative indicators are better). Therefore, different algorithms are needed for data standardization processing of positive and negative indicators.

[0029] Positive indicators:

[0030] Negative indicators:

[0031] For convenience, the normalized data Still recorded as .

[0032] S203. Calculate the proportion of the i-th sample value under the j-th indicator to that indicator. :

[0033] S204. Calculate the entropy value of the j-th index. :

[0034] S205. Calculate information entropy redundancy. (difference):

[0035] S206. Calculate the weights of each indicator. :

[0036] Therefore, the entropy weight factor corresponding to each feature is calculated. The contribution of the i-th feature index to lifespan is calculated to be... ;in This represents the average insulator lifetime from the historical database.

[0037] S3. Simultaneously, to consider the interaction between the output results and various indicators, a random forest algorithm is used to calculate feature importance. The random forest algorithm effectively evaluates the importance of indicators. Its principle is to construct multiple decision trees and analyze the role of each feature in the construction process of these trees. During the construction of each tree, the reduction in impurity for each feature at node splits is recorded, thus obtaining the contribution of each feature to the overall model performance. Furthermore, feature importance can also be assessed by observing changes in model performance after feature values ​​are shuffled. If model performance significantly decreases, it indicates that the feature is crucial to the model's predictive ability. This method comprehensively considers both the direct and indirect contributions of features to the model, thus providing a comprehensive assessment of feature importance.

[0038] The specific process of calculating feature importance using the random forest algorithm is as follows: S301. Constructing a Random Forest Model: Use the random forest algorithm to construct a model consisting of multiple decision trees. Each tree is trained on a random subset of the original data, and the feature selection for each split node is also random.

[0039] S302. Calculate the feature importance of each tree: During the construction of each decision tree, record the contribution of each feature to the reduction of impurity at each node split. This is typically achieved by calculating the average reduction of impurity for each feature across all node splits.

[0040] S303. Summarize Feature Importance: Summarize the feature importance of all decision trees to obtain the overall importance of each feature. This can be done by averaging the reduction in impurity of each feature across all trees.

[0041] S304. Evaluate Feature Importance: Based on the aggregated feature importance scores, evaluate the contribution of each feature to the model's predictive ability. Features with higher importance scores have a greater impact on the model's prediction results.

[0042] S305. Visualize feature importance: Display feature importance in the form of charts to facilitate understanding and interpretation.

[0043] The importance of each feature is then calculated. And thus obtain the first i The contribution of each characteristic index to lifespan is .

[0044] S4. A comprehensive lifetime prediction model is constructed by integrating the results of the entropy weight method and the random forest algorithm using a novel improved method. While the entropy weight method can objectively determine weights based on the dispersion of data and reflect the variation information of indicators, it does not consider the interaction between multiple indicators. Random forest can effectively assess the importance of indicators, but it may be sensitive to outliers and may overestimate the importance of certain features in some cases. Therefore, the entropy weight method is needed to adjust weights and mitigate the impact of outliers. A single algorithm may have shortcomings, thus requiring a comprehensive evaluation of indicators. By integrating the results of different algorithms, the model's accuracy in identifying feature importance can be improved. The comprehensive feature scoring allows for adjusting the algorithm weights according to different data characteristics and problem requirements, making the model more adaptable and flexible. The lifetime formula for the basin insulator is designed as follows:

[0045] in T Lifetime under the combined effect of various characteristics For the first i The lifetime was calculated using the entropy weighting method based on several characteristic indicators. For the first i The lifetime of each feature index is calculated using random forest. a , b To adjust the parameters, based on expert experience, a、b All values ​​are assumed to be 0.5. n The number of characteristic indicators; It is a descriptive feature i and characteristics j Functions that interact between them Interaction factors are assigned values ​​based on general physical principles: a positive correlation is assigned a value of 1, a negative correlation is assigned a value of -1, and no significant interaction is assigned a value of 0. Assignment diagram as follows Figure 4 As shown.

[0046] Due to the above The assignment of values ​​is relatively static and may be subjective. Therefore, in the early stages of training the prediction model, historical datasets are used, and the PC algorithm (Peter-Clark algorithm) or a constraint-based causal discovery method is employed to automatically analyze the potential causal direction and strength between various feature indicators (voltage, temperature, air pressure, and SF6 content) at the data level. This step aims to reduce complete reliance on prior physical knowledge and objectively identify data-driven causal relationships from the operational data, such as "whether an increase in temperature directly leads to an intensification of SF6 decomposition" or "whether voltage fluctuations precede the activity of partial discharge," generating a directional causal correlation matrix. to replace the original static Assign a value.

[0047] Finally, example calculations are performed to verify the effectiveness of the method. The lifetime prediction process is as follows: Figure 1 As shown: Historical data on various characteristic indicators of charge accumulation on the insulator surface are compiled, and the data is imported into MATLAB to obtain the entropy weight factor weight Ki (e.g., Figure 2 As shown), the processed data is then imported into a random forest model to obtain the feature importance Mi (e.g., Figure 3 (As shown). Substituting the obtained feature entropy weighting factors K and importance M into the corresponding lifetime contribution formulas, we can obtain the i-index. and Six samples with different characteristic values ​​were selected, and the characteristic values ​​were substituted into the insulator life formula to obtain the estimated life. The estimated life was then compared with the actual life. Figure 5 As shown, the maximum error range is no more than 6%, which indicates that it has high accuracy.

[0048] This invention constructs a basin insulator life prediction model that comprehensively reflects the influence of multiple factors and the interaction between indicators by integrating the objective weight evaluation of the entropy weight method with the feature importance analysis of the random forest algorithm. This method overcomes the limitations of single weighting methods that neglect the correlation between indicators or are sensitive to outliers, achieving in-depth utilization and collaborative interpretation of multi-source monitoring data such as operating voltage, temperature, air pressure, and SF6 content. Practical verification shows that the model's prediction error does not exceed 6%, significantly improving the accuracy and reliability of life assessment. It can provide quantitative and refined decision support for insulation condition assessment, maintenance strategy formulation, and life cycle management of ultra-high voltage and extra-high voltage GIS equipment.

[0049] The present invention also discloses a computer program product, comprising a computer program that, when executed by a processor, performs the steps of the method described above.

[0050] The present invention further discloses a computer-readable storage medium having a computer program stored thereon, the computer program executing the steps of the method described above when run by a processor.

[0051] The present invention also discloses a computer system including a memory and a processor interconnected thereon, wherein the memory stores a computer program that, when run by the processor, performs the steps of the method described above.

[0052] The products, media, and systems of the present invention, corresponding to the methods described above, also possess the advantages described above.

[0053] The present invention can implement all or part of the processes in the methods of the above embodiments, or it can be implemented by hardware related to computer program instructions. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium includes: any entity or device capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. The memory is used to store computer programs and / or modules. The processor implements various functions by running or executing the computer programs and / or modules stored in the memory, and by calling data stored in the memory. The memory may include high-speed random access memory, as well as non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital (SD) cards, flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0054] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for predicting the life of GIS basin insulators based on combined weighting, characterized in that, Including the following steps: Multiple characteristic indicators affecting the lifespan of basin insulators were identified and collected, and an initial sample evaluation matrix was constructed. Based on the initial sample evaluation matrix, the objective weight of each feature index is calculated using the entropy weight method, denoted as the entropy weight factor. And calculate the lifetime contribution value of each feature index based on the entropy weight method. ; Based on the initial sample evaluation matrix, the random forest algorithm is used to calculate the feature importance of each feature index to the life assessment result, denoted as […]. And calculate the lifetime contribution value of each feature index based on the random forest algorithm. ; Based on the entropy weight factor of each feature index Importance of each feature indicator A comprehensive lifespan prediction model is constructed; the model is fused using a weighted method. and And introduce the interaction function between feature indices. The estimated lifespan of the basin-type insulator under the combined effect of multiple characteristic indicators was calculated. T ; The characteristic index data of the pot insulator to be evaluated are input into the comprehensive life prediction model, and the predicted life value of the insulator is output to achieve a quantitative assessment of the remaining service life.

2. The method for predicting the life of GIS basin-type insulators based on combined weighting as described in claim 1, characterized in that, Based on the initial sample evaluation matrix, the objective weight of each feature index is calculated using the entropy weight method, denoted as the entropy weight factor. And calculate the lifetime contribution value of each feature index based on the entropy weight method. The specific process is as follows: The initial sample evaluation matrix is ​​normalized, and different standardization algorithms are used to distinguish between positive and negative indicators; Calculate the proportion of the i-th sample value under the j-th indicator to that indicator. ; Calculate the information entropy value of the j-th indicator. and information entropy redundancy ; Calculate the entropy weight factor weights of each indicator. Then, the lifespan contribution value of each characteristic indicator is calculated. .

3. The method for predicting the life of GIS basin-type insulators based on combined weighting according to claim 2, characterized in that, Lifetime contribution value of each characteristic indicator The calculation formula is: in This represents the average insulator lifetime from historical databases. ； ； ； in Let be the value of the j-th indicator for the i-th sample.

4. The method for predicting the life of GIS basin-type insulators based on combined weighting according to claim 1, 2, or 3, characterized in that, Based on the initial sample evaluation matrix, the random forest algorithm is used to calculate the feature importance of each feature index to the life assessment result, denoted as […]. And calculate the lifetime contribution value of each feature index based on the random forest algorithm. The specific process is as follows: A random forest model consisting of multiple decision trees is constructed using sample data, with each tree trained based on a random subset and random feature selection; During the construction of each decision tree, the contribution of each feature to the reduction of impurity is recorded when a node splits; Summarize the reduction in impurity for each feature across all decision trees, calculate the average, and obtain the overall importance of each feature indicator. ; Based on the overall importance of each feature indicator The lifespan contribution value of each characteristic indicator was calculated. .

5. The method for predicting the life of GIS basin-type insulators based on combined weighting according to claim 1, 2, or 3, characterized in that, The specific calculation formula for the integrated life prediction model is as follows: in T Lifetime under the combined effect of various characteristics a , b To adjust the parameters; n The number of characteristic indicators; To describe features i and characteristics j Functions that interact with each other.

6. The method for predicting the life of GIS basin-type insulators based on combined weighting according to claim 5, characterized in that, Interaction factors are assigned values ​​based on general physical principles: positive correlation is assigned a value of 1, negative correlation is assigned a value of -1, and no significant interaction is assigned a value of 0.

7. The method for estimating the life of GIS basin insulators based on combined weighting according to claim 1, 2, or 3, characterized in that, The characteristic indicators include the operating voltage amplitude, operating ambient temperature, gas pressure, and SF6 gas content.

8. A computer program product, comprising a computer program, characterized in that, The computer program is executed by the processor to perform the steps of the method as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1-7.

10. A computer system comprising a memory and a processor interconnected thereon, the memory storing a computer program, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1-7.

Images