Electric power facility emergency processing system based on earthquake early warning

Abstract

The invention relates to the technical field of electric power emergency control, and discloses an electric power facility emergency processing system based on earthquake early warning, and the system comprises a region division module which is used for obtaining the data of an electric power topological structure, and generating a plurality of region sections; the data acquisition module is used for acquiring earthquake early warning data, equipment operation data and equipment environment data and representing the data through a multi-source data set; the fault risk assessment module is used for calculating the failure probability of each power device; the isolation decision and control module is used for calculating a region reliability index of the sub-region and determining a target sub-region; the off-grid operation regulation and control module is used for switching the target sub-region to off-grid operation and adjusting the control parameters; the operation state evaluation module is used for generating an operation health score; and the reconnection recovery module is used for recovering the connection of the target sub-region to the main power grid. According to the invention, intelligent isolation and safe reconnection control of the power subareas under earthquake early warning driving are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of power emergency control, and particularly relates to a power facility emergency handling system based on earthquake early warning. Background Art

[0002] In areas where natural disasters such as earthquakes occur frequently, the power system faces severe challenges in operation stability and emergency handling. The mechanical vibrations, equipment damage, and communication interruptions caused by earthquakes may lead to the expansion of power facility failures, local frequency fluctuations, and voltage instability, seriously affecting the continuity of power supply. In recent years, in order to improve the earthquake response efficiency, earthquake early warning systems have begun to attempt to integrate with the power system to sense the magnitude, epicenter location, and wave propagation characteristics in advance, providing an opportunity for power facilities to respond in advance.

[0003] Although some existing technologies have introduced earthquake early warning information to trigger tripping protection, the current power facility emergency handling methods still have deficiencies: First, there is a lack of multi-source data fusion, and the electrical and environmental states of each region cannot be comprehensively characterized; second, the judgment of the regional island operation ability is insufficient, and the isolation strategy lacks adaptability; third, the reclosing condition judgment mechanism is imperfect, and the stability of the restoration process cannot be guaranteed. Summary of the Invention

[0004] The present invention provides a power facility emergency handling system based on earthquake early warning, which solves the technical problems of untimely emergency response, inaccurate isolation strategy, and unstable reclosing process in related technologies.

[0005] The present invention provides a power facility emergency handling system based on earthquake early warning, including:

[0006] A regional division module, configured to obtain power topology structure data, generate multiple regional sections through the multi - minimum cut algorithm, and the regional sections are used to divide the first power supply area into M sub - areas;

[0007] A data acquisition module, configured to collect earthquake early warning data, equipment operation data, and equipment environment data in a first preset time period at a preset time interval, and represent them by a multi - source data set;

[0008] A fault risk assessment module, configured to calculate the failure probability of each power equipment based on the multi - source data set and a logistic regression model;

[0009] An isolation decision and control module, configured to calculate the regional reliability index of the sub - areas isolated by each regional section according to the regional section and the failure probability of the power equipment, select the regional section with the largest regional reliability index as the optimal regional section, determine the corresponding sub - area as the target sub - area, and send a tripping command to the target sub - area;

[0010] The off-grid operation control module is used to switch the target sub-region to off-grid operation and adjust the control parameters of power equipment in the target sub-region according to a preset ratio;

[0011] The operation status evaluation module is used to calculate the standard deviation of frequency deviation, the ratio of over-voltage duration, and the load power volatility based on the equipment operation data in the multi-source dataset of the target sub-region, and generate the operation health score of the target sub-region through linear weighting;

[0012] The reconnection recovery module is used to determine whether the operation health score of the target sub-region reaches the first preset reconnection threshold, and when the operation health score is greater than the first preset reconnection threshold, reconnect the target sub-region to the main power grid.

[0013] Furthermore, the earthquake early warning data includes: magnitude, epicenter location, wave velocity, P-wave arrival time; the equipment operation data includes: voltage, current, frequency, phase angle, and active power; the equipment environment data includes: equipment location, installation method, seismic resistance level, and load level.

[0014] Furthermore, the power topology structure data includes: substation number, feeder branch point number, microgrid access point number, circuit breaker number, and circuit breaker status;

[0015] The area division module is used to pre-divide the first power supply area in the pre-earthquake stage. The specific steps include:

[0016] S201, construct a power undirected graph according to the power topology structure data. Among them, the power undirected graph includes nodes and edges. The nodes are represented by substation numbers, feeder branch point numbers, and microgrid access point numbers, and the edges are represented by the line connection relationship between the nodes. The line connection relationship is determined by the line number;

[0017] S202, perform weighted calculation on each edge according to the preset line load capacity and line length to obtain the weight value of the edge;

[0018] S203, use the minimum cut algorithm on the power undirected graph. Each time, cut a group of edges with the smallest total weight value from the current graph to generate an area section, and repeat this operation on the remaining part of the power undirected graph until M predetermined sub-regions are obtained.

[0019] Furthermore, the determination process of the failure probability is as follows:

[0020] Extract all the characteristic values corresponding to each power equipment from the multi-source dataset, and the number of characteristics is fixed at n;

[0021] Perform a weighted process on each eigenvalue, multiply the first eigenvalue by the first pre-trained regression coefficient, the second eigenvalue by the second regression coefficient, until the nth eigenvalue is multiplied by the nth regression coefficient, add up all the product results, and superimpose a pre-trained first bias parameter on the sum to obtain a weighted sum result;

[0022] Take the negative value of the weighted sum result and calculate the natural exponential function value of the negative value;

[0023] Process the natural exponential function value, then take the reciprocal, and finally output a device failure probability value between 0 and 1.

[0024] Further, the determination process of the regional reliability index is as follows:

[0025] According to the selected regional section, identify all b power devices contained in the sub-region isolated by the regional section; for the first device in the sub-region: process the failure probability of this device to obtain its reliability; repeat this operation and calculate the reliability of the second to the bth devices in turn;

[0026] Add up the reliabilities of all b devices to obtain the total reliability;

[0027] Divide the total reliability by the total number of devices b and output the regional reliability index of this sub-region.

[0028] Further, the control parameters include: frequency-power droop coefficient and voltage-excitation droop coefficient. Among them, the frequency-power droop coefficient is calculated by the ratio between the preset frequency offset and the rated power of the power device, and the voltage-excitation droop coefficient is calculated by the ratio between the preset voltage offset and the change amount of the maximum excitation current of the power device.

[0029] Further, the control parameters are adjusted by the multiplication scaling method with the first preset adjustment ratio and the second preset adjustment ratio for the frequency-power droop coefficient and the voltage-excitation droop coefficient respectively.

[0030] Further, calculate the standard deviation of the frequency deviation by squaring the difference between the frequency of each sampling point and the average frequency and then taking the square root of the average of all results;

[0031] Calculate the voltage overrun duration ratio by counting the cumulative time when the voltage exceeds the preset voltage threshold and taking the ratio with the total duration of the first preset time period;

[0032] Calculate the load power volatility by squaring the difference in active power between adjacent sampling points and then taking the square root of the average of all results.

[0033] Further, after the reclosing recovery module reaches the first preset reclosing threshold in terms of the operation health score, it is further configured to determine whether the frequency deviation between the main power grid and the target sub-region is lower than the second preset frequency deviation threshold, and whether the phase angle deviation is lower than the third preset phase angle deviation threshold. If both conditions are met, the target sub-region is restored to be connected to the main power grid; wherein, the frequency deviation is the difference between the main power grid frequency and the target sub-region frequency, and the phase angle deviation is the difference between the main power grid phase angle and the target sub-region phase angle.

[0034] The beneficial effects of the present invention are as follows: The present invention integrates earthquake early warning data, equipment operation data, and equipment environment data to construct a multi-source data set, realizing comprehensive perception and dynamic analysis of pre-earthquake risks. By constructing an undirected power graph and executing the minimum cut algorithm, a structured division of sub-regions is achieved, and regions with high operation reliability are preferentially identified for pre-isolation, enhancing the stability of the island and the resilience of the system. During operation, the present invention evaluates the sub-region state through the operation health score and combines the frequency and phase angle deviation criteria to intelligently determine whether the grid connection condition is met, realizing adaptive and safe reclosing control. The overall solution supports the closed-loop regulation of "prediction - isolation - operation - reclosing", which can significantly improve the rapid response ability and post-disaster recovery efficiency of the distribution system in the earthquake scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the modules of the power facility emergency handling system based on earthquake early warning of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] Now, the subject matter described herein will be discussed with reference to example embodiments. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the protection scope of the content of this specification, changes can be made to the functions and arrangements of the elements discussed. Each example can omit, substitute, or add various processes or components as needed. Additionally, the features described relative to some examples can also be combined in other examples.

[0037] As Figure 1 shown, the power facility emergency handling system based on earthquake early warning includes:

[0038] A regional division module 101, configured to obtain power topology structure data and generate multiple regional sections through the minimum cut algorithm multiple times, where the regional sections are used to divide the first power supply area into M sub-regions;

[0039] A data acquisition module 102, configured to collect earthquake early warning data, equipment operation data, and equipment environment data of the first preset time period at a preset time interval and represent them through a multi-source data set;

[0040] A fault risk assessment module 103, which is used to calculate the failure probability of each power device based on a logistic regression model according to a multi-source data set;

[0041] An isolation decision and control module 104, which is used to calculate the regional reliability index of the sub-regions isolated by each regional section according to the regional section and the failure probability of the power device, select the regional section with the largest regional reliability index as the optimal regional section, and determine the corresponding sub-region as the target sub-region, and send a tripping command to the target sub-region;

[0042] An off-grid operation regulation module 105, which is used to switch the target sub-region to off-grid operation and adjust the control parameters of the power devices in the target sub-region according to a preset ratio;

[0043] An operation status assessment module 106, which is used to calculate the standard deviation of frequency deviation, the ratio of voltage over-limit duration, and the load power volatility according to the device operation data in the multi-source data set of the target sub-region, and generate the operation health score of the target sub-region through linear weighting;

[0044] A reconnecting and restoring module 107, which is used to judge whether the operation health score of the target sub-region reaches a first preset reconnecting threshold, and when the operation health score is greater than the first preset reconnecting threshold, reconnect the target sub-region to the main power grid.

[0045] In an embodiment of the present invention, the earthquake early warning data includes: magnitude, epicenter location, wave velocity, P-wave arrival time; the device operation data includes: voltage, current, frequency, phase angle, and active power; the device environment data includes: device location, installation method, seismic resistance level, and load level.

[0046] Specifically, the earthquake early warning data is obtained through the data interface accessing the China Earthquake Networks Center, the device operation data is collected by the smart meters installed on-site, the phase angle is represented in the form of voltage phase angle, and the device environment data is obtained through the data center database; the unit of magnitude is level, the epicenter location is represented by longitude and latitude coordinates, the unit of wave velocity is m / s, and the unit of P-wave arrival time is s.

[0047] In an embodiment of the present invention, in order to ensure that various indicators in the multi-source data set can be effectively fused and calculated, the earthquake early warning data, the device operation data, and the device environment data are normalized by the maximum-minimum normalization method.

[0048] In an embodiment of the present invention, the regional section represents a trippable circuit breaker, and after being cut off, the first power supply area can be divided into two non-connected sub-regions.

[0049] In one embodiment of the present invention, the power topology structure data includes: substation number, feeder branch point number, microgrid access point number, circuit breaker number, and circuit breaker status; among them, the substation number is used to identify each substation node in the area; the feeder branch point number is used to describe the connection nodes and branch positions in the feeder topology; the microgrid access point number is used to identify the access positions of various distributed energy sources; the circuit breaker number and circuit breaker status are used to represent the connection status between nodes, and the circuit breaker status includes connected and disconnected; the power topology structure data is obtained through accessing the distribution management system.

[0050] The area division module is used to pre-divide the first power supply area in the pre-earthquake stage, and the specific steps include:

[0051] S201, construct a power undirected graph according to the power topology structure data, where the power undirected graph includes nodes and edges. The nodes are represented by substation numbers, feeder branch point numbers, and microgrid access point numbers, and the edges are represented by the line connection relationship between nodes. The line connection relationship is determined by the line number. The edges are used to reflect the physical connectivity between devices, and each edge corresponds to a physical line; the power undirected graph is organized and stored in a graph database structure to improve the access efficiency;

[0052] S202, perform weighted calculation on each edge according to the preset line load capacity and line length to obtain the weight value of the edge. The calculation formula of the weight value is: , where represents the weight value of the edge for subsequent area division, represents the preset line load capacity, d represents the line length, and respectively represent the first weight coefficient and the second weight coefficient;

[0053] S203, use the minimum cut algorithm on the power undirected graph. Each time, cut a group of edges with the smallest total weight value from the current graph to generate an area section, and repeat this operation on the remaining part of the power undirected graph until M predetermined sub-areas are obtained;

[0054] Specifically, based on the weight value of the edge calculated in step S202, use the minimum cut algorithm to identify a set of edges with the smallest total weight value in the current power undirected graph, and use the connected branch separated by this set as a new sub-area, and its boundary is the area section. After completing one minimum cut operation, in this embodiment, the nodes and associated edges corresponding to this sub-area are removed from the original power undirected graph, and the minimum cut operation is repeated on the remaining graph until the power undirected graph is divided into M predetermined sub-areas; through this method, it is possible to achieve the target sub-areas with good boundary characteristics and controllable division without overly affecting the connectivity of the main power grid.

[0055] In one embodiment of the present invention, the process of determining the failure probability is as follows:

[0056] Extract all the characteristic values corresponding to each power device from the multi-source dataset, and the number of characteristics is fixed at n;

[0057] Perform weighted processing on each characteristic value, multiply the first characteristic value by the first pre-trained regression coefficient, the second characteristic value by the second regression coefficient, until the nth characteristic value is multiplied by the nth regression coefficient, add up all the product results, and superimpose a pre-trained first bias parameter on the sum to obtain the weighted sum result;

[0058] Take the negative value of the weighted sum result and calculate the natural exponential function value of the negative value;

[0059] Process the natural exponential function value, take the reciprocal, and finally output the device failure probability value between 0 and 1.

[0060] When specifically applied, the above formula for calculating the failure probability can be implemented by the following formula, for example: ;

[0061] Where P represents the failure probability, which is used to evaluate the possibility of a power device having a functional failure under seismic disturbance, exp represents the exponential function, represents the first bias parameter, n represents the number of characteristics in the multi-source dataset, in this embodiment, n is 12, j represents the index of the characteristic, represents the value of the jth characteristic in the multi-source dataset, represents the jth regression coefficient, and the regression coefficient is obtained through offline training.

[0062] In one embodiment of the present invention, the process of determining the regional reliability index is as follows:

[0063] According to the selected regional section, identify all b power devices included in the sub-region isolated by the regional section; for the first device in the sub-region: process the failure probability of this device to obtain its reliability; repeat this operation, and calculate the reliability of the second to bth devices in turn;

[0064] Add up the reliabilities of all b devices to obtain the total reliability;

[0065] Divide the total reliability by the total number of devices b and output the regional reliability index of this sub-region.

[0066] When specifically applied, the above formula for calculating the regional reliability index can be implemented by the following formula, for example: ;

[0067] Among them, represents the regional reliability index of the sub-region isolated by the regional section, b represents the number of power equipment in the sub-region, and i represents the index of the power equipment. represents the failure probability of the i-th power equipment; the regional reliability index is used to quantify the overall operation reliability of the power equipment in the sub-region under seismic disturbances, facilitating the preferential selection of sub-regions with high operation stability for isolation, ensuring that the isolated sub-regions have strong off-grid operation capabilities and self-recovery capabilities, so as to maintain local power supply stability and prevent faults from spreading to stable regions when large-scale disturbances or instability occur in the main power grid.

[0068] According to the calculated regional reliability indices of the sub-regions isolated by each regional section, select the regional section with the largest regional reliability index as the optimal regional section, and use the sub-region isolated by it as the target sub-region; further determine the boundary circuit breaker between the target sub-region and the main power grid, and issue a tripping command to the boundary circuit breaker, thereby achieving the electrical isolation between the target sub-region and the main power grid.

[0069] In an embodiment of the present invention, the control parameters include: frequency-power droop coefficient and voltage-excitation droop coefficient. Among them, the frequency-power droop coefficient is calculated by the ratio between the preset frequency offset and the rated power of the power equipment, and the voltage-excitation droop coefficient is calculated by the ratio between the preset voltage offset and the change amount of the maximum excitation current of the power equipment.

[0070] The frequency offset is the maximum allowable frequency offset value of the target sub-region in the off-grid operation state, with the unit of Hz, and the rated power is the rated active power output of the power equipment in the normal operation state, with the unit of kW; the voltage offset is the allowable bus voltage fluctuation range, with the unit of V, and the change amount of the maximum excitation current is the maximum value within the voltage offset range of the power equipment, with the unit of A.

[0071] After the target sub-region is disconnected from the main power grid and enters off-grid operation, its frequency and voltage stability will no longer depend on the main grid inertia and reactive power support. Therefore, it is necessary to improve the control sensitivity of local equipment to make it have a faster frequency-voltage response ability; that is, the control parameters of the power equipment in the target sub-region are adjusted by the multiplication scaling method; the frequency-power droop coefficient is used to represent the response sensitivity of the power equipment to frequency deviation in the off-grid operation state, and the voltage-excitation droop coefficient is used to control the response degree of the synchronous motor or reactive power regulation device to voltage deviation.

[0072] The control parameters are adjusted for the frequency-power droop coefficient and the voltage-excitation droop coefficient respectively by means of multiplicative scaling using a first preset adjustment ratio and a second preset adjustment ratio; specifically, the frequency-power droop coefficient and the voltage-excitation droop coefficient are multiplied by the first preset adjustment ratio and the second preset adjustment ratio respectively to obtain the adjusted control parameters; preferably, the first preset adjustment ratio is 0.7 and the second preset adjustment ratio is 0.75.

[0073] In an embodiment of the present invention, the standard deviation of frequency deviation is calculated by squaring the difference between the frequency of each sampling point and the average frequency, and then taking the square root of the average of all results, which is used to represent the degree of frequency fluctuation within the target sub-region. The calculation formula for the standard deviation of frequency deviation is as follows: , where represents the standard deviation of frequency deviation, N represents the number of sampling points, k represents the sampling point index, represents the frequency of the k-th sampling point, represents the average value of the frequencies of each sampling point;

[0074] By statistically calculating the cumulative time when the voltage exceeds the preset voltage threshold and taking the ratio with the total duration of the first preset time period, the voltage overrun duration ratio is obtained, which is used to reflect the degree of long-term voltage instability. The calculation formula for the voltage overrun duration ratio is as follows: , where represents the voltage overrun duration ratio, represents the cumulative voltage overrun time, that is, the total time when the voltage exceeds the preset voltage threshold, represents the total duration of the first preset time period;

[0075] By squaring the difference in active power between adjacent sampling points and then taking the square root of the average of all results, the load power volatility is obtained, which is used to measure the degree of change in the active load. The calculation formula for the load power volatility is as follows: , where represents the load power volatility, o represents the sampling point index, and respectively represent the active power of the (o + 1)-th and the o-th sampling points.

[0076] The operating health score is obtained by linearly weighting the standard deviation of frequency deviation, the voltage overrun duration ratio, and the load power volatility.

[0077] In an embodiment of the present invention, after the reclosing recovery module reaches the first preset reclosing threshold in terms of the running health score, it is further configured to determine whether the frequency deviation between the main power grid and the target sub-region is lower than the second preset frequency deviation threshold, and whether the phase angle deviation is lower than the third preset phase angle deviation threshold. If both conditions are met, the target sub-region is restored to be connected to the main power grid by controlling the circuit breaker to close; wherein, the frequency deviation is the difference between the main power grid frequency and the target sub-region frequency, and the phase angle deviation is the difference between the main power grid phase angle and the target sub-region phase angle.

[0078] In an embodiment of the present invention, when the running health score of the target sub-region does not reach the first preset reclosing threshold, this embodiment determines that the target sub-region currently does not have the operating conditions for safe grid connection, and the reclosing recovery module does not perform the grid connection operation. At this time, the target sub-region will continue to maintain the offline operating state. This embodiment sets a re-evaluation period, for example, recalculating the running health score every 10 seconds until the running health score reaches the first preset reclosing threshold or other recovery conditions are met;

[0079] On the premise of reaching the first preset reclosing threshold, if the frequency deviation and phase angle deviation between the main power grid and the target sub-region do not meet the requirements of being lower than the second preset frequency deviation threshold and the third preset phase angle deviation threshold, the reclosing operation will be suspended and the reclosing waiting state will be entered; similarly, the system can set a monitoring duration threshold and a maximum waiting number, for example, continuously monitoring for 30 seconds and waiting at most 5 times; during this period, the system keeps real-time monitoring of the frequencies and phase angles of the main power grid and the sub-region. Once it is detected that the deviation enters the allowable range, the grid connection instruction is triggered. If the criterion is still not met after exceeding the maximum waiting number, an alarm can be generated and transferred to the manual intervention mode or the reclosing plan can be postponed to ensure that the overall operation safety of the system is not affected.

[0080] In an embodiment of the present invention, by introducing the running health score and the frequency and phase angle deviations as the combined reclosing criterion, the stability and safety of the reclosing of the target sub-region are improved.

[0081] It should be noted that the setting of the interval and threshold values is for the convenience of comparison. Among them, the size of the threshold depends on the amount of sample data and the base quantity set by those skilled in the art for each group of sample data, as long as the proportional relationship between the parameters and the quantified values is not affected. And the above formulas are all calculations of taking the numerical value without dimension. The formulas are all obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0082] The embodiments of the present invention have been described above, but the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative rather than restrictive. Under the inspiration of this embodiment, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this embodiment.

Claims

1. An emergency handling system for power facilities based on earthquake early warning, characterized in that, Including: A regional division module, configured to obtain power topology structure data, generate multiple regional cross-sections through the minimum cut algorithm multiple times, and the regional cross-sections are used to divide the first power supply area into M sub-areas; A data acquisition module, configured to collect earthquake early warning data, equipment operation data, and equipment environment data in a first preset time period at a preset time interval, and represent them through a multi-source data set; A fault risk assessment module, configured to calculate the failure probability of each power equipment based on the multi-source data set and a logistic regression model; An isolation decision and control module, configured to calculate the regional reliability index of the sub-areas isolated by each regional cross-section according to the regional cross-section and the failure probability of the power equipment, select the regional cross-section with the largest regional reliability index as the optimal regional cross-section, and determine the corresponding sub-area as the target sub-area, and send a tripping command to the target sub-area; An off-grid operation regulation module, configured to switch the target sub-area to off-grid operation and adjust the control parameters of the power equipment in the target sub-area according to a preset ratio; An operation status assessment module, configured to calculate the standard deviation of frequency deviation, the ratio of voltage over-limit duration, and the load power volatility according to the equipment operation data in the multi-source data set of the target sub-area, and generate the operation health score of the target sub-area through linear weighting; A reconnecting and restoring module, configured to determine whether the operation health score of the target sub-area reaches a first preset reconnecting threshold, and when the operation health score is greater than the first preset reconnecting threshold, reconnect the target sub-area to the main power grid.

2. The emergency treatment system for power facilities based on earthquake early warning according to claim 1, wherein The earthquake early warning data includes: magnitude, epicenter location, wave velocity, P-wave arrival time; the equipment operation data includes: voltage, current, frequency, phase angle, and active power; the equipment environment data includes: equipment location, installation method, seismic resistance level, and load level.

3. The emergency treatment system for power facilities based on earthquake early warning according to claim 1, wherein The power topology structure data includes: substation number, feeder branch point number, microgrid access point number, circuit breaker number, and circuit breaker status; The regional division module is used to pre-divide the first power supply area in the pre-earthquake stage, and the specific steps include: S201, construct a power undirected graph according to the power topology structure data, where the power undirected graph includes nodes and edges, the nodes are represented by substation numbers, feeder branch point numbers, and microgrid access point numbers, the edges are represented by the line connection relationship between the nodes, and the line connection relationship is determined by the line number; S202, perform weighted calculation on each edge according to the preset line load capacity and line length to obtain the weight value of the edge; 4. The emergency processing system for power facilities based on earthquake early warning according to claim 1, characterized in that, S203, use the minimum cut algorithm on the power undirected graph, each time cut a group of edges with the smallest total weight value from the current graph, generate a regional cross-section, and repeat this operation on the remaining part of the power undirected graph until M predetermined sub-areas are obtained. The determination process of the failure probability is: Extract all the characteristic values corresponding to each power equipment from the multi-source data set, and the number of characteristics is fixed at n; Weight each eigenvalue by multiplying the first eigenvalue by the first pre-trained regression coefficient, the second eigenvalue by the second regression coefficient, until the nth eigenvalue by the nth regression coefficient, sum all the product results, and superimpose a pre-trained first bias parameter on the sum to obtain a weighted sum result; Take the negative value of the weighted sum result and calculate the natural exponential function value of the negative value; Process the natural exponential function value, take the reciprocal, and finally output a device failure probability value between 0 and 1.

5. The emergency processing system for power facilities based on earthquake early warning according to claim 1, characterized in that, The determination process of the regional reliability index is as follows: According to the selected regional section, identify all b power devices included in the sub-region isolated by the regional section; for the first device in the sub-region: process the failure probability of the device to obtain its reliability; repeat this operation and calculate the reliability of the second to bth devices in turn; Sum the reliabilities of all b devices to obtain the total reliability; Divide the total reliability by the total number of devices b and output the regional reliability index of the sub-region.

6. The emergency processing system for power facilities based on earthquake early warning according to claim 1, wherein The control parameters include: frequency-power droop coefficient and voltage-excitation droop coefficient, where the frequency-power droop coefficient is calculated by the ratio between a preset frequency offset and the rated power of the power device, and the voltage-excitation droop coefficient is calculated by the ratio between a preset voltage offset and the change in the maximum excitation current of the power device.

7. The emergency processing system for power facilities based on earthquake early warning according to claim 1, characterized in that, The control parameters are adjusted by a multiplicative scaling method using a first preset adjustment ratio and a second preset adjustment ratio for the frequency-power droop coefficient and the voltage-excitation droop coefficient respectively.

8. The emergency processing system for power facilities based on earthquake early warning according to claim 1, wherein, Calculate the standard deviation of the frequency deviation by squaring the difference between the frequency at each sampling point and the average frequency and taking the square root after averaging all the results; Calculate the voltage overrun duration ratio by counting the cumulative time when the voltage exceeds a preset voltage threshold and taking the ratio with the total duration of a first preset time period; Calculate the load power volatility by squaring the difference in active power between adjacent sampling points and taking the square root after averaging all the results.

9. The emergency processing system for power facilities based on earthquake early warning according to claim 1, wherein, After the operation health score of the reclosing recovery module reaches a first preset reclosing threshold, it is also used to determine whether the frequency deviation between the main power grid and the target sub-region is lower than a second preset frequency deviation threshold and whether the phase angle deviation is lower than a third preset phase angle deviation threshold. If both are satisfied, the target sub-region is restored to be connected to the main power grid; where the frequency deviation is the difference between the main power grid frequency and the target sub-region frequency, and the phase angle deviation is the difference between the main power grid phase angle and the target sub-region phase angle.

Images