A method and device for prioritizing the reconstruction of a distribution network problem line, a terminal device and a storage medium

Abstract

The application discloses a kind of transformation priority sequencing method, device, terminal equipment and storage medium of distribution network problem line, belong to line transformation technical field, above-mentioned method is: obtaining the original power operation data of distribution network, original meteorological data, equipment account data and topological connection relationship, and combining preset line explicit problem diagnosis rule base and preset line implicit problem identification model, determine problem line, while calculating the value of each index of each problem line;According to the value of all indexes of all problem lines, the objective weight vector of all indexes is calculated, and the subjective weight vector of all indexes is obtained based on the analytic hierarchy process, and then the relative closeness of each problem line is calculated;Finally, according to the size order of relative closeness, transformation priority sequencing is carried out for all problem lines. By implementing the present application, the transformation priority sequencing method of the prior art relies on personal experience, so there is a problem of unreasonable resource allocation in the transformation process.

Description

Technical Field

[0001] This invention relates to the field of line renovation technology, and in particular to a method, apparatus, terminal equipment, and storage medium for prioritizing the renovation of problematic distribution network lines. Background Technology

[0002] As the goal of achieving carbon peaking and carbon neutrality is further advanced, the scale of new energy sources, represented by distributed photovoltaic and wind power, in the distribution network continues to expand. As a result, the distribution network structure has become more complex, and its operating characteristics have shown obvious uncertainty and strong nonlinearity. This change has brought severe challenges to the management of traditional distribution networks.

[0003] Existing technologies typically rely on expert scoring to determine the priority of processing and modifying each problem line. This prioritization method depends on personal experience, and urgent problems may be delayed while less important problems are prioritized. Therefore, there is a problem of unreasonable resource allocation in the final modification process. Summary of the Invention

[0004] This invention provides a method, apparatus, terminal equipment, and storage medium for prioritizing the transformation of problematic distribution network lines. It can solve the problem that the prioritization method of the prioritization of transformation in the prioritization of transformation of existing technology relies on personal experience, and therefore has the problem of unreasonable resource allocation in the final transformation process.

[0005] An embodiment of the present invention provides a method for prioritizing the upgrading of problematic distribution network lines, comprising: Obtain raw power operation data, raw meteorological data, equipment ledger data, and topology connections of the power distribution network; Based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships, a digital model of the power distribution network is constructed. Based on the digital model of the distribution network, the preset rule base for diagnosing explicit line problems, and the preset rule base for identifying implicit line problems, the problematic lines of the distribution network are determined. Based on the digital model of the distribution network, the values ​​of various indicators for each problematic line are calculated; Based on the values ​​of all indicators for all problematic routes, the original data matrix is ​​constructed. Based on the original data matrix, the entropy weight of each indicator is calculated, and an objective weight vector of all indicators is generated based on all entropy weights. The subjective weight vector of all indicators is obtained based on the analytic hierarchy process. Based on the original data matrix, objective weight vector, and subjective weight vector, the relative proximity of each problem route is calculated. All problematic lines are prioritized for modification according to their relative proximity from highest to lowest, and modifications are then carried out based on the priority ranking.

[0006] Furthermore, the process of constructing a digital model of the distribution network based on the original power operation data, original meteorological data, equipment ledger data, and topological connections includes: Based on the equipment ledger data and topology connection relationships, determine the power equipment of the power distribution network, the electrical connection relationships between each power equipment, the equipment attribute information of the power equipment, and the connection attribute information of the electrical connection relationships. Using the power equipment as nodes, the connection relationships between the power equipment as edges, the equipment attribute information of each power equipment as node attribute information, and the connection attribute information as edge attribute information, a power distribution network topology model is constructed. Identify the target power equipment corresponding to the original power operation data and original meteorological data, and associate the original power operation data and original meteorological data with the node corresponding to the target power equipment based on the target power equipment; All node attribute information, corresponding raw power operation data, and raw meteorological data are processed to unify units, semantics, and naming to obtain the digital model of the power distribution network.

[0007] Furthermore, based on the digital model of the distribution network, the preset rule base for diagnosing explicit line problems, and the preset rule base for identifying implicit line problems, the problematic lines of the distribution network are determined, including: Line information data for each line is extracted from the digital model of the power distribution network. Based on the route information data, determine the monitoring index values ​​of the preset monitoring indicators for each route; For each route, the monitoring index values ​​are compared with a preset rule base for diagnosing explicit problems in routes to identify routes with explicit risk issues. Based on the line information data, construct fault statistics feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector; The fault statistical feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector are input into a preset line hidden problem identification model to obtain the confidence level that the line has hidden risks. Lines with a confidence level not less than a preset confidence threshold are considered as lines with hidden risks. The explicit problem path and the implicit problem path are collectively referred to as the problem path.

[0008] Furthermore, the training of the preset line hidden problem identification model includes: Obtain first-line information data for each historically problematic line and second-line information data for each historically unproblematic line; Based on the first route information data and the second route information data, a historical multidimensional feature vector is constructed for the corresponding historical route; wherein, the historical route includes: historical routes with hidden problems and historical routes without hidden problems; A training sample set is generated based on the historical multidimensional feature vectors with real labels; wherein, the real labels are used to indicate whether the corresponding historical line is a historical line with hidden problems or a historical line without hidden problems. The training sample set is input into the hidden line problem identification model to be trained, so that the hidden line problem identification model to be trained is iteratively trained based on the gradient boosting algorithm until the preset number of iterations is reached, and the preset hidden line problem identification model is obtained after training. In each iteration, the training sample set is input into the hidden problem identification model of the railway line to be trained to generate the current predicted value; a new decision tree is added with the goal of fitting the residual between the current predicted value and the corresponding true label, and the structure and leaf node weights of the newly added decision tree are determined by minimizing the objective function; it is determined whether the current iteration round has reached the preset iteration round; if so, the preset hidden problem identification model of the railway line is generated after training; otherwise, the next iteration round is continued based on the training sample set. The objective function is constructed based on the current model prediction value; the initial prediction value is a preset initial prediction value.

[0009] Furthermore, based on the original data matrix, the entropy weight of each indicator is calculated, including: The original data matrix is ​​standardized to obtain a standardized matrix; Based on the standardized matrix, the characteristic weight of each problem line under each indicator is calculated; Based on the weight of the features, the information entropy of each indicator is calculated; Based on the information entropy, the entropy weight of each indicator is calculated.

[0010] Furthermore, based on the analytic hierarchy process (AHP), the subjective weight vectors of all indicators are obtained, including: Obtain the judgment matrix constructed by each expert for all indicators; For each expert, calculate the maximum eigenvalue and eigenvector of the judgment matrix constructed by the current expert; The feature vectors are normalized to obtain the initial subjective weight vectors of all indicators, and then the initial subjective weight vectors of each expert are obtained. For each initial subjective weight vector, the consistency index value is calculated based on the maximum eigenvalue corresponding to the current initial subjective weight vector. The consistency ratio is calculated based on the consistency index value; The initial subjective weight vectors with a consistency ratio less than a preset ratio threshold are removed to obtain the effective subjective weight vectors. Calculate the arithmetic mean of all valid subjective weight vectors to obtain the subjective weight vector for each indicator.

[0011] Furthermore, based on the original data matrix, objective weight vector, and subjective weight vector, the relative proximity of each problem path is calculated, including: The values ​​of each indicator in the original data matrix are subjected to trend normalization and standardization to obtain a normalized matrix; Based on the objective weight vector and the subjective weight vector, the comprehensive weight vector of each indicator is calculated. Calculate the product of the normalized matrix and the comprehensive weight vector to obtain the weighted normalized matrix; The maximum and minimum values ​​of each index are extracted from the weighted normalized matrix; Construct a positive ideal solution based on the maximum value of all indicators, and construct a negative ideal solution based on the minimum value of all indicators; Based on the positive ideal solution and the weighted normalization matrix, the first Euclidean distance between each problem path and the positive ideal solution is calculated. Based on the negative ideal solution and the weighted normalization matrix, the second Euclidean distance between each problem path and the negative ideal solution is calculated. The relative proximity of each problem route is calculated based on the first Euclidean distance and the second Euclidean distance.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments; This invention provides a prioritization device for upgrading problematic distribution network lines, comprising: The module includes: data acquisition module, digital model construction module, problem route identification module, indicator calculation module, matrix construction module, objective weight calculation module, subjective weight calculation module, relative proximity calculation module, and route modification ranking module. The data acquisition module is used to acquire raw power operation data, raw meteorological data, equipment ledger data and topology connection relationships of the power distribution network; The digital model building module is used to construct a digital model of the power distribution network based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships. The problematic line identification module is used to identify problematic lines in the distribution network based on the digital model of the distribution network, a preset rule library for diagnosing explicit line problems, and a preset model for identifying implicit line problems. The index calculation module is used to calculate the values ​​of various indicators for each problematic line based on the digital model of the power distribution network. The matrix construction module is used to construct the original data matrix based on the values ​​of all indicators of all problematic routes; The objective weight calculation module is used to calculate the entropy weight of each indicator based on the original data matrix, and generate an objective weight vector of all indicators based on all entropy weights. The subjective weight calculation module is used to obtain the subjective weight vector of all indicators based on the analytic hierarchy process. The relative proximity calculation module is used to calculate the relative proximity of each problem line based on the original data matrix, objective weight vector, and subjective weight vector. The line modification sorting module is used to sort all problematic lines according to their relative proximity from largest to smallest, and to perform line modifications based on the sorting results.

[0013] Based on the above method embodiments, the present invention provides a corresponding terminal device embodiment; The present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the above-described method for prioritizing the transformation of problematic distribution network lines according to any embodiment of the present invention.

[0014] Based on the above method embodiments, the present invention provides a corresponding storage medium embodiment; The present invention provides a storage medium including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the above-described method for prioritizing the transformation of problematic distribution network lines according to any embodiment of the present invention.

[0015] The embodiments of the present invention have the following beneficial effects: This invention provides a method, apparatus, terminal equipment, and storage medium for prioritizing the renovation of problematic power distribution network lines. The method includes: acquiring raw power operation data, raw meteorological data, equipment ledger data, and topological connections of the power distribution network; constructing a digital model of the power distribution network based on the raw power operation data, raw meteorological data, equipment ledger data, and topological connections; identifying problematic power distribution network lines based on the digital model, a preset rule base for diagnosing explicit problems, and a preset rule base for identifying implicit problems; calculating the values ​​of various indicators for each problematic power distribution network line based on the digital model; constructing a raw data matrix based on the values ​​of all indicators for all problematic power distribution network lines; calculating the entropy weight of each indicator based on the raw data matrix, and generating an objective weight vector for all indicators based on all entropy weights; obtaining a subjective weight vector for all indicators based on the analytic hierarchy process (AHP); calculating the relative proximity of each problematic power distribution network line based on the raw data matrix, the objective weight vector, and the subjective weight vector; prioritizing the renovation of all problematic power distribution network lines according to their relative proximity from largest to smallest, and renovating the lines based on the priority ranking results. In this invention, firstly, multiple indicators are set, and the objective weight of each indicator is determined based on the entropy weight method. In the entropy weight method, the greater the numerical difference between lines under the same indicator, the more distinguishing information the indicator provides, and the entropy weight method will automatically assign a higher weight to this indicator. Secondly, the TOPSIS method is used to calculate the relative proximity to reflect the overall quality of the lines. The final ranking process is transformed from the subjective experience of experts in existing technologies into a standardized decision-making process based on objective data calculation. This ensures that the priority ranking of each line is completely determined by the quantitative performance of its indicator data, fundamentally eliminating subjective arbitrariness and improving the rationality of resource allocation. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for prioritizing the upgrading of problematic distribution network lines, provided in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of a priority sorting device for upgrading problematic distribution network lines according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0021] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0024] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0025] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0026] See Figure 1 To address the issue that existing methods for prioritizing upgrades rely on personal experience, leading to unreasonable resource allocation during the final upgrade process, this invention provides a method for prioritizing upgrades of problematic distribution network lines, comprising: Step S101: Obtain the original power operation data, original meteorological data, equipment ledger data, and topology connection relationships of the distribution network; Specifically, the aforementioned raw power operation data consists of telemetry data such as line current, voltage, active power, and reactive power obtained from the SCADA system, as well as remote signaling data such as switch changes and protection actions.

[0027] Specifically, the raw meteorological data is obtained from the meteorological department's data service interface, including meteorological data such as temperature, humidity, wind speed, and icing thickness in the areas where each line is located.

[0028] Specifically, the aforementioned equipment ledger data and topology connection data can both be exported from the PMS system.

[0029] Preferably, in this invention, the aforementioned original power operation data, original meteorological data, equipment ledger data, and topology connection relationships are all data from the previous quarter, and are all collected uniformly at the beginning of each quarter.

[0030] Step S102: Based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships, a digital model of the power distribution network is constructed; Specifically, the aforementioned digital distribution network model is an integrated system encompassing topology, equipment inventory, operational data, and environmental information. This model serves as a unified digital foundation, providing comprehensive, standardized, and high-quality data support for subsequent identification and prioritization of problematic lines. The digital distribution network model is updated at the beginning of each quarter based on relevant data from the previous quarter.

[0031] In a preferred embodiment, the step of constructing a digital distribution network model based on the original power operation data, original meteorological data, equipment ledger data, and topological connectivity includes: Based on the equipment ledger data and topology connection relationships, determine the power equipment of the power distribution network, the electrical connection relationships between each power equipment, the equipment attribute information of the power equipment, and the connection attribute information of the electrical connection relationships. Using the power equipment as nodes, the connection relationships between the power equipment as edges, the equipment attribute information of each power equipment as node attribute information, and the connection attribute information as edge attribute information, a power distribution network topology model is constructed. Specifically, for the complex mesh structure of distribution networks, the Neo4j graph database can be used to construct the distribution network topology model. The specific implementation involves abstracting power equipment such as feeders, sectionalizing switches, tie switches, distribution transformers, and towers as nodes in the distribution network topology model, and abstracting the electrical connections between equipment as edges between nodes.

[0032] Subsequently, a detailed attribute set is defined for each node, namely the node attribute information mentioned above. For example, the node attribute information for the node corresponding to the feeder includes attributes such as line number, model, length, commissioning date, and rated current carrying capacity; the node attribute information for the node corresponding to the switch includes attributes such as switch number, type, normal opening and closing status, and number of operations; the node attribute information for the node corresponding to the distribution transformer includes attributes such as transformer number, capacity, turns ratio, wiring group, and load type. Similarly, attributes are also defined for the connection edge, namely the edge attribute information mentioned above, such as connection type, conductor type, and connection method.

[0033] Preferably, this node-edge structure allows the distribution network topology model to intuitively recreate the actual connectivity of the network structure, providing efficient query capabilities for subsequent topology analysis (such as power supply range tracing and transfer path search). The graph database construction process is as follows: export equipment ledger data and topology connection relationships from the PMS system; after data cleaning and format conversion, nodes and edges are created in batches using Cypher statements; simultaneously, a mapping relationship between equipment IDs and nodes is established to ensure the accuracy of subsequent data association; based on this, corresponding node attribute information and edge attribute information are added.

[0034] Identify the target power equipment corresponding to the original power operation data and original meteorological data, and associate the original power operation data and original meteorological data with the node corresponding to the target power equipment based on the target power equipment; Specifically, based on the completed topology modeling, raw power operation data and raw meteorological data from different sources are linked and integrated with the graph database. First, an index is created based on the timestamps corresponding to the raw power operation data, storing it as time-series information for the corresponding nodes. Considering the time scale differences between the raw meteorological data and the raw power operation data, an interpolation algorithm is used to align the hourly meteorological data with the minute-level operation data. Finally, after outlier detection and missing value completion, measurement errors and communication interference are eliminated. Using the device ID as the primary key, the three preprocessed data items are linked to the corresponding nodes, achieving an integrated representation of the network structure, operating status, and environmental factors.

[0035] Preferably, after fusion processing, each node not only has static ledger attributes, but also carries dynamic runtime time series data and environmental tags, laying a data foundation for subsequent multi-dimensional feature extraction.

[0036] All node attribute information, corresponding raw power operation data, and raw meteorological data are processed to unify units, semantics, and naming to obtain the digital model of the power distribution network.

[0037] Specifically, to ensure semantic consistency and interoperability of data from different sources, the data is standardized according to the Common Information Model (CIM) standard. First, a mapping relationship is established between the CIM model and the node attribute information of each node, mapping core CIM packages such as equipment, measurement, and topology classes to node types and attribute definitions. Second, measurement data is named and unit-converted according to CIM specifications to ensure that data reported by different systems has unified semantics, units, and naming. Finally, topology connections are standardized according to CIM specifications, clarifying the electrical node numbers and connection types of equipment connection nodes.

[0038] In this preferred embodiment, a digital model of the power distribution network is constructed based on raw power operation data, raw meteorological data, equipment ledger data, and topological connection relationships.

[0039] Step S103: Based on the digital model of the distribution network, the preset rule base for diagnosing explicit problems of lines, and the preset rule base for identifying implicit problems of lines, determine the problematic lines of the distribution network; Specifically, based on distribution network operation procedures and long-term accumulated operation and maintenance experience, a series of quantifiable and simple judgment rules are extracted to generate the aforementioned preset rule base for diagnosing explicit problems in the lines. Once the relevant data in the distribution network digital model triggers a rule in the rule base, it indicates that the corresponding line has an explicit risk problem. Here, explicit risk problems refer to common problems that can be directly judged based on rule thresholds.

[0040] As an illustration, the rule base can be divided into several categories according to the type of question: The first category is the bottleneck rule for the communication channel: including situations where the line load rate exceeds 80% and the communication line load rate also exceeds 80%, a single power supply with no backup communication channel, and a communication switch that is disconnected for a long time and cannot be used for power transfer.

[0041] The second category is low-voltage risk rules, which include end-user voltage below 9.3kV for more than 1 hour, daily voltage fluctuation exceeding 15%, and voltage compliance rate below 98%.

[0042] The third category is rules for hidden faults, including more than 3 fault trips in the past three months, more than 2 failed reclosing attempts after the protection device has been activated, and frequent abnormal reports from the fault indicator.

[0043] The fourth category is rules related to excessively long power supply radii, including line lengths exceeding 15 kilometers, end-user distances exceeding 10 kilometers from the power source, and line voltage drops exceeding 8%.

[0044] Specifically, the aforementioned preset line hidden problem identification model is an XGBoost model that can identify hidden risks in the line. Hidden risks refer to complex hidden dangers that are difficult to determine directly through simple rule thresholds, including but not limited to: potential high failure risks caused by the coupling of multiple factors, voltage over-limit or protection adaptation risks caused by the access of new energy sources, early power quality degradation risks, and health risks coupled with equipment aging and the environment.

[0045] In a preferred embodiment, the problematic lines of the distribution network are determined based on the digital model of the distribution network, a preset rule base for diagnosing explicit line problems, and a preset rule base for identifying implicit line problems, including: Line information data for each line is extracted from the digital model of the power distribution network. Specifically, the aforementioned line information data refers to the power operation data and meteorological data corresponding to each line extracted from the digital model of the distribution network and processed by unit unification, semantic unification and naming unification.

[0046] Based on the route information data, determine the monitoring index values ​​of the preset monitoring indicators for each route; Specifically, the monitoring indicators are determined based on a pre-set rule base for diagnosing explicit problems in the line. For example, if there is a rule in the rule base that says "daily voltage fluctuation exceeds 15%", then the corresponding monitoring indicator is "daily voltage fluctuation", which can be directly determined based on the voltage value in the line information data.

[0047] For each route, the monitoring index values ​​are compared with a preset rule base for diagnosing explicit problems in routes to identify routes with explicit risk issues. Specifically, for any line, its state vector can be defined as: In the formula, For n different monitoring indicators, This represents the state vector of line L.

[0048] Subsequently, the rule set in the preset line explicit problem diagnosis rule base is defined as follows: In the formula, Representing a set of rules, This represents m distinct rules.

[0049] Specifically, the j-th rule can be represented as a conditional judgment function: ,when At that time, line L was marked as having The corresponding problem type is that route L is a route with explicit risk issues.

[0050] Based on the line information data, construct fault statistics feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector; Specifically, the fault statistical feature vector includes: annual fault frequency, monthly fault frequency, fault type distribution vector, and reclosing success rate, etc.

[0051] Specifically, the operational status feature vector includes: average load rate, load rate volatility, daily maximum load rate, daily minimum load rate, and peak-to-valley difference rate. The peak-to-valley difference rate is calculated based on the following formula: In the formula, Indicates the peak-to-valley difference rate. Indicates the maximum daily load rate. This indicates the minimum daily load rate.

[0052] Specifically, power quality feature vectors include: daily voltage deviation, voltage fluctuation rate, three-phase unbalance, and harmonic distortion rate. Among these, daily voltage deviation, voltage fluctuation rate, three-phase unbalance, and harmonic distortion rate are calculated based on the following formula: In the formula, Indicates the daily voltage deviation. This represents the voltage value at time t. Indicates the rated voltage of the line. Indicates time period, Indicates voltage fluctuation rate. Indicates the maximum voltage value. Indicates the minimum voltage value. Indicates the degree of three-phase imbalance. Indicates the maximum current value. Indicates the minimum current value. This represents the average current value. Indicates harmonic distortion rate. This represents the root mean square value of the h-th harmonic voltage. This represents the root mean square value of the fundamental voltage.

[0053] Specifically, the feature vectors related to the impact of new energy sources include: photovoltaic grid connection capacity, reverse power transmission period, and maximum reverse power transmission value.

[0054] Specifically, environmental factor feature vectors include: number of days with lightning strikes, number of days with icing, and number of times typhoons affect the area.

[0055] Preferably, all feature vectors of each line can be integrated and represented as: In the formula, This represents the integrated feature vector of the line. Let n' represent the n' distinct eigenvectors of the line.

[0056] The fault statistical feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector are input into a preset line hidden problem identification model to obtain the confidence level that the line has hidden risks. Specifically, after inputting the above feature vectors into the preset line hidden problem identification model, the model will output the problem probability of the line, that is, the confidence level that the line has hidden risk problems.

[0057] Lines with a confidence level not less than a preset confidence threshold are considered as lines with hidden risks. Preferably, for lines with a confidence level lower than a preset confidence threshold but not lower than another preset confidence threshold, the line is marked as pending review for a second manual review to determine whether it is a problem line. The value of the aforementioned "another preset confidence threshold" is smaller than the preset confidence threshold. Lines with a confidence level lower than the other preset confidence threshold are not considered as hidden problem lines with latent risks.

[0058] The explicit problem path and the implicit problem path are collectively referred to as the problem path.

[0059] Preferably, if a certain line is both an explicit problem line and a implicit problem line, it indicates that the line is a high-risk line, and it can be combined and recorded as a single problem line during the recording process.

[0060] Preferably, by combining the rule base with the preset hidden problem identification model of the line, this step not only ensures efficient screening of common problems, but also enhances the ability to perceive complex problems, realizing a leap from single-point alarm to intelligent diagnosis.

[0061] In this preferred embodiment, the problematic lines of the distribution network are identified based on the digital model of the distribution network, the preset rule base for diagnosing explicit line problems, and the preset rule base for identifying implicit line problems.

[0062] In another preferred embodiment, the training of the preset line latent problem identification model includes: Obtain first-line information data for each historically problematic line and second-line information data for each historically unproblematic line; Specifically, based on historical fault records and manual troubleshooting records, we can identify historically problematic lines and lines without historically problematic lines, and then further obtain the corresponding line information data.

[0063] Based on the first route information data and the second route information data, a historical multidimensional feature vector is constructed for the corresponding historical route; wherein, the historical route includes: historical routes with hidden problems and historical routes without hidden problems; Specifically, the features contained in the aforementioned historical multidimensional feature vector are the same as those in the fault statistics feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector.

[0064] A training sample set is generated based on the historical multidimensional feature vectors with real labels; wherein, the real labels are used to indicate whether the corresponding historical line is a historical line with hidden problems or a historical line without hidden problems. Specifically, the true label is used to characterize whether the historical route corresponding to the historical multidimensional feature vector is a hidden problem route. The true label for a historical route without hidden problems can be marked as 0, and the true label for a historical route with hidden problems can be marked as 1.

[0065] The training sample set is input into the hidden line problem identification model to be trained, so that the hidden line problem identification model to be trained is iteratively trained based on the gradient boosting algorithm until the preset number of iterations is reached, and the preset hidden line problem identification model is obtained after training. Preferably, before inputting the training sample set into the model, Z-score normalization is first applied to normalize the training samples, followed by SMOTE algorithm for sample balancing to avoid class imbalance affecting model performance. The formula for Z-score normalization is: In the formula, Let j represent the j-th standardized feature vector. This represents the j-th feature vector before standardization. Let represent the mean of the j-th eigenvector. Let represent the standard deviation of the j-th eigenvector.

[0066] In each iteration, the training sample set is input into the hidden problem identification model of the railway line to be trained to generate the current predicted value; a new decision tree is added with the goal of fitting the residual between the current predicted value and the corresponding true label, and the structure and leaf node weights of the newly added decision tree are determined by minimizing the objective function; it is determined whether the current iteration round has reached the preset iteration round; if so, the preset hidden problem identification model of the railway line is generated after training; otherwise, the next iteration round is continued based on the training sample set. The objective function is constructed based on the current model prediction value; the initial prediction value is a preset initial prediction value.

[0067] Specifically, during the training phase, a strong classifier is constructed by integrating multiple decision tree models based on the XGBoost algorithm, thereby obtaining a pre-defined hidden problem identification model for railway lines. The core idea is to add a new tree in each iteration to fit the residual of the previous tree. The pre-defined hidden problem identification model for railway lines can be represented as: In the formula, This represents the predicted output of the preset hidden problem identification model for training sample i, where i represents training sample i, and K represents the number of decision trees. Indicates based on input feature vector The predicted output of the k-th decision tree, where F represents the decision tree function space. This represents the predicted output of the k-th decision tree.

[0068] Specifically, the objective function described above consists of two parts: a loss function and a regularization term, expressed as follows: In the formula, Describe the objective function. This represents the set of parameters for the model, and N represents the total number of training samples. This represents the loss function corresponding to training sample i. This represents the true label of training sample i. Let represent the regularization term of the k-th decision tree.

[0069] In this invention, the loss function is a logarithmic loss function: In the formula, This represents the value of the loss function. Indicates the true label, This represents the model's predicted output.

[0070] The regularization term is used to control the model complexity, and its expression is: In the formula, and Represents the regularization parameter. This represents the number of leaf nodes in the decision tree. Indicates the first The weight of each leaf node.

[0071] Specifically, the model employs an additive training strategy during training. First, the training sample set is input into the hidden problem identification model to be trained, and iterative training begins. Before the first iteration, a preset initial prediction value is set for all training samples. This initial prediction value is typically set as the log probability of the proportion of positive samples in the training sample set. In each subsequent iteration, all training samples are input into the current model to obtain the current prediction value for each sample. A new decision tree is then added with the goal of fitting the residual between the current prediction value and the corresponding true label.

[0072] Specifically, the newly added decision tree determines its structure and the weights of each leaf node during its growth process by minimizing the objective function. When determining the tree structure, a greedy algorithm is used to traverse each feature and its candidate split points, selecting the feature and threshold that maximize the split gain as the node splitting condition. The split gain is calculated based on the sample sets of the left and right child nodes after the split, and the optimal weight of each leaf node is calculated based on the first and second gradients of all samples within that node.

[0073] Specifically, after each decision tree is constructed, it is determined whether the current iteration round has reached the preset number of iteration rounds. If not, the newly added decision tree is added to the current model, the current predicted value of each sample is updated, and the residual is recalculated based on the updated predicted value, and the next iteration round is entered. This process is repeated until the preset number of iteration rounds is reached. At this point, the initial predicted value and the cumulative result of all the decision trees generated in the iterations constitute the trained hidden problem identification model of the line.

[0074] Specifically, during training, the optimal combination of hyperparameters for the model is determined through k-fold cross-validation. The combination of hyperparameters includes at least the maximum depth of the tree, the learning rate, the minimum leaf node weight, and the column sampling ratio. After training, the trained model for identifying hidden problems in railway lines and the feature importance score of each feature are output. The feature importance score is calculated by weighting the number of times each feature is selected as a split point in all decision trees, and is used to characterize the contribution of each feature to identifying hidden problems in railway lines.

[0075] Indicative, For the first Training samples after round of iterations The predicted value, then the first The goal of the wheel is: In the formula, This represents the objective function value in the t-th iteration. This represents the predicted value of training sample i at iteration t-1. This indicates that in the t-th iteration, based on the input feature vector... The decision tree output, This represents the regularization term at iteration t.

[0076] Approximating the loss function with a second-order Taylor expansion, we get: In the formula, This represents the loss function value in the t-th iteration. This represents the first-order gradient of training sample i. Let represent the second gradient of training sample i.

[0077] Reorganizing the samples according to the leaf nodes of the decision tree, the objective function can be rewritten as: In the formula, Indicates the division to the th A sample set of 'leaf nodes' Functions representing tree structures.

[0078] For a fixed tree structure The optimal weight for each leaf node is: In the formula, Indicates the first The optimal weight of each leaf node.

[0079] Substituting into the objective function yields the optimal value: In the formula, This represents the optimal objective function value.

[0080] Specifically, the splitting gain is calculated based on the following formula: In the formula, Indicates split gain. Represents the sample set of the left child node. This represents the sample set of the right child node. This represents the complete set of samples for the parent node to be split.

[0081] Specifically, after training, the feature importance score is calculated based on the following formula: In the formula, This represents the feature importance score of the j-th feature, used to measure the contribution of this feature to the prediction in the XGBoost model. j represents the index of the feature, and k represents the index of the decision tree. Represents the nodes in the decision tree. Let k be the decision tree. This indicates an indicator function that takes a value of 1 when the selected feature for a node split is the j-th feature, and a value of 0 otherwise. This represents the segmentation feature selected during a node split. This represents the split gain of the node.

[0082] In this preferred embodiment, a preset line latent problem identification model based on XGBoost is trained and generated based on the first line information data of each historical latent problem line and the second line information data of each historical line without latent problems.

[0083] Step S104: Based on the digital model of the distribution network, calculate the values ​​of various indicators for each problematic line. Specifically, this invention constructs a multi-level evaluation index system from three dimensions: safety, economy, and feasibility. Each dimension is further subdivided into quantifiable underlying indicators to ensure the comprehensiveness and operability of the evaluation.

[0084] Therefore, the above indicators include safety dimension indicators, economic dimension indicators, and feasibility dimension indicators. Among them, the safety dimension indicators mainly reflect the degree of impact of problematic lines on the safe operation of the power grid and the power supply to users. Specifically, they include: the number of affected users, which refers to the total number of users supplied by the problematic line, reflecting the breadth of the fault's impact; the proportion of important users, which refers to the proportion of important users such as government agencies, hospitals, and schools to the total number of users; the historical fault rate, which refers to the average number of line faults over the past three years converted to the value per 100 kilometers per year; the power outage loss, which is the power loss estimated based on the historical average power outage time and average load; and the voltage over-limit duration, which refers to the cumulative number of hours of voltage over-limit per year, characterizing power quality issues.

[0085] The economic dimension indicators mainly reflect the economic benefits and input-output ratio of the problem line renovation, specifically including: estimated renovation cost, which is the total investment of the project estimated based on line length, number of equipment, etc.; unit investment loss reduction benefit, which refers to the ratio of the expected reduction in power outage losses after renovation to the investment; line loss improvement space, which refers to the expected reduction in annual line loss after renovation; power supply capacity improvement, which refers to the increased power supply capacity after renovation; and investment payback period, which refers to the expected number of years to recover the investment through the reduction of power outage losses.

[0086] The feasibility dimension indicators mainly reflect the implementation difficulty and external conditions of the renovation project, specifically including: construction difficulty, which is a quantitative score considering factors such as terrain conditions, number of crossings, and difficulty of compensation; policy support, which is a score for whether it is included in the government's key projects or has obtained planning permission; material procurement cycle, which is the expected procurement time of major equipment and materials; and power outage coordination difficulty, which is a score for the degree of impact of power outages required for the renovation on users.

[0087] It should be noted that the above indicators should be calculated or assigned according to a unified standard based on data obtained from the digital model of the distribution network for each problematic line.

[0088] Step S105: Based on the values ​​of all indicators for all problematic routes, construct the original data matrix: Specifically, based on the values ​​of various indicators for all problematic routes, the original data matrix is ​​constructed as follows: In the formula, Represents the original data matrix. Indicates the first The first of the problematic routes 'n' represents the number of problematic lines, and 'm' represents the number of indicators.

[0089] Step S106: Calculate the entropy weight of each indicator based on the original data matrix, and generate an objective weight vector for all indicators based on all entropy weights. Specifically, this invention uses the entropy weight method to first calculate the entropy weight of each indicator, and then uses the entropy weight as the corresponding objective weight vector.

[0090] In a preferred embodiment, the entropy weight of each index is calculated based on the original data matrix, including: The original data matrix is ​​standardized to obtain a standardized matrix; Specifically, the data in the original data matrix is ​​standardized to eliminate the influence of dimensions, resulting in a standardized matrix. Different standardization formulas are used for positive indicators (the larger the better) and negative indicators (the smaller the better). For positive indicators, the following formula is used for standardization: In the formula, Represents the first in the normalized matrix The first of the '' routes ''Indicator value, Represents the first element in the original data matrix. The minimum value of each indicator. Represents the first element in the original data matrix. The maximum value of each indicator.

[0091] For negative indicators, standardization is performed using the following formula: Based on the standardized matrix, the characteristic weight of each problem line under each indicator is calculated; Specifically, the characteristic weight of each problem line under each indicator is calculated using the following formula: In the formula, Indicates the first The first of the problematic routes The characteristic proportion of each indicator.

[0092] Based on the weight of the features, the information entropy of each indicator is calculated; Specifically, the information entropy is calculated using the following formula: In the formula, Indicates the first The first of the '' routes The information entropy of each indicator.

[0093] It should be noted that the regulations stipulate that when hour, Information entropy The smaller the value, the greater the data variation of the indicator, and the more information it provides.

[0094] Based on the information entropy, the entropy weight of each indicator is calculated.

[0095] Specifically, the entropy weight is calculated using the following formula: In the formula, Indicates the first 'Indicator entropy weight.'

[0096] In this preferred embodiment, the entropy weight of each index is calculated based on the original data matrix.

[0097] Step S107: Obtain the subjective weight vector of all indicators based on the analytic hierarchy process; Specifically, this invention invites several experts in the fields of distribution network planning and operation to compare indicators at the same level in pairs, construct a judgment matrix, and then determine the subjective weight vector based on the judgment matrix.

[0098] In a preferred embodiment, the subjective weight vector of all indicators is obtained based on the analytic hierarchy process (AHP), including: Obtain the judgment matrix constructed by each expert for all indicators; Specifically, the judgment matrix uses a 1-9 scale, where 1 indicates that the two indicators are equally important, and 9 indicates that one indicator is absolutely more important than the other. Each element in the judgment matrix represents the comparison result of the importance between the corresponding two indicators.

[0099] For each expert, calculate the maximum eigenvalue and eigenvector of the judgment matrix constructed by the current expert; Specifically, the calculation of the maximum eigenvalue and eigenvector is based on existing technology and will not be elaborated here.

[0100] The feature vectors are normalized to obtain the initial subjective weight vectors of all indicators, and then the initial subjective weight vectors of each expert are obtained. Specifically, the resulting initial subjective weight vector can be represented as: In the formula, Let represent the initial subjective weight vector corresponding to the k'-th expert. This represents the initial subjective weight of the k'-th expert for each indicator.

[0101] For each initial subjective weight vector, the consistency index value is calculated based on the maximum eigenvalue corresponding to the current initial subjective weight vector. Specifically, the consistency index value is calculated based on the following formula: In the formula, This represents the consistency index corresponding to the initial subjective weight vector obtained by the k'-th expert. Let represent the largest eigenvalue of the judgment matrix constructed by the k'th expert.

[0102] The consistency ratio is calculated based on the consistency index value; Specifically, the consistency ratio is calculated using the following formula: In the formula, This represents the consensus ratio corresponding to the k'th expert. This represents the random consistency index (which can be obtained from a table).

[0103] The initial subjective weight vectors with a consistency ratio less than a preset ratio threshold are removed to obtain the effective subjective weight vectors. Specifically, the preset ratio threshold is set to 0.1. That is, if the consistency ratio is less than 0.1, the consistency of the judgment matrix constructed by the corresponding expert is acceptable; otherwise, the judgment matrix of the corresponding expert is removed, that is, the corresponding initial subjective weight vector is removed.

[0104] Calculate the arithmetic mean of all valid subjective weight vectors to obtain the subjective weight vector for each indicator.

[0105] Specifically, the subjective weight vector is calculated using the following formula: In the formula, Represents the subjective weight vector. This indicates the number of experts who passed the consistency test.

[0106] In this preferred embodiment, the subjective weight vector of all indicators was obtained based on the analytic hierarchy process.

[0107] Step S108: Calculate the relative proximity of each problem route based on the original data matrix, objective weight vector, and subjective weight vector; Specifically, this invention uses the TOPSIS method to calculate the relative proximity of each problem path and achieve priority ranking.

[0108] In a preferred embodiment, the relative proximity of each problem path is calculated based on the original data matrix, the objective weight vector, and the subjective weight vector, including: The values ​​of each indicator in the original data matrix are subjected to trend normalization and standardization to obtain a normalized matrix; Specifically, through trend-following processing, all indicators are transformed into positive indicators. Indicators already positive remain unchanged, while inverse indicators are transformed using the reciprocal method or by taking the negative value. Then, the range method is used to normalize the data, resulting in a normalized matrix. The elements in the normalized matrix are determined based on the following formula: In the formula, The normalized matrix represents the first... The first of the '' routes The element value corresponding to each indicator. This indicates that after the trend-following process, the first... The first of the '' routes The element value corresponding to each indicator. This indicates the first trend after the trending process. The minimum value of each indicator. This indicates the first trend after the trending process. The maximum value of each indicator.

[0109] Based on the objective weight vector and the subjective weight vector, the comprehensive weight vector of each indicator is calculated. Specifically, this invention obtains a comprehensive weight vector by linearly weighting the objective weight vector and the subjective weight vector: In the formula, Represents the overall weight vector. Represents the empirical weighting coefficient. This represents the objective weight vector.

[0110] Preferred, The value range is generally 0.3 to 0.7, and can be adjusted according to the actual scenario. This solution recommends using a value of [value missing]. This achieves a balance between subjective experience and objective information.

[0111] Calculate the product of the normalized matrix and the comprehensive weight vector to obtain the weighted normalized matrix; Specifically, the normalized matrix is ​​multiplied by the comprehensive weight vector to obtain the weighted normalized matrix. The elements of the weighted normalized matrix are obtained by the following formula: In the formula, In the weighted normalized matrix, the first... The first of the '' routes The value of the element corresponding to each index value. Represents the first element in the comprehensive weight vector. The comprehensive weight corresponding to each indicator value.

[0112] The maximum and minimum values ​​of each index are extracted from the weighted normalized matrix; Construct a positive ideal solution based on the maximum value of all indicators, and construct a negative ideal solution based on the minimum value of all indicators; Specifically, the ideal solution is composed of the maximum values ​​of each index in the weighted normalization matrix, expressed as: In the formula, This represents the ideal solution. This represents the maximum value of each of the m'' indicators in the weighted normalized matrix. In the weighted normalized matrix, the first... The first of the problematic routes The indicator value corresponding to each indicator.

[0113] Specifically, the negative ideal solution is composed of the minimum values ​​of each index in the weighted normalization matrix, expressed as: In the formula, This represents a negative ideal solution. This represents the minimum value of m'' indicators in the weighted normalized matrix.

[0114] Based on the positive ideal solution and the weighted normalization matrix, the first Euclidean distance between each problem path and the positive ideal solution is calculated. Specifically, the first Euclidean distance is calculated using the following formula: In the formula, Indicates the first The first Euclidean distance of the problematic route.

[0115] Based on the negative ideal solution and the weighted normalization matrix, the second Euclidean distance between each problem path and the negative ideal solution is calculated. Specifically, the second Euclidean distance is calculated using the following formula: In the formula, Indicates the first The second Euclidean distance of the problematic route.

[0116] The relative proximity of each problem route is calculated based on the first Euclidean distance and the second Euclidean distance.

[0117] Specifically, the relative similarity is calculated using the following formula: In the formula, Indicates the first The relative proximity of the problematic routes.

[0118] Preferred, The value ranges from 0 to 1. The closer it is to 1, the closer the problem line is to the ideal optimal state, that is, the more urgent and worth prioritizing the modification; the closer it is to 0, the lower the priority of modification.

[0119] In this preferred embodiment, the relative proximity of each problem path is calculated based on the original data matrix, the objective weight vector, and the subjective weight vector.

[0120] Step S109: Sort all problematic lines according to their relative proximity from largest to smallest, and then modify the lines according to the priority ranking results.

[0121] Specifically, according to The problematic routes are sorted from highest to lowest value to generate a priority list for modification. In addition to the sorting results, the list can also include the key indicator values, weighting coefficients, and proximity calculation process for each route, facilitating review and adjustment by management personnel.

[0122] Preferably, this step transforms the complex information of multiple problematic routes into an intuitive priority ranking result, providing a scientific and quantifiable decision-making basis for project approval and funding allocation.

[0123] Preferably, the newly generated priority list can be automatically pushed to the review module of the visual dashboard and presented in a "pending confirmation" status, highlighting the updated content with color or a red dot to alert managers. The review interface uses a side-by-side comparison method, simultaneously displaying the ranking, relevance, key indicators, and historical data from the previous period for the recommended transformation priorities in this round, making it easy for managers to intuitively compare changes.

[0124] Preferably, managers can review and confirm the results in the list item by item or in batches. If the current sorting result is approved, clicking "Confirm" will officially add the record to the current priority list and mark it as reviewed. If a discrepancy is found between the current sorting result and the actual situation on site, clicking "Correct" will enter correction mode, allowing manual adjustment of the problem type, modification of indicator values, or adjustment of the ranking order. Simultaneously, a correction reason note should be filled in, such as changes in construction conditions discovered during on-site inspections or the addition of important users in user repair requests. The corrected data will not directly overwrite the original calculation results but will be stored as an independent review record for subsequent optimization of related models. All review operations are recorded to ensure that every decision adjustment is traceable.

[0125] Meanwhile, the visual dashboard assists managers in daily monitoring and decision-making with three core modules: the grid problem map module uses a geographic information system as the base map, displaying the spatial distribution and priority level of various problems with markers or line segments of different colors. Clicking on any marker will pop up a details window, displaying key indicators such as problem type, identification time, load rate, number of failures, and their historical trends; the priority list module displays a complete list of lines to be upgraded in tabular form, sorted by proximity, including line name, region, problem type, proximity value, number of affected users, estimated upgrade cost, construction difficulty, and other core decision-making indicators. It supports filtering by problem type, region, and time range, and can export reports; the key indicator trend module displays the changing trends of key indicators such as power supply reliability, comprehensive voltage qualification rate, average line load rate, and number of problem lines over time in the form of curves. It supports viewing at different granularities of daily, weekly, monthly, and yearly, automatically calculates year-on-year and month-on-month changes, and automatically generates comparison curves of key indicators before and after the upgrade project is completed.

[0126] Specifically, once a line renovation project is officially approved, the management personnel will associate the project with the corresponding problematic line and enter the expected start time, planned completion time, and expected target indicators, such as the expected reduction in the number of faults, the improvement in voltage qualification rate, or the reduction in line loss, to generate a renovation tracking file for the line.

[0127] Preferably, after project completion, the system tracks the line's operational data for three consecutive months and compares it with historical data from the same period before the upgrade, generating a detailed upgrade effect evaluation report. The report includes: a comparison table of key indicators before and after the upgrade, clearly showing changes in indicators such as the number of faults, voltage over-limit duration, line loss rate, and average user outage time; a proximity change analysis, which visually reflects the improvement in the grid health by comparing the proximity values ​​of the line in the two priority rankings before and after the upgrade; and a comprehensive effect evaluation, where the system automatically labels indicators as "significant effect," "moderate effect," or "effect below expectations" based on the degree of improvement, and calculates the investment benefit ratio.

[0128] Finally, based on the effect evaluation results, further feedback optimization is promoted: For renovation cases with significant effects, they are included in the excellent case library for reference in subsequent renovation plans for similar problems. At the same time, the data before and after the renovation of these cases are used as positive samples to enhance the training dataset of the preset line hidden problem identification model; For renovation projects with unsatisfactory results, managers review and analyze the reasons. Managers can record the specific reasons, such as design changes, construction quality problems, changes in the external environment, etc. This information will be fed back to the rule base or used as negative samples for the optimization of the preset line hidden problem identification model; According to the effect tracking results, the weight of the corresponding problem in the priority ranking can also be appropriately adjusted. For example, if it is found that the effect of renovation of a certain type of problem is generally poor, the system can appropriately increase the weight of this type of problem in the priority ranking or optimize the judgment threshold of the relevant rules.

[0129] Preferably, through the above process, this method not only provides managers with an intuitive decision support tool, but more importantly, it constructs a closed-loop solution for continuous learning: the effects of each identification, each review, and each transformation are recorded and fed back, driving the continuous optimization of the preset line hidden problem identification model, the continuous improvement of the priority ranking rules, and ultimately achieving a spiral increase in the level of distribution network management.

[0130] Preferably, compared with the current technical solutions that rely on human experience, the present invention achieves a fundamental innovation in the power distribution network management mode by constructing four interconnected stages: data fusion, intelligent identification, quantitative sorting, and dynamic closed loop.

[0131] First, this invention overcomes the shortcomings of data silos and single analysis dimensions in existing technologies. By constructing a digital model based on graph databases, it deeply integrates multi-source data such as SCADA, PMS, and meteorology, providing a unified digital foundation for a comprehensive and accurate assessment of the health status of the grid structure, rather than making partial judgments based solely on reports from a single dimension.

[0132] Secondly, this invention changes the passive and lagging situation of the traditional manual analysis mode. It innovatively adopts a rule engine based on a preset line explicit problem diagnosis rule library to identify explicit problems, and a preset line implicit problem identification model. This can not only quickly screen out explicit problems such as communication channel bottlenecks, but also deeply explore implicit risks that are difficult to judge through simple thresholds, realizing the transformation from "passive emergency repair" to "proactive operation and maintenance".

[0133] Furthermore, in response to the problem of strong subjectivity and arbitrariness in existing project decision-making, this invention constructs a multi-level evaluation index system covering safety, economy, and feasibility, and adopts a combination of AHP-entropy weighting method and TOPSIS method to transform complex information into an objective and quantifiable priority ranking list, providing a scientific decision-making basis for precise resource allocation.

[0134] Ultimately, this invention constructs a closed-loop system for continuous learning, which achieves dynamic updates and manual review through a visual dashboard, and feeds the transformation effect back to the identification model and sorting rules. This solves the problems of fragmented traditional management processes and lack of effect verification, realizes a spiral increase in grid management capabilities, and significantly improves the level of lean management of distribution networks.

[0135] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0136] like Figure 2 As shown, an embodiment of the present invention provides a priority ranking device for upgrading problematic distribution network lines, comprising: The module includes: data acquisition module, digital model construction module, problem route identification module, indicator calculation module, matrix construction module, objective weight calculation module, subjective weight calculation module, relative proximity calculation module, and route modification ranking module. The data acquisition module is used to acquire raw power operation data, raw meteorological data, equipment ledger data and topology connection relationships of the power distribution network; The digital model building module is used to construct a digital model of the power distribution network based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships. The problematic line identification module is used to identify problematic lines in the distribution network based on the digital model of the distribution network, a preset rule library for diagnosing explicit line problems, and a preset model for identifying implicit line problems. The index calculation module is used to calculate the values ​​of various indicators for each problematic line based on the digital model of the power distribution network. The matrix construction module is used to construct the original data matrix based on the values ​​of all indicators of all problematic routes; The objective weight calculation module is used to calculate the entropy weight of each indicator based on the original data matrix, and generate an objective weight vector of all indicators based on all entropy weights. The subjective weight calculation module is used to obtain the subjective weight vector of all indicators based on the analytic hierarchy process. The relative proximity calculation module is used to calculate the relative proximity of each problem line based on the original data matrix, objective weight vector, and subjective weight vector. The line modification sorting module is used to sort all problematic lines according to their relative proximity from largest to smallest, and to perform line modifications based on the sorting results.

[0137] It should be noted that the device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort. The above schematic diagram is merely an example of a priority ranking device for the renovation of problematic distribution network lines and does not constitute a limitation on a priority ranking device for the renovation of problematic distribution network lines. It may include more or fewer components than shown, or combine certain components, or use different components.

[0138] Based on the above method embodiments, the present invention provides corresponding terminal device embodiments.

[0139] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the priority ranking method for upgrading distribution network problem lines described in any embodiment of the present invention.

[0140] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the device. The aforementioned terminal devices can be computing devices such as desktop computers, laptops, handheld computers, and cloud servers. These devices may include, but are not limited to, processors and memory. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the device, connecting various parts of the device via various interfaces and lines. The aforementioned memory can be used to store the aforementioned computer programs and / or modules. The aforementioned processor implements various functions of the aforementioned device by running or executing the computer programs and / or modules stored in the aforementioned memory, and by calling data stored in the memory. The aforementioned memory may mainly include a program storage area and a data storage area, wherein the program storage area may store the operating system, at least one application program required for a function, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0141] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0142] Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is running, it controls the device where the storage medium is located to execute the priority ranking method for upgrading problematic distribution network lines as described in any embodiment of the present invention.

[0143] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include any entity or device capable of carrying the 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.

[0144] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for prioritizing the upgrading of problematic distribution network lines, characterized in that, include: Obtain raw power operation data, raw meteorological data, equipment ledger data, and topology connections of the power distribution network; Based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships, a digital model of the power distribution network is constructed. Based on the digital model of the distribution network, the preset rule base for diagnosing explicit line problems, and the preset rule base for identifying implicit line problems, the problematic lines of the distribution network are determined. Based on the digital model of the distribution network, the values ​​of various indicators for each problematic line are calculated; Based on the values ​​of all indicators for all problematic routes, the original data matrix is ​​constructed. Based on the original data matrix, the entropy weight of each indicator is calculated, and an objective weight vector of all indicators is generated based on all entropy weights. The subjective weight vector of all indicators is obtained based on the analytic hierarchy process. Based on the original data matrix, objective weight vector, and subjective weight vector, the relative proximity of each problem route is calculated. All problematic lines are prioritized for modification according to their relative proximity from highest to lowest, and modifications are then carried out based on the priority ranking.

2. The method for prioritizing the upgrading of problematic distribution network lines according to claim 1, characterized in that, The process of constructing a digital model of the power distribution network based on the original power operation data, original meteorological data, equipment ledger data, and topological connections includes: Based on the equipment ledger data and topology connection relationships, determine the power equipment of the power distribution network, the electrical connection relationships between each power equipment, the equipment attribute information of the power equipment, and the connection attribute information of the electrical connection relationships. Using the power equipment as nodes, the connection relationships between the power equipment as edges, the equipment attribute information of each power equipment as node attribute information, and the connection attribute information as edge attribute information, a power distribution network topology model is constructed. Identify the target power equipment corresponding to the original power operation data and original meteorological data, and associate the original power operation data and original meteorological data with the node corresponding to the target power equipment based on the target power equipment; All node attribute information, corresponding raw power operation data, and raw meteorological data are processed to unify units, semantics, and naming to obtain the digital model of the power distribution network.

3. The method for prioritizing the renovation of problematic distribution network lines according to claim 2, characterized in that, Based on the digital model of the distribution network, the preset rule base for diagnosing explicit line problems, and the preset rule base for identifying implicit line problems, the problematic lines of the distribution network are determined, including: Line information data for each line is extracted from the digital model of the power distribution network. Based on the route information data, determine the monitoring index values ​​of the preset monitoring indicators for each route; For each route, the monitoring index values ​​are compared with a preset rule base for diagnosing explicit problems in routes to identify routes with explicit risk issues. Based on the line information data, construct fault statistics feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector; The fault statistical feature vector, operating status feature vector, power quality feature vector, new energy impact feature vector, and environmental factor feature vector are input into a preset line hidden problem identification model to obtain the confidence level that the line has hidden risks. Lines with a confidence level not less than a preset confidence threshold are considered as lines with hidden risks. The explicit problem path and the implicit problem path are combined as the problem path.

4. The method for prioritizing the upgrading of problematic distribution network lines according to claim 3, characterized in that, The training of the preset line hidden problem identification model includes: Obtain first-line information data for each historically problematic line and second-line information data for each historically unproblematic line; Based on the first route information data and the second route information data, a historical multidimensional feature vector is constructed for the corresponding historical route; wherein, the historical route includes: historical routes with hidden problems and historical routes without hidden problems; A training sample set is generated based on the historical multidimensional feature vectors with real labels; wherein, the real labels are used to indicate whether the corresponding historical line is a historical line with hidden problems or a historical line without hidden problems. The training sample set is input into the hidden line problem identification model to be trained, so that the hidden line problem identification model to be trained is iteratively trained based on the gradient boosting algorithm until the preset number of iterations is reached, and the preset hidden line problem identification model is obtained after training. In each iteration, the training sample set is input into the hidden problem identification model of the railway line to be trained to generate the current predicted value; a new decision tree is added with the goal of fitting the residual between the current predicted value and the corresponding true label, and the structure and leaf node weights of the newly added decision tree are determined by minimizing the objective function; it is determined whether the current iteration round has reached the preset iteration round; if so, the preset hidden problem identification model of the railway line is generated after training; otherwise, the next iteration round is continued based on the training sample set. The objective function is constructed based on the current model prediction value; the initial prediction value is a preset initial prediction value.

5. The method for prioritizing the renovation of problematic distribution network lines according to claim 4, characterized in that, Based on the original data matrix, the entropy weight of each indicator is calculated, including: The original data matrix is ​​standardized to obtain a standardized matrix; Based on the standardized matrix, the characteristic weight of each problem line under each indicator is calculated; Based on the weight of the features, the information entropy of each indicator is calculated; Based on the information entropy, the entropy weight of each indicator is calculated.

6. The method for prioritizing the renovation of problematic distribution network lines according to claim 5, characterized in that, The subjective weight vectors of all indicators were obtained based on the analytic hierarchy process, including: Obtain the judgment matrix constructed by each expert for all indicators; For each expert, calculate the maximum eigenvalue and eigenvector of the judgment matrix constructed by the current expert; The feature vectors are normalized to obtain the initial subjective weight vectors of all indicators, and then the initial subjective weight vectors of each expert are obtained. For each initial subjective weight vector, the consistency index value is calculated based on the maximum eigenvalue corresponding to the current initial subjective weight vector. The consistency ratio is calculated based on the consistency index value; The initial subjective weight vectors with a consistency ratio less than a preset ratio threshold are removed to obtain the effective subjective weight vectors. Calculate the arithmetic mean of all valid subjective weight vectors to obtain the subjective weight vector for each indicator.

7. The method for prioritizing the renovation of problematic distribution network lines according to claim 6, characterized in that, Based on the original data matrix, objective weight vector, and subjective weight vector, the relative proximity of each problem route is calculated, including: The values ​​of each indicator in the original data matrix are subjected to trend normalization and standardization to obtain a normalized matrix; Based on the objective weight vector and the subjective weight vector, the comprehensive weight vector of each indicator is calculated. Calculate the product of the normalized matrix and the comprehensive weight vector to obtain the weighted normalized matrix; The maximum and minimum values ​​of each index are extracted from the weighted normalized matrix; Construct a positive ideal solution based on the maximum value of all indicators, and construct a negative ideal solution based on the minimum value of all indicators; Based on the positive ideal solution and the weighted normalization matrix, the first Euclidean distance between each problem path and the positive ideal solution is calculated. Based on the negative ideal solution and the weighted normalization matrix, the second Euclidean distance between each problem path and the negative ideal solution is calculated. The relative proximity of each problem route is calculated based on the first Euclidean distance and the second Euclidean distance.

8. A priority ranking device for upgrading problematic distribution network lines, characterized in that, include: The module includes: data acquisition module, digital model construction module, problem route identification module, indicator calculation module, matrix construction module, objective weight calculation module, subjective weight calculation module, relative proximity calculation module, and route modification ranking module. The data acquisition module is used to acquire raw power operation data, raw meteorological data, equipment ledger data and topology connection relationships of the power distribution network; The digital model building module is used to construct a digital model of the power distribution network based on the original power operation data, original meteorological data, equipment ledger data, and topological connection relationships. The problematic line identification module is used to identify problematic lines in the distribution network based on the digital model of the distribution network, a preset rule library for diagnosing explicit line problems, and a preset model for identifying implicit line problems. The index calculation module is used to calculate the values ​​of various indicators for each problematic line based on the digital model of the power distribution network. The matrix construction module is used to construct the original data matrix based on the values ​​of all indicators of all problematic routes; The objective weight calculation module is used to calculate the entropy weight of each indicator based on the original data matrix, and generate an objective weight vector of all indicators based on all entropy weights. The subjective weight calculation module is used to obtain the subjective weight vector of all indicators based on the analytic hierarchy process. The relative proximity calculation module is used to calculate the relative proximity of each problem line based on the original data matrix, objective weight vector, and subjective weight vector. The line modification sorting module is used to sort all problematic lines according to their relative proximity from largest to smallest, and to perform line modifications based on the sorting results.

9. A terminal device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement a priority sorting method for upgrading problematic distribution network lines as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is running, it controls the device where the storage medium is located to execute a priority sorting method for upgrading problematic distribution network lines as described in any one of claims 1 to 7.

Images