A power distribution network self-healing reconfiguration control method and system based on switch health perception

Abstract

This invention discloses a distribution network self-healing reconfiguration control method and system based on switch health perception, belonging to the field of power system automation and distribution network intelligent control technology. The method comprises: acquiring several candidate reconfiguration schemes and corresponding candidate switch sets corresponding to the current fault scenario of the distribution network; constructing a health feature vector for each candidate switch based on multi-source monitoring quantities, inputting the health feature vector into a switch health model for switch health assessment to obtain a switch health execution profile for each candidate switch; calculating the comprehensive evaluation results corresponding to the candidate reconfiguration schemes based on the several switch health execution profiles and a pre-constructed candidate scheme evaluation model; obtaining several comprehensive evaluation results corresponding to the several candidate reconfiguration schemes; selecting a target reconfiguration scheme from the several candidate reconfiguration schemes based on the several comprehensive evaluation results; and executing the distribution network self-healing reconfiguration action based on the target reconfiguration scheme, significantly improving the success rate and reliability of distribution network fault recovery.

Description

Technical Field

[0001] This invention relates to the field of power system automation and distribution network intelligent control technology, and in particular to a distribution network self-healing reconfiguration control method and system based on switch health perception. Background Technology

[0002] Existing power distribution network self-healing and fault recovery technologies typically follow the basic logic of "fault detection, fault isolation, interconnection and transfer, and load restoration." Public information indicates that the goal of fault location, isolation, and service restoration methods is generally to restore power to as many lost customers as possible while meeting operational constraints, maximize the margin of feeders after refactoring, and minimize the number of switch operations to reduce switch wear. These methods often rely on remotely controllable switches and reliable communication to achieve automatic isolation and restoration after a fault.

[0003] However, in practical engineering, traditional self-healing algorithms typically assume that the sectionalizing switches, tie switches, reclosers, and distribution switchgear involved in self-healing can all operate correctly according to instructions, i.e., they are considered "remotely controllable," which is equivalent to "reliably executed." This assumption holds approximately true when the equipment is in good condition, the control power supply is normal, and communication is stable. However, in scenarios with long operating years, numerous accumulated operations, significant differences in maintenance strategies, or complex environmental conditions, issues such as mechanical wear of the switchgear, abnormal auxiliary control circuits, deterioration of contact, undervoltage of the control power supply, and drift in arrival time directly affect its opening success rate, closing success rate, operation delay, and continued load-bearing capacity after operation. Furthermore, existing publicly available research has found that a significant portion of circuit breaker failures in the field are related to mechanical and auxiliary control circuit problems. Moreover, the information acquired by the monitoring system should not be limited to maintenance purposes but can also be used to control the operating sequence of circuit breakers, assist in system topology analysis, and improve system operational reliability. On the other hand, research on condition monitoring and life management of power transmission and distribution switchgear has formed a relatively mature technical approach. Publicly available information also indicates that establishing switchgear health indices or asset health indices based on condition monitoring data has become an important direction for switchgear lifecycle management. The key indicators include operational health, action time, switching time, closing time, and temperature rise, and it emphasizes the need to extract key status indicators from heterogeneous data sources. However, most of these technologies still serve asset management, condition-based maintenance, and lifespan assessment, and have not yet formed a closed-loop control mechanism that directly embeds switchgear health results into fault self-healing reconfiguration, action path screening, and action sequence optimization.

[0004] In summary, existing technologies have the following shortcomings: self-healing reconfiguration schemes usually only judge theoretical feasibility from the perspective of network topology and power flow constraints, lacking constraints on the actual operability of switches; even if a switch can operate, its operation delay and the load risk after operation are not included in the recovery decision; among multiple candidate self-healing schemes, existing methods usually only compare the recovery load, the number of switches, and the network loss level, without incorporating switch health degradation, operation failure probability, and failure consequences into a unified optimization objective. Summary of the Invention

[0005] This invention provides a distribution network self-healing reconfiguration control method and system based on switch health perception. It can solve the technical problem that in the prior art, all switch devices involved in the operation are assumed to operate reliably, which means that although the self-healing scheme can theoretically restore more load, it may fail in actual implementation due to switch failure, maloperation, delayed operation, deterioration of contact, or insufficient load-carrying capacity, resulting in interruption of restoration, repeated power outages, or increased equipment risks. This invention significantly improves the success rate and reliability of distribution network fault restoration.

[0006] This invention provides a self-healing reconfiguration control method for distribution networks based on switch health sensing, comprising: Obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and obtain a set of candidate switches corresponding to each candidate reconfiguration scheme; For any candidate reconstruction scheme, the candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector corresponding to each candidate switch is constructed. The health feature vector corresponding to each candidate switch is input into the pre-constructed switch health model to perform switch health evaluation and obtain the switch health execution profile corresponding to each candidate switch. The comprehensive evaluation result corresponding to the candidate reconstruction scheme is calculated based on the health execution profile of the candidate switches and the pre-built candidate scheme evaluation model. Obtain several comprehensive evaluation results corresponding to the candidate reconstruction schemes; Based on the comprehensive evaluation results, a target reconfiguration scheme is selected from the candidate reconfiguration schemes, and the distribution network self-healing reconfiguration action is performed based on the target reconfiguration scheme.

[0007] The above-mentioned scheme constructs a health feature vector through multi-source monitoring data of switches and completes a comprehensive switch health assessment to obtain a switch health execution profile. Then, using each switch health execution profile as input, a comprehensive benefit and risk assessment is conducted on each reconfiguration scheme. Finally, based on the assessment results, the optimal target self-healing reconfiguration scheme is selected and executed. By introducing the switch health status into the self-healing decision-making process, fault recovery is improved from topologically feasible to topologically feasible and reliable. This reduces the probability of self-healing failure caused by switch refusal to operate and overload, significantly improving the success rate and reliability of distribution network fault recovery. It solves the technical problem in existing technologies where traditional distribution network self-healing reconfiguration only selects recovery schemes based on network topology and power flow constraints, assuming that all switches to be operated can stably and reliably complete opening and closing operations, without incorporating the switch's own operating health status into the reconfiguration decision. This can easily lead to theoretically feasible reconfiguration schemes failing to operate due to switch deterioration, communication anomalies, insufficient carrying capacity, etc., thereby causing secondary power outages and interruptions in load recovery.

[0008] Furthermore, the switch health assessment includes: For several health features in the health feature vector, the degradation degree normalization processing is performed on each of the several health features to obtain several normalized feature values ​​corresponding to each of the several health features. A normalized health feature vector is constructed based on several of the normalized feature values; Degradation calculations are performed on several normalized eigenvalues ​​to obtain several degradation rates corresponding to several health features; A health score is calculated based on the health feature vector, several normalized feature values, and several degradation rates. The normalized health feature vector is input into a pre-constructed probability prediction model to calculate the probability of successful action. The predicted action delay is obtained based on a preset constant term, a preset regression parameter vector, and a normalized health feature vector; The carrying margin is calculated based on the health score and health feature vector. Based on the health score, action success probability, predicted action delay, and load margin, a switch health execution profile corresponding to the candidate switch is obtained.

[0009] The above-mentioned scheme completes the following steps in the switch health assessment process: normalization of health characteristic degradation, solution of degradation rate, calculation of health score, prediction of success probability of opening and closing action, estimation of action delay, and solution of switch load margin. Finally, it integrates four key indicators to form a complete switch health execution profile, which quantitatively represents the current degree of switch degradation, reliability of remote control action, speed of action response, and load-carrying capacity of fault transfer. This provides a refined and multi-dimensional quantitative basis for switch health for the comprehensive benefit and risk calculation of each subsequent self-healing scheme.

[0010] Furthermore, in the process of normalizing the degradation degree of several health features in the health feature vector to obtain several normalized feature values ​​corresponding to several health features, for any health feature, the degradation degree normalization process includes: Obtain the upper limit alarm threshold, lower limit alarm threshold, and reference health value corresponding to the health feature; The upper limit of degradation range is calculated based on the upper limit alarm threshold and the reference health value, and the lower limit of degradation range is calculated based on the lower limit alarm threshold and the reference health value. When the health feature is less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the lower limit of the degradation range. When the health feature is not less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the upper limit of the degradation range.

[0011] The above scheme sets up a unified normalization calculation logic for any health characteristic, and uniformly realizes the quantitative standard that "the larger the value, the more serious the deterioration" for all positive and negative deterioration health indicators, thus eliminating the evaluation bias caused by inconsistencies in various monitoring dimensions and normal ranges.

[0012] Further, the calculation of the health score based on the health feature vector, several normalized feature values, and several degradation rates includes: The static degradation degree is obtained by summing several of the normalized eigenvalues. For each of the aforementioned health characteristics corresponding to several degradation rates, a degradation trend value is selected from the degradation rate and the preset degradation threshold based on preset degradation conditions; A trend penalty term is obtained by summing the deterioration trend values ​​corresponding to the aforementioned health characteristics; Obtain the cumulative number of actions in the health feature vector, and calculate the lifespan reduction term by comparing the cumulative number of actions with the rated lifespan number of actions; The overall degradation degree is obtained by weighting and summing the preset health assessment weight coefficients, static degradation trend penalty term and life reduction term; The overall degradation degree is subjected to an exponential decay transformation to obtain a health score.

[0013] The above scheme calculates the comprehensive degradation degree step by step by normalizing characteristics, degradation rate, and switch life loss, and then converts it into a health score. It quantitatively outputs the switch health score in the range of 0 to 1 to intuitively reflect the overall integrity of the switchgear.

[0014] Further, the step of inputting the normalized health feature vector into a pre-constructed probability prediction model to calculate the probability of successful action includes: The communication health score is obtained based on the health feature vector. When the switch action corresponding to the candidate switch is a closing action, a closing prediction term is calculated based on a preset closing bias term, a preset closing model parameter vector, and a normalized health feature vector. The closing prediction term is then input into a pre-constructed probability prediction model to obtain the closing success probability. Finally, the closing action success probability is obtained based on the closing success probability and the communication health. When the switch action corresponding to the candidate switch is the gate opening action, the gate opening prediction term is calculated based on the preset gate opening bias term, the preset gate opening model parameter vector and the normalized health feature vector, and the gate opening prediction term is input into the probability prediction model to obtain the gate opening success probability. Then, the gate opening action success probability is obtained based on the gate opening success probability and the communication health. The success probability of the opening or closing action is taken as the success probability of the action.

[0015] The above scheme distinguishes between opening and closing actions, establishes independent probability prediction branches, calculates the inherent success probability of the corresponding actions, and then combines the communication availability level to obtain the final action success probability. It also integrates the availability level of the communication link and the control loop to correct the action reliability results, and accurately quantifies the actual remote control success rate when the switch performs different operations.

[0016] Further, the calculation of carrying margin based on the health score and health feature vector includes: The contact resistance is obtained from the health feature vector; The allowable current carrying capacity is calculated based on the health score, contact resistance, preset health reduction factor, and switch rated current. The expected current is obtained based on the candidate reconstruction scheme; The carrying capacity margin is obtained by calculating the difference between the allowable current carrying value and the expected current.

[0017] The above solution quantifies the spare capacity of the switch after self-healing and long-term load operation, accurately identifies switches that can complete opening and closing actions but have deteriorated contacts and insufficient current carrying capacity, thereby incorporating the risk of long-term load overload into the switch health profile and avoiding contact overheating and equipment damage caused by long-term overload operation of the switch after reconfiguration.

[0018] Further, the step of calculating the comprehensive evaluation result corresponding to the candidate reconstruction scheme based on the health execution profiles of the candidate switches and the pre-built candidate scheme evaluation model includes: Obtain the health execution profiles of the candidate switches corresponding to the candidate switches; The health performance profiles of several switches are input into a pre-built candidate scheme evaluation model for comprehensive benefit and risk assessment, and the comprehensive evaluation results corresponding to the candidate reconstruction schemes are obtained. The comprehensive benefit risk assessment includes: From the health execution profiles of the aforementioned switches, we obtain the success probabilities of several actions, the predicted delays of several actions, and the carrying margins of several candidate switches, respectively. The overall success probability is calculated based on the success probability of the current action corresponding to each of the candidate switches; Calculate the weights of several failure consequences corresponding to several candidate switches based on the candidate reconstruction scheme; The total risk of the scheme is obtained by weighting the failure consequence weights, success probabilities, predicted action delays, and carrying margins corresponding to the candidate switches. A comprehensive evaluation result is obtained based on the pre-acquired load power, total risk of the scheme, overall success probability, and preset weighting coefficients.

[0019] The above schemes take into account multiple evaluation dimensions, including load restoration benefits, switch failure risk, action delay impact, and potential overload hazards, to achieve a comprehensive quantitative comparison of the advantages and disadvantages of each self-healing scheme, providing a complete quantitative evaluation basis for selecting the optimal target reconfiguration scheme.

[0020] Further, the step of selecting a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and performing distribution network self-healing reconfiguration actions based on the target reconfiguration scheme, includes: Based on the comprehensive evaluation results, a target reconstruction scheme that meets the preset reconstruction conditions is selected from the candidate reconstruction schemes. For any switch to be operated in the target reconstruction scheme, obtain the current switch health execution profile corresponding to the switch to be operated, and perform an admission judgment on the switch to be operated based on the preset admission function and the current switch health execution profile to obtain the admission judgment result; When the access judgment result corresponding to each of the switches to be operated meets the preset access conditions, the distribution network self-healing reconfiguration action is executed based on the target reconfiguration scheme. Otherwise, obtain a backup self-healing reconfiguration scheme and execute the distribution network self-healing reconfiguration action based on the backup self-healing reconfiguration scheme.

[0021] After selecting the optimal self-healing reconfiguration scheme based on the comprehensive benefit and risk results, the above scheme adds an access verification step based on the health execution profile of the switches before execution. The current scheme is only executed if all switches to be operated meet the health, communication, and power access conditions. If any switch fails to meet the standards, the backup reconfiguration scheme is automatically switched. Unqualified switches with deteriorated health, communication interruption, or abnormal control power supply can be eliminated in advance before issuing remote control commands, avoiding interruptions in self-healing due to non-compliance of switch access. At the same time, the backup scheme ensures that the faulty load has a continuous recovery path, further improving the reliability of the distribution network fault self-healing execution.

[0022] Furthermore, after selecting a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and performing a distribution network self-healing reconfiguration action based on the target reconfiguration scheme, the method further includes: Obtain the actual action time, measured peak coil current, measured integral coil current, measured displacement deviation, and measured contact rebound related indicators for all switches to be operated in the target reconstruction scheme; Based on the actual action time, measured coil current peak value, measured coil current integral value, measured displacement deviation, and measured contact rebound related indicators corresponding to all the switches to be operated, the multi-source monitoring quantities are updated to obtain real-time multi-source monitoring quantities, and the real-time multi-source monitoring quantities are used as the multi-source monitoring quantities for the next round of distribution network self-healing reconfiguration control.

[0023] After the above scheme completes the self-healing reconfiguration action, it collects the measured mechanical and electrical characteristics of all operating switches for this action, updates the multi-source monitoring quantities stored in the system in real time, and uses them as input for the next round of self-healing decision-making, forming a closed-loop update mechanism for the equipment status after self-healing execution. This ensures that the switch health evaluation and scheme risk assessment results for each subsequent fault self-healing continuously match the latest deterioration state of the equipment as the actual operation and wear of the equipment changes.

[0024] This invention also provides a distribution network self-healing reconfiguration control system based on switch health sensing, comprising: a candidate scheme generation module, a switch health assessment module, a comprehensive assessment module, and a reconfiguration control module, wherein: The candidate scheme generation module is used to obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and to obtain a set of candidate switches corresponding to each candidate reconfiguration scheme; The switch health assessment module is used to assess the switch health of any candidate reconstruction scheme. The candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector is constructed for each candidate switch. The health feature vector for each candidate switch is then input into the pre-constructed switch health model to perform switch health assessment and obtain a switch health execution profile for each candidate switch. The comprehensive evaluation module is used to calculate the comprehensive evaluation results corresponding to the candidate reconstruction schemes based on the health execution profiles of the candidate switches and the pre-built candidate scheme evaluation model; and to obtain the comprehensive evaluation results corresponding to the candidate reconstruction schemes. The reconfiguration control module is used to select a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and to perform distribution network self-healing reconfiguration actions based on the target reconfiguration scheme. Attached Figure Description

[0025] 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.

[0026] Figure 1 This embodiment provides a schematic diagram of a distribution network self-healing reconfiguration control method based on switch health sensing. Figure 2 This embodiment provides a schematic diagram of the specific implementation process of distribution network self-healing reconfiguration control based on switch health perception. Detailed Implementation

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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).

[0033] 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.

[0034] Example 1: This embodiment provides a self-healing reconfiguration control method for distribution networks based on switch health sensing, such as... Figure 1 As shown, it includes: S1. Obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and obtain the candidate switch set corresponding to each candidate reconfiguration scheme; S2. For any candidate reconstruction scheme, the candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector corresponding to each candidate switch is constructed. The health feature vector corresponding to each candidate switch is input into the pre-constructed switch health model to perform switch health evaluation and obtain the switch health execution profile corresponding to each candidate switch. S3. Calculate the comprehensive evaluation result corresponding to the candidate reconstruction scheme based on the health execution profile of the candidate switches and the pre-built candidate scheme evaluation model. S4. Obtain several comprehensive evaluation results corresponding to the candidate reconstruction schemes respectively; S5. Based on the comprehensive evaluation results, select a target reconfiguration scheme from the candidate reconfiguration schemes, and perform a self-healing reconfiguration action of the distribution network based on the target reconfiguration scheme.

[0035] In practical implementation, this embodiment provides a distribution network self-healing reconfiguration control method based on switch health perception, specifically targeting the fault self-healing scenario of medium-voltage distribution networks. It incorporates the health status, operational reliability, operational delay, and post-operation carrying capacity of distribution switches into the fault isolation and power restoration decision-making process to achieve distribution network self-healing reconfiguration based on switch health perception. In practical applications, this method is applied to the distribution network's master station system, combined with edge control devices for controlling distribution switches and switch monitoring terminals for acquiring multi-source detection quantities.

[0036] After a fault occurs in the distribution network, the main station system or edge controller of the distribution network first completes the fault location, fault section isolation, and power outage area identification. Based on available tie paths, operable switches, and load restoration requirements, it generates multiple candidate tie-forward or network reconfiguration schemes under the current network topology, resulting in several candidate reconfiguration schemes. Each candidate reconfiguration scheme includes a sequence of switch actions, specifying the switch operations and their order of execution for each candidate switch. Switch operations include closing and opening actions. The candidate switches involved in each scheme form a corresponding candidate switch set. It should be noted that candidate switches include sectionalizing switches, tie switches, reclosers, and ring main unit switches, among other distribution switches.

[0037] While obtaining several candidate reconfiguration schemes, the system collects the opening or closing coil current of the candidate switches (such as sectionalizing switches, tie switches, reclosers, and ring main unit switches) during the last switch operation, as well as the opening or closing action time corresponding to the switch operation. It also obtains the cumulative number of switch operations and the number of abnormal alarms. Simultaneously, it collects multi-source monitoring quantities in real time, including the current control power supply voltage, contact temperature rise, contact resistance, current actual contact position, auxiliary contact status, cumulative number of operations, communication quality, and historical anomalies.

[0038] It should be noted that the health feature vector obtained based on the multi-source monitoring data includes the current operating data and historical operating data of the candidate switches. The historical operating data includes the time taken for opening the circuit breaker, the time taken for closing the circuit breaker, the peak value of the opening coil current, the peak value of the closing coil current, the integral value of the opening coil current, the integral value of the closing coil current, displacement deviation, contact rebound index, cumulative number of actions, and number of abnormal alarms. The current operating data includes contact temperature rise, contact resistance, control power supply voltage, and communication health. Let the set of switches participating in the self-healing decision-making process in the distribution network be... The switch set includes all candidate switches from several candidate switch sets. For any candidate switch... Construct its health feature vector It can be expressed as follows: ; in, Indicates switch The time consumed during the previous tripping action. Indicates switch The closing action time used in the previous closing action; Indicates switch The peak value of the trip coil current during the last tripping operation. They represent switches respectively. The peak value of the closing coil current during the last closing operation; This represents the integral of the trip coil current during the last tripping operation. This represents the integral of the closing coil current during the last tripping operation; This indicates the contact temperature rise at the current data collection moment. This data can be either the contact point temperature rise or the cabinet temperature rise, which can be selected according to the actual application scenario. This refers to the contact resistance, which is the additional resistance generated when current flows through the contact parts such as the moving and stationary contacts of a switch, and conductive connection terminals. This indicates the displacement deviation of the mechanism stroke or the contact, representing the amount of deviation between the current actual position of the contact and the factory standard position. Indicators related to contact rebound are used to characterize the rebound oscillation characteristics of the contacts after collision during the previous switch opening and closing process; Indicates the control power supply voltage; Indicates switch The cumulative number of actions is specifically the sum of the number of opening actions and closing actions completed by the switch since it was put into operation or within a preset statistical period (such as the most recent 7 days, 30 days or the most recent action records). Each opening or closing action is counted as one action. If the same switch completes one opening and one closing action in a self-healing process, it is counted as two actions. Indicates the health status of communication; Indicates switch The number of abnormal alarms occurring within a preset recent statistical window, including alarms related to the reliability of switch action, control reliability, communication reliability, or load-bearing capacity such as failure to operate, maloperation, incomplete operation, operation timeout, inconsistent auxiliary contact status, abnormal opening and closing coil current, undervoltage or power failure of control power supply, abnormal communication link, contact temperature rise or contact resistance exceeding the limit, abnormal mechanism stroke or contact rebound, and failure of post-operation status verification. For alarms continuously sent during the period when the same abnormality has not been resolved, they are counted as one abnormal event; if they reappear after the abnormality has been resolved, they are counted again in the number of abnormal alarms.

[0039] In this embodiment, the communication health level is used to characterize the availability of control commands issued by the master station system or edge controller to the distribution automation terminal via the communication link, and the availability of the terminal control loop to drive the switch actuator to complete the corresponding action. The status can be determined jointly based on the communication link status and the control loop status. The communication link status includes terminal online status, heartbeat message success rate, communication packet loss rate, communication delay, remote control command response success rate, and recent communication anomaly records. The control loop status includes control power supply voltage, remote / local control position, interlocking or unlocking status, integrity of the opening and closing control loop, terminal output relay status, mechanism energy storage status, and recent remote control failure records. When only status-based judgment conditions are available on-site, if communication is normal, control power supply is normal, remote control is engaged, and there are no interlocking conditions, then [the following can be considered]. Otherwise take When historical statistical data and real-time monitoring data are available, the communication success rate, remote control success rate, and control loop health status can be calculated based on the communication success rate, remote control success rate, and control loop health status from the most recent preset time window or the most recent remote control records. Continuous probability values ​​within a range.

[0040] This health feature vector Used to characterize switches The current mechanical, electrical, control, and communication statuses are as follows: among them, the time taken for opening and closing actions, the mechanism travel deviation, and the contact rebound-related indicators mainly reflect the mechanical status of the switch; the peak coil current, the integral coil current, the contact temperature rise, and the contact resistance mainly reflect the electrical status of the switch; the control power supply voltage mainly reflects the control status of the switch; the communication health mainly reflects the communication status of the switch; and the cumulative number of actions and the number of recent abnormal alarms are used to reflect the historical deterioration level and recent abnormal level of the switch.

[0041] Optionally, step S1 involves obtaining several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, including: Obtain several initial self-healing reconfiguration schemes corresponding to the current fault scenario of the distribution network; For each of the initial self-healing reconfiguration schemes, a power flow balance relationship is constructed to obtain active power flow, reactive power flow, and node voltage data. Based on the active power flow, reactive power flow, and node voltage data, several candidate reconfiguration schemes that meet preset constraints are selected from the initial self-healing reconfiguration schemes.

[0042] In the specific implementation process, assuming the current fault scenario in the distribution network, the network after fault isolation is as follows: The initial set of self-healing reconstruction schemes is as follows: ; in, Indicates the first An initial self-healing refactoring scheme This represents the total number of candidate solutions. The following power flow balance relationship is established for each initial self-healing reconfiguration solution: ; ; ; And satisfy the following constraints: ; ; in, and Representing branches The active and reactive currents; and Representing nodes respectively The active and reactive loads; and These represent the output power of distributed power sources; , These represent the node voltage data for switch i and the node voltage data for switch j, respectively. , These represent the branch resistance and reactance, respectively. Indicates a branch The maximum allowable current. Only initial self-healing reconfiguration schemes that meet the voltage, current, and radial operation requirements can enter the health perception evaluation stage, that is, to obtain several candidate reconfiguration schemes that meet the above constraints from all initial self-healing reconfiguration schemes.

[0043] Optionally, the switch health assessment includes: For several health features in the health feature vector, the degradation degree normalization processing is performed on each of the several health features to obtain several normalized feature values ​​corresponding to each of the several health features. A normalized health feature vector is constructed based on several of the normalized feature values; Degradation calculations are performed on several normalized eigenvalues ​​to obtain several degradation rates corresponding to several health features; A health score is calculated based on the health feature vector, several normalized feature values, and several degradation rates. The normalized health feature vector is input into a pre-constructed probability prediction model to calculate the probability of successful action. The predicted action delay is obtained based on preset constant terms, preset regression parameter vectors, and normalized health feature vectors; The carrying margin is calculated based on the health score and health feature vector. Based on the health score, action success probability, predicted action delay, and load margin, a switch health execution profile corresponding to the candidate switch is obtained.

[0044] Optionally, in the process of normalizing the degradation degree of several health features in the health feature vector to obtain several normalized feature values ​​corresponding to several health features, for any health feature, the degradation degree normalization process includes: Obtain the upper limit alarm threshold, lower limit alarm threshold, and reference health value corresponding to the health feature; The upper limit of degradation range is calculated based on the upper limit alarm threshold and the reference health value, and the lower limit of degradation range is calculated based on the lower limit alarm threshold and the reference health value. When the health feature is less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the lower limit of the degradation range. When the health feature is not less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the upper limit of the degradation range.

[0045] In practical implementation, since the physical meanings and deterioration directions of different health characteristics are not entirely the same, in order to ensure that the normalized indicators all meet the unified evaluation direction of "the larger the value, the more serious the deviation from the healthy state," this embodiment divides health characteristics into positive deterioration type and reverse deterioration type according to the direction of the characteristic value deviating from the reference health value. For indicators with bidirectional anomalies, a bidirectional deviation processing rule is adopted. Positive deterioration type refers to indicators where the switch health state deteriorates when the original characteristic value increases relative to the reference health value. The limit value or alarm threshold corresponding to this type of indicator is usually greater than the reference health value, and the larger the indicator value, the more serious the deterioration. Positive deterioration type health characteristics include tripping action time, closing action time, contact temperature rise, contact resistance, mechanism stroke or contact displacement deviation, contact rebound related indicators, cumulative action count, and abnormal alarm count. Reverse deterioration type refers to indicators where the original characteristic value decreases relative to the reference health value, indicating a deterioration in the switch health state. The limit value or alarm threshold corresponding to this type of indicator is usually less than the reference health value, and the smaller the indicator value, the more serious the deterioration. Reverse degradation-type health characteristics include control power supply voltage and communication health.

[0046] For health characteristics exhibiting both excessively high and excessively low abnormalities, such as peak current of the trip coil, peak current of the closing coil, integral current of the trip coil, and integral current of the closing coil, these indicators may be excessively high due to mechanical jamming or coil malfunctions, or excessively low due to insufficient control power or poor circuit contact. Therefore, they are treated as bidirectional deviation indicators. The degree of deviation from both the upper and lower alarm thresholds is calculated separately, and the larger value or weighted value is taken as the normalized degradation value for that indicator. In practical applications, the classification of each health characteristic can be determined based on its physical meaning, manufacturer's technical specifications, equipment ledger parameters, maintenance alarm thresholds, and historical operating data. If an indicator exceeds the upper alarm threshold, it indicates degradation and is classified as a positive degradation type; if an indicator falls below the lower alarm threshold, it indicates degradation and is classified as a negative degradation type. Through the above classification and normalization process, it can be ensured that different types of health indicators, after normalization, all meet the unified evaluation direction that "the larger the value, the more serious the deviation from the healthy state."

[0047] Specifically, for the first in the health feature vector Health characteristics Normalized eigenvalues ​​are obtained by normalizing the degree of degradation. The specific process is shown in the following formula: ; in, This represents the m-th health characteristic of switch i; This represents a reference health value for that characteristic; This represents the corresponding limit value or alarm threshold, i.e., the upper or lower alarm threshold. When the health characteristic is in a positive deterioration state, Take the upper limit alarm threshold; when this health characteristic is in a reverse deterioration type... Take the lower limit alarm threshold; the normalized feature value obtained after normalizing the above health features. The larger the value, the more severely the characteristic deviates from a healthy state.

[0048] Health feature vector After each original feature is normalized, the normalized feature values ​​are recombined according to the original feature order to form a normalized health feature vector.

[0049] Next, the degradation rate is calculated. In this embodiment, the specific process of degradation calculation is defined as follows: ; in, This represents the rate of deterioration of the m-th health feature at the current sampling time. This represents the normalized feature value of the m-th health feature at adjacent sampling times, used to reflect the trend of health status changes, where k represents the current sampling time. This indicates the interval between adjacent sampling. It should be noted that the sampling time and the interval between adjacent sampling in this embodiment can be set according to the specific scenario requirements.

[0050] Optionally, the step of calculating the health score based on the health feature vector, several normalized feature values, and several degradation rates includes: The static degradation degree is obtained by summing several of the normalized eigenvalues. For each of the aforementioned health characteristics corresponding to several degradation rates, a degradation trend value is selected from the degradation rate and the preset degradation threshold based on preset degradation conditions; A trend penalty term is obtained by summing the deterioration trend values ​​corresponding to the aforementioned health characteristics; Obtain the cumulative number of actions in the health feature vector, and calculate the lifespan reduction term by comparing the cumulative number of actions with the rated lifespan number of actions; The overall degradation degree is obtained by weighting and summing the preset health assessment weight coefficients, static degradation trend penalty term and life reduction term; The overall degradation degree is subjected to an exponential decay transformation to obtain a health score.

[0051] In the specific implementation process, after completing the above-mentioned health characteristic normalization and degradation rate calculation, the overall degradation degree and health score of the switch are calculated to quantify the health status of the switch. This embodiment considers static degradation, trend degradation, and lifetime consumption factors, and defines the switch... Overall degradation The calculation process is shown in the following formula: ; in, Indicates the static degradation degree. Indicates the total number of health characteristics; The degradation rate is represented by the value between the degradation rate and the preset degradation threshold. The preset degradation condition is to select the maximum value between the degradation rate and the preset degradation threshold as the degradation trend value. In this embodiment, the preset degradation threshold is set to 0. This indicates a reduction in life expectancy. Indicates the rated lifespan of the switch in terms of the number of operations. Indicates the first Static weights of each health characteristic; Indicates the first Penalty weights for each characteristic trend item; This indicates the life expectancy reduction weight. Preset health assessment weighting coefficients include... , and .

[0052] Next, an exponential decay transformation is performed based on the overall degradation level to obtain the health score. The specific calculation process is shown in the following formula: ; in, The closer the value is to 1, the healthier the switch is; the closer it is to 0, the more severe the switch deterioration.

[0053] Optionally, the step of inputting the normalized health feature vector into a pre-constructed probability prediction model to calculate the probability of successful action includes: The communication health score is obtained based on the health feature vector. When the switch action corresponding to the candidate switch is a closing action, a closing prediction term is calculated based on a preset closing bias term, a preset closing model parameter vector, and a normalized health feature vector. The closing prediction term is then input into a pre-constructed probability prediction model to obtain the closing success probability. Finally, the closing action success probability is obtained based on the closing success probability and the communication health. When the switch action corresponding to the candidate switch is the gate opening action, the gate opening prediction term is calculated based on the preset gate opening bias term, the preset gate opening model parameter vector and the normalized health feature vector, and the gate opening prediction term is input into the probability prediction model to obtain the gate opening success probability. Then, the gate opening action success probability is obtained based on the gate opening success probability and the communication health. The success probability of the opening or closing action is taken as the success probability of the action.

[0054] In practical implementation, considering that the opening and closing performance of the same switch are not completely consistent, this embodiment designs a probability prediction model with two prediction branches: the probability of successful opening and the probability of successful closing. The process of determining the effective action probability based on the pre-constructed probability prediction model in this embodiment can be expressed as follows: ; in, , representing the Logistic function; For switch The normalized health feature vector. and Consistent dimensions Indicates the preset gate opening bias term, Indicates the preset closing bias; Represents the preset gate opening model parameter vector, This represents the preset closing model parameter vector. (Superscript) This indicates that the corresponding action of the switch in this round of self-healing is tripping, indicated by the superscript. This indicates that the corresponding action of the switch in this round of self-healing is closing; Indicates switch The probability of successfully opening the floodgates in this round of self-healing. Indicates switch The success rate of closing during this round of self-healing. Then, based on the success rate of opening and communication health, the switch... Success rate of actions in this round of self-healing The specific calculation process is as follows: .

[0055] After obtaining the success probability of the action, this embodiment predicts the action delay based on this data. This is to describe the actual execution speed of different switches, for each action type. The specific process of action delay prediction is shown in the following formula: ; in, Indicates switch The predicted action delay is used to measure whether a switch will disrupt the timing coordination of the action sequence due to slow action in the current reconfiguration round. This is a preset constant term; This is the preset regression parameter vector.

[0056] Optionally, the calculation of carrying margin based on the health score and health feature vector includes: The contact resistance is obtained from the health feature vector; The allowable current carrying capacity is calculated based on the health score, contact resistance, preset health reduction factor, and switch rated current. The expected current is obtained based on the candidate reconstruction scheme; The carrying capacity margin is obtained by calculating the difference between the allowable current carrying value and the expected current.

[0057] In practical implementation, for tie switches or sectionalizing switches that need to continue to carry current after self-healing, this embodiment calculates their post-operation carrying capacity and carrying margin, and first designs the allowable current carrying value. The specific calculation process is shown in the following formula: ; in, Used to characterize the permissible current carrying capacity after health reduction. Indicates switch The rated current of the switch; and Preset health reduction factor.

[0058] Candidate Reconstruction Schemes If switch The expected current after transfer is Then carrying margin The calculation process is shown in the following formula: in, A value greater than zero indicates that the switch has sufficient load-carrying margin. A value less than zero indicates that although the switch can operate, it is not suitable for long-term operation under this scheme.

[0059] Optionally, step S3 includes: Obtain the health execution profiles of the candidate switches corresponding to the candidate switches; The health performance profiles of several switches are input into a pre-built candidate scheme evaluation model for comprehensive benefit and risk assessment, and the comprehensive evaluation results corresponding to the candidate reconstruction schemes are obtained. The comprehensive benefit risk assessment includes: From the health execution profiles of the aforementioned switches, we obtain the success probabilities of several actions, the predicted delays of several actions, and the carrying margins of several candidate switches, respectively. The overall success probability is calculated based on the success probability of the current action corresponding to each of the candidate switches; Calculate the weights of several failure consequences corresponding to several candidate switches based on the candidate reconstruction scheme; The total risk of the scheme is obtained by weighting the failure consequence weights, success probabilities, predicted action delays, and carrying margins corresponding to the candidate switches. A comprehensive evaluation result is obtained based on the pre-acquired load power, total risk of the scheme, overall success probability, and preset weighting coefficients.

[0060] In the specific implementation process, candidate reconstruction schemes are set. The required set of actions is ,in, Indicates the candidate switch number participating in the operation. This indicates the action type corresponding to the candidate switch. Assuming that each action is approximately independent, the overall success probability of the candidate reconstruction scheme is calculated. The specific process can be represented as follows: ; To describe the impact of action failure on system recovery, this embodiment assigns weights to the consequences of individual action failures. Defined as follows: ; in, Indicates action Predicted load power that cannot be recovered after failure; Indicates action Number of users affected by the failure; Indicates action The number of key users affected by the failure; , , These are the preset weighting coefficients.

[0061] Next, a weighted calculation is performed to obtain the total risk of the candidate reconstruction scheme. The specific calculation process is shown in the following formula: ; in, Indicates the risk of action failure. This indicates a penalty for exceeding the action delay limit. This indicates a penalty for insufficient load margin after an action; and For the corresponding preset weights; This indicates the maximum allowed time for an action.

[0062] It should be noted that after fault location and isolation are completed, the distribution network system can determine the theoretically recoverable load range of the candidate schemes based on the distribution network topology model, the power outage area, candidate interconnection and transfer paths, load node information, and user profile information. For a certain action in candidate reconfiguration scheme k, a hypothetical scenario of the action's failure is first constructed, i.e., maintaining the network state after the action's unsuccessful execution. Based on this network topology, topology connectivity judgment, radial verification, and power flow constraint verification are performed again to determine the set of load nodes that cannot be recovered or need to continue to be out of power due to the failure of the action. Subsequently, the active power of each load node in this set is summed to obtain the predicted load power that cannot be recovered after the action failure; the number of users corresponding to each load node in this set is summed to obtain the number of affected users; and the number of important users, supply guarantee users, or high-priority users is summed to obtain the number of affected important users. The aforementioned predicted load power can be obtained from the distribution automation master station, the Supervisory Control and Data Acquisition (SCADA) system, the distribution management system (DMS), the electricity consumption information collection system, the load forecasting system, or real-time measurement data. The number of users and important user attributes can be obtained from the marketing system, user files, the ledger of the distribution geographic information system (GIS), or the distribution network operation management system.

[0063] After calculating the total risk of the proposed solution, a joint assessment and optimization of recovery benefits and health risks is performed. Specifically, in the current fault scenario, the set of load nodes consisting of all distribution nodes connected to user loads within the power outage area after the fault section is isolated is as follows: This includes various load access nodes such as distribution transformer nodes and feeder load branch nodes, covering all load nodes in the candidate reconfiguration schemes, both recoverable and unrecoverable. The included nodes... The recovery state variable is Obtain the load power of this node. (Typically, this represents the total active power load of the node). Define candidate schemes. The overall objective function is: ; in, This indicates the revenue from restoring load; the more critical loads restored, the larger this item becomes. Indicates the importance weight of the load. Indicates the cost of the number of switching operations. This is the cost coefficient for a single switching operation; Indicates the health risk penalty of the plan. This is the risk penalty coefficient; This represents the overall success probability reward. The preset weighting coefficient, which is the success probability reward coefficient, includes the load importance weight, the cost coefficient of a single switching operation, the risk penalty coefficient, and the success probability reward coefficient.

[0064] Optionally, step S5 includes: Based on the comprehensive evaluation results, a target reconstruction scheme that meets the preset reconstruction conditions is selected from the candidate reconstruction schemes. For any switch to be operated in the target reconstruction scheme, obtain the current switch health execution profile corresponding to the switch to be operated, and perform an admission judgment on the switch to be operated based on the preset admission function and the current switch health execution profile to obtain the admission judgment result; When the access judgment result corresponding to each of the switches to be operated meets the preset access conditions, the distribution network self-healing reconfiguration action is executed based on the target reconfiguration scheme. Otherwise, obtain a backup self-healing reconfiguration scheme and execute the distribution network self-healing reconfiguration action based on the backup self-healing reconfiguration scheme.

[0065] In the specific implementation process, after obtaining the comprehensive evaluation results of several candidate reconstruction schemes, the optimal candidate scheme is selected as the target reconstruction scheme from among the several candidate reconstruction schemes. In this embodiment, the candidate reconstruction scheme that meets the preset reconstruction conditions is the one with the largest comprehensive evaluation result among several candidate reconstruction schemes, as shown in the following formula: ; in, This refers to a restructuring plan that offers the highest overall recovery benefits while minimizing health risks.

[0066] After obtaining the optimal target reconstruction scheme, and before executing the optimal scheme, this embodiment also performs a health access determination on each switch to be operated in the target reconstruction scheme. For the switches to be operated... First, obtain the current switch health execution profile corresponding to the switch to be operated, so as to obtain the switch to be operated from the current switch health execution profile. Corresponding current communication health Current control power supply voltage and current health score The admission decision is obtained by performing an admission judgment based on communication health, control power supply voltage, health score, and a preset admission function. In this embodiment, the preset admission function is defined as follows: ; in, It is an indicator function; To control the lower limit of the power supply voltage; This represents the lower limit of communication health. This is the lower limit of the health score. The preset admission condition is that only if the admission judgment result is... At that time, the switch is ready to be operated. Only then were they allowed to participate in this round of self-healing reconstruction; if The switch to be operated Not allowed to participate in this round of self-healing refactoring, only when all pending switches in the target refactoring scheme meet the requirements. If the target reconfiguration scheme is selected, a self-healing reconfiguration action of the distribution network is triggered. Otherwise, the target reconfiguration scheme is removed from several candidate reconfiguration schemes and backup self-healing reconfiguration schemes are automatically re-selected, or a backup path is enabled.

[0067] Optionally, after selecting a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and performing distribution network self-healing reconfiguration based on the target reconfiguration scheme, the method further includes: Obtain the actual action time, measured peak coil current, measured integral coil current, measured displacement deviation, and measured contact rebound related indicators for all switches to be operated in the target reconstruction scheme; Based on the actual action time, measured coil current peak value, measured coil current integral value, measured displacement deviation, and measured contact rebound related indicators corresponding to all the switches to be operated, the multi-source monitoring quantities are updated to obtain real-time multi-source monitoring quantities, and the real-time multi-source monitoring quantities are used as the multi-source monitoring quantities for the next round of distribution network self-healing reconfiguration control.

[0068] In the specific implementation process, after the distribution network self-healing reconfiguration action is executed, that is, after the switches that need to be operated have completed their actual actions, their action signature vectors are extracted. : ; in, This refers to the actual time taken for the action; This represents the measured peak value of the coil current. This is the integral of the measured coil current. This represents the measured displacement deviation. This is for the actual measurement of relevant indicators of contact rebound.

[0069] Optionally, in this embodiment, the switch... Before actually performing the action, according to the switch The current switch health execution profile calls its normalized health feature vector, health score, and action type. Calculate the action signature predicted by the health model. Specifically, the switch... Current normalized health feature vector Health score before the activity and action type Input the action signature prediction model and get: ; in, Indicates switch Perform actions based on current health status. Health model predicts action signatures at that time; Indicates the predicted action time; Indicates the predicted peak coil current; This represents the integral of the predicted coil current; Indicates the predicted travel deviation; Indicators that predict a rebound; Representation and action type The corresponding action signature prediction model; This represents the parameters of the action signature prediction model. These parameters can be obtained through training or calibration based on historical action records of the same model of switch, factory test data, handover test data, periodic maintenance test data, online status monitoring data, and the historical action signature data of this switch.

[0070] After the switch completes this opening and closing action, collect the actual action signature corresponding to this action. The actual action signature is compared with the predicted action signature to construct the anomaly score. : ; in, For action type The corresponding weight matrix reflects the importance and normalized weights of different action signature components in anomaly detection. If... A larger deviation indicates that the actual action process deviates significantly from the predicted action process under the current health state, which can be used to determine that the health state of the switch has further deteriorated or that there is a potential abnormality.

[0071] Based on the abnormality level obtained from this action, the health score after the switch action is completed is updated according to the following recursive formula. : ; in, This serves as a memory factor. Relying on the aforementioned dynamic update mechanism for health scores, the complete health profile of a single switch can be continuously iterated and corrected with each switching action, providing accurate equipment health decision-making basis for subsequent evaluation of distribution network fault self-healing reconfiguration schemes and selection of optimal self-healing strategies.

[0072] In summary, the self-healing reconfiguration control method for distribution networks based on switch health awareness provided in this embodiment addresses the following technical problem: Existing self-healing reconfiguration methods for distribution networks primarily make recovery decisions based on network topology, interconnection relationships, power flow distribution, and load importance. They typically assume that all participating switches can operate reliably, without considering switch health status, success probability of operation, operation delay, and post-operation load risk as self-healing constraints. This leads to situations where, while the self-healing scheme can theoretically restore more load, it may fail in practice due to switch refusal to operate, maloperation, delayed operation, deteriorated contact, or insufficient load-carrying capacity, resulting in recovery interruptions, repeated power outages, or increased equipment risk. Another key issue addressed by this embodiment is that existing switch status monitoring and asset health management technologies mainly serve maintenance and life assessment, without directly participating in distribution network fault self-healing control. This makes it difficult to support real-time switch selection, operation sequencing, and online rollback in fault recovery scenarios. Therefore, a control method that deeply integrates multi-source monitoring information of switches with self-healing reconfiguration optimization is needed. This method would prevent the system from treating switches as simple binary switching elements, but instead model them as dynamic execution units with different opening and closing capabilities, action delays, and load-bearing capacities, thereby achieving more reliable automatic fault recovery.

[0073] In response, the self-healing reconfiguration method for distribution networks proposed in this embodiment first generates several candidate reconfiguration schemes by combining the available interconnection paths, operable switches, and power loss load recovery requirements of the current network after a fault occurs in the distribution network and the fault section is isolated. Simultaneously, it collects multi-source monitoring quantities corresponding to sectionalizing switches, interconnection switches, reclosers, and ring main unit switches, thereby constructing the health characteristics of each candidate switch based on the multi-source monitoring quantities. On this basis, it calculates the probability of successful opening, the probability of successful closing, the predicted action delay, and the load margin for each switch, and integrates the health score, the probability of successful action, the predicted action delay, and the load margin to generate a switch health execution profile adapted to the self-healing scenario.

[0074] Next, for each candidate reconfiguration scheme, a power flow constraint check is first performed to verify whether the scheme meets conventional constraints such as node voltage, branch current, and radial operation of the distribution network. If the check passes, the overall success probability, the weight of each switch failure consequence, the action delay penalty, and the load risk of the tie switch are calculated for the action set consisting of all switches to be operated in the candidate reconfiguration scheme. The load restoration benefit, important load restoration benefit, switch operation cost, total scheme risk, and overall success probability reward are incorporated into a unified comprehensive objective function to complete the comprehensive benefit and risk assessment. The target reconfiguration scheme with the highest comprehensive restoration benefit and the lowest equipment action risk is selected, and the optimal switch action execution sequence is determined simultaneously.

[0075] When entering the execution phase of the scheme, instead of directly issuing remote control commands for all switches, an access judgment is first performed on each switch to be operated. The current communication health, control power supply voltage, and health score of the switch are verified using a preset access function. Only switches whose three indicators all reach the corresponding thresholds can participate in this round of self-healing reconfiguration. After successful verification, the system issues opening and closing commands sequentially according to the optimized action sequence, simultaneously collecting coil current, action time, and contact position information to verify the actual action status of the switch in real time. If abnormal conditions such as switch refusal to operate, erroneous operation, action timeout, or insufficient load margin are detected during the process, the current action sequence is immediately terminated, and a backup reconfiguration scheme is retrieved or the self-healing recovery path is recalculated.

[0076] Finally, after the entire self-healing operation is completed, the actual action signature is formed by collecting the action time, coil current peak value, coil current integral value, displacement deviation, and contact rebound related indicators of all switches to be operated. The multi-source monitoring quantities corresponding to the switch health model are iteratively updated using the measured action signature, forming a complete closed-loop self-healing control system based on multi-source monitoring and acquisition, switch health prediction, reconfiguration scheme decision-making, switch action execution, action status verification, and equipment status update.

[0077] This invention deeply integrates the full-dimensional health status of switches into the entire process of distribution network fault self-healing, effectively reducing the risks of self-healing execution failure, secondary power outages, and equipment malfunctions caused by switch degradation. It significantly improves the actual success rate of fault load recovery and is widely applicable to smart medium-voltage distribution network scenarios with a large number of remotely controllable switches. The core innovation lies in constructing a complete distribution network control method based on deep coupling of switch health perception and self-healing reconfiguration. Equipment health assessment logic is embedded in all stages, including scheme generation, scheme evaluation, action sequencing, execution access, anomaly rollback, and model self-updating. This effectively reduces the risk of secondary power outages caused by switch refusal to operate, delayed operation, and overload, significantly improving the success rate of automatic fault recovery in distribution networks and demonstrating good engineering applicability.

[0078] Specifically, compared to traditional distribution network self-healing methods centered on topological connectivity, power flow constraints, and load recovery, the primary innovation of this invention lies in abandoning the presupposition of ideal and reliable switch operation. It introduces quantitative indicators of switch performance reliability into the entire self-healing process, ensuring that the theoretical calculation results of the reconfiguration scheme are consistent with the actual on-site operating conditions. Instead of treating switches as ideal components with only two states (open and closed), it combines multi-source monitoring data to quantify switch health scores, the probability of successful opening / closing actions, predicted action delays, and load margins, modeling switches as physical devices with differentiated execution capabilities. Compared to existing switch health index assessment technologies used only for asset management and condition-based maintenance, which only use health evaluation methods as offline maintenance references, this invention transforms the switch health performance profile into real-time self-healing decision variables, directly participating in candidate reconfiguration scheme selection, switch admission determination, and action sequence arrangement, fully leveraging the online control value of condition monitoring data. Meanwhile, this invention, for the first time, incorporates the success probability of switch action, action delay, and load margin into the candidate scheme evaluation model, constructing a unified objective function that considers both load restoration benefits and equipment action risks. Traditional self-healing schemes only compare the scale of restored load and the frequency of switch operations. This invention, however, comprehensively evaluates the restoration benefits, action failure risks, delay impacts, and potential overload risks of each scheme, selecting the optimal reconfiguration scheme that best fits the on-site conditions, rather than simply a theoretically optimal solution that satisfies power flow constraints. Furthermore, this invention adds a pre-execution switch health access mechanism and a post-action measured signature update mechanism: before issuing remote control commands, access verification is completed based on the switch health execution profile, preemptively filtering out high-risk switches with substandard communication, power supply, and health indicators; after the switch completes its action, measured action features are collected to update multi-source monitoring data, iteratively optimizing the switch health model, unlike traditional self-healing algorithms with fixed parameters and maintenance systems that only perform offline evaluations.

[0079] Example 2: This embodiment also provides a distribution network self-healing reconfiguration control system based on switch health perception, including: a candidate scheme generation module, a switch health assessment module, a comprehensive assessment module, and a reconfiguration control module, wherein: The candidate scheme generation module is used to obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and to obtain a set of candidate switches corresponding to each candidate reconfiguration scheme; The switch health assessment module is used to assess the switch health of any candidate reconstruction scheme. The candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector is constructed for each candidate switch. The health feature vector for each candidate switch is then input into the pre-constructed switch health model to perform switch health assessment and obtain a switch health execution profile for each candidate switch. The comprehensive evaluation module is used to calculate the comprehensive evaluation results corresponding to the candidate reconstruction schemes based on the health execution profiles of the candidate switches and the pre-built candidate scheme evaluation model; and to obtain the comprehensive evaluation results corresponding to the candidate reconstruction schemes. The reconfiguration control module is used to select a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and to perform distribution network self-healing reconfiguration actions based on the target reconfiguration scheme.

[0080] Example 3: like Figure 2 As shown, based on the above-described method embodiments, this embodiment provides a specific implementation process for distribution network self-healing reconfiguration control based on switch health perception, including: S21. Construction of switch health feature vector: Collect multi-source monitoring quantities corresponding to sectional switches, tie switches, reclosers, and ring main unit switches, integrate all status indicators of mechanical, electrical, control, and communication categories, and construct a dedicated health feature vector for each candidate switch to fully characterize the real-time operating status of the switch.

[0081] S22. Normalization of health characteristics and calculation of degradation rate: Distinguish between positive, negative and bidirectional degradation health indicators, and complete degradation normalization based on the upper and lower alarm thresholds and reference health values ​​of each indicator to obtain normalized characteristic values. Combine the normalized data at the sampling time before and after to solve the degradation rate of each characteristic and unify the degradation evaluation benchmark.

[0082] S23. Calculation of overall switch degradation degree and health score: Based on normalized eigenvalues, degradation rate, and cumulative operating life loss, the static degradation degree, trend penalty term, and life reduction term are solved respectively. After weighting to obtain the overall degradation degree, the switch health score in the range of 0 to 1 is generated by exponential transformation, which intuitively quantifies the overall degradation level of the switch.

[0083] S24. Calculation of successful opening, successful closing and effective action probability: Based on the normalized health feature vector, independent Logistic probability prediction branches for opening and closing are built. Combined with communication health correction, the effective action probability of opening and closing of the switch is obtained, and the reliability of remote control execution of different operations is accurately quantified.

[0084] S25. Action Delay Prediction: A linear regression model is used, with normalized health feature vectors as input and corresponding action type regression parameters matched, to predict the theoretical action delay of switch opening / closing, which is used to assess the risk of action timing coordination.

[0085] S26. Calculation of load capacity and load margin after operation: Combine the switch health score, measured contact resistance, and rated current to solve for the health reduction allowable current carrying value, compare it with the expected operating current after reconstruction to obtain the load margin, and evaluate the safety margin of the switch for long-term load supply.

[0086] S27. Generation of candidate self-healing schemes and verification of steady-state constraints: After the fault section is isolated, an initial reconfiguration scheme is generated based on the tie path and load demand. The node voltage, branch current and radial distribution network constraints are verified by the power flow equation, and candidate reconfiguration schemes that meet the steady-state operation conditions are selected.

[0087] S28. Calculation of success probability and action risk of candidate scheme: Traverse all switches to be operated in the scheme, combine the single-unit action probability to obtain the overall success probability of the whole scheme, and calculate the action delay and overload multi-dimensional scheme total risk by combining the load loss of each switch action failure and the user impact weight.

[0088] S29. Joint optimization of recovery benefits and health risks: Construct a unified comprehensive objective function that integrates load recovery benefits, switching operation costs, scheme health risks, and action success rewards, and quantify the comprehensive evaluation results of each candidate reconfiguration scheme.

[0089] S210. Optimal target reconstruction scheme selection: Compare the comprehensive evaluation results of all candidate reconstruction schemes, and select the scheme with the largest objective function value as the target reconstruction scheme with the best comprehensive benefits.

[0090] S211. Pre-execution health access judgment: Before issuing remote control commands, the control power supply, communication health and health score of all switches in the target scheme are checked based on the preset access function. The self-healing operation can be started only if all indicators meet the standards. If they do not meet the standards, the backup reconstruction scheme is switched.

[0091] S212. Post-action signature verification and health model update: After the self-healing action is completed, the actual action time of the switch, coil current, displacement, and contact rebound data are collected to form an action signature. The signature is compared with the model prediction signature to identify new equipment degradation and update the multi-source monitoring quantities of the system to realize the iterative correction of the switch health model.

[0092] Example 4: Based on the above embodiment of a distribution network self-healing reconfiguration control method based on switch health sensing, another embodiment of the present invention provides a terminal device, which includes 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 a distribution network self-healing reconfiguration control method based on switch health sensing according to any embodiment of the present invention.

[0093] 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 terminal device.

[0094] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0095] 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. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0096] Example 5: Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the distribution network self-healing reconfiguration control method based on switch health sensing described in any of the above-described method embodiments of the present invention.

[0097] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0098] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles 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 self-healing reconfiguration control method for distribution networks based on switch health sensing, characterized in that, include: Obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and obtain a set of candidate switches corresponding to each candidate reconfiguration scheme; For any candidate reconstruction scheme, the candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector corresponding to each candidate switch is constructed. The health feature vector corresponding to each candidate switch is input into the pre-constructed switch health model to perform switch health evaluation and obtain the switch health execution profile corresponding to each candidate switch. The comprehensive evaluation result corresponding to the candidate reconstruction scheme is calculated based on the health execution profile of the candidate switches and the pre-built candidate scheme evaluation model. Obtain several comprehensive evaluation results corresponding to the candidate reconstruction schemes; Based on the comprehensive evaluation results, a target reconfiguration scheme is selected from the candidate reconfiguration schemes, and the distribution network self-healing reconfiguration action is performed based on the target reconfiguration scheme.

2. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 1, characterized in that, The switch health assessment includes: For several health features in the health feature vector, the degradation degree normalization processing is performed on each of the several health features to obtain several normalized feature values ​​corresponding to each of the several health features. A normalized health feature vector is constructed based on several of the normalized feature values; Degradation calculations are performed on several normalized eigenvalues ​​to obtain several degradation rates corresponding to several health features; A health score is calculated based on the health feature vector, several normalized feature values, and several degradation rates. The normalized health feature vector is input into a pre-constructed probability prediction model to calculate the probability of successful action. The predicted action delay is obtained based on a preset constant term, a preset regression parameter vector, and a normalized health feature vector; The carrying margin is calculated based on the health score and health feature vector. Based on the health score, action success probability, predicted action delay, and load margin, a switch health execution profile corresponding to the candidate switch is obtained.

3. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 2, characterized in that, In the process of normalizing the degradation level of several health features in the health feature vector to obtain several normalized feature values ​​corresponding to several health features, for any health feature, the degradation level normalization process includes: Obtain the upper limit alarm threshold, lower limit alarm threshold, and reference health value corresponding to the health feature; The upper limit of degradation range is calculated based on the upper limit alarm threshold and the reference health value, and the lower limit of degradation range is calculated based on the lower limit alarm threshold and the reference health value. When the health feature is less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the lower limit of the degradation range. When the health feature is not less than the reference health value, the normalized feature value is obtained by calculating the ratio of the absolute value of the error between the health feature vector and the reference health value to the upper limit of the degradation range.

4. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 2, characterized in that, The calculation of the health score based on the health feature vector, several normalized feature values, and several degradation rates includes: The static degradation degree is obtained by summing several of the normalized eigenvalues. For each of the aforementioned health characteristics corresponding to several degradation rates, a degradation trend value is selected from the degradation rate and the preset degradation threshold based on preset degradation conditions; A trend penalty term is obtained by summing the deterioration trend values ​​corresponding to the aforementioned health characteristics; Obtain the cumulative number of actions in the health feature vector, and calculate the lifespan reduction term by comparing the cumulative number of actions with the rated lifespan number of actions; The overall degradation degree is obtained by weighting and summing the preset health assessment weight coefficients, static degradation trend penalty term and life reduction term; The overall degradation degree is subjected to an exponential decay transformation to obtain a health score.

5. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 2, characterized in that, The step of inputting the normalized health feature vector into a pre-constructed probability prediction model to calculate the probability of successful action includes: The communication health score is obtained based on the health feature vector. When the switch action corresponding to the candidate switch is a closing action, a closing prediction term is calculated based on a preset closing bias term, a preset closing model parameter vector, and a normalized health feature vector. The closing prediction term is then input into a pre-constructed probability prediction model to obtain the closing success probability. Finally, the closing action success probability is obtained based on the closing success probability and the communication health. When the switch action corresponding to the candidate switch is the gate opening action, the gate opening prediction term is calculated based on the preset gate opening bias term, the preset gate opening model parameter vector and the normalized health feature vector, and the gate opening prediction term is input into the probability prediction model to obtain the gate opening success probability. Then, the gate opening action success probability is obtained based on the gate opening success probability and the communication health. The success probability of the opening or closing action is taken as the success probability of the action.

6. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 2, characterized in that, The calculation of carrying margin based on the health score and health feature vector includes: The contact resistance is obtained from the health feature vector; The allowable current carrying capacity is calculated based on the health score, contact resistance, preset health reduction factor, and switch rated current. The expected current is obtained based on the candidate reconstruction scheme; The carrying capacity margin is obtained by calculating the difference between the allowable current carrying value and the expected current.

7. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 2, characterized in that, The calculation of the comprehensive evaluation result corresponding to the candidate reconstruction scheme based on the health execution profiles of the candidate switches and the pre-built candidate scheme evaluation model includes: Obtain the health execution profiles of the candidate switches corresponding to the candidate switches; The health performance profiles of several switches are input into a pre-built candidate scheme evaluation model for comprehensive benefit and risk assessment, and the comprehensive evaluation results corresponding to the candidate reconstruction schemes are obtained. The comprehensive benefit risk assessment includes: From the health execution profiles of the aforementioned switches, we obtain the success probabilities of several actions, the predicted delays of several actions, and the carrying margins of several candidate switches, respectively. The overall success probability is calculated based on the success probability of the current action corresponding to each of the candidate switches; Calculate the weights of several failure consequences corresponding to several candidate switches based on the candidate reconstruction scheme; The total risk of the scheme is obtained by weighting the failure consequence weights, success probabilities, predicted action delays, and carrying margins corresponding to the candidate switches. A comprehensive evaluation result is obtained based on the pre-acquired load power, total risk of the scheme, overall success probability, and preset weighting coefficients.

8. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 1, characterized in that, The step of selecting a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and performing distribution network self-healing reconfiguration actions based on the target reconfiguration scheme, includes: Based on the comprehensive evaluation results, a target reconstruction scheme that meets the preset reconstruction conditions is selected from the candidate reconstruction schemes. For any switch to be operated in the target reconstruction scheme, obtain the current switch health execution profile corresponding to the switch to be operated, and perform an admission judgment on the switch to be operated based on the preset admission function and the current switch health execution profile to obtain the admission judgment result; When the access judgment result corresponding to each of the switches to be operated meets the preset access conditions, the distribution network self-healing reconfiguration action is executed based on the target reconfiguration scheme. Otherwise, obtain a backup self-healing reconfiguration scheme and execute the distribution network self-healing reconfiguration action based on the backup self-healing reconfiguration scheme.

9. The distribution network self-healing reconfiguration control method based on switch health sensing as described in claim 1, characterized in that, After selecting a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and performing distribution network self-healing reconfiguration actions based on the target reconfiguration scheme, the method further includes: Obtain the actual action time, measured peak coil current, measured integral coil current, measured displacement deviation, and measured contact rebound related indicators for all switches to be operated in the target reconstruction scheme; Based on the actual action time, measured coil current peak value, measured coil current integral value, measured displacement deviation, and measured contact rebound related indicators corresponding to all the switches to be operated, the multi-source monitoring quantities are updated to obtain real-time multi-source monitoring quantities, and the real-time multi-source monitoring quantities are used as the multi-source monitoring quantities for the next round of distribution network self-healing reconfiguration control.

10. A self-healing reconfiguration control system for a distribution network based on switch health sensing, characterized in that, include: The module comprises a candidate solution generation module, a switch health assessment module, a comprehensive assessment module, and a reconfiguration control module, among which: The candidate scheme generation module is used to obtain several candidate reconfiguration schemes corresponding to the current fault scenario of the distribution network, and to obtain a set of candidate switches corresponding to each candidate reconfiguration scheme; The switch health assessment module is used to assess the switch health of any candidate reconstruction scheme. The candidate switch set includes several candidate switches. Based on the pre-acquired multi-source monitoring data, a health feature vector is constructed for each candidate switch. The health feature vector for each candidate switch is then input into the pre-constructed switch health model to perform switch health assessment and obtain a switch health execution profile for each candidate switch. The comprehensive evaluation module is used to calculate the comprehensive evaluation results corresponding to the candidate reconstruction schemes based on the health execution profiles of the candidate switches and the pre-built candidate scheme evaluation model; and to obtain the comprehensive evaluation results corresponding to the candidate reconstruction schemes. The reconfiguration control module is used to select a target reconfiguration scheme from several candidate reconfiguration schemes based on several comprehensive evaluation results, and to perform distribution network self-healing reconfiguration actions based on the target reconfiguration scheme.

Images