Fault adaptive recovery method in loss-of-mains mode of hybrid switch power distribution network

Abstract

This invention relates to the field of distribution network fault recovery technology, specifically an adaptive fault recovery method for hybrid switchgear distribution networks under power outage conditions. The method includes: collecting distribution network topology data, distributed generation operating parameters, and real-time power demands of load nodes under power outage conditions; unifying the format and aligning the time stamps to form an initial operating dataset; identifying disconnected tie switches and sectionalizing switches in the network, classifying them into mechanical switch groups and power electronic switch groups according to their physical type; constructing a candidate reconfiguration path set including switch action delay and transmission capacity constraints based on the differences in the operating characteristics of the two types of switches; performing radial topology verification on each path, eliminating paths that do not meet the unidirectional power supply constraint, and retaining usable recovery paths; and calculating the remaining transmission margin of each usable recovery path as the basis for path selection. This method can achieve accurate and efficient adaptive recovery of power outage faults in hybrid switchgear distribution networks.

Description

Technical Field

[0001] This invention relates to the field of power distribution network fault recovery technology, and in particular to an adaptive fault recovery method for hybrid switch distribution networks under power outage mode. Background Technology

[0002] Hybrid switchgear distribution networks, combining the advantages of both mechanical and power electronic switches, are increasingly widely used in distribution network systems. When a power outage occurs in the distribution network, fault recovery is a crucial step in ensuring power supply reliability. Existing power outage fault recovery methods typically collect data such as network topology, distributed generation operating parameters, and load demand to identify switches in the network that are in an open state. This allows for the construction of candidate reconfiguration paths, followed by topology verification to filter available paths and ultimately restore power supply.

[0003] Existing technologies, when handling power outage recovery in hybrid switchgear distribution networks, fail to differentiate between different physical types of switches, treating mechanical and power electronic switches as a single type for path construction, ignoring the fundamental differences in their operational characteristics. Furthermore, existing methods do not adequately consider the dual constraints of switch action delay and transmission capacity, relying solely on topology for path design. This results in a large number of invalid paths that do not meet actual operational requirements, increasing the workload of topology verification. Moreover, the selected usable paths are difficult to adapt to the operational characteristics of hybrid switches, failing to achieve efficient and accurate adaptive power outage recovery. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an adaptive fault recovery method for hybrid switchgear distribution networks under power outage conditions.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a fault adaptive recovery method for hybrid switchgear distribution networks under power outage mode, comprising: Data on the topology of the hybrid switchgear distribution network under power outage conditions, the operating parameters of distributed power sources, and the real-time power demand of load nodes are collected. The collected data is formatted and time-stamped to form an initial operating dataset under power outage scenarios. Based on the initial running dataset, identify the tie switches and sectionalizing switches that are currently disconnected in the network, and classify them into mechanical switch groups and power electronic switch groups according to the physical type of the switches. Based on the differences in the operating characteristics of the mechanical switch group and the power electronic switch group, a set of candidate reconstruction paths is constructed, which includes switching operation delay and transmission capacity constraints. For each path in the candidate reconstruction path set, a radial topology check is performed to eliminate paths that do not meet the unidirectional power supply constraint and retain available recovery paths that meet the operation rules. Based on the load distribution of the available recovery paths, the remaining transmission margin of each path is calculated, and the remaining transmission margin is used as the basis for subsequent path selection.

[0006] As a further aspect of the present invention, the step of unifying the format and aligning the time stamps of the collected data to form an initial operating dataset under a power failure scenario specifically includes: Read telemetry data and teleindication data uploaded by the smart terminal installed on the power distribution line. The telemetry data includes voltage amplitude and current amplitude, and the teleindication data includes the opening and closing status of the switch. Map the data fields reported by smart terminals from different manufacturers to a unified internal data dictionary, and perform unit conversion for parameters with inconsistent dimensions. Based on the timing signal from the Global Positioning System, the telemetry and teleindication data are timestamped and then processed by moving average within a time window to filter out transient interference. The processed data is indexed by node number and associated with a pre-stored static topology map of the distribution network to generate an initial operational dataset for a power outage scenario that includes spatial location information.

[0007] As a further aspect of the present invention, taking into account the differences in operating characteristics between the mechanical switch group and the power electronic switch group, a set of candidate reconstruction paths is constructed, including switching operation delay and transmission capacity constraints, specifically including: The state of the mechanical switch group is extracted from the initial running dataset. The mechanical switch group has a minute-level action delay and the operation is irreversible. The state of the power electronic switch group is extracted. The power electronic switch group has a response speed of seconds or less and supports multiple continuous adjustments. Starting from the power outage area, all power supply paths are searched in the adjacency matrix of the distribution network. The power supply paths are composed of alternating connections of the mechanical switch group and the power electronic switch group. A path attribute label is set for each power supply path. The path attribute label records the number of mechanical switches, the number of power electronic switches, and the theoretical maximum current carrying capacity of the entire path. The cumulative action delay value in the path attribute label is compared with the transmission capacity constraint. Paths that have timed out or overloaded are eliminated, and the remaining paths constitute a set of candidate reconstruction paths.

[0008] As a further aspect of the present invention, a radial topology check is performed on each path in the candidate reconstruction path set, specifically including: Traverse each path in the candidate reconstruction path set and perform a virtual closure operation on the switch states involved in the path in memory; The graph theory algorithm is called to detect whether there is a loop in the network structure after the closure operation. If a loop exists, the path is determined not to satisfy the radial topology constraint. Simultaneously, it detects whether there are isolated nodes in the network after the closure operation. An isolated node refers to a node that is connected to the power supply side but has no load connected. If a loop or isolated node is detected, the virtual closure operation is rolled back in memory, and the path is removed from the candidate reconstruction path set. Mark all paths that have not triggered a rollback operation as available recovery paths and record their connection relationships in the topology.

[0009] As a further aspect of the present invention, the remaining transmission margin of each path is calculated based on the load distribution of the available recovery paths, specifically including: Retrieve the output prediction curves of the distributed power sources connected to the currently available recovery paths, where the output prediction curves cover the entire fault recovery process. The historical load curves of the load nodes at the end of the available recovery path are read and corrected in conjunction with the current meteorological data to obtain the actual load power demand. The power output prediction curve and the load power demand are superimposed on the time axis to calculate the actual power flow distribution of the available recovery path at each time. The difference between the actual power flow distribution and the physical rated capacity of the available recovery path is the remaining transmission margin at the current moment. The minimum remaining transmission margin at all times is taken as the final remaining transmission margin value of the available recovery path.

[0010] As a further aspect of the present invention, it also includes the step of path optimization and switching based on the remaining transmission margin: Sort the set of available recovery paths from highest to lowest based on the remaining transmission margin value; Select the highest-ranked available recovery path as the primary recovery path and send its corresponding switch action command to the underlying control device. The voltage recovery status at the end of the main recovery path is monitored within a preset waiting time, and the voltage recovery status at the end is collected by a voltage transformer installed at the monitoring point. If a stable voltage recovery signal is not detected within the waiting period, the next available recovery path is selected as the backup path, and a switching operation is performed. Repeat the monitoring and switching process until all load nodes in the power-loss area are restored to power.

[0011] As a further aspect of the present invention, monitoring the voltage recovery status at the end of the main recovery path within a preset waiting time specifically includes: At the target load node of the main recovery path, a lower voltage threshold and an upper voltage threshold are set; The voltage sample values ​​from the target load node are received in real time, and their standard deviation over a period of time is calculated to evaluate the stability of the voltage. If the voltage sample value remains between the lower and upper limits of the voltage threshold for multiple consecutive cycles, and the standard deviation is less than the preset fluctuation threshold, then the voltage recovery is considered successful. If the voltage sample value drops below the lower limit of the voltage threshold, or if severe oscillations cause the standard deviation to exceed the limit, the recovery is deemed to have failed. The determination result is fed back to the path optimization module to control the subsequent backup path selection logic.

[0012] As a further aspect of the present invention, the step of setting a path attribute label for each power supply path also includes a quantitative assessment of the support capability of distributed power sources: Extract the type of the distributed power source and distinguish whether it is a photovoltaic power source, a wind power source, or an energy storage power source; Based on the type of the distributed power source, read its maximum active power output under the current light intensity or wind speed conditions; Based on the state of charge of the energy storage power source, estimate the maximum support duration it can provide during fault recovery; The upper limit of active power output and the maximum support duration are converted into an equivalent capacity support coefficient, and the equivalent capacity support coefficient is written into the path attribute tag. When calculating the remaining transmission margin, the equivalent capacity support coefficient is superimposed as a gain factor onto the output prediction curve of the distributed power source.

[0013] As a further aspect of the present invention, the power output prediction curve and the load power demand are superimposed on the time axis to calculate the actual power flow distribution of the available recovery path at each time point, specifically including: The output prediction curve is discretized into an output sequence with the same time step, and the load power demand is discretized into a load sequence. Subtract the values ​​at corresponding time points from the output sequence to obtain the net power flow sequence; Substitute the net power flow sequence into the admittance matrix of the available recovery path, solve the node voltage equations, and obtain the voltage phase angle of each node; The active power loss of the line is calculated based on the voltage phase angle of the node, and the active power loss is subtracted from the net power flow sequence to correct and obtain the actual power flow distribution. The actual power flow distribution data is stored in the power flow calculation result cache area for use in the calculation step of the remaining transmission margin.

[0014] As a further aspect of the present invention, the method of using graph theory algorithms to detect whether there are loops in the network structure after the closure operation includes: The network structure after the virtual closure operation is read from memory. The network structure consists of a node list and an edge list, and the edge list corresponds to the lines connected by the closed switch. The node list and edge list are converted into an adjacency matrix representation, where non-zero elements indicate an electrical connection between two corresponding nodes. In the topological graph represented by the adjacency matrix, any node is selected as the starting node for depth-first traversal, and the traversal begins from the starting node. During the depth-first traversal, a list of visited nodes is maintained to record the nodes that have been visited. When a new node is visited, the new node is marked as visited. When visiting all neighboring nodes of the new node, check whether the neighboring nodes of the new node already exist in the list of visited nodes and whether the neighboring nodes are not the direct parent nodes of the current traversal. If the conditions are met, it is determined that there is a loop in the network structure and all nodes contained in the loop are recorded.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Based on the physical type of the switches, tie switches and sectionalizing switches in the open state are divided into mechanical switch groups and power electronic switch groups. This breaks the conventional technology's undifferentiated and uniform treatment of all switches. It can accurately adapt to the inherent operating characteristics of the two types of switches, avoid the disconnect between path construction and actual equipment operation caused by confusion of switch types, make the construction of subsequent candidate reconfiguration paths more targeted, reduce the problem of invalid paths caused by mismatch of switch operating characteristics, and make path construction more in line with the actual operating state of hybrid switch distribution networks.

[0016] By combining the differences in operating characteristics between mechanical and power electronic switchgear, a candidate reconfiguration path set is constructed, incorporating switch action delay and transmission capacity constraints. This differs from conventional methods that lack differences in switch type and dual constraints, enabling the elimination of invalid paths due to action delay conflicts or insufficient transmission capacity. This reduces the workload of subsequent radial topology verification and makes the candidate reconfiguration paths more consistent with the operational constraints of hybrid switchgear distribution networks. This results in more rational and adaptable available recovery paths, aligning with distributed generation operating parameters and load node power requirements, thereby improving the adaptability and efficiency of power outage fault recovery. Attached Figure Description

[0017] Figure 1 This is a flowchart of the fault adaptive recovery method for hybrid switchgear distribution network under power loss mode according to the present invention; Figure 2 A flowchart for generating the initial runtime dataset under a power outage scenario; Figure 3 A flowchart for calculating the remaining transmission margin of each path; Figure 4 A power timing comparison diagram for the fault recovery phase of a hybrid switchgear distribution network; Figure 5 This is a comparison diagram of voltage recovery timing during fault recovery in a hybrid switchgear distribution network. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1The system collects topology data, distributed generation operating parameters, and real-time power demand of load nodes in a hybrid switchgear distribution network under power outage conditions. The collected data is formatted and time-stamped to form an initial operational dataset for the power outage scenario. Based on this initial operational dataset, disconnected tie switches and sectionalizing switches in the current network are identified and classified into mechanical switch groups and power electronic switch groups according to their physical type. Combining the differences in operating characteristics between the mechanical and power electronic switch groups, a candidate reconfiguration path set is constructed, incorporating switch operation delays and transmission capacity constraints. For each path in the candidate reconfiguration path set, a radial topology check is performed, eliminating paths that do not meet the unidirectional power supply constraint and retaining usable recovery paths that conform to the operating rules. Based on the load distribution of the usable recovery paths, the remaining transmission margin of each path is calculated, and this remaining transmission margin is used as the basis for subsequent path selection.

[0021] In one embodiment of the present invention, see [reference] Figure 2 The system reads telemetry and tele-signaling data uploaded by smart terminals installed on the power distribution lines. Telemetry data includes voltage and current amplitudes, while tele-signaling data includes the open / closed status of switches. The smart terminals come from different manufacturers, such as Manufacturer A and Manufacturer B. Manufacturer A's smart terminals report voltage data with the field name "Voltage," while Manufacturer B's smart terminals report voltage data with the field name "U." The system maps the data fields reported by smart terminals from different manufacturers to a unified internal data dictionary. This internal data dictionary defines standard field names such as "Voltage Amplitude," "Current Amplitude," and "Switch Status," and performs unit conversions for parameters with inconsistent dimensions. For example, the voltage unit kilovolt reported by Manufacturer A is converted to volt as defined in the internal data dictionary, and the current unit ampere reported by Manufacturer B is converted to kiloampere. Based on the GPS timing signal, the telemetry and tele-signaling data are timestamped, and a moving average is performed according to a time window to filter out instantaneous interference. The moving average is calculated using the following formula:

[0022] in: Indicates at a point in time Processed data values, Indicates the length of the sliding time window. Indicates at a point in time The raw data values ​​are processed and indexed by node number. The processed data is then associated with a pre-stored static topology map of the power distribution network to generate an initial operating dataset for a power outage scenario that includes spatial location information. The initial operating dataset is stored in matrix form, with rows corresponding to time series and columns corresponding to node parameters such as voltage amplitude, current amplitude, and switch status.

[0023] In some embodiments, data field mapping is achieved by querying a predefined mapping table. This table stores the correspondence between field names from different manufacturers and field names in the internal data dictionary. Unit conversion applies a conversion factor based on the parameter type; for example, multiplying by 1000 is used when converting voltage from kilovolts to volts. It can be understood that the GPS timing signal ensures that all smart terminal data has a unified time base, and the time window length in the moving average processing... The dataset is set to 10 sampling points, corresponding to a 10-second time window. Optionally, data comparison can be used to show instantaneous changes in voltage amplitude in the original telemetry data. For example, the voltage at node number 101 is recorded as 10.5 kV at time point t1 and 9.5 kV at time point t2. After moving average processing, the voltage at time point t2 is corrected to 10.0 kV, reducing voltage sequence fluctuations in the initial running dataset. In some embodiments, the pre-stored static topology map of the distribution network contains a mapping between node numbers and geographic coordinates. The association operation binds the processed data records with the topology node coordinates, generating an initial running dataset containing node number, timestamp, voltage amplitude, current amplitude, switch status, and geographic coordinate fields.

[0024] In one embodiment of the present invention, the state of mechanical switch groups is extracted from the initial operating dataset under a power outage scenario. The initial operating dataset contains state records of switches numbered S1, S2, and S3. The state records show that switches S1 and S2 are in the open state and are identified as mechanical. Mechanical switch groups have a minute-level action delay and the operation is irreversible. The state of power electronic switch groups is extracted. The initial operating dataset shows that switches numbered P1 and P2 are in the open state and are identified as power electronic. Power electronic switch groups have an action response speed of seconds or less and support multiple continuous adjustments. Starting from the power outage area, all power supply paths are searched in the adjacency matrix of the distribution network. The rows and columns of the adjacency matrix represent network nodes. A matrix element value of 1 indicates that there is an operable switch between two nodes. The search algorithm traverses the path from the power outage node to all possible power supply nodes. The power supply path is composed of alternating connections of mechanical switch groups and power electronic switch groups. For example, a power supply path is recorded as "Node A - Switch S1 (mechanical) - Node B - Switch P1 (power electronic) - Node C - Power Supply". Each power supply path is assigned a path attribute tag, which records the number of mechanical switches, the number of power electronic switches, and the theoretical maximum current carrying capacity of the entire path. During the path attribute tagging process, the type of distributed generation is extracted, distinguishing between photovoltaic (PV), wind, and energy storage power sources. Based on the type, the upper limit of active power output of the distributed generation under current sunlight or wind speed conditions is read. Combined with the state of charge of the energy storage power source, the maximum support duration provided by the energy storage power source during fault recovery is estimated. The upper limit of active power output and the maximum support duration are converted into an equivalent capacity support coefficient, which is then written into the path attribute tag. The conversion process uses a formula:

[0025] in: This represents the equivalent capacity support coefficient. This indicates the upper limit of the active power output of the distributed generation. This indicates the maximum duration of power storage, for both photovoltaic and wind power sources. The value is the preset fault recovery period length. Indicates the reference power capacity. This indicates the reference time length. The cumulative action delay value in the path attribute label is compared with the transmission capacity constraint. The cumulative action delay value is obtained by multiplying the number of mechanical switches in the path by the average action time of a single mechanical switch and the number of power electronic switches by the average action time of a single power electronic switch, and then summing the results. The transmission capacity constraint compares the theoretical maximum current carrying capacity of the path with the power demand after correction by the equivalent capacity support coefficient recorded in the path attribute label. Paths that have exceeded the time limit or are overloaded are eliminated, and the remaining paths constitute the candidate reconstruction path set.

[0026] In some embodiments, the adjacency matrix search for power supply paths employs a depth-first search algorithm, which records all switches and their types traversed during the process. It can be understood that the minute-level action delay of mechanical switch groups refers to a single switch action time within the range of 1 to 3 minutes, while the second-level or lower action response speed of power electronic switch groups refers to a single switch action time within the range of 100 milliseconds to 1 second. Optionally, path attribute tags are stored in a data table structure, with fields including path number, list of mechanical switches, list of power electronic switches, theoretical maximum current carrying capacity, and equivalent capacity support coefficient. In some embodiments, the upper limit of active power output of distributed power sources under the current irradiance is read by querying the power-irradiance curve of the photovoltaic power source, and the upper limit of active power output of distributed power sources under the current wind speed conditions is read by querying the power-wind speed curve of the wind power source. It can be understood that the state of charge of the energy storage power source is directly read from the battery management system, and the maximum support time is estimated by dividing the state of charge by the average discharge power during fault recovery. Optionally, when the data comparison does not consider the equivalent capacity support coefficient, a path containing photovoltaic power has a theoretical maximum current carrying capacity of 1000 kW and a load demand of 900 kW, so the path is retained; after considering the equivalent capacity support coefficient, since it is currently nighttime and the upper limit of the active power output of the photovoltaic power is 0, the power that can be transmitted by the path after the equivalent capacity support coefficient is corrected is reduced, and the path is removed from the candidate reconfiguration path set because it cannot meet the load demand.

[0027] In one embodiment of the present invention, each path in the candidate reconstruction path set is traversed. The candidate reconstruction path set includes records of path numbers Path001 and Path002. The switch states involved in the path are virtually closed in memory. For example, for path Path001, the switches involved include mechanical switches S1 and S2 and power electronic switch P1. The status flags of these three switches are changed from "open" to "closed" in memory. A graph theory algorithm is called to detect whether there is a loop in the network structure after the closure operation. If a loop exists, it is determined that the path does not satisfy the radial topology constraint. The specific process of calling the graph theory algorithm to detect loops includes: reading the network structure after the virtual closure operation from memory. The network structure consists of a node list and an edge list. The node list contains [nodeA, nodeB, nodeC, nodeD], and the edge list corresponds to the lines connected by the closed switches. For example, the edge list contains [(nodeA, nodeB, S1), (nodeB, nodeC, S2), (nodeC, nodeD, P1), (nodeD, nodeA, SWX)], representing four lines and their associated switches. Convert the node list and edge list into an adjacency matrix representation, where non-zero elements indicate an electrical connection between two corresponding nodes. For the edge list, the values ​​of elements Matrix[nodeA][nodeB], Matrix[nodeB][nodeC], Matrix[nodeC][nodeD], and Matrix[nodeD][nodeA] in the adjacency matrix are all 1. In the topology graph represented by the adjacency matrix, select any node as the starting node for depth-first traversal, for example, select nodeA as the starting node, and start traversing from the starting node. During the depth-first traversal, maintain a list of visited nodes to record the nodes that have been visited. Initially, the list of visited nodes is empty. When a new node is visited, mark the new node as visited. For example, after visiting nodeA, add nodeA to the list of visited nodes. When visiting all adjacent nodes of a new node, check whether the adjacent nodes of the new node already exist in the list of visited nodes, and whether the adjacent nodes are not the direct parent nodes of the current traversal. If the condition is met, it is determined that a loop exists in the network structure, and all nodes contained in the loop are recorded. The determination condition is expressed by a logical expression:

[0028] in: It is a Boolean variable indicating whether a loop exists. Represents a node Boolean status indicating whether it has been visited. Indicates the currently inspected adjacent nodes. Represents a node In a depth-first traversal of the tree, the parent node This indicates the node currently being visited.

[0029] In some embodiments, the edge list data structure includes three fields: start node number, end node number, and switch number. It can be understood that depth-first traversal is implemented through a recursive function, whose parameters include the current node, parent node, and a list of visited nodes. Simultaneously, it checks whether there are isolated nodes in the network after the closure operation. An isolated node refers to a node that is connected to the power supply but has no load connected to it. The detection method is to check whether each node in the connected subgraph formed by the virtual closure operation is associated with a record where the load power demand is greater than zero. If a node is connected to the power supply but its load power demand is zero, then the node is determined to be an isolated node. Optionally, data comparison shows that after performing a virtual closure operation on path Path001, the graph theory algorithm detected a loop, and path Path001 is marked as not satisfying the constraint; after performing a virtual closure operation on path Path002, no loop was detected, but node nodeE was detected as an isolated node, and path Path002 is also marked as not satisfying the constraint. If a loop or island node is detected, the virtual closure operation is rolled back in memory, and the path is removed from the candidate reconstruction path set. The rollback operation restores the status flag of the relevant switch to "open".

[0030] In one embodiment of the present invention, see [reference] Figure 3 The process involves retrieving the output prediction curves of the distributed power sources connected to the currently available recovery path, designated Path-A. One of the connected distributed power sources is a photovoltaic (PV) power source, whose output prediction curve covers the entire fault recovery process, e.g., from 08:00 to 10:00, with predicted values ​​at 15-minute intervals. The historical load curves of the load nodes at the end of Path-A (node ​​number L-5) are read and corrected using current meteorological data to obtain the actual load demand. The meteorological data includes a current temperature 3 degrees Celsius higher than the historical average for the same period, with a correction factor of 1.1 applied to the historical load curves. The output prediction curves and load demand are overlaid on the time axis to calculate the actual power flow distribution of Path-A at various times. This calculation involves discretizing the output prediction curves into output sequences with a time step consistent with a 15-minute interval, denoted as […]. ,in The load power demand is discretized into a load sequence, denoted as [a_t]. Subtracting the output sequence from the load sequence at corresponding time points yields the net power flow sequence. The net power flow sequence Substituting the admittance matrix into the available recovery path Path-A (which is established based on the path's line parameters), we solve the node voltage equations to obtain the voltage phase angles at each node. ,in Index the nodes. Based on the voltage phase angle of the nodes. Calculate the active power loss of the line and from the net power flow sequence Subtracting active power losses The actual power flow distribution is obtained after correction. The correction relationship is described by the formula:

[0031] in: Indicates at a point in time The actual power flow distribution value, Indicates at a point in time Net power flow value, Indicates at a point in time The calculated active power loss value is stored in the power flow calculation result cache for later use in the remaining transmission margin calculation step. See Table 1 for examples of output sequence, load sequence, and net power sequence.

[0032] Table 1: Comparison Table of Output Sequence, Load Sequence, and Net Power Sequence Time point (k) time Output sequence Load sequence Net power sequence 1 08:00 520 500 20 2 08:15 600 580 20 3 08:30 700 750 -50 4 08:45 750 800 -50 5 09:00 800 780 20 In some embodiments, the power output prediction curve is derived from the ultra-short-term prediction results provided by the photovoltaic power prediction system, with a time resolution of 15 minutes. It can be understood that the historical load curve is the average value of load node L-5 at the same time over the past seven days, and meteorological data correction is achieved by querying the temperature-load sensitivity coefficient table. The actual power flow distribution is then considered. Comparing the available recovery path Path-A to its physical rated capacity of 1000 kilowatts, the difference represents the remaining transmission margin at the current moment. Optionally, a data comparison display shows the calculation results of available recovery paths Path-A and Path-B, and the actual power flow distribution of Path-A at 08:30. With a capacity of -50 kW, its remaining transmission margin The power flow of Path-A is 950 kW; the actual power flow distribution of Path-B at the same time is -200 kW, and its remaining transmission margin is 800 kW. The minimum value of the remaining transmission margin at all times is taken as the final remaining transmission margin value of the available recovery path. For Path-A, its remaining transmission margin sequence at each time is [980,980,950,950,980] kW, and the minimum value is 950 kW. Therefore, the final remaining transmission margin value of Path-A is 950 kW.

[0033] See Figure 4 This is a power time-series comparison chart of the fault recovery phase in a hybrid switchgear distribution network, visually presenting the dynamic changes in photovoltaic (PV) power output, load demand, and net power flow under Path-A. From 08:00 to 08:15 and 09:00, PV output exceeds load demand, resulting in a positive net power, indicating a power surplus that can be supplied to other areas or used for energy storage. From 08:30 to 08:45, load demand exceeds PV output, resulting in a negative net power, indicating a power deficit that requires activation of backup paths or distributed generation. This time-series data is the core input for calculating the remaining transmission margin, and the deficit period from 08:30 to 08:45 is a key constraint for path optimization. Combined with Path-A's rated capacity of 1000kW, the calculated remaining margin for this period is 950kW, which is the minimum value for the final path margin. This clearly demonstrates the dynamic game between power supply and demand during fault recovery, providing an intuitive decision-making basis for switch action timing and distributed generation scheduling.

[0034] In one embodiment of the present invention, the set of available recovery paths is sorted from high to low according to the remaining transmission margin value. The set of available recovery paths includes path numbers R1, R2, and R3, with remaining transmission margin values ​​of 950 kW, 850 kW, and 700 kW, respectively, and the sorted order is R1, R2, R3. The highest-ranked available recovery path R1 is selected as the main recovery path, and the switching action command corresponding to the main recovery path R1 is sent to the underlying control device. The switching action command includes a list of switch numbers to be closed and the operation sequence. The voltage recovery status at the end of the main recovery path R1 is monitored within a preset waiting time of 180 seconds. The voltage recovery status at the end is collected by a voltage transformer installed at the monitoring point, which is located at the target load node LN-12 of the main recovery path R1. The specific process of monitoring the voltage recovery status at the end includes: at the target load node LN-12 of the main recovery path R1, a lower voltage threshold and an upper voltage threshold are set. The lower voltage threshold is 0.9 times the nominal voltage, and the upper voltage threshold is 1.1 times the nominal voltage. The system receives voltage samples from the target load node LN-12 in real time and calculates the standard deviation of the voltage samples over a period of time to assess voltage stability. The calculation uses the following formula:

[0035] in: This represents the standard deviation of the voltage sample value sequence. This indicates the number of sampling points within the calculation window. Indicates the first Each voltage sample value, This represents the average of all voltage samples within the calculation window. If the voltage sample value remains between the lower and upper limits of the voltage threshold for multiple consecutive periods, and the standard deviation is... If the voltage fluctuation is less than the preset fluctuation threshold, the voltage recovery is considered successful. Multiple consecutive cycles are set to 10 power frequency cycles, and the preset fluctuation threshold is set to 2% of the nominal voltage. If the voltage sample value drops below the lower limit of the voltage threshold, or if severe oscillations occur causing a deviation in the standard deviation... If the limit is exceeded, the restoration is deemed a failure. The result is fed back to the path selection module to control the subsequent backup path selection logic. The result is transmitted in Boolean form. If a stable voltage restoration signal is not detected within the preset waiting time of 180 seconds, the next available path is selected as the backup path, for example, R2 is selected as the backup path, and a switching operation is performed. The switching operation includes canceling the switch closing command for path R1 and initiating the switch action command for path R2. The monitoring and switching process is repeated until all load nodes in the power loss area are restored to power.

[0036] In some embodiments, the switching action command is issued to the field intelligent terminal in the standard IEC61850 GOOSE message format. It can be understood that the sorting operation is based on descending order of the final remaining transmission margin value field recorded in the available recovery path set data table. Optionally, the data comparison shows that in the first recovery attempt, the terminal voltage sample value of the main recovery path R1 remains at 0.92 times the nominal voltage for 150 seconds after startup, with a standard deviation of [missing information]. The calculated value is 0.008 per unit, which is below the fluctuation threshold of 0.02, indicating successful voltage recovery and the process terminates. In another data comparison, the voltage sample value at the end of the main recovery path R1 drops to 0.85 times the nominal voltage 100 seconds after startup, below the lower threshold of 0.9, indicating recovery failure, and the system automatically switches to the backup path R2. In some embodiments, the voltage sample value reception period is 20 milliseconds, and the standard deviation is calculated. window This includes the most recent 100 sampling points. It can be understood that the path selection module, based on the received recovery failure determination result, removes the failed path R1 from the sorting list and reselects the path R2 with the highest remaining margin in the current list as the new primary recovery path. This repeated monitoring and switching process continues until the set of available recovery paths is empty or the power supply status flag of all load nodes displays "restored".

[0037] See Figure 5 This is a voltage recovery timing comparison chart during fault recovery in a hybrid switchgear distribution network. It clearly shows the dynamic voltage recovery process of the main recovery path R1, the failure scenario of the main path R1, and the backup path R2. It is a visual representation of the "path optimization + switching" logic in adaptive fault recovery. The main path R1 reaches a stable voltage threshold at 150 seconds, which is the "successful recovery" determination point. The main path voltage drops sharply between 200 and 250 seconds, triggering the backup path switching, which is the core of fault tolerance. At 300 seconds, the backup path recovers to a per-unit value of 0.92, achieving stable power supply to the full load and verifying the effectiveness of the multi-path redundancy design. By comparing the voltage curves of successful / failed / switching scenarios, the time point of main path failure and the switching delay of the backup path can be located, providing data support for optimizing switch action timing and improving recovery speed.

[0038] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A fault adaptive recovery method for hybrid switchgear distribution networks under power outage mode, characterized in that, Includes the following steps: Data on the topology of the hybrid switchgear distribution network under power outage conditions, the operating parameters of distributed power sources, and the real-time power demand of load nodes are collected. The collected data is formatted and time-stamped to form an initial operating dataset under power outage scenarios. Based on the initial running dataset, identify the tie switches and sectionalizing switches that are currently disconnected in the network, and classify them into mechanical switch groups and power electronic switch groups according to the physical type of the switches. Based on the differences in the operating characteristics of the mechanical switch group and the power electronic switch group, a set of candidate reconstruction paths is constructed, which includes switching operation delay and transmission capacity constraints. For each path in the candidate reconstruction path set, a radial topology check is performed to eliminate paths that do not meet the unidirectional power supply constraint and retain available recovery paths that meet the operation rules. Based on the load distribution of the available recovery paths, the remaining transmission margin of each path is calculated, and the remaining transmission margin is used as the basis for subsequent path selection.

2. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 1, characterized in that, The process of unifying the format and aligning the time stamps of the collected data to form an initial operational dataset for a power outage scenario specifically includes: Read telemetry data and teleindication data uploaded by the smart terminal installed on the power distribution line. The telemetry data includes voltage amplitude and current amplitude, and the teleindication data includes the opening and closing status of the switch. Map the data fields reported by smart terminals from different manufacturers to a unified internal data dictionary, and perform unit conversion for parameters with inconsistent dimensions. Based on the timing signal from the Global Positioning System, the telemetry and teleindication data are timestamped and then processed by moving average within a time window to filter out transient interference. The processed data is indexed by node number and associated with a pre-stored static topology map of the distribution network to generate an initial operational dataset for a power outage scenario that includes spatial location information.

3. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 2, characterized in that, Based on the differences in operating characteristics between the mechanical switch group and the power electronic switch group, a set of candidate reconfiguration paths is constructed, including switching operation delay and transmission capacity constraints, specifically including: The state of the mechanical switch group is extracted from the initial running dataset. The mechanical switch group has a minute-level action delay and the operation is irreversible. The state of the power electronic switch group is extracted. The power electronic switch group has a response speed of seconds or less and supports multiple continuous adjustments. Starting from the power outage area, all power supply paths are searched in the adjacency matrix of the distribution network. The power supply paths are composed of alternating connections of the mechanical switch group and the power electronic switch group. A path attribute label is set for each power supply path. The path attribute label records the number of mechanical switches, the number of power electronic switches, and the theoretical maximum current carrying capacity of the entire path. The cumulative action delay value in the path attribute label is compared with the transmission capacity constraint. Paths that have timed out or overloaded are eliminated, and the remaining paths constitute a set of candidate reconstruction paths.

4. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 3, characterized in that, For each path in the candidate reconstruction path set, a radial topology check is performed, specifically including: Traverse each path in the candidate reconstruction path set and perform a virtual closure operation on the switch states involved in the path in memory; The graph theory algorithm is called to detect whether there is a loop in the network structure after the closure operation. If a loop exists, the path is determined not to satisfy the radial topology constraint. Simultaneously, it detects whether there are isolated nodes in the network after the closure operation. An isolated node refers to a node that is connected to the power supply side but has no load connected. If a loop or isolated node is detected, the virtual closure operation is rolled back in memory, and the path is removed from the candidate reconstruction path set. Mark all paths that have not triggered a rollback operation as available recovery paths and record their connection relationships in the topology.

5. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 4, characterized in that, Based on the load distribution of the available recovery paths, the remaining transmission margin of each path is calculated, specifically including: Retrieve the output prediction curves of the distributed power sources connected to the currently available recovery paths, where the output prediction curves cover the entire fault recovery process. The historical load curves of the load nodes at the end of the available recovery path are read and corrected in conjunction with the current meteorological data to obtain the actual load power demand. The power output prediction curve and the load power demand are superimposed on the time axis to calculate the actual power flow distribution of the available recovery path at each time. The difference between the actual power flow distribution and the physical rated capacity of the available recovery path is the remaining transmission margin at the current moment. The minimum remaining transmission margin at all times is taken as the final remaining transmission margin value of the available recovery path.

6. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 5, characterized in that, It also includes the steps of path optimization and switching based on the remaining transmission margin: Sort the set of available recovery paths from highest to lowest based on the remaining transmission margin value; Select the highest-ranked available recovery path as the primary recovery path and send its corresponding switch action command to the underlying control device. The voltage recovery status at the end of the main recovery path is monitored within a preset waiting time, and the voltage recovery status at the end is collected by a voltage transformer installed at the monitoring point. If a stable voltage recovery signal is not detected within the waiting period, the next available recovery path is selected as the backup path, and a switching operation is performed. Repeat the monitoring and switching process until all load nodes in the power-loss area are restored to power.

7. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 6, characterized in that, The monitoring of the end voltage recovery status of the main recovery path within a preset waiting time specifically includes: At the target load node of the main recovery path, a lower voltage threshold and an upper voltage threshold are set; The voltage sample values ​​from the target load node are received in real time, and their standard deviation over a period of time is calculated to evaluate the stability of the voltage. If the voltage sample value remains between the lower and upper limits of the voltage threshold for multiple consecutive cycles, and the standard deviation is less than the preset fluctuation threshold, then the voltage recovery is considered successful. If the voltage sample value drops below the lower limit of the voltage threshold, or if severe oscillations cause the standard deviation to exceed the limit, the recovery is deemed to have failed. The determination result is fed back to the path optimization module to control the subsequent backup path selection logic.

8. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 7, characterized in that, The process of setting path attribute labels for each power supply path also includes a quantitative assessment of the support capability of distributed power sources: Extract the type of the distributed power source and distinguish whether it is a photovoltaic power source, a wind power source, or an energy storage power source; Based on the type of the distributed power source, read its maximum active power output under the current light intensity or wind speed conditions; Based on the state of charge of the energy storage power source, estimate the maximum support duration it can provide during fault recovery; The upper limit of active power output and the maximum support duration are converted into an equivalent capacity support coefficient, and the equivalent capacity support coefficient is written into the path attribute tag. When calculating the remaining transmission margin, the equivalent capacity support coefficient is superimposed as a gain factor onto the output prediction curve of the distributed power source.

9. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 8, characterized in that, The power output prediction curve and the load power demand are superimposed on the time axis to calculate the actual power flow distribution of the available recovery path at each time point, specifically including: The output prediction curve is discretized into an output sequence with the same time step, and the load power demand is discretized into a load sequence. Subtract the values ​​at corresponding time points from the output sequence to obtain the net power flow sequence; Substitute the net power flow sequence into the admittance matrix of the available recovery path, solve the node voltage equations, and obtain the voltage phase angle of each node; The active power loss of the line is calculated based on the voltage phase angle of the node, and the active power loss is subtracted from the net power flow sequence to correct and obtain the actual power flow distribution. The actual power flow distribution data is stored in the power flow calculation result cache area for use in the calculation step of the remaining transmission margin.

10. The fault adaptive recovery method for hybrid switchgear distribution network under power outage mode according to claim 9, characterized in that, Use graph theory algorithms to detect whether there are loops in the network structure after a closure operation, including: The network structure after the virtual closure operation is read from memory. The network structure consists of a node list and an edge list, and the edge list corresponds to the lines connected by the closed switch. The node list and edge list are converted into an adjacency matrix representation, where non-zero elements indicate an electrical connection between two corresponding nodes. In the topological graph represented by the adjacency matrix, any node is selected as the starting node for depth-first traversal, and the traversal begins from the starting node. During the depth-first traversal, a list of visited nodes is maintained to record the nodes that have been visited. When a new node is visited, the new node is marked as visited. When visiting all neighboring nodes of the new node, check whether the neighboring nodes of the new node already exist in the list of visited nodes and whether the neighboring nodes are not the direct parent nodes of the current traversal. If the conditions are met, it is determined that there is a loop in the network structure and all nodes contained in the loop are recorded.

Images