A fault section fast isolation and recovery method of intelligent power distribution terminal

Abstract

This invention relates to the field of power distribution network relay protection technology and discloses a method for rapid isolation and recovery of fault sections in intelligent power distribution terminals. It addresses the problems in traditional methods where untimely updates to network connections lead to inconsistencies between fault section location and actual electrical connections, resulting in maloperation of disconnecting switches, expansion of fault isolation range, and accidental power outages in intact sections. The method first collects local electrical quantities, switch positions, and distributed power source status, and receives information from adjacent terminals to identify and update the current network connection relationships. It then adjusts overcurrent protection setting parameters and directional protection criteria based on distributed power source output and power flow direction. After fault initiation, it exchanges multi-terminal directional judgment information to determine the fault section and controls the isolation of switches at both ends. Finally, it performs recovery control based on the lost load and candidate power supply paths, achieving accurate location of fault sections, minimal isolation, and rapid power recovery of non-fault sections in active power distribution networks.

Description

Technical Field

[0001] This invention relates to the field of power distribution network relay protection technology, specifically to a method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal. Background Technology

[0002] With the advancement of the "dual carbon" target and the accelerated construction of new power systems, distributed power sources such as distributed photovoltaics, energy storage devices, and electric vehicle charging facilities are being connected to medium and low voltage distribution networks in large quantities. This is gradually transforming the distribution network from a traditional single-source radial power supply structure to an active distribution network with multiple power sources and bidirectional power flow. In such distribution networks, distribution terminals are typically used to collect line electrical quantities and switch status, and to cooperate in locating fault sections, isolating faults, and restoring power supply to non-faulty sections after a fault occurs. Chinese authorized invention patent CN110988596B discloses a novel active distribution network fault section location method, which establishes an active distribution network fault section location model and combines multi-source information to assist in the division of fault areas to obtain fault section location results. Chinese authorized invention patent CN109193582B discloses an intelligent distribution network area protection control system and control method, which realizes distribution network fault location, fault isolation, and self-healing recovery control through intelligent distributed devices, flexible DC coordinated controllers, and GOOSE real-time information interaction.

[0003] The aforementioned technologies can improve the fault handling capabilities of distribution networks to some extent, but they still have shortcomings in application scenarios with a high proportion of distributed power sources. On the one hand, the grid connection relationship may change during the commissioning or decommissioning of distributed power sources, load transfer, or maintenance operations. If the intelligent distribution terminal still relies on pre-set grid connection relationships, fixed switch logic, or manually maintained operating information for fault judgment, it is easy to cause inconsistencies between the fault location result and the actual electrical connection relationship due to untimely updates of the grid connection relationship, which may lead to malfunctions of disconnecting switches, expansion of the fault isolation range, and false power outages in intact sections. On the other hand, distributed photovoltaic, energy storage, and other equipment are usually connected to the grid through power electronic devices. The short-circuit current amplitude is limited, and the current direction fluctuates with changes in operating mode and output status. If a fixed overcurrent setting value or a single-direction criterion is still used for fault identification, overcurrent protection may fail to operate or directional protection may malfunction, thus prolonging the fault clearing and power restoration time. Therefore, it is urgent to propose a method for rapid isolation and restoration of fault sections in intelligent distribution terminals suitable for active distribution networks. By identifying changes in the grid connection relationship in real time and adaptively correcting fault criteria and setting parameters in combination with bidirectional power flow conditions, the method can achieve accurate location of fault sections, minimize the isolation of fault range, and rapidly restore power supply to non-fault sections, thereby improving the reliability of fault handling and power restoration efficiency of active distribution networks. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for rapid isolation and recovery of fault sections in intelligent power distribution terminals. This method solves the problems in traditional methods where untimely updates to the grid connection relationship lead to inconsistencies between the fault section location results and the actual electrical connection relationship, resulting in malfunctions of disconnecting switches, expansion of the fault isolation range, and accidental power outages in the affected sections.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal includes:

[0007] S1. The intelligent power distribution terminal collects local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position and distributed power supply operation status, and receives electrical quantities, switch status and grid verification information sent by adjacent terminals.

[0008] S2. Identify the current grid connection relationship based on locally collected information and adjacent terminal information, and update the grid connection relationship record when the switch changes position, distributed power supply is put into or taken out, or load is transferred.

[0009] S3. Adjust the overcurrent protection setting parameters and directional protection criteria based on the current grid connection relationship, distributed power output status and power flow direction;

[0010] S4. After detecting the fault start, extract the electrical quantity data before and after the fault, generate local fault direction determination information, and exchange fault direction determination information with adjacent terminals.

[0011] S5. Determine the fault section based on the fault direction determination information of each terminal and the current network connection relationship, control the switches at both ends of the fault section to open and isolate, and control the tie switch or backup power supply access switch to close and restore based on the load of the power-loss section and the candidate power supply path.

[0012] Prioritizing the collection of local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position, and distributed power source operating status, including:

[0013] During local information collection, the collection channel and communication receiving channel are uniformly time-marked, and voltage, current and zero-sequence quantities are collected according to the power frequency cycle. The collected data is written into an electrical quantity record containing channel number, collection time, quality flag and source flag.

[0014] Debouncing confirmation is performed on switch position changes, and power type identifier and fault current contribution range identifier are generated based on the operating status of distributed power sources.

[0015] When an abnormality is detected in the local acquisition channel, the abnormal channel is marked with a quality indicator and verified by means of adjacent phase calculation, zero-sequence component verification, or comparison of similar electrical quantities between adjacent terminals.

[0016] Prioritize receiving electrical quantities, switch status, and network structure verification information sent by adjacent terminals, including:

[0017] When receiving information sent by neighboring terminals, perform source verification, timestamp verification, sequence number continuity verification, and field integrity verification on the messages sent by neighboring terminals. Write messages that pass the verification into the neighboring terminal information table, mark messages with timestamp deviation as delayed messages, and record messages that fail the verification as communication abnormal events.

[0018] The adjacent terminal information table records voltage amplitude, voltage phase angle, current amplitude, current direction, switch status, verification timestamp, and trust level, and adjusts the trust level of adjacent terminals based on the message verification results.

[0019] Prioritize identifying the current network connection relationships based on locally collected information and information from adjacent terminals, including:

[0020] When identifying the current network connection relationship, a connection relationship matrix and an adjacency relationship table are used to jointly record the connection status, connection strength, adjacency relationship and update time of the terminal pairs;

[0021] The connection status is categorized into direct connection, indirect connection, electrical isolation, and pending confirmation.

[0022] Within the statistical window, the connection status of the terminal pair is determined based on the voltage amplitude difference, voltage phase angle difference, current change trend, active power change trend, and zero-sequence quantity synchronization change degree.

[0023] Prioritize updating the grid connection records during switch changes, distributed power generation commissioning / decommissioning, or load transfer, including:

[0024] When updating the connection relationship record, the connection status is verified by using weak characteristic current verification or operation status linkage verification after the switch status changes, the operating status of the distributed power source changes, or the load transfer is triggered.

[0025] Based on the differences between the old and new connection relationships, a connection change index is generated, and a connection relationship summary is sent to adjacent terminals, power distribution master stations, or higher-level terminals.

[0026] Based on the returned summary matching results, perform partial updates, mark pending confirmation, retain rollback versions, or rebuild the connection records.

[0027] Prioritize adjusting overcurrent protection setting parameters and directional protection criteria based on the current network connection relationships, distributed power output status, and power flow direction, including:

[0028] When adjusting protection parameters, protection parameter update tasks are created based on periodic triggering and event triggering.

[0029] The power flow sensitivity state is determined based on the current network connection version, distributed power source operating status, local load current, power flow direction, and forward and reverse short-circuit current contributions.

[0030] Load avoidance and fault sensitivity checks were performed on the overcurrent protection parameters.

[0031] Based on the polarity relationship between the zero-order component, the negative-order component, and the mutation amount, a directional protection criterion, a low margin flag, and a protection parameter version number are formed.

[0032] Prioritizes extracting electrical quantity data before and after the fault is detected, and generating local fault direction determination information, including:

[0033] During fault start identification, the overcurrent start condition and transient energy start condition are monitored based on the current protection setting parameters. The fault start time is determined by the sampling judgment window that first meets the start condition, and the voltage, current, zero sequence quantity and switch status data before and after the fault are captured.

[0034] After processing the intercepted data by checking for missing channels, saturated sampling, data completion, and low-confidence marking, the direction protection criterion is invoked.

[0035] Local fault direction determination information is generated based on zero-order direction characteristics, negative-order direction characteristics, and transient mutation direction characteristics.

[0036] Prioritizes and exchanges fault direction determination information with adjacent terminals, including:

[0037] When exchanging fault direction information, the local fault direction determination information is sent to the adjacent terminal through a strong real-time communication message. The strong real-time communication message records the terminal number, switch number, fault start time, direction determination result, direction matching degree, protection parameter version number, network connection relationship version number and data quality flag.

[0038] The received fault direction determination information is checked for version consistency. The message trust level is adjusted according to the check result. The checked fault direction determination information is written into the fault coordination cache for minimum direction combination judgment.

[0039] Prioritize determining the faulty section based on the fault direction information of each terminal and the current network connection relationship, and control the switches at both ends of the faulty section to trip and isolate it, including:

[0040] When confirming the fault section, adjacent terminal pairs are used as the judgment unit. Candidate scores are generated based on the direction of the relative orientation, the reliability of the connection relationship, the consistency of the fault start time, and the data quality score.

[0041] For candidate sections whose candidate scores meet the predetermined confirmation conditions, fault recording data is used to verify current energy. After the verification is passed, the terminals at both ends of the section perform double-end handshake confirmation and synchronous tripping control.

[0042] If any section-end switch does not report the open position, an open retry is performed, and an expanded isolation request is sent after the retry fails.

[0043] Prioritize restoring power by closing the tie switch or backup power supply switch based on the load of the power-loss section and candidate power supply paths, including:

[0044] When restoring power supply control, the section to be restored is determined based on the power loss node after the circuit breaker trips, the estimated load before the fault, and the load importance level. The candidate power supply paths are scored for power capacity margin, line thermal stability margin, end voltage drop, loop closing safety, and islanding risk. The candidate power supply path that meets the safety boundary is selected to perform the circuit breaker closing control.

[0045] When there is no single candidate power supply path that meets the safety boundary, the section to be restored is restored in sections, and voltage verification, islanding risk removal verification and abnormal rollback processing are performed after the circuit is closed.

[0046] Compared with the prior art, the present invention provides a method for rapid isolation and recovery of fault sections in intelligent power distribution terminals, which has the following beneficial effects:

[0047] 1. This invention, by uniformly time-marking, quality-verifying, and managing the reliability level of local operating data and adjacent terminal information, and by combining the connection relationship matrix and adjacency relationship table to identify and update the grid connection relationship, enables intelligent power distribution terminals to obtain grid information consistent with the actual operating state when distributed power sources are put into operation or deactivated, loads are transferred, or switches are changed. Simultaneously, based on the output status of distributed power sources, the contribution of forward and reverse short-circuit currents, and the power flow direction, the overcurrent protection setting parameters and directional protection criteria are dynamically adjusted. After a fault occurs, the fault section is determined by combining multi-terminal fault direction determination information, version consistency verification, and fault waveform review. After double-ended disconnection and isolation, recovery control is performed according to the power loss load, the safety boundary of candidate power supply paths, and islanding risk. This solves the problems of inaccurate fault location and expanded isolation range caused by delayed grid connection relationship updates, as well as protection failure or maloperation caused by reverse feedin from distributed power sources. It achieves accurate fault section location, minimized fault range isolation, and rapid power restoration to non-faulty power loss sections.

[0048] 2. This invention performs consistency verification on collected data, adjacent terminal messages, protection parameter versions, and network connection relationship versions. When data is missing, channel saturation occurs, communication delays occur, version deviations occur, or candidate paths do not meet safety boundaries, a low-confidence flag, a pending confirmation status, a degradation judgment, and an abnormal rollback mechanism are set. This prevents abnormal data, expired parameters, or inconsistent messages from directly participating in switch control. Simultaneously, through fault collaborative caching, minimum direction combination judgment, circuit breaker retry, expanded isolation request, voltage verification after closing, and islanding risk removal verification, a closed-loop verification of fault location, isolation actions, and power restoration processes is performed. This solves the problems of misjudgment, miscontrol, and recovery failure caused by unstable data quality, inconsistent terminal coordination, or lack of verification of control actions in existing fault handling processes, thereby improving the reliability, anti-interference capability, and operational safety of fault handling. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of a method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to the present invention.

[0050] Figure 2 This is a schematic diagram illustrating local data acquisition and adjacent terminal information archiving in this invention;

[0051] Figure 3 This is a schematic diagram of the grid connection relationship identification and updating mechanism of the present invention;

[0052] Figure 4 This is a schematic diagram illustrating the dynamic adjustment of the protection setting parameters and direction protection criteria of the present invention;

[0053] Figure 5 This is a schematic diagram of the collaborative fault direction determination and fault segment location of the present invention;

[0054] Figure 6 This is a schematic diagram of the fault isolation and power restoration control closed loop of the present invention;

[0055] Figure 7 This is a schematic diagram illustrating the deployment and information interaction of the intelligent power distribution terminal for the active power distribution network according to the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1: Figure 1 - Figure 6A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal is presented, including:

[0058] S1. The intelligent power distribution terminal collects local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position and distributed power supply operation status, and receives electrical quantities, switch status and grid verification information sent by adjacent terminals.

[0059] S2. Identify the current grid connection relationship based on locally collected information and adjacent terminal information, and update the grid connection relationship record when the switch changes position, distributed power supply is put into or taken out, or load is transferred.

[0060] S3. Adjust the overcurrent protection setting parameters and directional protection criteria based on the current grid connection relationship, distributed power output status and power flow direction;

[0061] S4. After detecting the fault start, extract the electrical quantity data before and after the fault, generate local fault direction determination information, and exchange fault direction determination information with adjacent terminals.

[0062] S5. Determine the fault section based on the fault direction determination information of each terminal and the current network connection relationship, control the switches at both ends of the fault section to open and isolate, and control the tie switch or backup power supply access switch to close and restore based on the load of the power-loss section and the candidate power supply path.

[0063] This method is applied to feeder fault handling scenarios in a 10 kV medium-voltage active distribution network. Multiple intelligent distribution terminals are installed along the distribution line. Each intelligent distribution terminal is associated with a sectionalizing switch, tie switch, backup power access switch, and distributed power grid connection point, and interacts with adjacent terminals via peer-to-peer communication. The objects involved in fault handling include at least the intelligent distribution terminals, adjacent terminals, distributed photovoltaic grid-connected devices, energy storage converters, the distribution master station, and field switch actuators. The information accessed by the intelligent distribution terminals includes at least three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position, distributed power output status, electrical quantities of adjacent terminals, switch status of adjacent terminals, and grid verification information. The information generated during the processing includes at least... The system records network connection relationships, connection reliability, power flow direction, protection setting parameters, directional protection criteria, fault direction determination information, terminal consistency determination information, fault section location results, and recovery path scores. The control information output by the intelligent distribution terminal includes at least the switching control commands for the switches at both ends of the fault section, isolation status confirmation information, switching control commands for the tie switch or backup power supply access switch, fault handling event records, and power restoration confirmation information. The overall processing follows the logical sequence of data collection and filing, network identification, protection criterion updating, fault direction determination, and section isolation and restoration. The intermediate results generated in the previous processing stage serve as the input for the next processing stage, enabling the fault handling process to adapt to distributed power source commissioning and decommissioning, load transfer, and bidirectional power flow changes.

[0064] Specifically, such as Figure 2 As shown: After the intelligent power distribution terminal enters the operating state, it performs unified time marking on the local acquisition channel and the communication receiving channel; the time reference prioritizes Beidou time synchronization or GPS time synchronization; when satellite time synchronization is unavailable, it switches to the station network time synchronization mode; during normal operation, the time synchronization error between multiple intelligent power distribution terminals is controlled within 1 millisecond; during degraded operation, the time synchronization error is controlled within 10 milliseconds; the above synchronization error range is based on the following: when the 10 kV power distribution network uses a 50 Hz power frequency, one power frequency cycle is 20 milliseconds, and 10 milliseconds is equivalent to half a power frequency cycle; if the time deviation between multiple terminals exceeds 10 milliseconds, it is easy to affect the comparison of the order of fault start time and the direction coordination judgment between adjacent terminals. Therefore, 10 milliseconds is taken as the maximum allowable synchronization error during degraded operation, and 1 millisecond is taken as the preferred synchronization target during normal operation;

[0065] After time stamping, the intelligent power distribution terminal collects three-phase voltage, three-phase current, zero-sequence voltage, and zero-sequence current through local voltage transformers, current transformers, and zero-sequence acquisition circuits. The sampling frequency is no less than 80 points per power frequency cycle; preferably, the sampling frequency is 256 points per power frequency cycle, corresponding to a 12.8 kHz sampling rate at a 50 Hz power frequency. 80 points per power frequency cycle corresponds to a 4 kHz sampling rate, which can cover the fundamental frequency and major low-order harmonics, and is suitable for monitoring routine operating conditions. 256 points per power frequency cycle corresponds to a 12.8 kHz sampling rate, which can further acquire transient components at the initial stage of fault initiation, and is suitable for subsequent fault initiation identification and fault direction determination.

[0066] After the sampled data enters the local data buffer, electrical quantity records are formed according to the channel number, acquisition time, effective value, phase angle, instantaneous peak value, quality flag, and source flag. The quality flag includes at least the following states: normal, out of limit, saturated, missing, suspected crosstalk, and pending verification, which are used to distinguish data that can be directly involved in subsequent processing, data that needs to be reviewed, and data that cannot be involved in real-time judgment.

[0067] The switch position is provided by feedback signals from the switch auxiliary contacts, remote signaling nodes, or actuators. The intelligent power distribution terminal sets an anti-jitter confirmation window of 20 to 60 milliseconds for switch position changes, preferably 40 milliseconds. If the same switch position state remains consistent within the anti-jitter confirmation window, it is recorded as a valid switch position. If the switch position state changes repeatedly within the anti-jitter confirmation window, it is recorded as an unstable position and will not be used as the basis for subsequent grid connection relationship updates. This window range matches the common mechanical jitter duration of the power distribution switch auxiliary contacts. When it is less than 20 milliseconds, transient jitter is easily mistaken for a valid change, and when it is more than 60 milliseconds, it will reduce the timeliness of the state update before fault isolation.

[0068] The operating status of distributed power sources is uploaded by the grid-connected controller, photovoltaic inverter, energy storage converter, or synchronous generator control unit; the collected content includes at least the following fields: grid connection status, rated capacity, rated current, current active power, current reactive power, operating mode, fault ride-through status, charging and discharging status, and grid connection point number; the intelligent distribution terminal generates a power source type identifier and a fault current contribution range identifier based on the distributed power source type, which can be used for subsequent protection parameter adjustment.

[0069] Among them, the fault current contribution range of photovoltaic inverter power supply is recorded as 1.2 to 2 times the rated current, the fault current contribution range of energy storage converter power supply is recorded as 1.5 to 2 times the rated current, and the fault current contribution range of synchronous generator power supply is recorded as 6 to 8 times the rated current. The above range is used to reflect the engineering characteristics of power electronic power supply affected by current limiting control and rotating motor power supply with large initial short-circuit current during fault, and serves as the engineering estimation boundary for subsequent protection parameter adjustment. The specific values ​​can be updated according to the equipment nameplate parameters, grid connection test report, low voltage ride-through control strategy, or equipment parameter table issued by the distribution master station.

[0070] The intelligent distribution terminal receives electrical quantities, switch status, and grid verification information from adjacent terminals via peer-to-peer communication. Protection-related real-time information is transmitted using GOOSE messages with a transmission delay target of no more than 10 milliseconds, used for collaborative fault direction determination between adjacent terminals. Status-related and verification-related non-real-time information is transmitted using MMS messages or equivalent distribution automation communication messages with a transmission delay target of no more than 100 milliseconds, used for grid status verification and operational status synchronization. When receiving data, the intelligent distribution terminal performs source verification, timestamp verification, sequence number continuity verification, and field integrity verification on the messages.

[0071] If the difference between the message timestamp and the local clock is no more than 10 milliseconds and the message sequence number is consecutive, it is written into the adjacent terminal information table; if the timestamp difference is greater than 10 milliseconds but no more than 100 milliseconds, it is marked as a delayed message and is only used for slow network connection relationship verification; if the timestamp difference exceeds 100 milliseconds or the message verification fails, it will not enter the real-time fault judgment and will be recorded as a communication abnormal event; 10 milliseconds matches the upper limit of time synchronization degradation and the GOOSE strong real-time transmission requirement, and 100 milliseconds matches the allowable update delay of status information and verification information;

[0072] The adjacent terminal information table includes at least the following fields: terminal number, adjacent relationship number, voltage amplitude, voltage phase angle, current amplitude, current direction, switch status, verification timestamp, and trust level. Among them, the current direction is defined as the feeder direction from the local terminal to the adjacent terminal, in order to avoid inconsistencies in the direction definition between different terminals.

[0073] The initial trust level is set to 0.8. When three consecutive message verifications are successful, the trust level increases by 0.02 each time, up to a maximum of 1.0. When message loss, out-of-order delivery, or verification failure occurs, the trust level decreases by 0.05 each time, down to a minimum of 0.3. The reason for setting the above trust level rules is as follows: the initial value of 0.8 indicates that the information of adjacent terminals is referable in the initial access phase but has not reached full trust; gradually increasing the trust level after three consecutive successful verifications avoids directly increasing the judgment weight of a single successful message; the abnormal decrease is greater than the normal increase, which is used to reduce the impact of abnormal communication on subsequent network connection relationship identification and fault direction collaborative judgment; the minimum value of 0.3 is used to retain the auxiliary reference value of abnormal terminal information, but to prevent abnormal terminal information from dominating subsequent judgments.

[0074] If an anomaly occurs in the local data acquisition channel, the intelligent power distribution terminal will perform quality marking and data processing on the abnormal channel. Abnormal situations include, for example, the voltage amplitude being lower than 5% of the rated voltage for 200 milliseconds for 2 consecutive milliseconds, and no corresponding drop in the same-phase voltage of adjacent terminals; the current measurement value reaching the upper limit of the full-scale range of the analog-to-digital converter, or exceeding the rated input range of the secondary side of the current transformer and occurring continuously for no less than 2 sampling points; the deviation between the three-phase current vector sum and the zero-sequence current measurement value exceeding 3% to 5% of the rated current and remaining for more than 3 sampling windows; or changes in voltage, current, and zero-sequence quantities that do not meet the basic electrical relationships of the power distribution line.

[0075] Among them, a sampling window can be selected as 1 to 2 power frequency cycles, corresponding to 20 to 40 milliseconds at 50 Hz power frequency; three consecutive sampling windows correspond to 60 to 120 milliseconds, used to eliminate single-point sampling glitches and transient electromagnetic interference; 200 milliseconds is used to distinguish between transient sampling fluctuations and continuous acquisition anomalies; 5% of the rated voltage is used to determine whether the voltage acquisition channel is at an abnormally low value; 3% to 5% of the rated current is used to cover the range of transformer error, sampling error and normal zero-sequence imbalance, while identifying inconsistent acquisition states;

[0076] For non-fault monitoring and slow grid verification, intelligent distribution terminals can use adjacent phase estimation, zero-sequence component verification, or comparison of similar electrical quantities between adjacent terminals to repair abnormal data. Adjacent phase estimation is only used when the three-phase load is in a basically balanced state. When the amplitude difference between two healthy phases does not exceed 10% of the rated value and the phase angle difference conforms to the three-phase symmetry relationship, the abnormal phase is estimated based on the three-phase symmetry relationship. If the three-phase imbalance is obvious, the adjacent phase estimation method is not used.

[0077] During zero-sequence component verification, the zero-sequence current calculated from the three-phase current vector sum is compared with the zero-sequence current measured by the zero-sequence acquisition circuit. The aforementioned deviation range of 3% to 5% is used to determine whether the three-phase acquisition data and the zero-sequence acquisition data are consistent. If the deviation range is exceeded, the corresponding three-phase current channel or zero-sequence current channel is marked as to be verified. When comparing similar electrical quantities of adjacent terminals, the changing trends of local in-phase voltage and in-phase current are compared with the changing trends of corresponding electrical quantities of adjacent terminals. If the changing directions are consistent and the timestamp difference is within the communication verification allowable range, the comparison result is used as a reference for abnormal channel repair or reliability level adjustment. For real-time protection judgment after fault start, repair data that has not been verified and confirmed does not directly participate in switch control.

[0078] After the above collection, verification, labeling and archiving processes, a local running dataset and adjacent terminal dataset with timestamps, quality marks, source marks and trust levels are formed, and this dataset is used as input for subsequent identification of the current network connection relationship.

[0079] Specifically, such as Figure 3 As shown: After obtaining the local running dataset and the adjacent terminal dataset, the intelligent power distribution terminal establishes a record of the current grid connection relationship. The record of the current grid connection relationship is jointly stored using a connection relationship matrix and an adjacency relationship table. The connection relationship matrix is ​​used to record the electrical connection status and connection strength between terminal pairs participating in grid identification, and the adjacency relationship table is used to record information such as the direct adjacent terminals, indirect adjacent terminals, electrically isolated terminals, and corresponding update times of the local terminal.

[0080] Connection status includes at least four categories: direct connection, indirect connection, electrical isolation, and pending confirmation. Direct connection indicates that there are no other sectionalizing switches, tie switches, branch nodes, or intermediate terminals between the two terminals. Indirect connection indicates that the two terminals are located in the same power supply channel, but there are other sectionalizing switches, tie switches, branch nodes, or intermediate terminals in between. Electrical isolation indicates that the two terminals are disconnected by a switch or are not in the same power supply path. Pending confirmation indicates that the existing information is insufficient to make a stable judgment. Through the above data structure, the intelligent distribution terminal can quickly query adjacent terminal pairs and their connection status, and update the relevant local connection relationships when switches are changed, distributed power sources are put into operation or deactivated, or loads are transferred, avoiding the need to rebuild the entire network connection relationship every time the operating mode changes.

[0081] Connection strength is generated based on connection status and identification confidence level; the base value for direct connection is 1, the base value for indirect connection is 0.5, the base value for electrical isolation is 0, and the base value for pending confirmation is 0.25; the base value for pending confirmation status is lower than that for indirect connection but higher than that for electrical isolation, in order to avoid equating pending confirmation status directly with disconnection status in difference assessment; connection strength can be obtained by multiplying the base value by identification confidence level, which is determined based on the consistency between passive identification results and active verification results, and is updated after subsequent consistency verification is completed;

[0082] The identification of grid connection relationships adopts a combination of passive identification and active verification. Passive identification is used for slow grid connection relationship identification under normal operating conditions. Its input includes three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, active power, reactive power in the local operating dataset, and similar electrical quantities in the adjacent terminal datasets. The intelligent distribution terminal calculates the voltage amplitude difference, voltage phase angle difference, current change trend correlation coefficient, active power change correlation coefficient, and zero-sequence quantity synchronization degree between adjacent terminals within a 5-minute statistical window, with a statistical step size of 1 minute. The 5-minute statistical window is used to filter out short-term load fluctuations, transient disturbances in distributed power output, and short-term communication jitter. The 1-minute statistical step size is used to maintain the minute-level connection relationship update capability. The passive identification results are used to form the basic judgment of grid connection relationships, but are not the sole source of judgment for fault instantaneous isolation actions.

[0083] The correlation coefficient of current change trend can be calculated from the effective value sequence of in-phase current of two terminals within the same statistical window; the correlation coefficient of active power change can be calculated from the active power change sequence of two terminals, wherein the active power change sequence is formed by the difference of active power at adjacent sampling times within the statistical window, or by the change in the average active power within consecutive sub-windows; the synchronization degree of zero-sequence quantity change can be calculated from the zero-sequence voltage change sequence or zero-sequence current change sequence of two terminals; the Pearson correlation coefficient is preferably used for the above correlation coefficients, with a value range of -1 to 1; the closer the value is to 1, the more consistent the change trend of the electrical quantities of the two terminals is.

[0084] For adjacent terminals A and B, if the voltage amplitude difference is less than 5% of the rated voltage, the voltage phase angle difference is less than 10 degrees, the correlation coefficient of current change trend is not less than 0.8, and the correlation coefficient of active power change is not less than 0.8, then A and B are recorded as direct connection candidates; if the voltage amplitude difference is less than 10% of the rated voltage, the voltage phase angle difference is between 10 and 30 degrees, and the correlation coefficient of current change trend is between 0.5 and 0.8, then A and B are recorded as indirect connection candidates; if the voltage phase angle difference exceeds 45 degrees, or the switch status indicates that there is a disconnection point between A and B, then A and B are recorded as electrical isolation candidates.

[0085] The above numerical ranges are based on the following: a 5% amplitude difference in rated voltage is used to characterize the common voltage deviation range of adjacent nodes on the same power supply path; a 10% amplitude difference in rated voltage is used to cover the voltage drop range caused by transmission through intermediate nodes or longer feeder sections; 10 degrees is used to determine the phase angle consistency of directly adjacent terminals on the same power supply path; 30 degrees is used to determine indirect connections when transmission is through intermediate nodes or when the load flow is large; 45 degrees is used as the exclusion boundary for weakened electrical connection or not being on the same power supply path; a correlation coefficient of 0.8 indicates a high correlation between the electrical quantities at both ends; a correlation coefficient of 0.5 to 0.8 indicates a moderate correlation between the electrical quantities at both ends; and a correlation coefficient below 0.5 should not be used alone to determine a stable connection relationship.

[0086] When there are changes in switch position, distributed power supply activation / deactivation, load transfer, maintenance status change, adjacent terminal sending connection relationship change messages, or passive identification results deviating for two consecutive statistical windows, the intelligent distribution terminal initiates active verification. Two consecutive statistical windows deviating means that the same adjacent terminal does not meet the original connection status criteria in two consecutive 1-minute step-size updates. This is used to exclude false triggers caused by single load fluctuations, instantaneous output changes of distributed power sources, or communication delays. Active verification includes two methods: weak characteristic current verification and operation status linkage verification.

[0087] When the intelligent power distribution terminal is equipped with a verification signal injection circuit, active verification is preferably performed using a weak characteristic current method. The intelligent power distribution terminal applies a weak characteristic current to the local measurement circuit through the verification signal injection circuit, or applies a weak characteristic current to the local bus through a coupling device. The characteristic current frequency can be selected as 75 Hz, 125 Hz, or 175 Hz, with the amplitude controlled below 0.5% of the local rated current, and the injection duration is 100 ms to 300 ms. 75 Hz, 125 Hz, and 175 Hz all avoid the 50 Hz fundamental wave and integer multiples of harmonics such as 100 Hz and 150 Hz, which can reduce spectral aliasing with the inherent harmonics of the distribution network. When the amplitude is less than 0.5% of the rated current, it is usually below the protection starting current level and the load disturbance sensitive range, and will not trigger the protection device or have a significant impact on the user load. 100 ms to 300 ms corresponds to 5 to 15 power frequency cycles, which can take into account both the stability of spectrum identification and the control of operating disturbances.

[0088] Adjacent terminals perform sliding spectrum analysis on the injected frequency band under the same time reference. The spectrum analysis window can be selected as 2 to 4 power frequency cycles, preferably 4 power frequency cycles, with a sliding step size of 1 power frequency cycle. 4 power frequency cycles correspond to 80 milliseconds at a 50 Hz power frequency. For a 75 Hz characteristic current, this covers approximately 6 characteristic cycles, and for a 125 Hz characteristic current, it covers approximately 10 characteristic cycles, which can improve the stability of non-integer multiple frequency component extraction. Adjacent terminals extract the characteristic frequency band current amplitude, phase, and receiving timestamp within the same time window at the same injected frequency from the spectrum analysis results, and establish a correspondence with the transmitting terminal number and the injection timestamp.

[0089] If the ratio of the received amplitude to the transmitted amplitude is not less than 0.7 and the phase difference does not exceed 15 degrees, the corresponding terminal pair is determined to be directly connected; if the ratio of the received amplitude to the transmitted amplitude is between 0.4 and 0.7, or the phase difference is between 15 and 45 degrees, the corresponding terminal pair is determined to be indirectly connected; if the ratio of the received amplitude to the transmitted amplitude is less than 0.4, or no corresponding frequency component is detected, the corresponding terminal pair is determined to be electrically isolated. The basis for setting the above numerical range is as follows: 0.7 indicates that the characteristic current transmission attenuation does not exceed 30%, which is suitable for connection judgment of short-distance direct feeder segments; 0.4 serves as the weak reception boundary when receiving at a remote end or transmitting through an intermediate node; 15 degrees is used as the merging margin for line transmission delay, transformer error, and sampling error in the direct connection state; above 45 degrees usually indicates that the electrical connection between the two ends is weak and should not be considered as a stable connection relationship.

[0090] When the conditions for injecting verification signals are not available on-site, active verification is performed using the linkage change method of operating status. The intelligent power distribution terminal uses switch status changes, active power changes of distributed power sources, energy storage charging and discharging switching, or natural load fluctuations as linkage events. It compares the voltage amplitude, current direction, active power changes, and zero-sequence quantity changes of the local terminal and adjacent terminals before and after the linkage event. If, after the linkage event occurs, the same type of electrical quantity of adjacent terminals changes in the same direction within 100 milliseconds to 500 milliseconds, and the correlation coefficient of the change trend is not less than 0.8, then the adjacent terminal pair is regarded as a connection candidate. If no corresponding change occurs, or the correlation coefficient of the change trend is less than 0.5, then the adjacent terminal pair is regarded as an electrical isolation candidate. 100 milliseconds to 500 milliseconds is used to cover the communication delay of power distribution automation, the sampling window update delay, and the on-site switch status feedback delay. If it is less than 100 milliseconds, it is easy to miss the lagging feedback, and if it is more than 500 milliseconds, it is easy to introduce irrelevant load fluctuations.

[0091] When the passive identification result matches the active verification result, the matching result is used as the current connection state, and the connection strength is updated according to the corresponding connection state. When the passive identification result does not match the active verification result, the active verification result is used as the temporary connection state, and the connection state is marked as pending confirmation, while the identification confidence of the connection relationship is reduced. If the active verification is unavailable, the passive identification result is used as the temporary connection state, and the system waits for confirmation from the next round of active verification or consistency check. This processing method can avoid incorrect updates of the network connection relationship caused by relying solely on slow statistical results or a single active verification result.

[0092] After completing passive identification and active verification, the intelligent power distribution terminal evaluates the difference between the newly identified connection relationship matrix and the historical connection relationship matrix according to the aforementioned connection status base values. The intelligent power distribution terminal calculates the connection status difference for each pair of terminals participating in the comparison, sums the absolute values ​​of the connection status differences for each pair of terminals, and then divides by the number of terminal pairs participating in the comparison to obtain the connection change index D.

[0093] If there are offline terminals or terminals with data quality flags indicating pending verification, the corresponding terminal pairs will not participate in this connection change index calculation and will remain in a pending confirmation state in the connection relationship record. If the number of terminal pairs participating in the comparison is less than 50% of the historical number of terminal pairs, the global connection change index judgment will not be performed, and only the terminal pairs that have obtained valid data will be partially updated or marked as pending confirmation. The 50% threshold is set based on the following: when the number of valid terminal pairs participating in the comparison is less than half, the global difference assessment is easily amplified by communication interruption or data loss, and it is not advisable to reconstruct the entire network connection relationship based on this.

[0094] When D < 0.05, it is determined to be measurement noise, minor load fluctuations, or short-term communication disturbances, and the network connection relationship record is not updated; when 0.05 ≤ D < 0.2, it is determined to be a local connection relationship change, and only the involved switches, tie points, or distributed power access points are updated; when 0.2 ≤ D < 0.5, it is determined to be a medium-range network adjustment, the local connection relationship record is updated, and a connection relationship change notification is sent to adjacent terminals; when D ≥ 0.5, it is determined to be a large-scale change in operating mode, the old connection relationship record is frozen as a rollback version, and the current network connection relationship record is re-established based on the results of this verification. The old record is replaced after the consistency verification is passed. The above numerical ranges are based on the following: 0.05 corresponds to small differences that may be caused by acquisition errors, load fluctuations, and message delays; 0.2 usually corresponds to changes in the status of a single segment switch, tie switch, or distributed power access point; 0.5 usually corresponds to multiple switch power transfers, maintenance mode switching, or large-scale operating mode adjustments.

[0095] To avoid incorrect updates to the network connection relationships due to misjudgment by a single terminal, the intelligent distribution terminal performs a consistency check before updating. The intelligent distribution terminal sends a summary of the connection relationship to be updated to at least two adjacent terminals. The summary includes at least the changed node, the connection status before the change, the connection status after the change, the change timestamp, the identification method, and the identification reliability. The adjacent terminals return a matching score based on their local connection relationship records and locally collected data. The matching score ranges from 0 to 1. When a local terminal can actually communicate with fewer than two adjacent terminals, the intelligent distribution terminal sends the summary of the connection relationship to be updated to the distribution master station or the next-level terminal for verification, and includes the matching score returned by the distribution master station or the next-level terminal in the average matching score calculation.

[0096] The intelligent power distribution terminal collects the matching scores returned by each checker and calculates the average matching degree. If the average matching degree is not lower than 0.8, the update of the grid connection relationship record is confirmed. If the average matching degree is between 0.5 and 0.8, the corresponding connection relationship is marked as pending confirmation, and a second active verification is performed after 30 to 60 seconds. If the average matching degree is lower than 0.5, the update is abandoned, and the identification result is recorded as an abnormal event. 0.8 indicates that most checkers agree, which is suitable as the confirmation update boundary. 0.5 indicates that there is a significant information discrepancy or communication abnormality among checkers, and it is not advisable to update directly when the value is lower than this. The 30 to 60 seconds are used to wait for the communication link to recover, the switch status to stabilize, and the distributed power supply operation status to stabilize before re-verifying, which can avoid erroneous updates caused by short-term disturbances.

[0097] After the above identification, verification, difference assessment and consistency confirmation, the current grid connection relationship record, connection relationship reliability, list of changed nodes and update timestamp are formed, and these are used as the basis input for subsequent adjustment of protection setting parameters and directional protection criteria.

[0098] Specifically, such as Figure 4As shown: After obtaining the current grid connection relationship record, the intelligent distribution terminal calls the power type identifier, fault current contribution range identifier, local load current record, adjacent terminal electrical quantity record, and distributed power supply operating status to establish a protection parameter update task corresponding to the local switch. The protection parameter update task is executed in parallel using periodic triggering and event triggering. The periodic triggering interval can be selected as 10 seconds. The event triggering conditions include at least the update of the current grid connection relationship record, the local load change exceeding 20%, the distributed power supply output change exceeding 30%, the energy storage device switching between charging and discharging states, the distribution master station issuing a setting verification command, or an adjacent terminal sending a protection parameter abnormality notification. The 10-second cycle is used for protection under normal operating conditions. The rolling refresh of parameters does not replace real-time protection judgment after fault initiation; this cycle can track the slow changes in photovoltaic output, energy storage operation mode and load level, while avoiding excessively frequent updates of protection parameters that would cause excessive consumption of computing resources at the edge terminal; the local load change amplitude can be calculated as the ratio of the change in the average load current of the most recent 1 minute to the average load current of the previous protection parameter update cycle; the distributed power output change amplitude can be calculated as the ratio of the change in the current active power to the change in the active power of the previous protection parameter update cycle; 20% is usually higher than the range of small short-term load disturbances for general users, and is suitable as a threshold for significant changes in load status; 30% can cover changes in operating conditions that have a significant impact on the contribution of fault current, such as photovoltaic cloud shadows and energy storage charging and discharging power scheduling;

[0099] The power flow direction is determined based on the local active power symbol, the relationship between the current phase angle and the voltage phase angle, and the current grid connection relationship. The feeder direction from the local terminal to the adjacent terminal is taken as the positive direction. When active power flows in this direction and the relationship between the current phase angle and the voltage phase angle conforms to the load-side power intake characteristics, it is recorded as a positive power flow. When active power flows in the opposite direction, or when distributed power sources feed power upstream, it is recorded as a reverse power flow. If the absolute value of the local active power is less than 3% to 5% of the rated capacity, and the power flow direction changes within two consecutive sampling windows, the power flow direction is marked as pending confirmation. The 3% to 5% range is used to cover power measurement errors, power factor fluctuations, and low load disturbances under weak power flow conditions, and two consecutive sampling windows are used to exclude single sampling jumps.

[0100] When adjusting the setting parameters, the intelligent distribution terminal determines the power supply sources on the forward and reverse sides of the local switch based on the current grid connection records. The power supply sources include at least the substation main power supply, photovoltaic inverter-type distributed power supply, energy storage converter-type distributed power supply, synchronous generator-type distributed power supply, and diesel backup power supply. The short-circuit current contribution value of the substation main power supply is determined based on the bus short-circuit capacity, the equivalent impedance of the upstream line, and the impedance of the feeder section where the local switch is located. The short-circuit current contribution value of the distributed power supply calls the established fault current contribution range identifier and is corrected in combination with the current grid connection status, rated current, output status, and operating mode. When the equipment nameplate parameters, grid connection test report, low voltage ride-through control strategy, or equipment parameter table issued by the distribution master station provide more specific fault current capabilities, the actual parameters are used to update the corresponding contribution range.

[0101] The intelligent power distribution terminal summarizes the short-circuit current contribution ranges on both the forward and reverse sides of the local switch. The maximum forward short-circuit current is the sum of the maximum short-circuit current contributions from each power source on the forward side, and the minimum forward short-circuit current is the sum of the minimum short-circuit current contributions from each power source on the forward side. The maximum reverse short-circuit current is the sum of the maximum short-circuit current contributions from each power source on the reverse side, and the minimum reverse short-circuit current is the sum of the minimum short-circuit current contributions from each power source on the reverse side. For distributed power sources that are in a state of withdrawal, blocking, or communication loss, the maximum short-circuit current contribution value is not included. For distributed power sources whose status is pending confirmation, only the minimum contribution value is used for fault sensitivity verification.

[0102] When performing setting calculations, the current maximum possible short-circuit current and the current minimum identifiable fault current are selected from the aforementioned forward or reverse short-circuit current ranges, and corrected in conjunction with the fault judgment direction and the lower limit of the observable current of adjacent terminals. If there are unconfirmed connection states in the current grid connection relationship, or if the connection relationship confidence is lower than 0.5, the intelligent distribution terminal does not generate new setting parameters based on the connection relationship alone. Instead, it uses the setting parameters corresponding to the most recent reliable grid connection relationship for transition and requests verification from adjacent terminals or the distribution master station. 0.5 is used as a confidence boundary, indicating that the connection relationship is in a state of information divergence or insufficient verification, and should not be directly used as the basis for updating protection parameters.

[0103] If the ratio of the maximum forward short-circuit current to the maximum reverse short-circuit current is less than 2, or if the maximum reverse short-circuit current exceeds 0.3 times the local rated current, the current switch is determined to be in a bidirectional power flow sensitive state. If the ratio is not less than 2, and the maximum reverse short-circuit current is less than 0.3 times the local rated current, the current switch is determined to be in a unidirectional power flow dominant state. The basis for setting the above numerical range is: if the ratio of the maximum forward short-circuit current to the maximum reverse short-circuit current is less than 2, it means that the reverse short-circuit current has reached more than 50% of the forward short-circuit current, and the reverse feed is sufficient to affect overcurrent start-up and direction judgment. When the maximum reverse short-circuit current exceeds 0.3 times the local rated current, it is significantly higher than the small reverse component caused by the current transformer measurement error and normal load fluctuation, and belongs to the reverse feed current that may affect the protection criteria.

[0104] For bidirectional power flow sensitive conditions, the intelligent distribution terminal does not directly use the pre-fixed setting value, but generates a dynamic setting value based on the current short-circuit current contribution range, the maximum load current, and the minimum identifiable fault current; the overcurrent protection setting parameters are checked at least by load avoidance verification and fault sensitivity verification; the load avoidance verification is used to make the setting value higher than the maximum load current, and the fault sensitivity verification is used to make the setting value lower than the minimum identifiable fault current, thereby avoiding false tripping caused by normal load fluctuations, and also avoiding failure to trip due to excessively high setting values;

[0105] The overcurrent stage I setting value can be selected within the range of 0.5 to 0.7 times the current maximum possible short-circuit current, and not less than 1.2 times the maximum load current; the overcurrent stage II setting value can be selected within the range of 0.3 to 0.5 times the current minimum identifiable fault current, and not less than 1.1 times the maximum load current; the overcurrent stage III setting value can be selected within the range of 0.1 to 0.3 times the current minimum identifiable fault current, and not less than 1.05 times the maximum load current; when there are backup protection coordination parameters for adjacent terminals, the overcurrent stage III setting value is also checked against the upper limit of the corresponding backup stage setting value of the adjacent upstream terminal.

[0106] The above setting range is based on the following: 0.5 to 0.7 times is used to maintain high operating sensitivity of the first-stage protection and avoid approaching the upper limit of the short-circuit current; 0.3 to 0.5 times is used to balance the sensitivity and selectivity of the second-stage protection; 0.1 to 0.3 times is used to meet the identification requirements of the third-stage backup protection for high-resistance or remote faults; 1.2 times, 1.1 times and 1.05 times are used to avoid the maximum load current and its short-term fluctuations, respectively; the maximum load current is taken as the maximum value of the effective value of the local load current in the most recent 5 minutes, and the 5-minute window can cover the general load fluctuation cycle, while not introducing the historical load state too early;

[0107] If, after load avoidance verification, any setting value exceeds 0.8 times the corresponding minimum identifiable fault current, the protection segment is marked as having a low margin. The basis for setting 0.8 times as the margin boundary is that when the setting value is close to the minimum identifiable fault current, the protection action sensitivity is significantly reduced. Setting it to 0.8 times allows for early identification of the risk of failure to operate before the setting value exceeds the minimum identifiable fault current. The low margin flag is used to reduce the impact of a single overcurrent criterion in subsequent fault determination and to increase the participation of directional protection criteria and adjacent terminal collaborative criteria.

[0108] For a unidirectional power flow dominant state, the intelligent distribution terminal can retain the original protection setting parameters and perform a verification based on the current maximum load current, minimum identifiable fault current, and current grid connection relationship. If the verification result meets the load avoidance verification and fault sensitivity verification, the original setting parameters will continue to be used. If the verification result does not meet any verification condition, the setting parameters will be regenerated according to the dynamic setting method for a bidirectional power flow sensitive state.

[0109] The directional protection criterion is formed using a multi-electrical quantity polarity matching method; the intelligent distribution terminal calculates the phase relationship between the zero-sequence current and the zero-sequence voltage, the phase relationship between the negative-sequence current and the negative-sequence voltage, and the direction of change between the three-phase voltage surge and the three-phase current surge in each protection parameter update cycle; the three characteristic quantities form polarity marks respectively, with positive marked as +1, negative marked as -1, and indeterminate marked as 0; the definitions of positive and negative directions follow the aforementioned power flow direction definition;

[0110] When the phase relationship falls into the positive discrimination interval, the corresponding polarity flag is recorded as +1; when the phase relationship falls into the negative discrimination interval, the corresponding polarity flag is recorded as -1; when the phase relationship is in the boundary interval, or when the amplitude of the zero-sequence, negative-sequence, or abrupt change characteristic is lower than 1% to 3% of the corresponding rated value, the corresponding polarity flag is recorded as 0; the basis for setting 1% to 3% is that this range is higher than the common sampling quantization error and the small measurement deviation of the transformer, which can avoid the weak characteristic quantity from misleading the direction judgment;

[0111] When all three polarity flags are +1, the direction protection criterion is recorded as positive; when all three polarity flags are -1, the direction protection criterion is recorded as negative; when two of the three polarity flags are consistent and the other is opposite or 0, the direction protection criterion is formed according to the two consistent flags, and the direction matching degree is set to 0.67; when the three polarity flags are inconsistent, or two or more polarity flags are 0, the direction protection criterion is marked as pending review, and the direction matching degree is set to 0.33; the setting of 0.67 and 0.33 is based on the three-feature voting logic. 0.67 corresponds to the support degree of two consistent features among the three features, and 0.33 corresponds to the low support state of only one feature being valid or having insufficient consistency, which is used to avoid a single feature abnormality directly determining the direction judgment;

[0112] To reduce the impact of harmonics and noise on the updating of directional protection criteria, the intelligent distribution terminal performs frequency band decomposition on the three-phase current sequence within the most recent second to determine whether the current sampling environment is suitable for updating the directional protection criteria. Frequency band decomposition can be achieved using bandpass filter banks, short-time Fourier transform, or variational mode decomposition. The 1-second window corresponds to 50 power frequency cycles, reflecting the harmonic and noise levels under the current operating environment. Frequency band decomposition generates five types of components: fundamental component, low-order harmonic components, inverter harmonic components, switching transient components, and high-frequency noise components. Among these, the components from 500 Hz to 2 kHz are used as transient-sensitive components to reflect transient changes during switching operations and the initial stage of fault initiation; components above 2 kHz are used as high-frequency noise components to identify noise in the acquisition system and electromagnetic interference in the field; the components from 500 Hz to 2 kHz... The kilohertz frequency band is higher than the conventional low-order harmonic frequency band and lower than the high-frequency noise-dominated frequency band of most acquisition systems, making it suitable for extracting transient change information between low-order harmonics and high-frequency noise. If the energy of the high-frequency noise component exceeds 30% of the total energy, the direction protection criterion for this cycle will not be updated, and the reliable direction protection criterion for the previous cycle will be retained. The 30% setting is based on the fact that when high-frequency noise accounts for a significant proportion of the total energy, the zero-sequence phase, negative-sequence phase, and abrupt change direction may all be affected by interference, and continuing to update the direction criterion may easily introduce misjudgment. If the state of high-frequency noise energy exceeding 30% of the total energy continues for 3 consecutive protection parameter update cycles, protection parameter confirmation information will be sent to adjacent terminals. When the cycle trigger interval is 10 seconds, the 3 protection parameter update cycles correspond to approximately 30 seconds, which can distinguish between short-term electromagnetic interference and continuous sampling environment anomalies.

[0113] After the protection parameters are updated, the intelligent distribution terminal establishes a setting parameter version management mechanism. After each update of the overcurrent protection setting parameters and directional protection criteria, a protection parameter version number is generated. The version number includes at least the terminal number, update time, grid connection relationship version, distributed power source status summary, setting value summary, and directional criterion summary. If the current grid connection relationship record is updated continuously within a short period of time, the intelligent distribution terminal only uses the latest grid connection relationship version for setting and marks the old version as invalid to avoid mixing different versions of grid connection relationships and protection parameters.

[0114] If communication interruption prevents adjacent terminals from confirming the latest protection parameters, the intelligent distribution terminal enters a local protection degradation state. In the local protection degradation state, the most recent reliable setting parameters and reliable directional protection criteria are used, and the time window for collaborative judgment between adjacent terminals is expanded during subsequent fault initiation detection. The most recent reliable setting parameters refer to the setting parameters that have passed the grid connection relationship version verification, distributed power source status verification, and adjacent terminal consistency confirmation. This maintains basic protection capabilities during communication anomalies and prevents unconfirmed parameters from participating in switch control.

[0115] After the above processing, the current protection setting parameters, directional protection criteria, directional matching degree, protection parameter version number and low margin flag are generated, which serve as inputs for subsequent fault start detection and local fault direction determination.

[0116] Specifically, such as Figure 5 As shown: The intelligent distribution terminal continuously monitors fault initiation events based on the current protection setting parameters, directional protection criteria, protection parameter version number, and low margin flag; the fault initiation triggering conditions include at least overcurrent initiation conditions and transient energy initiation conditions; the overcurrent initiation condition is that the current of any phase in the local area exceeds the current overcurrent stage I setting value within 3 consecutive sampling judgment windows, and the transient energy initiation condition is that the transient component energy exceeds 5 times the dynamic baseline;

[0117] The sampling decision window can be selected from 1 / 4 of a power frequency cycle to 1 power frequency cycle, corresponding to 5 milliseconds to 20 milliseconds at a 50 Hz power frequency; three consecutive sampling decision windows correspond to 15 milliseconds to 60 milliseconds, used to exclude single-point sampling spikes and maintain the speed of fault start-up identification; the transient component energy is obtained by the square integral of the current component in the 500 Hz to 2 kHz frequency band, which corresponds to the transient sensitive component and can reflect the transient changes in the switching action and the initial stage of fault start-up;

[0118] The dynamic baseline is generated using the median of transient component energy in the non-faulty operating state within the last 10 seconds. The non-faulty operating state refers to the operating state that does not meet the overcurrent start-up condition, does not meet the transient energy start-up condition, and is not within the switch opening and closing action confirmation window. The 10-second window matches the rolling refresh cycle of protection parameters, which can reflect the current load level and the output status of distributed power sources, while avoiding the introduction of premature historical operating states. Under normal operating conditions, the peak-to-average ratio of transient component energy is usually less than 3 times. Setting the start-up threshold to 5 times the dynamic baseline can reduce false triggering caused by load switching, inverter harmonic fluctuations, and normal switch operation. When the dynamic baseline is lower than the lower limit of the background noise energy of the acquisition system, the lower limit of the background noise energy is used as the dynamic baseline for comparison to avoid false triggering caused by an excessively low baseline under low load or no-load conditions.

[0119] When any fault initiation trigger condition is met, the intelligent power distribution terminal records the starting point of the sampling judgment window where the trigger condition is first met as the fault initiation time T0. If the overcurrent initiation condition and the transient energy initiation condition are met simultaneously within the same judgment period, the starting time of the transient energy surge is preferred as T0. If the difference between the two starting times does not exceed one power frequency cycle, the earlier time is taken as T0 to improve the integrity of the waveform data interception before and after the fault.

[0120] The intelligent power distribution terminal takes the fault initiation time T0 as the center and extracts the three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position, and distributed power source operation status data from the local cache for the two power frequency cycles before T0 and the two power frequency cycles after T0. The next power frequency cycle of the 50 Hz system is 20 milliseconds, and the four power frequency cycles before and after T0 correspond to 80 milliseconds. This time length can cover the steady state before the fault, the transient state after the fault initiation, and the initial response after the fault, and is suitable for local direction determination and adjacent terminal collaborative determination.

[0121] The intelligent power distribution terminal performs quality verification on the data within the fault time window; the key channels involved in the verification include at least the three-phase voltage channel, three-phase current channel, zero-sequence voltage channel, zero-sequence current channel involved in this direction determination, and the corresponding channel used for negative sequence component calculation; the missing rate is calculated as the ratio of the number of missing sampling points within the fault time window to the number of theoretical sampling points; continuous saturation refers to the sampling value continuously reaching the upper or lower limit of the full scale of the analog-to-digital converter, or exceeding the rated input range of the secondary side of the current transformer, and lasting for more than one sampling determination window;

[0122] If the critical channel missing rate does not exceed 5% and the three-phase current does not show continuous saturation, the main waveform within the fault time window remains continuous, and the system proceeds to local direction determination. If the critical channel missing rate exceeds 5% but does not exceed 20%, the waveform within the fault time window is still repairable, and the intelligent distribution terminal uses adjacent phase interpolation, zero-sequence quantity back-calculation, or similar electrical quantity data from adjacent terminals within the same time window to complete the missing data. The completed data must pass zero-sequence consistency verification or adjacent terminal similar electrical quantity change trend verification before it can participate in direction determination. If the critical channel missing rate exceeds 20%, or the critical channel shows continuous saturation, the system proceeds to local direction determination. If saturation occurs, the local direction determination information will be marked as low confidence and will not be used independently as the basis for subsequent fault segment determination. The 5% and 20% settings are based on the following: when the missing rate does not exceed 5%, the continuity and phase relationship of the fault waveform can be basically maintained; when the missing rate exceeds 5% but does not exceed 20%, it can still be repaired by data from adjacent phases or adjacent terminals; when the missing rate exceeds 20%, the continuity of key waveforms within the fault time window is insufficient, which can easily affect the judgment of zero-sequence phase, negative-sequence phase and abrupt change direction; the supplementary data that has not been verified and confirmed is only used for low confidence identification and is not directly used as the basis for switching action.

[0123] During zero-sequence consistency verification, the zero-sequence current calculated from the three-phase current vector sum is compared with the zero-sequence current measured by the zero-sequence acquisition circuit. If the deviation between the two does not exceed 3% to 5% of the rated current, the zero-sequence consistency verification is considered to have passed. During the verification of the changing trend of similar electrical quantities in adjacent terminals, the changing direction of the same-phase voltage, same-phase current, or zero-sequence quantity of the local and adjacent terminals within the same time window is compared. If the changing direction is consistent and the timestamp difference is within the allowable range of communication verification, the verification is considered to have passed. The 3% to 5% range is used to cover the range of transformer error, sampling error, and normal zero-sequence imbalance, and at the same time, to identify the inconsistency between the three-phase acquisition data and the zero-sequence acquisition data.

[0124] The local direction determination calls upon the established direction protection criteria, and uses the zero-sequence direction characteristics, negative-sequence direction characteristics, and transient change direction characteristics within the fault time window as inputs; the zero-sequence direction characteristics are formed by the phase relationship between zero-sequence voltage and zero-sequence current, the negative-sequence direction characteristics are formed by the phase relationship between negative-sequence voltage and negative-sequence current, and the transient change direction characteristics are formed by the change direction between the three-phase voltage change and the three-phase current change within the first power frequency cycle after T0;

[0125] The three directional features are voted on according to the direction protection criterion. When all three directional features point to the same direction, the local direction determination information is recorded as positive or negative, and the local direction matching degree is 1.0. When two directional features are consistent, the majority direction is output, and the local direction matching degree is 0.67. When only one directional feature is valid, the local direction determination information is recorded as pending confirmation, and the local direction matching degree is 0.33. When all three directional features are invalid, the local direction determination information is recorded as uncertain, and the local direction matching degree is 0. 1.0 indicates that the three directional features are completely consistent, 0.67 indicates that two directional features form a majority support, 0.33 indicates that only low support directional information is available, and 0 indicates that the local data is insufficient to form a direction determination.

[0126] When the protection parameter update result has been marked as a bidirectional power flow sensitive state or a low margin state, the local direction determination does not determine the fault direction solely based on the overcurrent amplitude direction. If the negative sequence direction characteristic or transient change direction characteristic is inconsistent with the overcurrent amplitude direction, the local direction determination information is marked as pending confirmation, or the local direction matching degree is reduced by one level to reduce the misleading current amplitude direction caused by reverse feed-in from distributed power sources such as photovoltaics and energy storage.

[0127] The initial fault type is determined based on the overcurrent phase, the zero-sequence component exceeding the limit, and the negative-sequence component change state, at least distinguishing between phase-to-phase faults, single-phase ground faults, and fault types to be confirmed. When two or three phase currents simultaneously exceed the corresponding setting values ​​and the negative-sequence component changes significantly, it is initially judged as a phase-to-phase fault. When the zero-sequence voltage or zero-sequence current exceeds the corresponding starting threshold and the phase current change is not completely symmetrical, it is initially judged as a single-phase ground fault. When the data quality is insufficient, the directional characteristics are inconsistent, or the fault type characteristics do not meet the above conditions, it is initially judged as a fault type to be confirmed.

[0128] After generating local fault direction determination information, the intelligent power distribution terminal sends it to adjacent terminals via strong real-time communication messages. The fault direction determination information includes at least the following fields: terminal number, switch number, fault start time T0, fault start trigger type, initial fault type judgment, direction determination result, direction matching degree, current amplitude, zero-sequence component characteristics, negative-sequence component characteristics, transient change characteristics, protection parameter version number, grid connection relationship version number, data quality flag, and local reliability level. When sending low-reliability fault start information, the following fields are added in addition to the above fields: low-reliability reason, available electrical quantity channels, and suggested review flag.

[0129] The local fault direction determination information can be sent within a window of no more than 20 milliseconds; 20 milliseconds is approximately the next power frequency cycle of 50 Hz, which can complete the transmission of local determination information within one power frequency cycle, leaving time margin for subsequent multi-terminal collaborative determination, fault section determination and switching action; if the local direction determination information is low confidence, the smart distribution terminal will still send fault start information and attach a low confidence flag and low confidence reason to the message for verification by adjacent terminals;

[0130] After receiving the fault direction determination information, adjacent terminals first perform version consistency verification. If the protection parameter version number is consistent with the local version and the network connection relationship version number is consistent, the message is recorded as a version consistent message. If the versions are different but the update time difference is less than 60 seconds, the message is recorded as a slightly version-deviation message, and the message's credibility level is reduced by 0.1 to 0.2 before participating in collaborative determination. If the version difference exceeds 60 seconds, or the network connection relationship version number is inconsistent, the message is set to pending confirmation, and the sending terminal is requested to resend the current network connection relationship summary or protection parameter summary. The 60-second interval corresponds to the upper limit of the secondary active verification time after the aforementioned pending confirmation of the network connection relationship. If this time is exceeded, different terminals may make judgments based on different network connection relationship versions or different protection parameter versions, and it is not advisable to directly use the message as high-credibility direction determination information. The message credibility level is reduced by 0.1 to 0.2 to retain the auxiliary reference value of slightly version-deviation messages while reducing their impact on the final collaborative determination.

[0131] During the exchange of fault direction determination information between adjacent terminals, the intelligent power distribution terminal establishes a fault collaborative buffer. The fault collaborative buffer is used to temporarily store local fault direction determination information, adjacent terminal fault direction determination information, message trust level, version consistency flag and data quality flag. The buffer retention time can be selected from 30 milliseconds to 50 milliseconds.

[0132] If the fault direction determination messages of directly adjacent terminals are received within 30 milliseconds, the intelligent power distribution terminal will output the information in the fault collaborative buffer to the subsequent segment judgment process; if not all the messages of directly adjacent terminals are received within 30 milliseconds, but the minimum direction combination has been obtained, the subsequent segment judgment process can also be entered, and the late message will be used as verification data; the minimum direction combination means that the direction determination information of at least the local terminal and one directly adjacent terminal is obtained, and the directions of both point to the same adjacent segment; or the determination information that the directions of the terminals at both ends of the suspected fault segment are opposite.

[0133] If the minimum directional combination cannot be formed after 50 milliseconds, the intelligent power distribution terminal will activate the degradation logic and use the local direction determination information and the direction determination information received from the adjacent terminals to form a temporary segment judgment condition. The temporary segment judgment condition is only used as the initial input for the subsequent segment judgment process. It needs to be confirmed by secondary verification through fault recording or supplementary messages from more adjacent terminals in the subsequent process before the switch tripping control command can be generated. 30 milliseconds corresponds to three times the margin under the typical peer-to-peer communication delay of 10 milliseconds, and 50 milliseconds is used to take into account the short-term jitter and multi-hop forwarding delay of the backup communication link.

[0134] After the above-mentioned fault initiation monitoring, fault time window extraction, data quality review, local direction determination, initial fault type judgment, version consistency verification, and fault direction information exchange, a multi-terminal fault direction determination information set, message trust level, version consistency flag, and temporary segment judgment conditions are formed, which are then used as inputs for subsequent fault segment determination, isolation, and power restoration control.

[0135] Specifically, such as Figure 6 As shown: The intelligent power distribution terminal receives a set of fault direction determination information from multiple terminals and calls upon the current network connection relationship record, connection relationship reliability, message reliability level, version consistency flag, and temporary segment judgment conditions to determine the suspected fault segment. The fault segment judgment is based on adjacent terminal pairs as the basic unit, and is traversed sequentially from the power supply side to the load side according to the power supply path direction in the current network connection relationship record; among them, upstream and downstream are determined according to the current power supply path direction.

[0136] For any adjacent terminal pair A and B, if the fault direction determination result of A points to the downstream direction of B and the fault direction determination result of B points to the upstream direction of A, then the line segment between A and B is marked as a candidate fault segment; if A is determined to be downstream but B has not started, and A and B are directly connected, and the voltage on the B side drops or loses voltage, then the line segment between A and B is also marked as a candidate fault segment.

[0137] Voltage sag can be defined as voltage falling below 0.7 times the rated voltage for more than one power frequency cycle; voltage loss can be defined as voltage falling below 0.1 times the rated voltage for more than one power frequency cycle. 0.7 times the rated voltage is significantly lower than the normal voltage fluctuation range and can be used to identify severe undervoltage; 0.1 times the rated voltage is close to a state of no effective voltage and can be used to identify voltage loss; a voltage loss lasting more than one power frequency cycle is used to exclude instantaneous sampling spikes or single-point telemetry anomalies.

[0138] If multiple adjacent terminal pairs simultaneously meet the candidate conditions for a fault section, the intelligent distribution terminal calculates a candidate score C based on directional matching degree, connection relationship reliability, consistency of fault initiation time, and data quality score; the candidate score C can be calculated as follows:

[0139]

[0140] The mean directional matching degree is obtained by averaging the local directional matching degrees of the terminals at both ends of the candidate segment; the reliability of the connection relationship is based on the connection strength or consistency verification result of the corresponding terminal in the current network connection relationship record, with a value range of 0 to 1; the temporal consistency score is determined based on the difference between the fault start times of the two ends; and the data quality score is determined based on the missing rate, continuous saturation state, version consistency flag, and low reliability flag.

[0141] If the fault direction determination information at either end of the candidate segment carries a low confidence flag, the direction matching degree at that end is calculated using a value reduced by 0.1 to 0.2; if the direction determination result at that end is uncertain, the direction matching degree at that end is 0; when determining the data quality score, the lower data quality score among the messages at both ends of the candidate segment is preferred as the data quality score of the candidate segment.

[0142] The weights are set as follows: Directional matching degree directly reflects the degree of support of the fault current direction for the section boundary, so it is set to 0.4; Connection relationship reliability is used to avoid positioning based on incorrect network connection relationships, so it is set to 0.3; Timing consistency is used to exclude different fault events or delayed triggering, so it is set to 0.2; Data quality score is used to correct the impact of missing, saturated, and version deviations on the positioning results, so it is set to 0.1.

[0143] The time consistency score is calculated as follows: when the difference between the fault start times of the two ends is no more than 5 milliseconds, it is 1; when the difference is greater than 5 milliseconds but no more than 20 milliseconds, it decreases linearly between 1 and 0.2; when the difference exceeds 20 milliseconds, it is 0.2; 5 milliseconds is close to the high consistency range of the fault start times of adjacent terminals under unified time synchronization and strong real-time communication conditions, and 20 milliseconds corresponds to the next power frequency cycle of 50 Hz power frequency; when it exceeds 20 milliseconds, the start information of the two ends may be affected by communication delay, weak fault characteristics or different disturbances, so only the low score is retained.

[0144] Data quality scores can be determined as follows: 1 is given when the data is complete and the protection parameter version and network connection relationship version are consistent; 0.6 to 0.8 is given when there are low reliability indicators or slight version deviations; 0.3 to 0.5 is given when there are missing key channels, continuous saturation, version pending confirmation, or message pending confirmation status. This scoring method ensures that candidate segments with clear direction, reliable network structure, consistent timing, and high data quality receive higher scores.

[0145] When the segment with the highest candidate score is unique and its score is not lower than 0.75, and it also meets the conditions of valid switch positions at both ends, consistent protection parameter versions or passing the slight deviation weighting verification, and normal self-test of the operating mechanism, the intelligent distribution terminal identifies the segment as the fault segment. If the highest score is between 0.6 and 0.75, a secondary verification is performed. If the highest score is lower than 0.6, or there are multiple candidate segments with similar scores that cannot be distinguished, automatic tripping is suspended, and the adjacent terminal is requested to send fault direction determination information or the distribution master station is requested to participate in the confirmation. Among these, multiple similar scores mean that the difference between the highest score and the second highest score is less than 0.05. A score above 0.75 indicates that the core factors such as direction matching degree, connection relationship reliability, and time consistency are generally in a highly reliable state, suitable as an automatic confirmation boundary. A score between 0.6 and 0.75 indicates that at least one core factor is insufficient, requiring secondary verification. A score below 0.6 indicates a high risk of automatic isolation, and it is not advisable to directly execute tripping control. A score of 0.05 is used to distinguish small differences in the scoring model to avoid mistakenly selecting a fault segment when the candidate segment scores are similar.

[0146] The secondary verification calls the waveform data recorded by the terminals at both ends of the suspected fault section from the four power frequency cycles before T0 to the four power frequency cycles after T0 at the fault initiation time; at a power frequency of 50 Hz, eight power frequency cycles correspond to 160 milliseconds, which can cover the state before the fault, the fault initiation process, and the initial current change after the fault; the intelligent power distribution terminal calculates the current energy integral of the current flowing into the candidate section at both ends respectively. The current energy integral can be the square integral of the fault phase current or the square integral of the effective value of the three-phase current;

[0147] If the current energy directions at both ends are opposite and the energy difference does not exceed 10% of the sum of the absolute values ​​of the energy at both ends, the secondary verification passes. Opposite current energy directions mean that the current energy calculated at both ends according to the direction pointing into the candidate section is in the inflow direction, or the signs are opposite when calculated according to a unified reference direction. If the current energy directions at both ends are the same, or the energy difference exceeds 10%, the current location result is abandoned, and a new request is made to the adjacent terminal to resend fault direction determination information or to request confirmation from the distribution master station. The 10% error range is used to cover current transformer measurement errors, sampling clock deviations, line losses, and short-term transient attenuation. If this range is too small, it may lead to a false fault being misjudged as a verification failure; if it is too large, it will reduce the fault section location constraint capability.

[0148] After the faulty section is confirmed, the intelligent power distribution terminal controls the switches at both ends of the faulty section to perform trip isolation. Before tripping, the terminals at both ends perform a two-way handshake confirmation. The downstream end sends a trip preparation message to the upstream end. The trip preparation message includes at least the fields of switch position, operating mechanism energy storage status, operating mechanism self-test result, trip coil status, and suggested action time. The upstream end returns a confirmation message after verifying that the local switch meets the action conditions. Meeting the action conditions means that the current switch position does not conflict with the proposed action, the operating mechanism energy storage status is normal, the trip coil circuit is not locked, the communication version is consistent, or it has passed the weighted verification. The terminals at both ends send trip pulses to their respective trip coils under a unified time reference.

[0149] The tripping pulse width can be selected from 40 milliseconds to 80 milliseconds; this range is suitable for the operation requirements of common power distribution switch tripping coils; less than 40 milliseconds may lead to insufficient coil engagement, while more than 80 milliseconds will increase the risk of coil overheating; after the tripping command is issued, the intelligent power distribution terminal monitors the switch position remote signal within 100 milliseconds. 100 milliseconds is used to cover the operation time of common power distribution switch tripping mechanisms and remote signal return delay, while not significantly delaying fault clearing.

[0150] If both ends report the tripping position, the isolation is recorded as successful. If neither end reports the tripping position, the tripping pulse is sent repeatedly, with a maximum of 3 retries and an interval of 100 milliseconds between each retrieval. The 3 retries are used to balance the intermittent failure of the switch to operate and the fault clearing time limit requirements. Retrying more than 3 times will significantly prolong the fault duration. If the 3 retries still fail, an expanded isolation request or a cascading trip request is sent to the corresponding terminal of the next higher-level switch, and confirmation is awaited from the next higher-level terminal. At the same time, the fault section status is marked as expanded isolation.

[0151] After the faulty section is isolated, the intelligent power distribution terminal identifies the power-loss section and selects a recovery path. The power-loss section is determined by the downstream node whose voltage is 0.7 times lower than the rated voltage after the circuit breaker is tripped and lasts for more than 200 milliseconds. The 200 milliseconds are used to exclude voltage dips at the moment of circuit breaker tripping, telemetry refresh delays, and short-term recovery processes. The load estimate for the power-loss section is generated based on the average active power and load importance level in the minute before the fault. The 1-minute window can reflect the short-term load level before the fault and reduce the impact of instantaneous load surges on the judgment of recovery capacity.

[0152] Candidate power supply paths are derived from the current grid connection records. The intelligent power distribution terminal calculates a power supply capacity score R for each candidate path. The power supply capacity score R can be calculated as follows:

[0153]

[0154] Among them, the power supply capacity margin is the ratio of the difference between the available capacity of the candidate power supply and the restored load to the restored load; the line thermal stability margin is the ratio of the difference between the allowable current carrying capacity and the expected restoration current in the candidate path to the allowable current carrying capacity; the terminal voltage drop is the ratio of the decrease in the voltage of the terminal node after restoration to the rated voltage; the loop safety score is determined based on the voltage phase angle difference on both sides of the tie switch; the islanding risk score is determined based on the islanding risk assessment results described below, and if there is unresolved islanding risk, the islanding risk score is 0;

[0155] Before a candidate path is entered into the ranking process, it must meet the following safety boundary conditions: power capacity margin of no less than 20%, line thermal stability margin of no less than 15%, voltage drop at the end of the path not exceeding 8%, voltage phase angle difference between the two sides of the tie switch not exceeding 30 degrees, and no unresolved distributed power islanding risk. Candidate paths that do not meet any of the safety boundary conditions will not be ranked. The 20% power capacity margin is used to prevent the power supply from approaching full load after restoration, the 15% line thermal stability margin is used to avoid line overload after transfer, the 8% voltage drop at the end of the path is used to ensure the quality of the restored voltage at the user end, and the 30-degree phase angle difference is used to reduce the impact of the tie switch closing. The 8% voltage drop boundary is used in the candidate path ranking stage, which is stricter than the minimum recovery criterion of 0.9 times the rated voltage after closing, and is used to reserve calculation error and load fluctuation margin in the path selection stage.

[0156] For candidate paths that meet the safety boundary conditions, the sub-scores range from 0 to 1; the sub-score corresponding to the safety boundary is no less than 0.6, and the larger the margin, the higher the score; the power capacity margin score and the line thermal stability margin score can be normalized and truncated to 1 according to the ratio of the actual margin to the corresponding safety boundary; the voltage quality score is generated in the manner that the smaller the voltage drop at the end, the higher the score; the loop closure safety score is generated in the manner that the smaller the phase angle difference on both sides of the tie switch, the higher the score; the islanding risk score is generated according to whether the islanding risk has been resolved; the basis for setting the above weights is as follows: the power capacity margin directly determines whether there is power supply capability after restoration, so it is set to 0.3; the line thermal stability margin is related to the line overload risk after restoration, so it is set to 0.25; the voltage quality affects the restoration effect on the user side, so it is set to 0.2; the loop closure safety is used to reduce the closing impact, so it is set to 0.15; the islanding risk is a safety constraint before restoration, so it is set to 0.1;

[0157] The intelligent power distribution terminal selects the candidate path with the highest power supply capacity score and a score of not less than 0.6 to perform the closing and restoration. 0.6 means that the candidate path meets the basic restoration conditions in terms of capacity, thermal stability, voltage quality, loop safety, and islanding risk. If there is no single path with a score of not less than 0.6, then the power outage section is divided into multiple sub-sections according to the load importance level and available power supply paths. First-level loads and second-level loads are restored first, while third-level loads are temporarily deferred or restored after load reduction.

[0158] Before performing the closing restoration, the intelligent power distribution terminal verifies the closing allowable conditions of the switch corresponding to the candidate path. The closing allowable conditions include at least successful isolation of the fault section, the voltage on both sides of the tie switch meeting the synchronization check, the candidate path still meeting the aforementioned safety boundary conditions, the risk of distributed power source islanding being eliminated, and the switch operating mechanism self-testing being normal. The closing pulse width can be selected from 80 milliseconds to 120 milliseconds. The closing coil usually needs a longer energy holding time than the opening coil, so the closing pulse width is higher than the opening pulse width. A pulse width lower than 80 milliseconds may lead to insufficient action of the closing mechanism, while a pulse width higher than 120 milliseconds will increase the risk of coil overheating.

[0159] After the circuit breaker is closed, the terminal of the power-loss section checks the three-phase voltage within 300 milliseconds. The 300 milliseconds are used to cover the voltage transient stabilization process after the circuit breaker is closed, the telemetry refresh cycle, and the short-term impact decay process. If the voltage recovers to more than 0.9 times the rated voltage and the three-phase voltage imbalance does not exceed 2%, the recovery is confirmed to be successful. 0.9 times the rated voltage can be used as the minimum voltage boundary after power supply is restored, and the 2% three-phase voltage imbalance is used to identify phase loss, wiring abnormality, or imbalance recovery status.

[0160] If the voltage does not meet the standard or the imbalance exceeds 2%, the current closing operation will be cancelled and a new candidate path will be selected. Cancelling the current closing operation means opening the control switch or backup power supply connection switch that has been closed and restoring the switch status record before closing. If the restoration criteria cannot be met after two consecutive path switching operations, manual takeover will be initiated.

[0161] During the recovery process, if the difference between the output of the distributed generation and the load of the section is less than 10% of the total load of the section, and the section is disconnected from the main grid or in a state of waiting to be reconnected, and a frequency difference of more than 0.5 Hz or a phase difference of more than 30 degrees is detected with respect to the main grid, then it is determined that there is a risk of distributed generation islanding. 10% is used to identify the state where the output of the distributed generation and the load of the section are close to balance and may maintain islanded operation; 0.5 Hz frequency difference and 30-degree phase difference are used to identify the risk of reconnection that is out of sync with the main grid.

[0162] When there is an islanding risk, the smart distribution terminal first sends a disconnection command or a grid-connection blocking command to the relevant distributed power source. After the islanding risk is eliminated, the terminal will perform the closing and restoration. The elimination of the islanding risk means that the relevant distributed power source confirms disconnection or grid-connection blocking, and the frequency difference and phase difference between the section to be restored and the main grid are restored to the allowable closing range.

[0163] After the above processing, the following results are generated: fault section confirmation result, isolation control result, recovery path selection result, and power restoration confirmation result.

[0164] Example 2: Based on Example 1, the specific application process of a method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal is further explained:

[0165] Specifically, such as Figure 7 The following description uses a 10 kV active distribution network ring feeder as an example. This feeder is powered by the substation busbar and is equipped with a first, second, third, fourth, and fifth section switch sequentially along its length. The feeder ends in connection with adjacent feeders via a tie switch. Each section switch and tie switch is equipped with a smart distribution terminal, which exchanges local electrical quantities, switch status, distributed power supply operating status, and fault direction determination information via a peer-to-peer communication link. A distributed photovoltaic grid-connected unit is connected in the middle of the feeder, and an energy storage converter is connected to the downstream user side. A backup power supply path is provided on the tie side.

[0166] During normal operation of the distribution network, all switches from the first to the fifth section are in the closed state, and the tie switch is in the open standby state; distributed photovoltaic power generation is in the grid-connected power generation state, and energy storage converters are in the discharging state or standby state; each intelligent distribution terminal timestamps the local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position, and distributed power source operation status according to a unified time base, and receives electrical quantities and switch status information sent by adjacent terminals; each terminal forms an operation data cache, an adjacent terminal information table, and a network connection relationship record locally, enabling each terminal to identify the line operation status on the same time axis;

[0167] During operation, when the energy storage converter switches from standby to discharge mode, or when the output of distributed photovoltaic power generation decreases significantly due to cloud cover, the intelligent distribution terminal re-identifies the current grid connection relationship based on changes in the status of distributed power sources and changes in electrical quantities of adjacent terminals. For the connection relationship between adjacent terminals, the intelligent distribution terminal first performs passive identification using voltage amplitude, voltage phase angle, current change trend, and active power change trend. When the passive identification result is inconsistent with the original connection relationship, or when events such as switch change, load transfer, or distributed power source commissioning / decommissioning are detected, the intelligent distribution terminal initiates active verification. Active verification can use weak characteristic current verification or operational status linkage change verification to determine whether the adjacent terminals are directly connected, indirectly connected, electrically isolated, or in a pending confirmation state, and updates the connection relationship reliability.

[0168] At a certain operating moment, the line between the third and fourth section switches is still in a direct connection state, but the energy storage converter downstream of the fourth section switch switches switches from standby to discharge state, causing reverse power flow in this section. The intelligent distribution terminal at the third section switch determines the power supply source on the forward and reverse sides of the local switch based on the current grid connection record, and includes the substation main power supply, photovoltaic inverter type distributed power supply, energy storage converter type distributed power supply and backup power supply in the short-circuit current contribution calculation. When the terminal determines that the maximum short-circuit current on the reverse side has reached the level that affects the overcurrent protection start-up and direction judgment, it marks the third section switch and its adjacent sections as bidirectional power flow sensitive state.

[0169] Under bidirectional power flow sensitive conditions, the intelligent distribution terminal at the third section switch does not directly use the original fixed overcurrent setting value. Instead, it dynamically verifies the setting values ​​of each section of the overcurrent protection by combining the current maximum load current, the contribution range of forward and reverse short-circuit currents, and the minimum identifiable fault current. At the same time, the terminal forms a directional protection criterion based on the zero-sequence component, negative-sequence component, and the direction of the sudden change. When the high-frequency noise is large, or the amplitude of the negative-sequence and zero-sequence characteristics is low, the directional criterion is marked as pending verification, or the confidence level of the directional criterion is reduced. Thus, before a fault occurs, each terminal can obtain a protection parameter version and a directional criterion version that are adapted to the current grid connection relationship and bidirectional power flow state.

[0170] When a phase-to-phase short-circuit fault occurs in the line section between the third and fourth section switches, the fault initiation time is recorded as T0. After the fault occurs, the intelligent distribution terminal at the third section switch detects that the local phase current exceeds the current overcurrent setting value within multiple consecutive sampling judgment windows, and at the same time, the energy of the transient component from 500 Hz to 2 kHz is significantly higher than the dynamic baseline, thus confirming the fault initiation. The intelligent distribution terminal at the fourth section switch also detects the transient change of the fault within a similar time. The terminals at both ends respectively extract the three-phase voltage, three-phase current, zero-sequence voltage, and zero-sequence current data before and after the fault, centered at T0. After reviewing the missing key channels, sampling saturation, and version consistency, local direction determination information is generated.

[0171] The intelligent distribution terminal at the third section switch votes based on the zero-sequence direction characteristics, negative-sequence direction characteristics, and transient change direction characteristics within the fault time window, determining that the fault direction points downstream of the fourth section switch, and the direction matching degree is high; the intelligent distribution terminal at the fourth section switch determines that the fault direction points upstream of the third section switch based on similar characteristics; since the direction determination results at both ends are opposite, and the protection parameter versions and grid connection relationship versions at both ends are consistent, each terminal marks the line section between the third section switch and the fourth section switch as a candidate fault section, and sends the local direction determination information to the adjacent terminal through strong real-time communication messages;

[0172] After receiving fault direction determination information from multiple terminals, the intelligent distribution terminal at the third section switch or the terminal responsible for collaborative determination of this section traverses adjacent terminal pairs according to the current grid connection relationship record. For the terminal corresponding to the third section switch and the terminal corresponding to the fourth section switch, since the direction determination result of the third terminal points downstream and the direction determination result of the fourth terminal points upstream, and the connection status of the two ends is direct connection, this section obtains a higher candidate score. The candidate score comprehensively considers the average direction matching degree of the two ends, the reliability of the connection relationship, the consistency of the fault start time, and the data quality score. When the candidate score of this section is higher than the automatic confirmation boundary, and the switch positions at both ends are valid, the self-test of the operating mechanism is normal, and the protection version is consistent, the line between the third section switch and the fourth section switch is confirmed as the fault section.

[0173] After confirming the faulty section, the two intelligent distribution terminals first perform a two-way handshake confirmation. The intelligent distribution terminal at the fourth section switch sends a tripping preparation message to the intelligent distribution terminal at the third section switch. The tripping preparation message includes at least the following fields: switch position, energy storage status of the operating mechanism, tripping coil status, suggested action time, and self-test result of the operating mechanism. After verifying that the local switch meets the action conditions, the intelligent distribution terminal at the third section switch returns a confirmation message to the intelligent distribution terminal at the fourth section switch. Subsequently, the two terminals send tripping pulses to the tripping coils of the third and fourth section switches respectively under a unified time reference. If both switches report the tripping position within the specified monitoring time, the faulty section is successfully isolated. If neither terminal reports the tripping position, the corresponding terminal retryes the process a preset number of times. If the retry fails, an expanded isolation request is sent to the corresponding terminal of the next higher-level switch.

[0174] After the faulty section is isolated, the non-faulty user section downstream of the fourth section switch may experience power loss. When the terminal corresponding to this downstream section detects that the voltage is continuously lower than the power loss judgment threshold, it is marked as a section to be restored, and an estimated value of the restored load is generated based on the short-term average active power before the fault and the load importance level. The intelligent distribution terminal searches for candidate power supply paths according to the current grid connection relationship record. The candidate power supply paths may include power transfer from adjacent feeders via tie switches, or power restoration to some loads via backup power access switches.

[0175] For each candidate power supply path, the intelligent distribution terminal calculates the power capacity margin, line thermal stability margin, terminal voltage drop, phase angle difference on both sides of the tie switch, and distributed generation islanding risk. If the power capacity margin, line thermal stability margin, terminal voltage quality, and synchronization conditions of the candidate path all meet the safety boundary, and there is no unresolved islanding risk, then the candidate path is entered into the ranking. The intelligent distribution terminal calculates the power supply capacity score for the candidate paths that have entered the ranking, and prioritizes the candidate path with the highest score that is not lower than the recovery boundary. For example, when the adjacent feeder still has sufficient capacity margin, the phase angle difference on both sides of the tie switch meets the synchronization requirements, and the downstream distributed generation has not formed an island, the intelligent distribution terminal selects the tie switch as the recovery switch.

[0176] Before performing the closing and restoration, the intelligent distribution terminal re-verifies whether the faulty section has been successfully isolated, whether the candidate path still meets the capacity and thermal stability safety boundaries, whether the voltage on both sides of the tie switch meets the synchronization check, whether the distributed power supply islanding risk has been eliminated, and whether the switch operating mechanism is normal. After the verification is passed, the intelligent distribution terminal sends a closing pulse to the tie switch. After the closing is completed, the terminal corresponding to the downstream power loss section verifies the three-phase voltage recovery status and three-phase voltage imbalance within a set time. If the voltage is restored to the rated voltage range and the three-phase voltage imbalance meets the requirements, the power supply is confirmed to be successfully restored, and the isolation result, restoration path, switch action result, and restoration confirmation information are sent to the distribution master station.

[0177] If, during the restoration process, it is detected that the output of the distributed power source is nearly balanced with the load of the power-loss section, and that the section is disconnected from the main grid or awaiting reconnection, while the frequency difference or phase difference exceeds the limit, the intelligent distribution terminal determines that there is a risk of distributed power source islanding. In this case, the intelligent distribution terminal first sends a disconnection command or a grid-connection blocking command to the relevant distributed power source. After the distributed power source confirms disconnection or grid-connection blocking, and after the frequency difference and phase difference between the restored section and the main grid are restored to the allowable range, the reconnection restoration is then performed. If the voltage does not reach the restoration criteria after the selected path is reconnected, the intelligent distribution terminal cancels the current reconnection and reselects a candidate path. If continuous path switching still cannot meet the restoration criteria, the system enters manual takeover mode.

[0178] Through the above operation process, the intelligent distribution terminal does not simply perform tripping based on fixed overcurrent settings or pre-fixed grid information. Instead, it dynamically updates protection criteria based on the status of distributed power sources and grid connection relationships before a fault occurs. After a fault occurs, it determines the fault section through multi-terminal direction information and waveform verification. After isolation is completed, it performs recovery control according to the power supply capacity and safety boundary of the candidate power supply path. This can reduce misjudgment of fault sections caused by the lag in grid connection relationships, reduce the interference of distributed power source reverse feed-in on direction judgment, narrow the fault isolation range, and improve the power restoration efficiency of non-faulty power loss sections.

[0179] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0180] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, the above embodiments can be implemented in whole or in part by a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions of the embodiments of this application are implemented in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted wirelessly or wiredly from one website, computer, server, or data center to another website, computer, server, or data center. Wired methods include optical fiber, twisted pair, coaxial cable, etc. Wireless methods include infrared, microwave, etc. Available media include any available media that can be accessed by a computer or data storage devices such as servers and data centers that contain one or more sets of available media. Available media can be magnetic media (floppy disks, hard disks, magnetic tapes), optical media (DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0181] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0182] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal, characterized in that, include: S1. The intelligent power distribution terminal collects local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position and distributed power supply operation status, and receives electrical quantities, switch status and grid verification information sent by adjacent terminals. S2. Identify the current grid connection relationship based on locally collected information and adjacent terminal information, and update the grid connection relationship record when the switch changes position, distributed power supply is put into or taken out, or load is transferred. S3. Adjust the overcurrent protection setting parameters and directional protection criteria based on the current grid connection relationship, distributed power output status and power flow direction; S4. After detecting the fault start, extract the electrical quantity data before and after the fault, generate local fault direction determination information, and exchange fault direction determination information with adjacent terminals. S5. Determine the fault section based on the fault direction determination information of each terminal and the current network connection relationship, control the switches at both ends of the fault section to open and isolate, and control the tie switch or backup power supply access switch to close and restore based on the load of the power-loss section and the candidate power supply path.

2. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, The intelligent power distribution terminal collects local three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current, switch position, and distributed power source operating status, including: During local information collection, the collection channel and communication receiving channel are uniformly time-marked, and voltage, current and zero-sequence quantities are collected according to the power frequency cycle. The collected data is written into an electrical quantity record containing channel number, collection time, quality flag and source flag. Debouncing confirmation is performed on switch position changes, and power type identifier and fault current contribution range identifier are generated based on the operating status of distributed power sources. When an abnormality is detected in the local acquisition channel, the abnormal channel is marked with a quality indicator and verified by means of adjacent phase calculation, zero-sequence component verification, or comparison of similar electrical quantities between adjacent terminals.

3. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Receive electrical quantities, switch status, and network structure verification information sent by adjacent terminals, including: When receiving information sent by neighboring terminals, perform source verification, timestamp verification, sequence number continuity verification, and field integrity verification on the messages sent by neighboring terminals. Write messages that pass the verification into the neighboring terminal information table, mark messages with timestamp deviation as delayed messages, and record messages that fail the verification as communication abnormal events. The adjacent terminal information table records voltage amplitude, voltage phase angle, current amplitude, current direction, switch status, verification timestamp, and trust level, and adjusts the trust level of adjacent terminals based on the message verification results.

4. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Identify the current network connection relationships based on locally collected information and information from adjacent terminals, including: When identifying the current network connection relationship, a connection relationship matrix and an adjacency relationship table are used to jointly record the connection status, connection strength, adjacency relationship and update time of the terminal pairs; The connection status is categorized into direct connection, indirect connection, electrical isolation, and pending confirmation. Within the statistical window, the connection status of the terminal pair is determined based on the voltage amplitude difference, voltage phase angle difference, current change trend, active power change trend, and zero-sequence quantity synchronization change degree.

5. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Update the grid connection records when switches are changed, distributed power sources are put into operation or deactivated, or loads are transferred, including: When updating the connection relationship record, the connection status is verified by using weak characteristic current verification or operation status linkage verification after the switch status changes, the operating status of the distributed power source changes, or the load transfer is triggered. Based on the differences between the old and new connection relationships, a connection change index is generated, and a connection relationship summary is sent to adjacent terminals, power distribution master stations, or higher-level terminals. Based on the returned summary matching results, perform partial updates, mark pending confirmation, retain rollback versions, or rebuild the connection records.

6. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Based on the current network connection relationship, distributed power output status, and power flow direction, adjust the overcurrent protection setting parameters and directional protection criteria, including: When adjusting protection parameters, protection parameter update tasks are created based on periodic triggering and event triggering. The power flow sensitivity state is determined based on the current network connection version, distributed power source operating status, local load current, power flow direction, and forward and reverse short-circuit current contributions. Load avoidance and fault sensitivity checks were performed on the overcurrent protection parameters. Based on the polarity relationship between the zero-order component, the negative-order component, and the mutation amount, a directional protection criterion, a low margin flag, and a protection parameter version number are formed.

7. The method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, After a fault is detected, electrical quantity data before and after the fault is extracted to generate local fault direction determination information, including: During fault start identification, the overcurrent start condition and transient energy start condition are monitored based on the current protection setting parameters. The fault start time is determined by the sampling judgment window that first meets the start condition, and the voltage, current, zero sequence quantity and switch status data before and after the fault are captured. After processing the intercepted data by checking for missing channels, saturated sampling, data completion, and low-confidence marking, the direction protection criterion is invoked. Local fault direction determination information is generated based on zero-order direction characteristics, negative-order direction characteristics, and transient mutation direction characteristics.

8. A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, It also exchanges fault direction determination information with adjacent terminals, including: When exchanging fault direction information, the local fault direction determination information is sent to the adjacent terminal through a strong real-time communication message. The strong real-time communication message records the terminal number, switch number, fault start time, direction determination result, direction matching degree, protection parameter version number, network connection relationship version number and data quality flag. The received fault direction determination information is checked for version consistency. The message trust level is adjusted according to the check result. The checked fault direction determination information is written into the fault coordination cache for minimum direction combination judgment.

9. A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Based on the fault direction determination information of each terminal and the current network connection relationship, the fault section is determined, and the switches at both ends of the fault section are controlled to trip and isolate the fault, including: When confirming the fault section, adjacent terminal pairs are used as the judgment unit. Candidate scores are generated based on the direction of the relative orientation, the reliability of the connection relationship, the consistency of the fault start time, and the data quality score. For candidate sections whose candidate scores meet the predetermined confirmation conditions, fault recording data is used to verify current energy. After the verification is passed, the terminals at both ends of the section perform double-end handshake confirmation and synchronous tripping control. If any section-end switch does not report the open position, an open retry is performed, and an expanded isolation request is sent after the retry fails.

10. A method for rapid isolation and recovery of fault sections in an intelligent power distribution terminal according to claim 1, characterized in that, Based on the load of the power outage section and the candidate power supply path, the tie switch or backup power access switch is closed to restore power, including: When restoring power supply control, the section to be restored is determined based on the power loss node after the circuit breaker trips, the estimated load before the fault, and the load importance level. The candidate power supply paths are scored for power capacity margin, line thermal stability margin, end voltage drop, loop closing safety, and islanding risk. The candidate power supply path that meets the safety boundary is selected to perform the circuit breaker closing control. When there is no single candidate power supply path that meets the safety boundary, the section to be restored is restored in sections, and voltage verification, islanding risk removal verification and abnormal rollback processing are performed after the circuit is closed.

Images