CN119994813B - A method for intelligent power distribution hierarchical self-healing fault isolation and power supply restoration - Google Patents

Abstract

The present invention discloses a method for intelligent power distribution hierarchical self-healing fault isolation and power supply restoration, which is suitable for quickly isolating faults and restoring power supply to non-fault areas when a fault occurs in a power distribution system. The method adopts a hierarchical control architecture, and the intelligent switch equipment at the lower layer is responsible for real-time acquisition of line current and voltage status, and immediately completes local identification and removal of the faulty section after detecting a preset fault signal. The fault information is then uploaded to the upper layer distribution automation master station, which analyzes the recoverable power supply area and generates a network reconstruction plan in combination with the current network topology, and switches the non-fault area to a backup power supply path by controlling the interconnection switch. The system supports retrospective analysis of historical data such as waveform distortion and partial discharge signals before the fault, identifies potential abnormal trends and optimizes operation strategies, and realizes an adaptive early warning function. The method has the advantages of rapid response, clear structure, and strong autonomy, and can significantly improve the reliability and continuity of the power distribution system.

Description

Intelligent power distribution layered self-healing fault isolation and power supply recovery method

Technical Field

The invention relates to the technical field of electric power, in particular to an intelligent power distribution layered self-healing fault isolation and power supply recovery method.

Background

In the prior art, power distribution networks often rely on power distribution automation systems to achieve fault detection, isolation, and power restoration. The system adopts a centralized control structure, and the master station analyzes the fault position through the data of the acquisition terminal and issues an operation instruction to the field switch equipment to complete fault isolation and network reconstruction. In addition, part of the system integrates feeder automation or small-range self-healing functions, fault response and recovery control of a local area can be realized under specific conditions, and power supply reliability is improved.

However, the prior art has the problems of slow response speed, insufficient information closed loop and single recovery strategy when dealing with complex topological structure or multi-point faults. The centralized control mode has high communication dependence, is limited by data transmission and processing time delay, and is difficult to complete large-scale self-adaptive power supply recovery in a short time. Meanwhile, due to the fact that an effective running state early warning mechanism is lacked, potential risks cannot be identified in time before a fault occurs in the system, and further improvement of self-healing capacity is limited.

Therefore, it is needed to provide an intelligent power distribution fault isolation and power supply recovery method with faster response, more flexible structure and self-adaptive early warning capability.

Disclosure of Invention

The application provides an intelligent power distribution layered self-healing fault isolation and power supply recovery method, which is used for improving the reliability and continuity of a power distribution system.

The application provides an intelligent power distribution layered self-healing fault isolation and power supply recovery method, which comprises the following steps:

establishing a layered control architecture in a power distribution system, wherein the layered control architecture comprises a lower intelligent switch device and an upper power distribution automation master station;

the intelligent switching equipment collects the current and voltage states of the connected lines, and when signals meeting preset fault criteria are detected, fault section identification is completed locally, corresponding line sections are cut off, and fault isolation at the first time is achieved;

the intelligent switch equipment uploads fault information and key electrical data to a distribution automation master station, and the distribution automation master station judges a recoverable power supply area by combining network topology information and dynamically generates a reconstruction scheme;

The distribution automation master station controls a group of tie switches to execute operation according to the reconstruction scheme, and switches the non-fault area to other power supply paths so as to realize power supply recovery;

And when the power supply recovery is executed, current waveform, harmonic characteristic or partial discharge signal before the fault occurs are retrospectively analyzed, if an abnormal trend is identified, an abnormal event is recorded and an operation strategy is updated for early warning and distinguishing of similar faults in the future.

The technical scheme provided by the application has the beneficial effects that:

(1) Through establishing the hierarchical control architecture, the cooperative coordination of the intelligent switch equipment and the distribution automation master station is realized, fault identification and isolation can be completed locally and rapidly, the fault response time is effectively shortened, and the instant processing capacity of the distribution system is improved. (2) The automatic master station is adopted to perform network topology analysis and reconstruction path generation, and the rapid power supply recovery of a non-fault area is completed without manual intervention, so that the power failure range and duration are obviously reduced, and the power supply continuity is ensured. (3) A backtracking analysis mechanism for the electrical abnormal signal before the fault is introduced, so that the system has self-adaptive early warning capability, potential risks can be identified in advance, and an operation strategy can be optimized, and therefore the operation safety and the fault resistance of the power distribution network are improved. (4) The whole method has the characteristic of high automation, can realize distributed cooperative control and flexible power supply path switching under various complex working conditions, and improves the intelligent level and the operation efficiency of the system.

Drawings

Fig. 1 is a flowchart of an intelligent power distribution layered self-healing fault isolation and power restoration method according to a first embodiment of the present application.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. The present application may be embodied in many other forms than those herein described, and those skilled in the art will readily appreciate that the present application may be similarly embodied without departing from the spirit or essential characteristics thereof, and therefore the present application is not limited to the specific embodiments disclosed below.

The first embodiment of the application provides an intelligent power distribution layered self-healing fault isolation and power supply recovery method. Referring to fig. 1, a schematic diagram of a first embodiment of the present application is shown. The following describes in detail a first embodiment of the present application with reference to fig. 1, which provides an intelligent power distribution layered self-healing fault isolation and power restoration method.

Step S101, a layered control architecture is established in a power distribution system, wherein the layered control architecture comprises a lower intelligent switch device and an upper power distribution automation master station.

In the intelligent power distribution layered self-healing fault isolation and power supply recovery method provided by the embodiment, step S101 is a basic preparation step of the whole method, and aims to construct a set of power distribution system structure with layered sensing, quick response and coordination control capability so as to support efficient execution of subsequent fault processing and power supply recovery operations. Step S101 is described in detail below.

In step S101, the step of establishing a hierarchical control architecture in the power distribution system refers to dividing the control architecture into at least two logic levels according to the difference of function division and information processing capability in an existing or newly-built medium-voltage or low-voltage power distribution network, wherein the lower layer is a field intelligent control layer, and the upper layer is a centralized coordination control layer so as to realize functional decoupling and distributed deployment of data acquisition, fault detection, decision analysis and control execution.

The lower intelligent switch equipment refers to power switch equipment arranged at each key node (such as a feeder switch, a branch switch, a ring main unit and the like), has the function of collecting basic electric parameters such as current, voltage and the like in real time, is internally provided with edge computing capability and fault recognition logic, and can independently complete fault feature recognition and switching-on and switching-off actions. Preferably, the device supports the IEC 61850 communication protocol and is equipped with a circuit Breaker Control Unit (BCU), a Current Transformer (CT), a voltage transformer (PT), a Programmable Logic Controller (PLC) and other components. The sampling frequency should be no less than 5 kHz to ensure accurate capture of short circuits, overloads, etc. rapidly changing events. The device should support at least one Digital Signal Processor (DSP) or equivalent embedded computing unit for implementing the operation of a local fault recognition algorithm, such as short circuit determination based on current slew rate Δi/Δt or zero sequence current analysis.

The upper-layer distribution automation master station is usually arranged in a dispatching center or a control room and is composed of a high-performance server or an industrial control computing platform, runs distribution automation master station software and has the capabilities of modeling a whole network topological structure, managing communication, generating fault processing strategies, controlling reconstruction of a power supply path and the like. The master station maintains a stable connection with each smart switching device through a communication network (e.g., fiber optic ethernet, wireless public network, loRa, NB-IoT, etc.). The master station system needs to integrate a network topology diagram, an SCADA interface, a real-time database and an expert control module, and supports remote measurement, remote signaling and remote control operation of external equipment.

The specific process of establishing the hierarchical control architecture includes the following implementation links. Firstly, digitally modeling a power distribution network structure, wherein the digital modeling comprises geographic information and electrical connection relations of each feeder line, each node, each transformer substation and each switch position. And secondly, installing intelligent switch equipment on each key node, and completing field debugging and communication network access. And then loading corresponding power distribution topological graph and equipment address information in the master station system, and configuring a fault response strategy template. The configuration of message format, communication period, heartbeat mechanism, encryption protocol and the like between the master station and each intelligent device should be completed so as to ensure data security and instantaneity.

The hierarchical control architecture established by the mode not only has the front-end distributed acquisition and local response capability, but also has the global coordination and control capability of the central end, and can realize a cooperative control mode of information up-concentration, decision down-distribution and execution at the edge. The architecture provides basic support for subsequent fault detection, isolation, reconstruction, early warning and other steps, and ensures that the whole system has remarkable advantages in the aspects of response speed, processing accuracy and operation reliability.

In summary, step S101 not only requires the establishment of a hierarchical structure at the hardware deployment level, but also emphasizes the hierarchical division of functional logic and control flow.

Still further, the establishing a hierarchical control architecture in a power distribution system includes:

In an initialization stage, a power distribution automation master station generates a logic layered structure of a multi-level power supply area based on a Geographic Information System (GIS) and historical operation topology data, and distributes corresponding switch equipment addresses, fault criterion templates and communication strategies for each level node to realize synchronous matching of the topology structure and functional configuration;

Through the issued configuration information, the operation roles and action authorities of all intelligent switch devices are synchronized, so that the intelligent switch devices have fault response priorities, coordination waiting time and reporting window parameters of different levels according to the network level, and self-adaptive cooperative response among multiple devices is realized in a concurrent fault scene;

After the distribution automation master station is deployed in a layered structure, chain type mutual recognition operation is carried out on intelligent switch equipment in each group of subareas, so that the adjacent switch equipment can finish identity authentication and communication handshake locally, and self-organizing boundary recognition capability among the equipment is established so as to support regional autonomous fault handling in an off-line state of the master station.

In this embodiment, a further limitation is set forth on how to build the hierarchical control architecture, and the focus is on a series of configurations and cooperations in the system initialization stage, so that the power distribution system not only has the hierarchical control capability, but also has the structural, self-organizing and self-adapting operation characteristics.

Before the power distribution system is put into operation, an initialization stage is firstly carried out, the stage is led by a power distribution automation main station, and the whole power distribution network is modeled in a layering way by utilizing a built-in Geographic Information System (GIS) platform and long-term accumulated historical operation topology data. GIS data typically includes spatial location information for substations, switchyard, feeder lines, branch lines, load points, communication nodes, etc., while topology data reflects electrical connection relationships, circuit breaker status, switching logic, and historical load paths. Based on the information, the master station automatically divides a multi-stage power supply area logic level, such as a transformer substation layer, a main feeder layer, a branch line layer, a terminal layer and the like, by taking area division, feeder line attribution, transformer nodes, load density and the like as references.

After the hierarchical definition is completed, the master station will assign dedicated configuration parameters to the critical nodes in each hierarchy. These parameters include unique address codes (e.g. logical node LN identification) bound to the intelligent switching device, preset fault criteria templates (e.g. short circuit current threshold, duration decision window, waveform distortion characteristics, etc.), communication mechanisms (e.g. reporting period, receive window size, congestion control policy, etc.). The configuration ensures that the logic positioning of each intelligent switch device in the structure is accurate, and the matching relation is formed between the intelligent switch device and the whole dispatching strategy of the system on function allocation.

The configuration information generated by the master station is issued to each level of intelligent switch equipment through a communication network (such as optical fiber Ethernet, wireless public network, industrial level 5G and the like). After receiving, the switching device automatically loads configuration content and updates local parameters, so that a multi-stage division work cooperation operation mode is formed. Specifically, the intelligent switch device at the upper layer level can obtain higher fault response priority, the fault section of the level can be judged and isolated firstly when multiple points of faults are concurrent, and the lower layer device needs to collect the state information of the adjacent devices in the coordination waiting time and then make action decisions. In addition, the reporting window of each device can also be set according to the hierarchy, for example, the reporting frequency of the main feeder node is higher to ensure the synchronization of the main path information, and the branch line device adopts an event triggering mode to reduce the communication pressure.

After all the configuration is issued and the deployment is confirmed to be effective, the master station further starts a chained mutual recognition mechanism so as to enhance the self-organizing control capability inside each partition. The operation is that the master station sends out a mutual-acknowledgement starting instruction to trigger the adjacent intelligent switch devices in the same area to sequentially execute the identity authentication and communication handshake flow. The identity authentication is based on equipment identification, key verification or consistency check of configuration version, and the communication handshake establishes a local low-delay communication channel, for example, an RS-485, CAN bus or a special wireless frequency hopping link is used for realizing local state exchange without central scheduling.

Through chain type mutual recognition operation, each group of intelligent switch equipment forms an adaptive small-sized autonomous domain. Under the condition that a communication link of a master station is interrupted or the master station operates offline, the autonomous domains can autonomously judge local faults and execute corresponding isolation and recovery strategies by virtue of the established cooperative relationship among devices, so that the communication fault resistance and the autonomous response efficiency of the system are remarkably improved.

In summary, the steps from topology modeling to parameter configuration to autonomous mutual recognition form a hierarchical control architecture with high self-adaption, reasonable structure and self-organizing capability. The architecture breaks through the excessive dependence on a main station in the traditional power distribution automation system, and also effectively supports distributed response and coordination control among the lower intelligent devices, so that the invention realizes remarkable improvement in the aspects of self-healing capacity, response speed, system elasticity and the like.

Still further, the chain mutual authentication operation includes:

After receiving a chain type mutual acknowledgement instruction issued by a power distribution automation master station, the intelligent switching equipment carries out mutual acknowledgement negotiation with other intelligent switching equipment physically adjacent to the intelligent switching equipment through a low-delay local communication protocol, and determines the fault response priority sequence of each intelligent switching equipment in a link based on equipment identification, geographical position information and historical response performance indexes;

After mutual recognition is completed, each intelligent switch device periodically broadcasts state abstract information of the intelligent switch device to adjacent intelligent switch devices, wherein the state abstract information comprises a current load state, a fault judgment abstract and a received control instruction identifier and is used for establishing an operation state consistency judgment mechanism in a local area;

under the condition that the communication of the power distribution automation master station is interrupted or not reachable, the intelligent switching equipment which completes mutual confirmation independently operates a reconstruction strategy in a local area according to a pre-negotiated response sequence, wherein the reconstruction strategy comprises delayed closing control, cascading fault isolation and communication chain integrity self-checking so as to realize autonomous self-healing operation within a limited range.

In this embodiment, a specific technical flow of the chain mutual recognition operation is further defined. The operation aims to enhance the local coordination capability between intelligent switch devices, so that the intelligent switch devices can still complete autonomous fault processing and power supply reconstruction by means of a local mechanism under the condition that a power distribution automation master station cannot timely instruct or communicate abnormally.

After the hierarchical control structure configuration is completed in the system initialization stage, the power distribution automation master station can issue a chain type mutual recognition instruction to intelligent switch equipment in a group of predefined areas. The instruction comprises a mutual recognition range identifier, a communication handshake rule, a response parameter template, mutual recognition algorithm configuration and the like. After receiving the instruction, the intelligent switch device can perform bidirectional identification and negotiation with other intelligent switch devices physically adjacent to the intelligent switch device through a local low-delay communication protocol. The communication protocol CAN be based on a CAN bus, RS-485, low-power wireless (such as Zigbee) or industrial wireless Ethernet, and the two-way handshake is required to be completed within 10-50 milliseconds.

In the mutual authentication process, the intelligent switch device actively exchanges respective identity identifiers (such as device addresses and unique serial numbers), geographic position information (provided by GIS data or a GPS module), and historical response performance indexes (such as recent action response time, misoperation rate, recovery success rate and the like). Based on these data, the devices will dynamically establish a priority list in the local link through a negotiation mechanism. The table defines which device preferably executes the isolation or reconstruction operation when the fault occurs in the future, and which device needs to wait for the other devices to respond after completing the preliminary judgment, thereby preventing the linkage conflict or misjudgment caused by the simultaneous actions of multiple devices.

After the mutual acknowledgement negotiation is completed, each intelligent switching device will enter an information synchronization state. The device will broadcast a set of status summary information to neighboring smart switching devices in its link at a set period (e.g., every 500 milliseconds to 2 seconds). The status summary information is generated locally by the device, and the content includes the current load status (such as current magnitude and load rate), the real-time fault judgment summary (such as fault type and confidence level), and the latest received or issued control instruction identification code. The mechanism can enable the equipment to judge whether local abnormality exists or whether the self-healing mode should be entered or not by fast comparing the judgment of the states of the neighbor equipment when the master station does not send a unified instruction.

The broadcast data should adopt a redundant compression coding mode, which not only ensures the communication flux, but also avoids transmission collision, and a polling broadcast or short frame communication mechanism based on frequency hopping is suggested to be used. And when any device in the link detects that the distribution automation master station is not reachable (for example, heartbeat packets are not received or master station address PING fails in a plurality of continuous communication periods), the distribution automation master station enters an autonomous control mode. In this case, the intelligent switching devices that have completed mutual authentication may independently start the local fault isolation and power restoration processes according to the response sequence table generated by the negotiation.

In the autonomous mode, the first operation includes a delayed switching-on mechanism, i.e. after detecting that the standby path has a switching-on condition, the operation is not immediately performed, but waiting for a preset delay time and confirming that the states of the neighbor devices are consistent, so that closed loop collision is avoided. Second, the device will implement a cascading failure isolation policy, for example, when the upstream device is not active, the downstream device presumes that the upstream may fail, and will actively extend the isolation scope to ensure power safety. In addition, the intelligent switch device also executes a communication chain self-checking operation, and whether the communication chain self-checking operation is still connected with other devices on the mutual recognition chain is checked through the broadcast response packet so as to judge whether the intelligent switch device has complete autonomous control conditions.

The whole autonomous flow is completed locally, and a data record synchronous interface with the master station is maintained, so that the state and control history of the equipment can be integrated seamlessly after the master station resumes communication.

In summary, the chained mutual recognition operation not only builds a set of dynamic updatable local response priority mechanism, but also introduces a device state sharing and fault emergency autonomous flow, and remarkably improves the robustness and recovery capability of the system in an abnormal communication environment.

On the basis of the chained mutual recognition mechanism provided by the embodiment, the intelligent switch equipment in the link not only can realize autonomous fault response in the traditional sense, but also can be further expanded to more complex and higher-level scenes, such as emergency dispatch linkage control, distributed power supply access management, load dynamic reconstruction negotiation and the like, so as to construct an decentralized and scene-self-adaptive power distribution intelligent cooperation system.

In the emergency scheduling scene, the intelligent switch equipment with chain type mutual recognition is not limited to responding to local fault information, but is used for dynamically evaluating the power supply safety margin based on the state coordination of the full link. When the distribution automation master station issues early warning instructions which possibly have extreme weather, short-term load impact or upper power interruption risks in advance, each intelligent switch device in the link can actively enter a pre-coordination state. In this state, parameters such as a risk predicted value, an adjustable load ratio, and a standby path feasibility score calculated based on a preset model are shared among the devices in addition to the conventional state information. The device can establish a distributed event voting mechanism on the chain, if a majority of nodes in the chain predict short-time power supply shortage to reach an early warning threshold value, the actions of open loop isolation, load sinking, standby contact pre-closing and the like are performed in a distributed mode in advance, regional defensive scheduling response is formed, and the overall reaction capacity to sudden risks is remarkably improved. The process is different from the centralized regulation and control of the existing master station, but is driven by link cooperation, and has stronger local rapidity and prejudgement accuracy.

In a distributed power access scenario, the chain structure may be extended to a power-aware mutual-acknowledgement chain. The intelligent switching device at the access point can automatically send a grid connection request to other devices in the chain once detecting that the distributed power supply has a grid connection enabling condition, such as stable output of the inverter, voltage frequency matching and enough harmonic tolerance of the power grid. The request will be accompanied by key parameters such as injectable power, maximum response delay, planned run length, etc. Other devices in the link dynamically feed back acceptable degrees based on current load conditions, voltage sensitivity of topological nodes, fault response priority and other data, and finally, a link consensus algorithm (such as minimum response delay path calculation based on device priority weighting) judges whether to accept the power supply access. If a local power grid connection is formed, the link can automatically generate a local energy flow reconstruction scheme, and a fault strategy in the link is adjusted, for example, the isolation priority of power nodes is set, so that the power sides can be quickly separated when in fault, and the self-healing logic of a main power supply path is not influenced. The mode breaks through the conventional fixed power supply access point or master station approval mechanism, allows the equipment to locally and cooperatively complete access negotiation and power supply mode switching, and improves the adaptability of the power distribution system to the dynamic access of the distributed power supply.

In addition, the chained cooperative mechanism can be expanded to multipath load distribution optimization. In the case where there are multiple backup paths available at the same time, the link device may actively initiate a path sharing negotiation based on the real-time load state. Each device constructs a local power supply path score based on the local load margin, the current bus voltage level, and the historical operational reliability. By propagating and converging on-chain, the scoring results will automatically form power optimization suggestions. For example, for the same load point, if two contact paths come from different feeder lines respectively, the link will recommend a load splitting mode, the load is shared by the two paths in proportion, and the distribution proportion is automatically adjusted according to real-time feedback, so that dynamic power supply balance of the link level is realized. The self-balancing logic is obviously different from the traditional failure reconstruction strategy based on the master station, and has stronger scene adaptability and equipment negotiation capability.

Through the mechanism, the real-time cooperation capability between the devices is expanded on the basis of chain mutual recognition, and the structural transition from a control command chain to an autonomous intelligent chain of the power distribution system is realized. The method is characterized in that the equipment has the functions of dynamic mutual identity recognition, context state propagation, self-establishment of operation consensus under the condition of missing a master station, and self-adaptive reconstruction of local control strategies according to different scene targets (such as safety, efficiency and flexibility) in a plurality of types of power events. The architecture has fundamental difference from the traditional hierarchical centralized master-slave control, does not depend on central computing or static configuration, and is an embedded, linked and continuously evolved chain type intelligent cooperative mechanism.

In an implementation aspect, the above functions may be implemented by deploying a controller with chained state management, dynamic consensus negotiation, and adaptive policy enforcement modules in the intelligent switching device. The mechanism can be directly embedded into the existing power distribution automation architecture by matching with a communication protocol stack supporting link state broadcasting and collaborative decision, and has good expandability and actual deployment prospect.

Step S102, collecting current and voltage states of the connected lines by the intelligent switching equipment, and when signals meeting preset fault criteria are detected, locally completing fault section identification and cutting off corresponding line sections to realize fault isolation at the first time.

In this embodiment, step S102 is a core link for realizing rapid fault isolation and reducing the power interruption range and duration. The basic goal of the step is to realize the on-site identification and removal of faults by virtue of intelligent switch equipment deployed on the lower layer when the power distribution system breaks down.

First, the intelligent switching apparatus should have the ability to monitor the current and voltage conditions in the connected distribution lines in real time. To ensure detection accuracy and response timeliness, the device should integrate a Current Transformer (CT) and a voltage transformer (PT) or a high-accuracy sensor module, and be equipped with a microprocessor or Digital Signal Processor (DSP) with edge computing capability. The device should acquire three-phase current and voltage data with a sampling frequency not lower than 5 kHz to ensure high-fidelity recording of the fault initial state, and simultaneously generate characteristic parameters such as effective values, phase angles, zero sequence components and the like in real time.

On the basis of collecting data in real time, the intelligent switch equipment needs to operate a set of preset fault identification criteria, and the criteria can be comprehensively judged according to parameters such as fault current amplitude mutation, duration time, current direction change, zero sequence current amplitude, phase voltage drop degree and the like. For example, a typical short circuit fault may be determined if the system current amplitude exceeds three times the rated current threshold for a very short period of time and continues for more than two sampling periods with a corresponding voltage dip or imbalance factor rise. In order to improve the recognition robustness, a multi-condition fusion criterion can be introduced and a priority is set, so that the distinction recognition of the instantaneous disturbance and the permanent fault is realized.

After determining that the fault exists, the device needs to perform positioning and determination of the fault section. The intelligent switching equipment can analyze the relation between the current direction and the current amplitude of the upstream and downstream switches by adopting a current direction method, a longitudinal difference method, a negative sequence current method or an artificial intelligent auxiliary algorithm so as to deduce whether the fault is located in the section of line. If the current at the outlet of the current switch is obviously increased and no obvious current is input at the inlet, the fault can be primarily judged to be positioned at the downstream of the current section, and if the fault current flows at the upstream and downstream, the comparison judgment is needed by combining the data of the adjacent switches.

Once the line of the section is determined to be a fault source, the equipment sends out a control instruction, and physical switching-off operation is implemented through an internal breaker driving mechanism. The circuit breaker can be in the form of a vacuum circuit breaker, an SF ₆ circuit breaker or a solid-state switch, and the like, and the opening time is required to be controlled within a millisecond range, so that fault current is ensured to be cut off before a zero transition point, and arc spreading or equipment damage is prevented. When the device is started and stopped, the device should record the current time stamp of the event, the current voltage curves before and after the fault, the action state of the switch and other information to form a complete local fault event record for further analysis of the subsequent uploading host station system.

It is worth noting that step S102 emphasizes that the fault identification and isolation operations are completed locally in the intelligent switch device and do not depend on the external instruction of the power distribution master station, so that the system has significant response speed advantage, and can still complete local self-healing processing at the first time when communication delay or the master station is not reachable.

In summary, step S102 achieves that the fault section can be rapidly identified and isolated without manual intervention after the fault occurs by deploying the intelligent switching device with the capability of real-time acquisition and local fault identification, lays a key foundation for subsequent power supply recovery and system reconfiguration, and remarkably improves the operation stability and reliability of the power distribution system.

Still further, the intelligent switching device collects current and voltage states of the connected lines, when detecting signals meeting preset fault criteria, locally completes fault section identification and cuts off corresponding line sections, and achieves fault isolation at a first time, including:

The intelligent switch equipment adopts a dynamic criterion adjusting mechanism, and according to the topological position of the intelligent switch equipment, the fluctuation characteristic of a historical load curve and the cooperative parameters of other intelligent switch equipment in a link, the threshold interval of local fault identification is adjusted in real time in operation, so that fault discrimination is closer to scene characteristics and misoperation or refusal is avoided;

when the appointed intelligent switch equipment detects that the electrical state of a circuit connected with the appointed intelligent switch equipment meets the primary fault characteristic, the fault primary judgment information is synchronized to other adjacent intelligent switch equipment in a link through a chain communication mechanism, the multiple equipment jointly completes multipoint synchronous verification judgment to confirm to be a consistency fault, then the equipment with the highest priority is selected to execute the isolation action, and the other equipment enters a waiting response state to prevent repeated excision;

after the cutting action is finished, the post-fault data caching mechanism is executed, the intelligent switching equipment executing the isolation action records complete electrical waveforms and criterion parameters before and after the fault, and the recorded data is stored in a chain type communication network in a structured mode for subsequent master station backtracking analysis or other equipment linkage strategy adjustment and use, so that event cascade closed-loop management is realized.

In order to realize rapid, accurate and misoperation-avoiding fault response, the embodiment provides a set of highly intelligent dynamic criterion adjustment and multi-equipment cooperative confirmation mechanism, and combines the structured caching and sharing of event data to construct a self-adaptive local fault processing logic which can adapt to the complex operation environment of the power distribution network.

Firstly, on the fault identification strategy of intelligent switch equipment, the method is different from the method adopting static fixed current and voltage threshold values as judgment standards in the existing distribution automation system, and the embodiment introduces a dynamic criterion adjustment mechanism. Specifically, the intelligent switch device refers to historical load fluctuation characteristics, such as early and late load change rules, periodic spikes and the like, in a long-term operation process of the intelligent switch device based on the current network topology position, such as a main feeder node, a branch node or an end node and the like, and receives operation state parameters, such as load rate change, voltage drop frequency, response delay conditions and the like, of other devices in a link in real time, so that a criterion window applicable to the current scene is dynamically calculated. The window includes, but is not limited to, a short circuit current decision threshold, a voltage sag criterion time window, zero sequence current amplitude, duration, etc. The dynamic adjustment mode enables the device to have stronger misoperation inhibition capability in a high-load period or a high-disturbance section, and can properly improve response sensitivity in a low-interference and sensitive area, thereby effectively avoiding misoperation and misoperation phenomena.

Secondly, after detecting a signal which possibly meets the primary fault characteristic, the intelligent switching equipment does not immediately execute the circuit breaking operation, and synchronously sends the primary judgment result to other adjacent intelligent switching equipment in the link through a low-delay chain communication mechanism. The initial judgment information comprises detection time, current and voltage abrupt change parameters, criterion triggering details, current state labels of the equipment and the like. After receiving the information, the adjacent equipment combines the self-monitoring data to perform one-time quick comparison to confirm whether similar fault trends are detected or judge whether faults are likely to be located in the section of line. Through the collaborative verification mechanism, multipoint synchronization judgment can be realized, and once a response consistency threshold set in a link is exceeded (for example, more than two devices report matching fault trends simultaneously), the device with the highest negotiation priority initiates an isolation action. The priority may be generated based on preset indicators such as device response delay, topology location weight, action stability record, etc., while other devices enter a waiting state and respond later when it is confirmed that isolation is successful or the link is abnormal. The chained consistency confirmation mode effectively prevents repeated multi-point cutting or false triggering caused by signal reflection, transient disturbance and the like, and is beneficial to maintaining the stability of the system and restoring the integrity of the path.

After the fault isolation action is completed, the intelligent switch device executing the action also automatically triggers a local data caching and reporting mechanism. The equipment not only records the action time and the switch state change, but also packages full waveform current, voltage data, dynamic criterion values, communication interaction records and the like in a plurality of seconds before and after the fault, writes the full waveform current, the voltage data, the dynamic criterion values, the communication interaction records and the like into a local data module in a structured format, and uploads the full waveform current, the voltage data, the dynamic criterion values, the communication interaction records and the like to a link sharing buffer area through a chain type communication interface. Other intelligent switching devices and upper master station systems may invoke this data for purposes of backtracking analysis, event verification, policy optimization, etc., during subsequent failure recovery or system reconfiguration. The recording mode of the structured data can adopt JSON or compressed tag format (such as CBOR), and the event tag and the link node number are embedded, so that the time sequence comparison and the behavior consistency audit of the cross-equipment and the cross-link are realized.

Through the mechanism, the embodiment realizes the substantial breakthrough of the prior art in three dimensions of intelligence of fault identification, cooperativity of execution actions and closed loop of data processing. The intelligent switch equipment is not a single-point static response unit any more, but an autonomous node with environment sensing, self-adaptive cooperative judgment and data trace remaining capacity, and an intelligent self-healing unit network with evolution capacity is formed. The functions of the step can be realized by embedding a dynamic threshold adjusting module, a lightweight chain type communication protocol stack, a fault collaborative judging algorithm and a structured data recording module on the basis of the control logic of the conventional intelligent switch equipment.

Still further, the dynamic criteria adjustment mechanism of the intelligent switching apparatus includes calculating a local short circuit current action threshold based on the following equation 1:

;

Wherein, the Maximum load current for the last 24 hours; standard deviation of load current in the last 1 hour is used for representing current load fluctuation; Is the average phase voltage over the current 15 minutes; a target voltage value issued by a power distribution automation master station; Representing topology level parameters of the intelligent switch equipment, wherein the value of the main feeder equipment is 1.0, the value of the branch equipment is 0.5, and the value of the end node equipment is 0; is an empirical coefficient.

In the method for isolating and recovering the fault of the intelligent power distribution layering self-healing type and the power supply, in order to realize a more sensitive and stable fault identification function, the intelligent switch equipment does not use a fixed short-circuit current threshold value for judgment, and a dynamic criterion adjusting mechanism is introduced. The mechanism combines the historical load operation condition, voltage stability and topological structure characteristics of the equipment to calculate the current short-circuit current action threshold in real time, so that the judgment standard is adaptively adjusted according to the actual condition of the site, and misoperation or refusal caused by different equipment positions or operation state changes are avoided.

The embodiment adopts the formula 1 to dynamically calculate the short-circuit current action threshold valueThe formula is constructed from physical meaning and practically available data, and has good engineering realizability. The following explanation is made for each parameter in the formula:

wherein, the Is the maximum load current in amperes (a) over approximately 24 hours, and is used to characterize the maximum power load condition of the current device access point over the past complete operating cycle. The value is recorded by an intelligent switch device built-in sampling module, and is generated by taking the maximum value in an integral point updating or sliding time window, and is used as the operation scale reference of the current device.

Parameters (parameters)Is the standard deviation of the load current (unit: a) in approximately 1 hour to reflect the extent of current fluctuation of the device access point in the short term, i.e., whether the load is stationary. By collecting the current data once per minute, the standard deviation can be calculated in a sliding window. The larger the standard deviation is, the more severe the load fluctuation is, so the action threshold of the equipment is correspondingly improved, and misoperation caused by short-time spikes is avoided.

Representing the average value of the phase voltage in volts (V) over the current 15 minutes, calculated by continuous sampling by the internal voltage measurement module of the device. The value reflects whether the current power supply voltage is stable or not, and is suitable for judging whether the system is in an under-voltage, light-load or local fluctuation state or not.

The target voltage value issued by the master station is also in volts (V), and is generally 220V, 230V or rated phase voltage set by the system, and is used as a reference standard when the equipment judges voltage deviation. The device obtains and periodically refreshes the parameter from the master station via the communication interface.

Is a topology level factor that reflects the structural location of the intelligent switching apparatus throughout the power distribution network. Considering that the main feeder device has a larger influence on the global power supply, the action of the main feeder device needs to be more careful, so the main feeder deviceSet to 1.0, and the spur device set to 0.5, the end user access node device set to 0. The parameters are generated by combining GIS and topology map when the system is initialized, and written into the configuration table of each device.

The empirical coefficients k1k_1, k2k_2, k3k_3 are weighting factors that can be set according to the system debugging result, and recommended values are respectively:

The sensitivity to load fluctuation is expressed, the recommended value is adjustable between 0.5 and 1.0, and the action threshold is higher when the fluctuation is more severe;

the sensitivity to voltage offset is expressed, the recommended value is adjustable between 1.0 and 1.5, and the tolerance threshold value of the equipment in undervoltage/overvoltage needs to be improved;

the importance weight of the topological position is represented, the recommended value is adjustable between 0.3 and 0.7, and the protection threshold of the trunk equipment is slightly adjusted upwards to avoid misoperation.

The short-circuit action current threshold value at the current moment is calculated according to the calculation result of the formulaThe value is correspondingly slightly higher than the historical maximum load current by one grade, and the value is dynamically adjusted according to the real-time load fluctuation, the voltage stability and the topology position. For example, when a device is at the end of a branch, the load is stationary, the voltage is normal, the threshold is close toThe device is convenient for quickly isolating faults, and when the device is positioned in the middle section of the main feeder line, the load fluctuation is large, and the voltage fluctuation is obvious, the threshold value is properly raised, so that misoperation is avoided. The action criterion based on the scene weight correction is quite different from the static setting mode commonly adopted in the prior art.

In conclusion, the dynamic criterion adjusting mechanism has remarkable engineering realization value. The parameters required by the formulas can be directly realized through standard modules such as sampling function, standard deviation operation, voltage average calculation, topology identification, master station communication interface and the like in the intelligent switch equipment, and the threshold value update is executed in an embedded logic or edge controller of the equipment in a timing task mode. The mechanism is particularly suitable for self-adaptive protection judgment of multi-level and heterogeneous nodes in a large-scale power distribution network, and the robustness and the misoperation inhibition capability of the system are effectively improved.

And step 103, uploading fault information and key electrical data to a distribution automation master station by the intelligent switch equipment, judging a recoverable power supply area by the distribution automation master station in combination with network topology information, and dynamically generating a reconstruction scheme.

In this embodiment, step S103 mainly involves information uploading, global situation awareness, topology structure analysis, and generation of a power supply reconstruction scheme, which is a key bridge link connecting local fault recognition and full-network recovery scheduling. The core goal of this step is to enable the distribution automation master station to accurately acquire fault related data at a first time and to intelligently analyze a viable power restoration path based on the current network operating state and device topology. Embodiments thereof are described in detail below.

After the intelligent switch equipment locally completes the identification and the excision of the fault section, the fault related information can be uploaded to a distribution automation master station through a distribution automation communication network. The information includes at least the time stamp of the fault occurrence, the type of fault identified (e.g., single phase to earth, phase to phase short, etc.), current and voltage waveform data before and after the action, unique identification codes of the equipment in which it is located, the current state of the switch, and the electrical state parameters of the adjacent lines. The communication mode can adopt standard protocols such as IEC 60870-5-104, DNP3 or IEC 61850 and the like, and realizes the stable transmission of information through networks such as optical fiber, ethernet, 4G/5G or LoRa and the like. In the high-reliability application scene, a redundant channel and time synchronization mechanism can be further arranged to ensure the timely and accurate reception of the fault information by the master station.

After the master station system receives the uploaded data, the master station system firstly updates the received equipment state in the distribution network information management module to finish the topology real-time refreshing of the distribution network. The master station system should be provided with a complete network topological graph and a node connection relation database, and can automatically judge the topological position and isolation range of the fault section and the influence of the fault section on other loads by comparing the original structure with the current state change.

Then, the master station system calls a topology analysis and path reconstruction algorithm, analyzes the current power supply arrangement, the line running condition and the accessibility of the standby power supply path, judges which power supply areas are not directly affected by faults but lose power supply due to network structure interruption, and further marks a 'recoverable power supply area'. In the analysis process, factors such as feeder load capacity, switch operation constraint, bus sectionalizing state, availability of a tie switch, capacity of a backup power supply, safety margin and the like need to be comprehensively considered, so that overload, reverse power transmission or unstable system caused by blind closing are avoided.

And then, the master station system dynamically generates a power supply reconstruction scheme according to the calculation result, wherein the scheme comprises operation instructions such as which tie switches are closed, which original feeder lines do not bear power supply tasks, how loads are redistributed, whether voltage control points need to be adjusted or not, and the like. In order to realize quick deployment, the scheme should be issued in the form of an instruction queue or scripted operation, and a safety logic check mechanism is arranged, so that further deterioration of the system caused by misoperation is avoided. When needed, the master station can also perform multi-objective optimization on the reconstruction scheme, for example, a scheme with the least operation steps, the least voltage fluctuation or the optimal load balance is selected from a plurality of feasible paths, and the network self-adaptive capacity and the operation efficiency are improved.

It should be noted that, the execution of the whole step must be performed on the premise of real-time, and it is generally required that the data uploading, topology identification and reconstruction scheme generation are completed within several seconds to ten seconds after the fault occurs. Therefore, the master station system should have multi-threaded processing capability and high performance data processing architecture, supporting asynchronous event response and dynamic visualization.

In summary, the key of the implementation of step S103 is that the intelligent switching device is relied on to completely upload fault information, and the master station system is relied on to globally sense and rapidly calculate the power distribution network. Through the step, the rapid reconstruction preparation of the power supply capacity of the non-fault area can be realized, and conditions are created for the automatic operation and power supply recovery of the follow-up tie switch, so that the system is ensured to have strong self-healing capacity and running toughness under most fault conditions.

Furthermore, the intelligent switch device uploads fault information and key electrical data to a distribution automation master station, and the distribution automation master station judges a recoverable power supply area by combining network topology information and dynamically generates a reconstruction scheme, which comprises the following steps:

Before uploading fault information, the intelligent switching equipment performs time sequence segmentation and feature extraction processing on current and voltage waveform data in appointed time before and after the fault, extracts low-dimensional feature vectors comprising core contents of fault trend, waveform disturbance amplitude and spectrum distortion features through principal component analysis and standard deviation filter compression algorithm, and packages the low-dimensional feature vectors, a switching state and equipment identification together to form structured compressed data, and uploads the structured compressed data to a distribution automation master station;

after receiving the structured compressed data, the power distribution automation master station constructs a weighted graph model based on the topological position of the fault node, the historical power supply path record and the current contact switch state, wherein the path resistance, the residual load capacity, the switch state, the power supply priority and the safety margin are comprehensively considered by the side weight, and the minimum cost path search is carried out on the graph to determine a plurality of alternative power supply path sets;

After a plurality of feasible power supply reconstruction schemes are generated, the master station calculates the comprehensive score of each scheme in the aspects of power supply coverage rate, load balance and control efficiency through a multi-objective optimization algorithm based on local operation evaluation data returned by each intelligent switch device in a link, including predicted switching load, switch response delay and adjustable pressure capability, selects the scheme with the highest score as a final power supply reconstruction strategy, and writes fault recovery operation logs before executing, so that traceable management of the decision process is realized.

Firstly, after fault identification and isolation operation are completed, the intelligent switch equipment does not directly upload the full amount of original sampling data, but pre-processes current and voltage waveforms in a designated time period before and after the fault. This time period typically contains sampling data from 1 second before the fault to 3 seconds after the fault, with a sampling frequency of not less than 2 kHz being recommended to cover common fault transients. In the data processing process, the device firstly divides waveform data into a plurality of equal-length fragments according to a time axis, and performs Principal Component Analysis (PCA) operation on each fragment to identify the most representative waveform change characteristics. On the basis, a local disturbance section, namely a waveform distortion mutation area, is identified through a standard deviation filter, and statistical characteristics such as disturbance amplitude, spectrum energy change and the like corresponding to the local disturbance section are extracted. Finally, a group of low-dimensional feature vectors containing fault trends, disturbance features, frequency domain information and the like are formed, and the low-dimensional feature vectors are further compressed into structured data blocks. Meanwhile, the device packages and combines the metadata of the current switch state (such as tripping, action delay and the like), the unique device identifier, the timestamp and the like to form a complete structured compressed data unit, and the complete structured compressed data unit is uploaded to the power distribution automation master station through a communication link.

After receiving the compressed data, the master station can quickly locate the fault node from the network topology model maintained by the master station based on the fault device position identified in the received data. The master station combines the historical power supply path record (comprising each time of contact switch operation log and success rate) with the current available real-time state (opening and closing position, remote signaling effectiveness and the like) of the contact switch, and constructs a regional weighted graph by taking the fault node as the root. The nodes of the graph represent the power supply units or switch positions, and the side weights comprehensively consider the following dimensions of line resistance values, the residual load capacity of equipment in a path (back-pushed by the real-time load rate of the equipment), the controllability of the switch state (whether automatic switching is allowed or not), the historical power supply priority of the path (such as the important load channel can be increased in grading), and the topological safety margin (such as whether branch linkage risks exist in a branch or not). Based on the graph structure, the master station adopts a minimum cost path searching algorithm (such as Dijkstra or improvementAlgorithm) traversing the graph structure to generate a plurality of alternative power supply reconstruction path sets with smaller path cost and feasible structure.

After completion of the alternative path generation, the master station further enters a path evaluation phase. In this process, the master station will query the local operation assessment data reported by all relevant intelligent switching devices in the link, including but not limited to, expected acceptable load capacity (calculated based on the current operating current and device rated capacity difference), expected switching response time (which can be dynamically predicted by combining the switching device response history average value and the current action queue), whether to have voltage regulation capability (such as being provided with on-load voltage regulation control or reactive compensation device), and score these data as path adaptability indexes.

After the reconstruction scheme is determined, the master station also writes information such as the scheme generation time, the scoring model, the selected path number, the state snapshot before execution and the like into a fault recovery operation log, and forms a traceable record for later system maintenance, strategy correction and operation tracking. The log can be used for the post-examination of operation and maintenance personnel, can also be used as a data source of a system self-learning mechanism, and promotes the evolution optimization of a future scheme selection model.

In summary, the embodiment combines intelligent compression of the edge equipment and global analysis of the central system, and constructs an intelligent power supply recovery framework which has the advantages of clear structure, flexible response, path evaluation and traceability. In actual deployment, the system only depends on conventional sampling hardware, an embedded controller and a master station control platform, and can be evolved on the basis of the current power distribution automation technology.

Furthermore, the comprehensive scoring function for evaluating the power supply reconstruction path quality in the multi-objective optimization algorithmGiven by the following equation 2:

;

The goal of the scoring function is to translate the behavior of multiple paths on key performance metrics into a unified scoring value, where The smaller the path, the better the path, and the more suitable as a power supply restoration path.

The number of loads that can be covered by the current candidate path, i.e., the number of users or nodes that can be powered back up once the path is put into operation, is represented. The units may be the number (number of nodes) or the total kw load. The value can be calculated by the master station based on the network topology diagram simulation power supply result.

Corresponding to itIndicating the total load quantity in all fault shut-off regions at present, measured in the same units. The ratio of this term measures the ability of the path to recover power coverage, the closer the value is to 1, the stronger the recovery, thus taking its complementIndicating an "unrecovered proportion", the smaller the proportion, the better the path.

The residual capacity of the equipment in which the load maximum node is located in the candidate path is expressed in kilovolts (kVA) and can be estimated by the difference between the current running current and the rated capacity of the equipment. For example, a device rated capacity of 100 kVA, a current operating load of 70 kVA, then30 KVA.The theoretical maximum power supply capacity of the path can be comprehensively determined by the capacity of equipment and the transmission capacity of lines in all power supply paths.

The average response delay in milliseconds (ms) returned by all intelligent switching devices participating in the liaison operation in the path is represented. This value may be derived by the master station issuing an inquiry command to the device or estimated from historical operating records, including the communication transmission delay, the time required for the device to internally execute the control command.

Is a reference maximum allowable response time defined in the system, recommended to be set in the range of 500 ms to 1000 ms.

For each device in the path to calculate the comprehensive reliability score according to the self local voltage regulation capability, the historical stability score and the device action accuracy, the value range is 0 to 1, for example, the following can be set:

;

Wherein the method comprises the steps of Is the action accuracy in the last 10 operations; is the stability score (generated by the master station operation and maintenance module according to the failure reporting rate) within three months of the device; A capability score (support of 1, not support of 0) indicating that the device supports voltage regulation or reactive compensation. The higher the score, the more reliable the device in the path.

The recommended values are 0.4,0.3,0.2 and 0.1 for the weighting factors, respectively.

The master station calculates a scoring value for each alternative path by using the scoring functionAnd finally, selecting the path with the lowest score as the current power supply restoration execution scheme. The master station also records the scoring result and parameters participating in calculation, and the scoring result and parameters participating in calculation are written into a operation log for later retrospective analysis or model optimization.

And step S104, the distribution automation master station controls a group of tie switches to execute operation according to the reconstruction scheme, and switches the non-fault area to other power supply paths so as to realize power supply recovery.

In the method for isolating and recovering power supply by using the intelligent power distribution layered self-healing fault provided by the embodiment, step S104 is a key link for realizing power supply reconstruction, and is characterized in that the power supply is recovered as soon as possible in the area where the fault is not involved by performing control on a reconstruction scheme through a power distribution automation master station, so that the influence range of power failure is reduced and the toughness of the system is improved.

In step S103, the power distribution automation master station determines a recoverable power supply region based on the fault information and the network topology, and generates a complete power supply reconstruction scheme. The scheme defines the topological structure of the standby power supply path, the control actions of each switch device, the load distribution strategy and the like. In step S104, the master station issues an instruction to the reconfiguration scheme, and completes automatic switching of the power supply path by precisely controlling the closing action of the tie switch.

The distribution automation master station first needs to establish stable connection with each field switch device through a communication system. The communication network may take the form of optical fiber, dedicated wireless, LTE public network, 5G or industrial ethernet, depending on the geographical distribution and communication requirements of the distribution network. The communication protocol preferably uses standard protocols such as IEC 60870-5-104, IEC 61850 or DNP3 to ensure efficient transmission of instructions and acknowledgement of execution.

On the basis of stable communication link, the master station will send control instructions to the target tie switch device in turn according to the reconstruction scheme. These tie switches are typically installed at interface locations between adjacent feeders or in the distribution ring network structure with remote control capability and status feedback capability. Each instruction typically includes a device identification address, a type of desired action (e.g., closing or holding a latch), execution timing control parameters, and the like. In order to prevent misoperation or electric shock, the master station also needs to carry out multiple checks before sending, including equipment on-line state confirmation, current electric parameter check, operation interlocking condition judgment and the like. For example, in a scenario that a plurality of tie switches need to be operated simultaneously, the master station can set time sequence according to the path logic relationship, so that parallel switching-on is avoided to cause parallel operation of power supplies.

Before the switching-on operation is executed, the master station can also judge whether the carrying capacity of the standby power supply is enough according to the data such as real-time voltage, current and the like, so that the power supply voltage drop or the system protection action caused by the sudden load increase is prevented. When the power supply path condition is confirmed to be met, the master station sends a closing instruction, and the contact switch executes a switching action after receiving the instruction to complete closing of the power supply path.

After each switching operation, the master station also needs to read back through remote signaling and telemetry data to confirm whether the operation is successfully executed. For example, after the tie switch is closed, the master station can read the "closed" state quantity of the switch and perform cross-validation in combination with the downstream current change condition. Once the successful establishment of the path and the normal restoration of the load are confirmed, the master station updates the network topology and the equipment state information to enable the system to be in a new stable running state.

In the scene that a plurality of contact points need to be operated simultaneously or in stages, the master station can also use an operation queue and a sequence control logic to ensure that the power supply recovery process accords with the electrical safety standard, such as avoiding improper operations such as no-voltage closing, closed-loop closing and the like. In addition, in the path switching process, the master station can control the voltage quality in cooperation with a voltage regulating device (such as reactive compensation equipment and a tapping transformer) to ensure the power supply stability after recovery.

In the whole process, the operation does not need manual intervention, can be automatically completed within a few seconds after the fault occurs, and obviously improves the power supply reliability and self-healing capacity of the power distribution system.

In summary, step S104 realizes the dynamic reconfiguration of the power supply path by the accurate control of the interconnection switch by the automation master station. The process not only depends on a stable communication and control mechanism, but also depends on the rationality of a reconstruction scheme and the rigor of operation logic, thereby realizing rapid, safe and efficient non-fault area power supply recovery.

Still further, the distribution automation master station controls a group of tie switches to perform operations according to the reconfiguration scheme, and switches the non-fault area to other power supply paths to achieve power restoration, including:

Before sending a power supply reconstruction command, the power distribution automation master station calculates operation risk factors in advance based on the current load state, voltage stability and action success history of equipment where all tie switches are located, which are determined by the comprehensive scoring function, of the optimal path, marks nodes with operation risks larger than a preset threshold as nodes to be confirmed, delays sending out an execution command for the nodes to be confirmed, and simultaneously broadcasts a confirmation request;

After receiving the confirmation request, the intelligent switch equipment marked as the node to be confirmed carries out state negotiation on the intelligent switch equipment adjacent to the intelligent switch equipment in the link, acquires voltage fluctuation at two ends, load prediction after switching, equipment temperature rise or response saturation state information through chain type mutual confirmation communication, and carries out local operability judgment;

After receiving feedback of all key nodes, the distribution automation master station dynamically corrects an original path scheme, sends control instructions to all contact switch groups according to updated path sequences, sequentially controls a power supply reconstruction process in a step-by-step delay closing mode, confirms contact node states in real time after each step of actions, and enables unsatisfied parts to enter a retry mechanism and write operation logs.

In order to ensure that safe, stable and controllable path switching can be realized in the power supply reconstruction process after fault isolation is completed, the embodiment also provides a control flow of the distribution automation master station to the communication switch. The process not only considers the scoring priority of path planning, but also introduces an operation risk assessment, local feasibility negotiation and state feedback mechanism of the contact nodes, thereby realizing the whole process management of orderly closing, sectionalized confirmation and dynamic adjustment of the contact switches and obviously improving the anti-risk capability and the intelligent level of the system in the fault recovery process.

First, before the master station generates a power supply reconfiguration path and prepares to issue a switching control instruction, it will preferentially perform state evaluation on all the tie switch devices involved in the selected path. The evaluation is based on current real-time load data of each intelligent switch device, stability trend of voltage at the device, and past action execution records of the device, including action success rate, misoperation rate, overtime frequency and the like. Through the multidimensional information, the master station calculates an operation risk factor for each device, wherein the factor represents the safety and reliability of the device for executing switching operation under the current working condition. If the operation risk factor of a certain device exceeds a preset threshold set by the system, the master station marks the operation risk factor as a node needing to be confirmed, and does not immediately send a closing command to the node needing to be confirmed, but broadcasts a confirmation request.

After receiving the confirmation request, the intelligent switch device marked as the node to be confirmed does not independently make a judgment, but exchanges state information with other adjacent intelligent switch devices through a chain type mutual confirmation mechanism. Specifically, the device will actively request the real-time voltage fluctuation amplitude at both ends of the upstream and downstream switch, whether transient waveform distortion exists, load change trend after switching is expected, whether equipment overheat or action queue congestion exists, and the like. These collaborative data are integrated to determine whether the device is provided with field conditions for performing a closing operation. If the electrical environment or the equipment state is found to be abnormal, the equipment returns to 'refusing to execute', and the equipment carries with negotiation reasons, such as 'voltage fluctuation exceeding limit value' or 'predicted overload', etc., for the main station to make decision.

And after collecting feedback of all nodes to be confirmed, the master station dynamically adjusts an original power supply reconstruction path scheme based on a return result. If the partial nodes refused to switch or do not respond after overtime, the master station will preferentially try to partially bypass, replace the nodes or adopt a path segmentation power supply strategy on the basis of the original path, and if the power supply target cannot be met, the master station automatically switches to the power supply path with the inferior comprehensive score ranking and restarting the control preparation flow. After the final path is confirmed, the master station sequentially issues control instructions to all the contact switches participating in switching, and the control instructions are executed in a step-delay closing mode, namely, the action of each switch is independently initiated after a set safety delay, and after the action is completed, the master station confirms state feedback of the switch in real time, wherein the state feedback comprises whether the switch is closed successfully, whether current rises to a load level, whether voltage is stable within an allowable range and the like.

If a certain node does not complete closing or returns to an abnormal state within a specified time, the master station pauses a subsequent control instruction, and the power supply path is prevented from being interrupted or parallel connection impact is avoided. Such an abnormal node will be automatically brought into the retry queue and retry the closing operation within the maximum number of retries set by the system. Each abnormal response or retry operation is recorded in the master station operation log in detail, including the device ID, the action time stamp, the type of the abnormality, the retry run and the final result, so that the subsequent operation and maintenance analysis and algorithm optimization are facilitated.

In summary, this embodiment constructs a highly reliable power supply reconfiguration control scheme with adaptive adjustment capability by introducing a risk-based control sequence planning, a link-level state negotiation feedback mechanism, and sequential delay control and failure retry logic. According to the scheme, under the condition of not depending on manual intervention, the intelligent cooperative closing control of the multipath communication switch is realized, and the recovery speed and the power supply safety of the system are obviously improved.

Further, the distribution automation master station generates a control sequence and a control delay according to the following rules when controlling the tie switch to execute operation:

The control sequence of the tie switches is comprehensively ordered according to the topology level of the equipment, the expected load increment and the operation risk level, and the equipment with lower level, small load change and lower risk is preferentially controlled;

Setting independent delay time by a control instruction of each tie switch, wherein the delay time is dynamically adjusted according to a basic safety interval and by combining the load mutation amplitude and the operation risk level of the equipment;

And the power distribution automation master station executes state confirmation on each closed operation, if the response is abnormal or overtime, the follow-up instruction is paused, the abnormal equipment is brought into a retry queue, and the retry operation is provided with the maximum try times and is automatically recorded into an operation log.

The embodiment also provides a closing sequence and a time control strategy adopted when the distribution automation master station controls the tie switch to execute the power supply path switching operation. The strategy aims to ensure that the action of the contact switch completes the closing operation one by one in an optimal sequence and at a reasonable rhythm on the premise of ensuring the safety and the power supply continuity, thereby realizing the accurate power supply recovery of a non-fault area and avoiding the phenomena of overvoltage, current impact or system oscillation and the like.

The master station will sequence all tie switch devices to be operated before determining the final power supply reconstruction path and preparing to issue tie switch closing control instructions. The basis of the ordering is not in the traditional physical location order, but comprehensively considers three key parameters. First, the hierarchical position of the devices in the topology, typically devices near the end or low voltage branch are considered to be at a lower level, as their actions have less impact on the overall system, and preferential closure of these devices can reduce the risk of system upsets. The next is the expected load increase, i.e., the amount of load change each device may take after closing. The system can preferentially close the nodes with smaller expected load jump amplitude so as to realize a distributed power supply strategy of 'gradual loading and gradual recovery'. Finally, it is the operational risk level for each device, which may be determined jointly by the device's action history, the device health assessment results, and the stability of the electrical environment. The master station weights the three parameters to obtain a sequencing priority, and generates a complete control instruction issuing sequence according to the priority.

The control commands for each tie switch are not issued simultaneously, but with independent delay times. This delay time setting is based on a system predefined base safety interval and is dynamically adjusted in combination with the load variation amplitude and the operational risk level of the device. For example, for a device that needs to take a large load transition after closing, or that has a history of malfunctions, the master station will allocate it a long delay time, thus ensuring that after the previous node is closed, the system has enough time to complete the voltage stabilization and self-regulation of the load distribution. The dynamic delay mechanism not only can effectively prevent system fluctuation caused by concurrent actions of equipment, but also can enable the power supply reconstruction process to be more flexible and controllable.

After each tie switch receives the closing command and executes the operation, the master station immediately monitors the state change of the tie switch, and closed loop confirmation is usually carried out through indexes such as remote signaling signals, switch position indication, load current rising trend and the like. If the device fails to return to the expected state within the specified time or feedback abnormality occurs, such as closing failure, load unresponsiveness, severe voltage fluctuation and the like, the master station immediately pauses the issuing of subsequent control instructions, and the continuous expansion of a fault path or the system interlocking error is avoided.

At this point, the exception node will be automatically added to the retry queue. The primary station will, depending on the current system load conditions and device stability, arrange for it to try again to perform a closing action at a later time period, typically with a maximum allowable number of retries per device, typically not more than three. The results of each attempt, including success and failure, feedback signal status, execution time, exception type, etc., are fully recorded into the system log. The logs can be used for operation and maintenance audit, equipment health tracking or iterative learning of a follow-up power supply strategy model, and a set of highly reliable control system of 'state sensing-action execution-feedback recording-risk closed loop' is constructed for the whole system.

Through the control sequence generation and delay strategy setting, the embodiment not only realizes a refined power supply recovery execution mechanism, but also remarkably improves the running stability, response flexibility and abnormality coping capability of the system. Unlike the conventional method of uniform command and no difference in centralized control, the embodiment utilizes a hierarchical topology sensing and equipment behavior prediction model to realize the self-healing control logic which is executed on demand, in sequence and in feedback in full-path communication, and has practical value.

And step 105, performing retrospective analysis on current waveforms, harmonic features or partial discharge signals before faults occur while performing power supply restoration, and if abnormal trends are identified, recording abnormal events and updating an operation strategy for early warning and distinguishing similar faults in the future.

In the intelligent power distribution layered self-healing fault isolation and power supply recovery method provided by the embodiment, step S105 not only bears the system backtracking analysis function after faults, but also builds a self-adaptive optimization foundation for the future, and the function is to endow the system with continuous learning and evolution capability through abnormal trend identification and strategy updating, so that the self-healing and prediction level of the whole power distribution network facing complex running environments is improved.

After the power supply recovery operation starts, the power distribution automation master station immediately starts a set of historical data backtracking mechanism for analyzing a key electrical parameter evolution process in a period of time before a fault occurs. These parameters mainly include current waveform, voltage waveform, harmonic distribution, zero sequence component, negative sequence current, partial discharge characteristics, power quality index, etc. The analysis time window can be flexibly set according to the fault type and the response characteristic of the system, and usually, backtracking of data for 5 to 30 seconds is recommended to cover the precursor change stage before the fault.

The sources of trace back data include data that is locally cached by the intelligent switching device and historical sample data that is stored in the master station database. To achieve high resolution analysis, the system should employ a sampling frequency of no less than 5kHz, supporting transient waveform extraction when necessary. For these data, the master station uses various algorithms for signal analysis and pattern recognition. For example, a Fast Fourier Transform (FFT) is adopted to analyze the variation trend of harmonic distribution so as to identify whether the higher harmonic surge is an unstable precursor before a fault, a sliding window zero sequence current mean value and a standard deviation are adopted to extract abnormal disturbance, and for a partial discharge signal, a high-frequency noise signal and pulse count statistics can be combined to realize the mining of early signs of cable insulation degradation.

After various feature extraction is completed, the system compares the current fault case with a historical known fault model or an expert knowledge base to judge whether the fault has identifiable precursor trend. If an abnormal trend with statistical significance is identified, the main station files the trend in association with the fault event, and records related key information such as lines, equipment, time points, parameter change curves and the like, so as to form a structured abnormal event record.

Further, the system may automatically adjust the operating strategy based on the characteristics of the anomaly event. For example, upon identifying that a feeder line has a type of harmonic distortion before a plurality of historical faults, the system may moderately increase the frequency of monitoring, decrease the abnormal waveform trigger threshold, or suggest that a dispatcher schedule an overhaul ahead of time for the line. Also, for cable segments where partial discharge signals are frequently abnormal, the system may incorporate an emphasis on routing or reconfigure the power paths to reduce the load.

The strategy updating not only acts on the treatment process of the fault, but also can be written into an operation strategy database of the master station, thereby becoming a knowledge base for judging similar faults in the future. The strategy can comprise multiple dimensions such as dynamic threshold adjustment, early warning level setting, equipment priority ordering or linkage switch strategy change, and the early warning accuracy and response efficiency of the system in the future facing similar working conditions are enhanced.

It is emphasized that the whole process does not need manual intervention, and the system automatically completes data extraction, analysis, matching and strategy updating, so that the early warning mechanism is ensured to have instantaneity, adaptability and sustainability.

In summary, step S105 is not only retrospective mining of the pre-fault state, but also is a future-oriented adaptive risk control mechanism, which enables the system to have self-learning, self-optimizing and self-defending capabilities by deeply analyzing historical waveforms, identifying abnormal trends and dynamically optimizing operation strategies, so as to greatly improve the early warning capability and self-healing level of the intelligent power distribution system.

Further, the retrospective analysis of the current waveform, the harmonic characteristic or the partial discharge signal before the fault occurs includes:

After detecting a fault, the distribution automation master station invokes the original sampling data cached by each related intelligent switching device in a pre-fault period, establishes a time continuous waveform set covering a plurality of devices, wherein each data sequence in the time continuous waveform set comprises a time stamp, a current value, a voltage value, a zero sequence quantity and each subharmonic amplitude, and uniformly aligns reference time nodes to form a comparable analysis basis;

The power distribution automation master station compares and analyzes characteristic parameters in the time continuous waveform set to identify whether a continuous change trend exists in the following three indexes before a fault, namely whether a specific subharmonic amplitude increases in a plurality of continuous sampling periods, whether a stable offset trend occurs in a zero sequence component of current, whether a voltage fluctuation amplitude repeatedly exceeds a set fluctuation threshold in a short period, and if any index meets a trend change condition, determining that the fault has an abnormal precursor characteristic;

The power distribution automation master station compares the fault event with the abnormal precursor characteristics with the current response process record and automatically updates the operation strategy data, wherein the operation strategy data comprises weight parameters added corresponding to the precursor characteristics and adjustment intervals of threshold sensitivity in future criteria, or issues early warning configuration parameter update instructions to related intelligent switch equipment for improving the advanced recognition and response capacity of subsequent similar events.

The embodiment also provides an electric abnormality pre-judging mechanism which can be implemented by the system and combines the judgment based on the trend of the sampling data and the strategy updating. The mechanism focuses on the reuse of historical data after the occurrence of faults, emphasizes the coherent coordination among the data organization structure, the trend judging method and the strategy updating logic, and can remarkably improve the identification capability of the power distribution system on potential faults and the evolution capability of an operation strategy.

When the system detects that a fault occurs in a line, the distribution automation master station can immediately call local cache data in each intelligent switch device related to the fault. Each intelligent switching device continuously records the electric quantity data of the connected line with a fixed sampling period during normal operation, and caches the electric quantity data in a local storage module. When issuing a data retrieval command, the master station will require each device to upload its original sampled data within seconds of the failure, including the time stamp of each sampling point, the three-phase current value, the three-phase voltage value, the zero sequence current component, and the voltage and current amplitude of at least the first five harmonics. In order to facilitate subsequent comparison and analysis, the master station performs unified time alignment processing on data sequences reported by all devices, selects failure triggering time as a unified reference point, converts time information in each data sequence into relative time relative to the reference point, and ensures that analysis windows have consistency and synchronism in time.

Based on the constructed time-continuous waveform set, the distribution automation master station then performs trend analysis of parameter levels. First, the master station will detect whether each of the subharmonic amplitudes, and in particular the third, fifth and seventh subharmonics, exhibit an increasing trend over a plurality of consecutive sampling periods before a fault occurs. If a harmonic component continues to rise in amplitude over two or more cycles, the potential for power quality disturbances may be initially identified. Second, the master station checks if the zero sequence current component starts to drift slowly from a zero or normal small amplitude state, with a significant asymmetric load characteristic, which is usually a precursor to a low level ground fault or insulation degradation. Again, the master station counts the frequency and amplitude of the voltage fluctuation in unit time, and determines whether the voltage fluctuation continuously exceeds a set short-period fluctuation threshold, for example, fluctuation exceeding a rated value ± 10% occurs more than three times in one second. When any one of the above indexes satisfies the trend change condition, the master station marks the fault as having an abnormal precursor characteristic, namely, not only the electric fault is generated, but also a definite physical precursor signal can be traced back from the history data before the fault.

In order to incorporate the current abnormal event into the knowledgeable operation policy system, the master station can correlate and compare the fault with the response process. For example, whether the present failure can be isolated quickly within a predetermined time, whether there is a malfunction or a malfunction rejection phenomenon, whether a specific contact path is used, or the like. And in combination with the response records, the master station adjusts the current operation strategy library. Specifically, for successfully identifying parameters of precursor features, the master station automatically increases weight parameters of the precursor features in future fault criteria, and improves sensitivity of the features in the local criteria of the intelligent switch equipment. Meanwhile, for the event that the trend abnormality occurs but is not responded in time, the master station can adapt to the action threshold in the down-regulating part criterion, so that the equipment intervenes in judgment earlier. In addition, the main station can directly send the early warning parameters to the related intelligent switch equipment, so that the main station can send out an alarm signal in advance or preset power-off action under the allowable condition once detecting similar trend in future operation.

Through the backtracking analysis and the strategy self-adaptive updating mechanism, the invention not only realizes the reverse modeling of the cause of the fault, but also establishes a self-learning closed loop taking trend identification-behavior association-strategy evolution as a core, and improves the protection capability of the system on hidden faults or marginal disturbance. Unlike traditional single-point waveform threshold judgment or purely manual experience rule-based, the embodiment provides a cross-node, cross-time and accumulative power distribution abnormality recognition logic, and provides a solid technical basis for realizing self-healing control under hierarchical coordination.

A second embodiment of the application provides an electronic device, including:

A processor;

And the memory is used for storing a program, and the program is used for executing the intelligent power distribution layered self-healing fault isolation and power supply recovery method provided in the first embodiment of the application when being read and executed by the processor.

A third embodiment of the present application provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs an intelligent power distribution layered self-healing fault isolation and power restoration method provided in the first embodiment of the present application.

While the application has been described in terms of preferred embodiments, it is not intended to be limiting, but rather, it will be apparent to those skilled in the art that various changes and modifications can be made herein without departing from the spirit and scope of the application as defined by the appended claims.

Claims (9)

1. An intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method, characterized by comprising: Establishing a hierarchical control architecture in the power distribution system, the hierarchical control architecture comprising a lower-layer intelligent switch device and an upper-layer distribution automation master station; The intelligent switch device collects the current and voltage status of the connected line. When a signal that meets the preset fault criteria is detected, the fault section is identified locally and the corresponding line section is cut off, so as to achieve fault isolation in the first time. The intelligent switch device uploads fault information and key electrical data to the distribution automation master station, which determines the recoverable power supply area based on the network topology information and dynamically generates a reconstruction plan; The distribution automation master station controls a group of tie switches to perform operations according to the reconstruction plan, switching the non-fault area to other power supply paths to achieve power supply recovery; While power restoration is being performed, the current waveform, harmonic characteristics or partial discharge signal before the fault occurs is retrospectively analyzed. If an abnormal trend is identified, the abnormal event is recorded and the operation strategy is updated for early warning and identification of similar faults in the future. The hierarchical control architecture is established in the power distribution system, including that after the hierarchical structure is deployed, the distribution automation master station performs a chain mutual recognition operation on the intelligent switch devices in each group of zones, so that the adjacent switch devices complete identity authentication and communication handshake locally, and establish a self-organizing boundary recognition capability between devices, so as to support regional autonomous fault handling when the master station is offline; The chain mutual recognition operation includes: After receiving the chain mutual recognition instruction issued by the distribution automation master station, the intelligent switch device conducts mutual recognition negotiation with other physically adjacent intelligent switch devices through a low-latency local communication protocol, and determines the fault response priority of each intelligent switch device in the link based on the device identification, geographic location information and historical response performance indicators; After completing mutual recognition, each intelligent switch device periodically broadcasts its own state summary information to adjacent intelligent switch devices. The state summary information includes the current load state, fault judgment summary and received control instruction identification, which is used to establish an operation state consistency judgment mechanism in a local area; In the event that communication with the distribution automation master station is interrupted or unreachable, the mutually recognized intelligent switching devices will independently operate the reconstruction strategy in the local area according to the pre-negotiated response sequence, including delayed closing control, cascade fault isolation and communication chain integrity self-check, to achieve autonomous self-healing operation within a limited range. 2. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 1 is characterized in that the establishment of a hierarchical control architecture in the power distribution system further comprises: In the initialization stage, the distribution automation master station generates a logical hierarchical structure of multi-level power supply areas based on the geographic information system GIS and historical operation topology data, and allocates corresponding switch equipment addresses, fault judgment templates and communication strategies to each level node to achieve synchronous matching of topological structure and functional configuration; By sending down configuration information, the operating roles and action permissions of each intelligent switch device are synchronized, so that it has different levels of fault response priority, coordination waiting time and reporting window parameters according to the network layer it is in, thereby realizing adaptive coordinated response among multiple devices in concurrent fault scenarios. 3. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 1 is characterized in that the intelligent switch device collects the current and voltage status of the connected line, and when a signal that meets the preset fault criterion is detected, the fault section is identified locally and the corresponding line section is cut off to achieve the first-time fault isolation, including: Intelligent switchgear adopts a dynamic judgment adjustment mechanism. According to the topological position of the device, the fluctuation characteristics of the historical load curve and the coordinated parameters of other intelligent switchgear in the link, the threshold range of local fault identification is adjusted in real time during operation, so that fault identification is closer to the scene characteristics and avoids false operation or refusal to operate. When a designated intelligent switch device detects that the electrical status of the line it is connected to meets the preliminary fault characteristics, it does not immediately remove the fault, but synchronizes the preliminary fault information to other adjacent intelligent switch devices in the link through a chain communication mechanism. Multiple devices jointly complete multi-point synchronous verification and judgment to confirm that it is a consistent fault, and then select the device with the highest priority to perform isolation action, and other devices enter a waiting response state to prevent repeated removal; After the removal action is completed, the post-fault data caching mechanism is executed. The intelligent switching device that performs the isolation action records the complete electrical waveform and judgment parameters before and after the fault, and saves the recorded data in a structured manner in the chain communication network for subsequent master station backtracking analysis or other equipment linkage strategy adjustment to achieve event cascade closed-loop management. 4. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 3 is characterized in that the dynamic criterion adjustment mechanism of the intelligent switch device includes calculating the local short-circuit current action threshold based on the following formula 1 : ; in, The maximum load current in the last 24 hours; is the standard deviation of the load current in the last hour, used to characterize the current load fluctuation; is the average phase voltage within the current 15 minutes; It is the target voltage value issued by the distribution automation master station; Indicates the topological level parameter of the intelligent switch device. The value of the main feeder device is 1.0, the value of the branch device is 0.5, and the value of the terminal node device is 0; is the empirical coefficient. 5. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 4 is characterized in that the intelligent switch device uploads fault information and key electrical data to the distribution automation master station, and the distribution automation master station determines the recoverable power supply area based on the network topology information and dynamically generates a reconstruction plan, including: Before uploading fault information, the intelligent switch device performs time series segmentation and feature extraction on the current and voltage waveform data within the specified time before and after the fault. Through principal component analysis and standard deviation filter compression algorithm, it extracts low-dimensional feature vectors including the core content of fault trend, waveform disturbance amplitude, and spectrum distortion characteristics, and packages them together with the switch status and equipment identification to form structured compressed data and upload them to the distribution automation master station; After receiving the structured compressed data, the distribution automation master station constructs a weighted graph model based on the topological location of the fault node, the historical power supply path record and the current contact switch status. The edge weight comprehensively considers the path resistance, remaining load capacity, switch status, power supply priority and safety margin, and searches for the minimum cost path on the graph to determine a set of multiple alternative power supply paths. After generating multiple feasible power supply reconstruction schemes, the master station calculates the comprehensive score of each scheme in terms of power supply coverage, load balancing and control efficiency through a multi-objective optimization algorithm based on the local operation evaluation data returned by each intelligent switching device in the link, including the expected switching load, switch response delay, and adjustable voltage capability. The scheme with the highest score is selected as the final power supply reconstruction strategy, and a fault recovery operation log is written before execution to achieve traceable management of the decision-making process. 6. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 5 is characterized in that the comprehensive scoring function used to evaluate the quality of the power supply reconstruction path in the multi-objective optimization algorithm is: Given by the following formula 2: ; in, Indicates the amount of load that the current candidate path can cover; Indicates the total load quantity in all current fault cut-off areas; The remaining available capacity of the device where the heaviest-loaded node in the path is located; The theoretical maximum power supply capacity of this path; Indicates the average response delay returned by all intelligent switch devices participating in the liaison operation in the path; is the reference maximum allowed response time defined in the system; It is the comprehensive reliability score of each device in the path calculated based on its local voltage regulation capability, historical stability score and device action accuracy, with a value range of 0 to 1; is the weighting factor. 7. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 6 is characterized in that the distribution automation master station controls a group of tie switches to perform operations according to the reconstruction plan to switch the non-fault area to other power supply paths to achieve power supply restoration, including: Before issuing a power supply reconstruction command, the distribution automation master station pre-calculates the operation risk factor based on the current load status, voltage stability and successful operation history of all the tie switches involved in the optimal path determined by the comprehensive scoring function, and marks the nodes with operation risks greater than a preset threshold as nodes requiring confirmation, delays issuing execution commands for the nodes requiring confirmation, and broadcasts confirmation requests at the same time; After receiving the confirmation request, the intelligent switch device marked as the node to be confirmed will negotiate the status with its adjacent intelligent switch devices in the link, obtain the voltage fluctuation at both ends, load prediction after switching, equipment temperature rise or response saturation status information through chain mutual recognition communication, and make local operability judgment; if it is confirmed that the switching conditions are met, it will return the switching status flag allowed, otherwise it will return the refusal to execute with the negotiation reason; After receiving feedback from all key nodes, the distribution automation master station dynamically modifies the original path plan, and issues control instructions to all interconnecting switch groups according to the updated path sequence, and controls the power supply reconstruction process in sequence using a step-by-step delayed closure method. After each step, the interconnection node status is confirmed in real time, and the unsatisfied part enters the retry mechanism and is written into the operation log. 8. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 7, characterized in that the distribution automation master station generates a control sequence and a control delay according to the following rules when controlling the tie switch to perform an operation: The control sequence of the tie switches is comprehensively sorted according to the topological level of the equipment, the expected load increment and the operational risk level, with priority given to controlling equipment with a lower level, small load change and lower risk; The control instruction of each tie switch sets an independent delay time, and the delay time is dynamically adjusted according to the basic safety interval and in combination with the load mutation amplitude and operation risk level of the equipment; The distribution automation master station performs status confirmation for each closing operation. If the response is abnormal or timed out, subsequent instructions are suspended and the abnormal device is included in the retry queue. The retry operation has a maximum number of attempts and is automatically recorded in the operation log. 9. The intelligent power distribution hierarchical self-healing fault isolation and power supply restoration method according to claim 8, characterized in that the retrospective analysis of the current waveform, harmonic characteristics or partial discharge signal before the fault occurs comprises: After a fault is detected, the distribution automation master station retrieves the original sampled data cached by each relevant intelligent switch device in the pre-fault period, and establishes a time-continuous waveform set covering multiple devices. Each data sequence in the time-continuous waveform set includes a timestamp, current value, voltage value, zero-sequence quantity, and amplitude of each harmonic, and uniformly aligns the reference time node to form a comparable analysis basis; The distribution automation master station compares and analyzes the characteristic parameters in the time-continuous waveform set to identify whether the following three types of indicators have a continuous change trend before the fault: first, whether the amplitude of a specific subharmonic increases in a number of consecutive sampling cycles; second, whether the zero-sequence component of the current shows a stable offset trend; third, whether the voltage fluctuation amplitude repeatedly exceeds the set fluctuation threshold in a short period; if any of the above indicators meets the trend change condition, it is determined that the fault has abnormal precursor characteristics; The distribution automation master station compares fault events with abnormal precursor characteristics with the current response process records, and automatically updates the operation strategy data, including: weight parameters added to the corresponding precursor characteristics, adjustment range of threshold sensitivity in future judgment criteria, or sends early warning configuration parameter update instructions to related intelligent switching devices to improve the early identification and response capabilities of subsequent similar events.