Device protection autonomous tour diagnosis method based on network diagram module

Abstract

This invention discloses an autonomous inspection method for equipment protection based on a distribution network diagram, belonging to the field of power equipment inspection technology. The method includes: obtaining an electrical coupling observation matrix and a state transition matrix based on historical electrical operation data and historical power grid topology; constructing a hidden state vector through the states of suspected changing nodes and suspected changing branches; using a hidden Markov model based on the state transition matrix and the electrical coupling observation matrix to obtain effective changing nodes and effective changing branches; reconstructing the power grid topology based on the effective changing nodes and effective changing branches; obtaining power grid equipment parameters and line parameters according to real-time power grid measurement data and the power grid topology to reconstruct a real-time distribution network diagram; and obtaining theoretical coordination protection constraint values ​​and theoretical protection settings across equipment, comparing them with actual coordination protection constraint values ​​and theoretical protection settings to obtain the inspection status of protection equipment. This method solves the technical problem that existing technologies struggle to improve the reliability and comprehensiveness of protection equipment inspection.

Description

Technical Field

[0001] This invention relates to the field of power equipment inspection technology, specifically to a method for autonomous inspection of equipment protection based on distribution network diagrams. Background Technology

[0002] The inspection of protection equipment directly affects the selectivity, speed, and reliability of protection actions under fault conditions. It is a core link in preventing power grid operation risks such as cascading trips, protection failures, and maloperations, and is of great significance to ensuring the safe and stable operation of the power grid. In existing technologies, the actual protection settings of field protection equipment are obtained by dispatching and maintenance personnel based on static power grid diagrams, set offline, and then sent to the field protection equipment for fixed operation. Current protection equipment inspection work is mainly carried out by inspecting these fixed actual protection settings. However, the distribution network topology is characterized by frequent changes, such as the addition or removal of equipment and the switching of switches, which occur routinely. The static power grid diagram updates at the dispatching master station have an inherent time delay, which can easily lead to inconsistencies with the actual wiring and actual operating conditions on site. Under these circumstances, using the actual protection settings obtained based on the static power grid diagrams as the basis for inspection will result in protection inspection results that are seriously out of sync with the actual operating state of the power grid. It is difficult to accurately assess the effective protection performance of protection equipment under the current power grid topology, to effectively prevent various protection anomaly risks, and thus to ensure the safety of power grid operation. Meanwhile, existing protection inspection technologies focus solely on numerical detection of protection settings, lacking inspection of cross-device protection coordination logic. When the power grid topology changes, situations often arise where the setting values ​​remain unchanged, but the coordination relationships have failed. In such cases, simply detecting setting values ​​is insufficient to uncover potential coordination issues between protection devices, thus failing to guarantee power grid operational safety. Therefore, improving the reliability and comprehensiveness of protection device inspections remains a challenging technical problem that current technologies struggle to solve. Summary of the Invention

[0003] To address the technical challenge of improving the reliability and comprehensiveness of protection equipment inspections using existing technologies, this invention provides an autonomous inspection method for equipment protection based on a distribution network diagram. This method constructs a hidden state vector by analyzing the states of suspected changing nodes and branches. Using a state transition matrix and an electrical coupling observation matrix, a Hidden Markov Model (HMM) is employed to obtain valid changing nodes and branches, thus reconstructing the power grid topology. Based on real-time power grid measurement data and the power grid topology, power grid equipment parameters and line parameters are obtained, resulting in a real-time distribution network diagram synchronized with the actual operating conditions of the power grid. Theoretical protection settings and theoretical cross-equipment coordination protection constraints are then obtained from the real-time distribution network diagram and compared with the actual protection settings and actual cross-equipment coordination protection constraints to obtain the inspection status of the protection equipment. This method solves the technical problem of improving the reliability and comprehensiveness of protection equipment inspections using existing technologies.

[0004] To address the aforementioned technical problems, this invention provides a method for autonomous equipment protection inspection based on a distribution network diagram, comprising the following steps: Based on real-time power grid measurement data, suspected change nodes and suspected change branches are identified, and electrical coupling observation matrix and state transition matrix are obtained based on historical electrical operation data and historical power grid topology. Hidden state vectors are constructed by the states of suspected changing nodes and suspected changing branches. Effective changing nodes and effective changing branches are obtained by using a hidden Markov model through the state transition matrix and the electrical coupling observation matrix. Reconstructing the power grid topology based on effective changing nodes and effective changing branches, obtaining power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology, and integrating the power grid topology, power grid equipment parameters and line parameters to form a real-time distribution network model; Based on the real-time distribution network diagram model, the theoretical coordination protection constraint values ​​of the protection devices in the power grid are obtained, and the actual coordination protection constraint values ​​of the protection devices are compared with the theoretical coordination protection constraint values ​​to obtain the coordination inspection results. The theoretical protection settings of the protection equipment are obtained based on the real-time distribution network diagram. The actual protection settings of the protection equipment are compared with the theoretical protection settings to obtain the setting inspection results. The inspection status of the protection equipment is obtained by combining the setting inspection results with the cooperative inspection results.

[0005] Preferably, the step of identifying suspected change nodes and suspected change branches based on real-time power grid measurement data includes: Based on the data representation status attributes, real-time power grid measurement data is divided into electrical connectivity data and electrical operation data. The electrical connectivity data is compared with the equipment connection data in the power grid ledger to identify the first suspected change node; The electrical operation data of the branch where the power grid node is located is combined with the electrical connectivity data of the power grid node for effective verification. The power grid node that fails the effective verification is designated as the second suspected change node. The first suspected change node and the second suspected change node constitute the suspected change node. Based on the electrical operation data of the power grid branch, determine whether the power grid branch has undergone a step change. The power grid branch that has undergone a step change and whose two ends are not suspected change nodes is regarded as a suspected change branch.

[0006] Preferably, the step of obtaining the electrical coupling observation matrix and state transition matrix based on historical electrical operation data and historical power grid topology includes: Obtain the mean and covariance matrix of historical electrical operation data corresponding to the topology state in the historical power grid topology; Conditional independence tests are performed on historical electrical operation data corresponding to topology states in the historical topology of the power grid to determine causal relationships. Based on the historical electrical operation data corresponding to topology states in the historical topology of the power grid, several historical operating conditions are identified, and the absolute deviation value of historical electrical operation data corresponding to each historical operating condition under the same topology state is obtained. Based on the absolute deviation value, the disturbance deviation corresponding to each historical operating condition under the same topology state is obtained, and the disturbance matrix of historical electrical operation data corresponding to topology states is obtained through the disturbance deviation. The final covariance matrix is ​​obtained based on the perturbation matrix, causal relationship and covariance matrix. The joint Gaussian distribution function is obtained based on the final covariance matrix and the mean. Typical historical electrical operation data in the historical electrical operation data are input into the joint Gaussian distribution function to obtain the conditional likelihood probability, and then the electrical coupling observation matrix is ​​obtained. State change records are extracted from the historical topology of the power grid and assigned to the corresponding operating conditions. The frequency of transitions between topology states in each operating condition is counted to obtain the state transition matrix corresponding to each operating condition.

[0007] Preferably, the step of constructing a hidden state vector through the suspected changing node states and suspected changing branch states, and obtaining the effective changing nodes and effective changing branches using a hidden Markov model through the state transition matrix and the electrical coupling observation matrix, includes: The current operating condition is determined based on real-time power grid measurement data. The target state transition matrix is ​​then selected from the state transition matrix based on the current operating condition. The real-time power grid measurement data is matched with typical historical electrical operation data to locate the corresponding area of ​​the electrical coupling observation matrix and obtain the observation condition probability. The optimal hidden state vector is derived using a hidden Markov model through the hidden state vector, the observation condition probability, and the target state transition matrix, thereby obtaining the effective change nodes and effective change branches.

[0008] Preferably, the reconstructing of the power grid topology based on effective changing nodes and effective changing branches includes: Based on the equipment connection relationships in the power grid ledger data, a basic topology framework is established. If the effective change node is a closing node, a connection between the effective change node and the corresponding electrical associated node is established in the basic topology framework. If the effective change node is a opening node, the connection between the effective change node and the corresponding electrical associated node is disconnected in the basic topology framework. If the effective change branch is an elimination branch, a connection between the nodes at both ends of the effective change branch is established in the basic topology framework. If the effective change branch is an addition branch, a node is added between the nodes at both ends of the effective change branch in the basic topology framework, and a connection is established between the nodes at both ends and the added node, thereby obtaining the power grid topology.

[0009] Preferably, the step of obtaining power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology includes: Starting from the power generation nodes of the power grid, power flow calculations are performed along the power grid topology to obtain predicted electrical operation data. The predicted electrical operation data is then matched with the actual electrical operation data. If the match is successful, the initial parameters of the power grid equipment and the initial parameters of the lines are used as the parameters of the power grid equipment and the lines, respectively. If the match is unsuccessful, anomaly areas are located based on the data deviation distribution characteristics between the predicted electrical operation data and the actual electrical operation data. The initial parameters of the power grid equipment and the initial parameters of the lines within the anomaly areas are iteratively corrected to obtain the parameters of the power grid equipment and the lines.

[0010] Preferably, the step of locating abnormal areas based on the data deviation distribution characteristics between the predicted electrical operation data and the actual electrical operation data includes: The data deviation between the predicted electrical operation data and the actual electrical operation data of the power grid equipment is obtained. The power grid equipment with the largest data deviation is identified as the core abnormal equipment. The power grid topology is traversed with the location of the core abnormal equipment as the center, and the abnormal area is formed with the power grid equipment whose data deviation is greater than the preset deviation as the boundary.

[0011] Preferably, the step of obtaining the cross-device theoretical coordination protection constraint values ​​of protection devices in the power grid based on the real-time distribution network diagram model, and comparing the actual cross-device coordination protection constraint values ​​of the protection devices with the cross-device theoretical coordination protection constraint values ​​to obtain the coordination inspection results includes: The final protection coordination pair is obtained based on the upstream and downstream connection relationship of the protection equipment in the real-time distribution network diagram, and the initial protection coordination pair is obtained through the equipment connection relationship in the power grid ledger data. The initial protection coordination pair and the final protection coordination pair are then matched across the entire domain. Perform single-domain matching between the initial protection pair that failed to match and the final protection pair to obtain the number of the final protection pair that were matched and the number of the initial protection pair; The theoretical coordination protection constraint value across devices is obtained based on the final number of protection coordination pairs and the preset time limit difference. If the theoretical coordination protection constraint value across devices is greater than or equal to the actual coordination protection constraint value of the matched initial protection coordination pair, it indicates that the coordination inspection is qualified; otherwise, it indicates that it is unqualified.

[0012] Preferably, the step of obtaining the theoretical protection settings of the protection device based on the real-time distribution network diagram includes: Based on the real-time distribution network model, the fault conditions of the power grid under the maximum and minimum operating modes are simulated respectively, so as to obtain the maximum fault current and minimum fault current of the fault point. The theoretical protection settings of the protection equipment for the main protection section are obtained based on the maximum and minimum fault currents. The theoretical protection settings of the protection equipment for the backup section of the entire line are obtained based on the setting matching relationship between the protection equipment and the minimum fault current. The maximum load current of the power grid is obtained based on the real-time distribution network diagram. The theoretical protection settings of the protection equipment for the remote backup section are obtained based on the maximum load current and the minimum fault current of the power grid.

[0013] Preferably, the step of comparing the actual protection settings of the protection equipment with the theoretical protection settings to obtain the setting inspection results, and combining the setting inspection results with the cooperative inspection results to obtain the protection equipment inspection status, includes: The theoretical protection settings of the protection equipment for the main protection section, the theoretical protection settings of the protection equipment for the entire backup section, and the theoretical protection settings of the protection equipment for the remote backup section are obtained respectively. The first, second, and third differences between these values ​​and the corresponding actual protection settings are then obtained. If the first, second, and third differences are all greater than the preset difference, the setting inspection is considered qualified; otherwise, the setting inspection is considered unqualified. When both the setpoint inspection result and the coordinated inspection result are qualified, it means that the protection equipment inspection is qualified; otherwise, it means that the protection equipment inspection is unqualified.

[0014] By adopting the above technical solution, the present invention has the following advantages: (1) A hidden state vector is constructed by the suspected changed node state and the suspected changed branch state. The effective changed nodes and effective changed branches are obtained by using the state transition matrix and the electrical coupling observation matrix to reconstruct the power grid topology. The power grid equipment parameters and line parameters are obtained based on the real-time power grid measurement data and the power grid topology, thereby obtaining a real-time distribution network model synchronized with the actual operating conditions of the power grid. The theoretical protection settings and cross-equipment theoretical coordination protection constraint values ​​are obtained based on the real-time distribution network model, and are compared with the actual protection settings and cross-equipment actual coordination protection constraint values ​​to obtain the inspection status of protection equipment. This solves the technical problem that the existing technology is difficult to improve the reliability and comprehensiveness of protection equipment inspection. (2) Specifically, considering that distribution network topology changes are implicit states that are difficult to observe quickly and easily, they can be indirectly identified by relying on time-series electrical measurement data. Therefore, this invention introduces a Hidden Markov Model (HMM) to realize the probabilistic inference of topology states. However, the direct application of the HMM to the identification of effective topology changes in distribution networks has obvious inherent defects: the model constructs a global implicit state vector based on the states of all nodes and branches in the entire network, which has problems such as redundant state dimensions, large computational overhead, and poor real-time performance of online inference. To this end, this invention first screens suspected change nodes and suspected change branches as candidate change objects based on real-time measurement data, and constructs an implicit state vector based only on the states of the candidate change objects, which greatly compresses the model dimension and improves the efficiency of topology identification. At the same time, the HMM uses a single fixed global state transition matrix, which is difficult to adapt to the differentiated topology jump patterns and time-series evolution characteristics under various operating conditions such as peak and valley of the distribution network, resulting in low state inference accuracy. This invention constructs a dedicated state transition matrix for each operating condition by statistically analyzing the historical jump frequency of topology states under different operating conditions, thereby accurately matching the real topology evolution patterns of different operating conditions of the power grid. Furthermore, the observation matrix in Hidden Markov Models strictly adheres to the observation independence assumption, assuming that various measurement data are independent of each other, ignoring the inherent coupling correlation between measurement data and the topological coupling constraints of switches and branches, which easily leads to topological misjudgments and omissions. This invention breaks the observation independence assumption by reconstructing the electrical coupling observation matrix, fully exploring the measurement coupling and topological correlation characteristics. Through lightweight hidden state modeling and dual-matrix collaborative optimization, this invention effectively overcomes the inherent defects of Hidden Markov Models, such as poor real-time performance, weak adaptability to operating conditions, insufficient utilization of coupling features, and low identification accuracy, significantly improving the accuracy of identifying effective changing nodes and effective changing branches in the power grid. (3) By reconstructing the real power grid topology and then correcting the parameters of power grid equipment and lines online, a real-time distribution network model synchronized with the field wiring and operating conditions is built. Theoretical protection settings and cross-equipment theoretical coordination protection constraints are obtained based on this synchronized power grid model. Theoretical protection settings and cross-equipment theoretical coordination protection constraints are used as objective evaluation benchmarks that conform to the actual operating conditions of the power grid. The effective protection performance of the protection equipment under the actual operating conditions of the power grid can be accurately determined by comparing the actual protection settings and cross-equipment actual coordination protection constraints with these benchmarks. By allowing the inspection evaluation benchmarks to be updated in real time with the dynamics of the power grid, the reliability of the inspection is fundamentally improved. This system provides protection settings reference for maintenance personnel, facilitating their review and distribution to various protection devices for optimization. It achieves comprehensive inspection through dual-dimensional verification: firstly, it compares the protection settings of the protection devices to identify explicit defects such as setting deviations; secondly, it compares the cross-device coordination logic to accurately identify implicit coordination problems caused by changes in the power grid topology, where the setting values ​​remain unchanged but the coordination relationship has failed. This breaks the limitations of traditional setting inspections that only target protection devices, upgrading from single-point setting inspections to systemic inspections. It effectively addresses the shortcomings of existing single-dimensional inspections and significantly improves the comprehensiveness of protection device inspections. Attached Figure Description

[0015] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0016] Figure 1 This is a flowchart illustrating the autonomous inspection method for equipment protection based on a distribution network diagram of the present invention. Figure 2 This is a schematic diagram illustrating the process of obtaining the electrical coupling observation matrix and the state transition matrix of the present invention. Detailed Implementation

[0017] 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 only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0019] Example 1: like Figure 1 As shown, the self-diagnosis method for equipment protection based on the distribution network diagram includes the following steps: S1: Identify suspected change nodes and suspected change branches based on real-time power grid measurement data, and obtain the electrical coupling observation matrix and state transition matrix based on historical electrical operation data and historical power grid topology; S2: Construct a hidden state vector by the suspected changing node state and the suspected changing branch state, and use a hidden Markov model to obtain the effective changing node and effective changing branch through the state transition matrix and the electrical coupling observation matrix. S3: Reconstruct the power grid topology based on effective changing nodes and effective changing branches. Obtain power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology. Integrate the power grid topology, power grid equipment parameters and line parameters to form a real-time distribution network model. S4: Based on the real-time distribution network diagram model, obtain the cross-device theoretical coordination protection constraint value of the protection equipment in the power grid equipment, and compare the actual cross-device coordination protection constraint value of the protection equipment with the cross-device theoretical coordination protection constraint value to obtain the coordination inspection result; S5: Based on the real-time distribution network diagram, obtain the theoretical protection settings of the protection equipment, compare the actual protection settings of the protection equipment with the theoretical protection settings to obtain the setting inspection results, and combine the setting inspection results with the cooperative inspection results to obtain the inspection status of the protection equipment.

[0020] Hidden Markov Models (HMMs) are classic time-series probabilistic models containing two core variables: hidden states, which are difficult to observe directly, and real-time observational data. They also include a state transition matrix and an observation matrix. The state transition matrix characterizes the temporal evolution of the hidden states, quantifying the probability that the hidden state at the previous moment will remain in its original state or switch to another state at the current moment. The observation matrix establishes a probabilistic mapping between the hidden states and the observational data, quantifying the matching probability of different real power grid topologies generating corresponding measurement data, i.e., the statistical distribution characteristics of various measurement data generated by different power grid topologies. Based on these dual probabilistic constraints, HMMs can deduce the hidden real power grid topology from the observational data, adapting to the characteristics of distribution network topologies that are difficult to monitor directly and require indirect identification through measurement data. This makes them suitable for scenarios requiring accurate identification of distribution network topology changes. In essence, the hidden state vector is formed by combining the states of suspected changed nodes and suspected changed branches, encompassing all possible combinations of these states. Suspected change node states include open and closed circuits, and suspected change branch states include added and removed equipment. The added and removed equipment can specifically be capacitors, transformers, and protection devices. The process of obtaining effective change nodes and effective change branches using a hidden Markov model through the state transition matrix and electrical coupling observation matrix involves probabilistically extrapolating all hidden state vectors and selecting the optimal hidden state vector that best matches the actual operating conditions of the power grid.

[0021] In this embodiment, considering that power grid topology reconstruction requires accurate understanding of the actual state changes of nodes and branches, but the actual topology changes are difficult to obtain directly and quickly, a hidden state vector is constructed based on the operating status of power grid nodes and branches, and the actual operating conditions of nodes and branches are deduced based on measurement data. Since the number of distribution network nodes and branches is large, performing full-domain modeling of all equipment and line states would significantly increase computational overhead and reduce topology reconstruction efficiency. Therefore, real-time measurement data is first used to screen out suspected changing nodes and branches, and the hidden state vector is constructed only based on the states of suspected changing nodes and branches, effectively improving the computational efficiency of topology identification and reconstruction. Meanwhile, the Hidden Markov Model has significant application limitations: using a single fixed global state transition matrix makes it difficult to adapt to the differentiated topology transition patterns and temporal evolution characteristics under different operating conditions such as peak and off-peak periods in the distribution network, resulting in insufficient accuracy in state deduction; its observation matrix follows the assumption of observation independence, assuming that various measurement data are independent of each other, ignoring both the inherent coupling relationship between measurement data and the topological coupling constraints between switches and branches, which easily leads to misjudgments and omissions of topology states. In view of this, this invention reconstructs a state transition matrix and an electrical coupling observation matrix adapted to multiple operating conditions, overcoming the inherent defects of Hidden Markov Models such as insufficient real-time performance, poor adaptability to operating conditions, insufficient exploitation of coupling features, and low state identification accuracy. Through targeted improvements to the state transition matrix and observation matrix, the accuracy of identifying effective changing nodes and branches in the power grid is effectively enhanced, laying the foundation for reliable reconstruction of the subsequent power grid topology.

[0022] A distribution network model is a digital, global power grid model that integrates the distribution network topology connections with the electrical parameters of equipment and lines. As is well known, based on the acquired real-time distribution network model, theoretical protection settings adapted to the current actual topology and operating conditions of the power grid can be accurately calculated. By comparing the theoretical protection settings with the theoretical protection constraint values ​​across equipment, the limitations of traditional settings inspections that only target protection devices are overcome. This upgrade from single-point settings inspection to global inspection effectively addresses the shortcomings of existing single-dimensional inspections and significantly improves the comprehensiveness of protection device inspections. The protection settings specifically refer to the protection current.

[0023] As an optional embodiment, the step of identifying suspected change nodes and suspected change branches based on real-time power grid measurement data includes: Based on the data representation status attributes, real-time power grid measurement data is divided into electrical connectivity data and electrical operation data. The electrical connectivity data is compared with the equipment connection data in the power grid ledger to identify the first suspected change node; The electrical operation data of the branch where the power grid node is located is combined with the electrical connectivity data of the power grid node for effective verification. The power grid node that fails the effective verification is designated as the second suspected change node. The first suspected change node and the second suspected change node constitute the suspected change node. Based on the electrical operation data of the power grid branch, determine whether the power grid branch has undergone a step change. The power grid branch that has undergone a step change and whose two ends are not suspected change nodes is regarded as a suspected change branch.

[0024] Electrical connectivity data specifically characterizes the on / off status of various types of switchgear in the power grid. If a switchgear is recorded as closed in the power grid ledger data, but its electrical connectivity data shows it as open, then that switchgear is designated as the first suspected change node. Electrical operation data specifically characterizes the power grid operating conditions, including voltage, current, and power flow. Effective verification uses the consistency between the switch on / off status and the actual electrical operating conditions of the branch as the core criterion to screen for the second suspected change node. When there is inconsistency between the switch on / off status and the branch electrical operating conditions, it indicates a possible distortion of remote signaling or abnormal electrical measurements. It is difficult to directly determine whether the node has actually undergone a topological change, possessing potential anomaly uncertainty. Therefore, this type of node is designated as the second suspected change node. Specifically: If the electrical connectivity data of a grid node shows an open state, theoretically, the branch where the switchgear is located should be in an open state, and the branch current and power flow should approach zero. If the real-time current and power flow of the branch are significantly not approaching zero and remain at a stable non-zero operating level, then the grid node is determined to be the second suspected change node. If the electrical connectivity data of a grid node shows a closed state, theoretically, the branch where the switchgear is located should be in a conducting state, and the branch current and power flow should match the current distribution network load conditions, presenting a stable non-zero operating condition. If the real-time current and power flow of the branch continuously approach zero and the branch voltage synchronously shows no power supply characteristics, then the grid node is determined to be the second suspected change node. Only when the open / closed state of the grid node shows a linkage consistency with the electrical operating data such as current, power flow, and voltage of the branch, that is, when the branch is open, the current and power flow approach zero, and when the branch is closed, the current and power flow are within the normal load range, and the voltage and power data match the open / closed state of the switchgear, is the valid test considered passed. Switching power grid equipment includes circuit breakers, sectionalizing switches, tie switches, and other equipment in the distribution network that can control the on / off state of lines.

[0025] The addition or removal of new equipment is equivalent to instantaneous switching of fixed loads, causing a sudden increase or decrease in the active power and current of the branch. This characteristic differs significantly from the smooth fluctuations of normal loads. However, besides the topology adjustments caused by equipment switching, other grid operating disturbances can also cause similar step-change characteristics in the electrical quantities of the branch. Therefore, it is difficult to determine the true topology change of the branch based solely on step changes. Specifically, electrical operating data of the branch are selected within a preset time period, and the average active power and average current are calculated. The average active power and average current are used as the reference steady-state values. When the change in the current active power and current of the branch relative to the corresponding reference steady-state value exceeds a preset percentage, such as 25% to 30%, and the subsequent electrical operating data does not quickly rebound to the steady-state operating level before the change, it indicates that a step change has occurred in the grid branch. For branches where a step change in electrical operating data is detected, the nodes at both ends of the branch are further examined. Branches whose nodes at both ends are not within the identified range of suspected change nodes are classified as suspected change branches, thus ruling out the possibility that the branch's operating condition change is caused by the opening or closing of switchgear. In this embodiment, identifying suspected change nodes and suspected change branches improves the efficiency of topology identification using hidden Markov models.

[0026] Specifically, such as Figure 2 As shown, the acquisition of the electrical coupling observation matrix and state transition matrix based on historical electrical operation data and historical power grid topology includes: S1a: Obtain the mean and covariance matrix of historical electrical operation data corresponding to the topology state in the historical power grid topology; S1b: Perform conditional independence tests on the historical electrical operation data corresponding to the topology states in the historical topology of the power grid to determine the causal relationship. Based on the historical electrical operation data corresponding to the topology states in the historical topology of the power grid, identify several historical operating conditions and obtain the absolute deviation value of the historical electrical operation data corresponding to each historical operating condition under the same topology state. Based on the absolute deviation value, obtain the disturbance deviation corresponding to each historical operating condition under the same topology state. Obtain the disturbance matrix of the historical electrical operation data corresponding to the topology state through the disturbance deviation. S1c: Obtain the final covariance matrix based on the disturbance matrix, causal relationship and covariance matrix. Obtain the joint Gaussian distribution function based on the final covariance matrix and mean. Input typical historical electrical operation data from historical electrical operation data into the joint Gaussian distribution function to obtain the conditional likelihood probability, and then obtain the electrical coupling observation matrix. S1d: Extract state change records from the historical topology of the power grid, and classify the state change records into the corresponding operating conditions. Calculate the jump frequency between the topology states in each operating condition to obtain the state transition matrix corresponding to each operating condition.

[0027] Specifically, historical electrical operation data includes historical voltage, historical current, and historical power flow, which includes active power flow and reactive power flow. It is understood that the topology state referred to here specifically refers to the switching state of grid nodes and the on / off state of branch equipment; each topology state corresponds to multiple sets of historical electrical operation data. The average historical electrical operation data corresponding to the i-th topology state... ,in, This represents the number of historical electrical operation data sets corresponding to the i-th topology state. Let represent the k-th set of historical electrical operation data under the i-th topology state. The elements of the covariance matrix of the i-th topology state can be uniformly denoted as . , Let m be the element in the m-th row and n-th column of the covariance matrix of the i-th topological state. This represents the covariance between the m-th type of electrical quantity and the n-th type of electrical quantity. This represents the m-th type of electrical quantity. Let m represent the nth type of electrical quantity. The values ​​of m and n both range from 1 to d, where d represents the dimension of the electrical operation data. In this embodiment, the historical electrical operation data is four-dimensional, and the electrical quantity categories are four types: historical voltage, historical current, historical active power flow, and historical reactive power flow. In this embodiment, the causal relationship between historical voltage, historical current, active power flow, and reactive power flow in the historical electrical operation data corresponding to the topology state is determined using partial correlation or LiNGAM algorithms. It can be understood that the causal relationship is a one-way influence relationship between historical voltage, historical current, active power flow, and reactive power flow. For example, if a change in historical voltage leads to a change in historical current, but a change in historical current does not conversely lead to a change in historical voltage, then there is no causal edge from historical current to historical voltage, but there is a causal edge from historical voltage to historical current. Specifically, based on the historical electrical operation data corresponding to the topology state, three indicators are obtained: historical average load rate, historical voltage deviation rate, and historical load fluctuation amplitude. These three indicators are then used to determine intervals using preset thresholds, thereby jointly identifying historical operating conditions.The process involves: obtaining the absolute deviation value of historical electrical operation data for each historical operating condition under the same topology; obtaining the disturbance deviation value for each historical operating condition under the same topology based on the absolute deviation value; and obtaining the disturbance matrix of historical electrical operation data for the same topology through the disturbance deviation value. This includes: obtaining the historical median voltage, historical median current, active power flow median, and reactive power flow median in the historical electrical operation data for each historical operating condition under the same topology; calculating the absolute voltage deviation value between the historical voltage median and the historical voltage in the historical electrical operation data for each historical operating condition under the same topology; summing the absolute voltage deviation values ​​to obtain the voltage disturbance deviation value for each historical operating condition under the same topology; and calculating the absolute current deviation value between the historical current median and the historical current in the historical electrical operation data for each historical operating condition under the same topology; summing the absolute current deviation values ​​to obtain the disturbance matrix of historical electrical operation data for each historical operating condition under the same topology. The current disturbance deviation is calculated by summing the current disturbance deviations corresponding to each historical operating condition under the same topology state to obtain the current disturbance value of the historical current under that same topology state. The active power flow median is calculated, and the absolute active power flow deviation between the median and the historical electrical operation data corresponding to each historical operating condition under the same topology state is calculated. The absolute active power flow deviation is then summed to obtain the active power disturbance deviation for each historical operating condition under the same topology state. The reactive power flow median is calculated, and the absolute reactive power flow deviation between the median and the historical electrical operation data corresponding to each historical operating condition under the same topology state is calculated. The absolute reactive power deviation is then summed to obtain the reactive power disturbance deviation for each historical operating condition under the same topology state. The reactive power disturbance deviation is then summed to obtain the reactive power disturbance value for the reactive power flow under that same topology state. Based on the voltage disturbance value, current disturbance value, active power disturbance value, and reactive power disturbance value, the disturbance matrix of the historical electrical operation data corresponding to the topology state is obtained. Obtaining the final covariance matrix based on the disturbance matrix, causal relationship, and covariance matrix specifically involves: setting the elements corresponding to non-causal edges in the covariance matrix to zero to obtain the causal covariance matrix; and summing the causal covariance matrix with the disturbance matrix to obtain the final covariance matrix. By abandoning the inherent limitations of using symmetric correlation modeling for electrical quantities such as voltage, current, and power flow based on the joint Gaussian distribution, this approach aligns with the actual causal driving mechanism of power flow driving voltage and voltage driving current within the power grid, effectively eliminating spurious statistical correlations caused by the symmetric correlation assumption. Simultaneously, it overcomes the modeling defect of traditional covariance matrices that assume isotropic measurement disturbances, closely reflecting the actual engineering characteristic that the amplitude of voltage measurement disturbances in the power grid is smaller than that of power flow measurement disturbances.Specifically, on the one hand, by eliminating invalid reverse correlations through causal relationships, the accuracy and reliability of the solved conditional likelihood probability are significantly improved; on the other hand, by calculating the absolute value deviation of various types of electrical operation data and constructing a perturbation matrix, the dilution of high-precision electrical operation data by the perturbation interference of low-precision electrical operation data is avoided. Simultaneously, the error influence of low-precision electrical operation data is reasonably constrained, preventing the error from being unreasonably amplified, further improving the accuracy of the conditional likelihood probability. The joint Gaussian distribution function of the i-th topological state is: ,in, Indicates that in the topological state At that time, the historical electrical operation data was The conditional likelihood probability. This represents the final covariance matrix of the i-th topological state. express The determinant of the matrix is ​​given. The typical historical electrical operation data corresponding to the i-th topological state is input into the joint Gaussian distribution function corresponding to the i-th topological state, and the corresponding conditional likelihood probability is obtained. The conditional likelihood probabilities of all topological states are then summarized to construct the electrical coupling observation matrix. Specifically, the historical electrical operation data corresponding to each topological state is clustered using the K-means++ clustering algorithm, and the main cluster centers are selected as the typical historical electrical operation data for that topological state. The expression for obtaining the state transition probability is: , This indicates that under operating condition c, from the topology state Jump to topology state The state transition probability, This indicates that under operating condition c, from the topology state Jump to topology state The jump frequency, M represents the total number of topological states that occur under condition c. This indicates that under operating condition c, from the topology state Jump to topology state The jump frequency, the state transition matrix corresponding to condition c In this embodiment, the operating conditions may include peak load conditions, flat load conditions, off-peak load conditions, daytime electricity consumption conditions, and nighttime electricity consumption conditions.

[0028] Specifically, the process of constructing a hidden state vector using suspected changing node states and suspected changing branch states, and obtaining effective changing nodes and effective changing branches using a hidden Markov model through a state transition matrix and an electrical coupling observation matrix, includes: The current operating condition is determined based on real-time power grid measurement data. The target state transition matrix is ​​then selected from the state transition matrix based on the current operating condition. The real-time power grid measurement data is matched with typical historical electrical operation data to locate the corresponding area of ​​the electrical coupling observation matrix and obtain the observation condition probability. The optimal hidden state vector is derived using a hidden Markov model through the hidden state vector, the observation condition probability, and the target state transition matrix, thereby obtaining the effective change nodes and effective change branches.

[0029] As is well known, after obtaining the state transition matrix and the electrical coupling observation matrix, the Hidden Markov Model (HMM) can specifically use the Viterbi algorithm to deduce the optimal hidden state vector. It is understandable that during the derivation process, the selected target state transition matrix and the obtained observation condition probabilities are dynamically updated with real-time power grid measurement data. In this embodiment, the average load rate, voltage deviation rate, and load fluctuation amplitude are obtained based on the real-time power grid measurement data. The average load rate defines the overall load level of the power grid, the voltage deviation rate characterizes the degree of deviation of node voltage from the rated value, and the load fluctuation amplitude reflects the short-term power surge characteristics. Preset thresholds are used to determine the intervals of these three indicators, thereby jointly identifying the current power grid operating condition. Specifically, the first electrical coupling relationship of the electrical operation data in the real-time power grid measurement data and the second electrical coupling relationship of each typical historical electrical operation data are calculated. The Euclidean distance between the eigenvectors corresponding to the first and second electrical coupling relationships is also calculated. It is understood that the electrical coupling relationship specifically refers to the inherent law of mutual influence and coordinated change among electrical quantities, such as how other electrical quantities change in tandem when one electrical quantity changes. The typical historical electrical operation data with the smallest Euclidean distance is used as the matched typical historical electrical operation data, and the corresponding region is located in the electrical coupling observation matrix to extract the observation condition probability. The optimal hidden state vector is based on the Hidden Markov Model, which integrates the state transition probability corresponding to the operating condition and the observation condition probability corresponding to the real-time measurement data of the power grid. Among all possible combinations of node and branch topology states, the hidden state vector with the largest posterior probability and the closest to the actual topology operation of the power grid is obtained. This vector contains the suspected changed node state and the suspected changed branch state. After decoupling, the effective changed nodes and effective changed branches can be determined.

[0030] In some embodiments, the reconstructing of the power grid topology based on effective changing nodes and effective changing branches includes: Based on the equipment connection relationships in the power grid ledger data, a basic topology framework is established. If the effective change node is a closing node, a connection between the effective change node and the corresponding electrical associated node is established in the basic topology framework. If the effective change node is a opening node, the connection between the effective change node and the corresponding electrical associated node is disconnected in the basic topology framework. If the effective change branch is an elimination branch, a connection between the nodes at both ends of the effective change branch is established in the basic topology framework. If the effective change branch is an addition branch, a node is added between the nodes at both ends of the effective change branch in the basic topology framework, and a connection is established between the nodes at both ends and the added node, thereby obtaining the power grid topology.

[0031] Understandably, a closing node refers to a node that is in an open state in the power grid ledger data but is in a closed state in the actual operating conditions of the power grid. A opening node refers to a node that is in a closed state in the power grid ledger data but is in an open state in the actual operating conditions of the power grid. An elimination branch refers to a topology branch in which power grid equipment A1 and C1 are connected via power grid equipment B1 in the power grid ledger data, but power grid equipment B1 is not in operation in the actual operating conditions of the power grid. An addition branch refers to a topology branch in which power grid equipment A1 and C1 are directly connected in the power grid ledger data, but power grid equipment B1 is newly added in the actual operating conditions and connected in series between power grid equipment A1 and C1. Using the equipment connection relationships in the power grid ledger data as the basic topology framework, effective changed nodes and effective changed branches are first identified, and then the basic topology framework is locally updated based on the effective changed nodes and effective changed branches. This significantly improves the efficiency of power grid topology reconstruction while ensuring a high degree of consistency between the power grid topology and the actual operating topology on site.

[0032] In some embodiments, obtaining power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology includes: Starting from the power generation nodes of the power grid, power flow calculations are performed along the power grid topology to obtain predicted electrical operation data. The predicted electrical operation data is then matched with the actual electrical operation data. If the match is successful, the initial parameters of the power grid equipment and the initial parameters of the lines are used as the parameters of the power grid equipment and the lines, respectively. If the match is unsuccessful, anomaly areas are located based on the data deviation distribution characteristics between the predicted electrical operation data and the actual electrical operation data. The initial parameters of the power grid equipment and the initial parameters of the lines within the anomaly areas are iteratively corrected to obtain the parameters of the power grid equipment and the lines.

[0033] Considering the multiple factors affecting power grid equipment and lines during long-term operation, such as equipment aging, changes in the external environment, and load fluctuations, actual electrical parameters are prone to deviating from the initial parameters recorded in the ledger. This solution compares and matches the predicted electrical operation data obtained from power flow calculations with the measured electrical operation data from the field. Based on the data deviation distribution characteristics, it accurately identifies abnormal parameter areas and only performs iterative optimization and correction on the initial parameters of power grid equipment and lines within these abnormal areas. This effectively improves the accuracy of power grid equipment and line parameters while avoiding blind correction of parameters across the entire network, significantly improving the efficiency of parameter correction. The iterative correction of the initial parameters of power grid equipment and lines within abnormal areas specifically involves: adjusting parameter values ​​using the gradient descent method; recalculating the predicted electrical operation data after each iteration and obtaining the root mean square error (RMSE) between it and the measured electrical operation data; terminating the iteration when any of the following conditions are met: the RMSE is less than a preset threshold (e.g., 0.05); the error change rate between two consecutive iterations is less than 1%; and the number of iterations reaches a preset maximum (e.g., 20). The parameter values ​​at the time of iteration termination are used as the corrected power grid equipment and line parameters.

[0034] Understandably, line parameters include positive-sequence impedance, zero-sequence impedance, capacitance to ground, and line structural parameters. Power grid equipment parameters include transformer impedance and turns ratio, power source equivalent internal impedance, equipment rated electrical parameters, and load equivalent impedance. The reconstruction of the power grid topology and the accurate acquisition of line parameters and power grid equipment parameters provide a reliable foundation for the subsequent derivation of theoretical protection settings and cross-equipment theoretical coordination protection constraint values.

[0035] Specifically, the step of locating abnormal areas based on the data deviation distribution characteristics between predicted electrical operation data and actual electrical operation data includes: The data deviation between the predicted electrical operation data and the actual electrical operation data of the power grid equipment is obtained. The power grid equipment with the largest data deviation is identified as the core abnormal equipment. The power grid topology is traversed with the location of the core abnormal equipment as the center, and the abnormal area is formed with the power grid equipment whose data deviation is greater than the preset deviation as the boundary.

[0036] In some embodiments, the step of obtaining the cross-device theoretical coordination protection constraint values ​​of protection devices in the power grid based on the real-time distribution network diagram model, and comparing the actual cross-device coordination protection constraint values ​​of the protection devices with the cross-device theoretical coordination protection constraint values ​​to obtain the coordination inspection results includes: The final protection coordination pair is obtained based on the upstream and downstream connection relationship of the protection equipment in the real-time distribution network diagram, and the initial protection coordination pair is obtained through the equipment connection relationship in the power grid ledger data. The initial protection coordination pair and the final protection coordination pair are then matched across the entire domain. Perform single-domain matching between the initial protection pair that failed to match and the final protection pair to obtain the number of the final protection pair that were matched and the number of the initial protection pair; The theoretical coordination protection constraint value across devices is obtained based on the final number of protection coordination pairs and the preset time limit difference. If the theoretical coordination protection constraint value across devices is greater than or equal to the actual coordination protection constraint value of the matched initial protection coordination pair, it indicates that the coordination inspection is qualified; otherwise, it indicates that it is unqualified.

[0037] The actual cross-device coordination protection constraint value refers to the actual set action time difference of the protection equipment, specifically read through the scheduling automation system or protection information substation. Taking the final protection coordination pair including ab, bc, and cd, and the initial protection coordination pair including ac and cd as an example, when matching the final protection coordination pair with the initial protection coordination pair, full-domain matching means comparing the first and last protection nodes of the two sets of coordination pairs one by one. If the first and last nodes are consistent, it is determined that the full-domain matching is successful; if only one end node is consistent, it is determined that the single-domain matching is successful. After completing the full-domain matching, the initial protection coordination pairs ac and ab and bc that failed to match were filtered out. Then, single-domain matching was performed on the remaining coordination pairs. It was found that the number of final protection coordination pairs matched in a single domain was 2. If the preset time difference is 0.02 seconds, the theoretical cross-device coordination protection constraint value is 0.04 seconds. For the initial protection coordination pair ac matched in a single domain, if its actual cross-device coordination protection constraint value is 0.03, then the theoretical cross-device coordination protection constraint value is greater than the actual cross-device coordination protection constraint value. Therefore, the coordination inspection is qualified. This method, by performing full-domain matching and single-domain matching in a layered manner and verifying the node correspondence and time constraint conditions of protection coordination pairs in a layered manner, can effectively avoid the problem of protection logic matching disorder caused by changes in power grid topology, and improve the accuracy and adaptability of protection coordination inspection.

[0038] Specifically, the acquisition of theoretical protection settings for protection devices based on real-time distribution network diagrams includes: Based on the real-time distribution network model, the fault conditions of the power grid under the maximum and minimum operating modes are simulated respectively, so as to obtain the maximum fault current and minimum fault current of the fault point. The theoretical protection settings of the protection equipment for the main protection section are obtained based on the maximum and minimum fault currents. The theoretical protection settings of the protection equipment for the backup section of the entire line are obtained based on the setting matching relationship between the protection equipment and the minimum fault current. The maximum load current of the power grid is obtained based on the real-time distribution network diagram. The theoretical protection settings of the protection equipment for the remote backup section are obtained based on the maximum load current and the minimum fault current of the power grid.

[0039] Understandably, the maximum operating mode of the power grid refers to the operating condition with the largest short-circuit current, while the minimum operating mode refers to the operating condition with the smallest short-circuit current. Based on the real-time distribution network diagram, the fault conditions of the power grid under the maximum and minimum operating modes are simulated respectively to obtain the maximum and minimum fault currents at the fault point. Specifically, based on the real-time distribution network diagram, the maximum fault current when a three-phase short circuit occurs at the end of this line under the maximum operating mode is simulated; based on the real-time distribution network diagram, the minimum fault current when a two-phase short circuit occurs at the end of this line under the minimum operating mode is simulated; based on the real-time distribution network diagram, the minimum fault current when a two-phase short circuit occurs at the end of the downstream line under the minimum operating mode is simulated. Taking the power grid link as A protection device - line 1 - B protection device - line 2 - C protection device as an example, if the inspection is performed on protection device A, then the end of this line refers to the end of line 1, and the end of the downstream line refers to the end of line 2. It is understandable that the main protection section is specifically the overcurrent section I, the backup section for the entire line is specifically the overcurrent section II, and the remote backup section is specifically the overcurrent section III. Since the overcurrent section I, overcurrent section II, and overcurrent section III are well-known concepts, they will not be elaborated here.

[0040] The theoretical protection settings for the main protection section of the protection equipment are obtained based on the maximum and minimum fault currents, including: The first initial theoretical protection setting is obtained by multiplying the first reliability coefficient by the maximum fault current. The first sensitivity value is obtained by dividing the minimum fault current at the end of the line when a two-phase short circuit occurs by the first initial theoretical protection setting. When the first sensitivity value is greater than or equal to the first sensitivity threshold, the first initial theoretical protection setting is used as the theoretical protection setting; otherwise, the first initial theoretical protection setting is reduced, and the first sensitivity value is recalculated until the first sensitivity value is greater than or equal to the first sensitivity threshold, at which point the theoretical protection setting is obtained. Specifically, the setting matching relationship between protection devices means that the theoretical protection setting of the upper-level protection device for overcurrent stage II must be greater than the theoretical protection setting of the lower-level protection device for overcurrent stage I. The theoretical protection settings for the backup section of the entire line are obtained based on the setting matching relationship between the protection devices and the minimum fault current. This includes: multiplying the theoretical protection settings for the overcurrent I section of the lower-level protection devices by the second reliability coefficient to obtain the second initial theoretical protection settings; dividing the minimum fault current when a two-phase short circuit occurs at the end of this line by the second initial theoretical protection settings to obtain the second sensitivity; when the second sensitivity value is greater than or equal to the second sensitivity threshold, the second initial theoretical protection settings are used as the theoretical protection settings; otherwise, the second initial theoretical protection settings are reduced, and the second sensitivity value is recalculated until the second sensitivity value is greater than or equal to the second sensitivity threshold, thus obtaining the theoretical protection settings. The theoretical protection settings for the remote backup section of the protection equipment are obtained based on the maximum load current and minimum fault current of the power grid. This includes: obtaining a third initial theoretical protection setting based on the maximum load current of the power grid; dividing the minimum fault current when a two-phase short circuit occurs at the end of this line by the third initial theoretical protection setting to obtain a third sensitivity value; and dividing the minimum fault current when a two-phase short circuit occurs at the end of the downstream line by the third initial theoretical protection setting to obtain a fourth sensitivity value. If both the third sensitivity value and the sensitivity threshold are greater than the corresponding sensitivity threshold, the third initial theoretical protection setting is used as the theoretical protection setting; otherwise, the third initial theoretical protection setting is reduced to obtain the theoretical protection setting. The sensitivity value reflects the detection margin of the protection equipment for faults at the end of this line under minimum operating conditions and is used to verify whether the protection equipment meets the requirement of not failing to operate. The reliability coefficient reflects the degree of safety compensation for uncertainties in the setting calculation and is used to verify whether the protection equipment meets the requirement of not maloperating. Through the joint setting of sensitivity and reliability coefficient, the protection equipment simultaneously meets the requirements of operational reliability and selectivity. As is well known, the reliability coefficient and sensitivity threshold are pre-set empirical constants, and in engineering applications, the values ​​specified in the regulations can be directly used.

[0041] Specifically, the process of comparing the actual protection settings of the protection equipment with the theoretical protection settings to obtain the setting inspection results, and combining the setting inspection results with the coordinated inspection results to obtain the protection equipment inspection status, includes: The theoretical protection settings of the protection equipment for the main protection section, the theoretical protection settings of the protection equipment for the entire backup section, and the theoretical protection settings of the protection equipment for the remote backup section are obtained respectively. The first, second, and third differences between these values ​​and the corresponding actual protection settings are then obtained. If the first, second, and third differences are all greater than the preset difference, the setting inspection is considered qualified; otherwise, the setting inspection is considered unqualified. When both the setpoint inspection result and the coordinated inspection result are qualified, it means that the protection equipment inspection is qualified; otherwise, it means that the protection equipment inspection is unqualified.

[0042] The specific embodiments described above are preferred embodiments of the device protection autonomous inspection method based on the distribution network diagram model of the present invention, and are not intended to limit the specific scope of the present invention. The scope of the present invention includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for autonomous equipment protection inspection based on a distribution network diagram, characterized in that, Includes the following steps: Based on real-time power grid measurement data, suspected change nodes and suspected change branches are identified, and electrical coupling observation matrix and state transition matrix are obtained based on historical electrical operation data and historical power grid topology. Hidden state vectors are constructed by the states of suspected changing nodes and suspected changing branches. Effective changing nodes and effective changing branches are obtained by using a hidden Markov model through the state transition matrix and the electrical coupling observation matrix. Reconstructing the power grid topology based on effective changing nodes and effective changing branches, obtaining power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology, and integrating the power grid topology, power grid equipment parameters and line parameters to form a real-time distribution network model; Based on the real-time distribution network diagram model, the theoretical coordination protection constraint values ​​of the protection devices in the power grid are obtained, and the actual coordination protection constraint values ​​of the protection devices are compared with the theoretical coordination protection constraint values ​​to obtain the coordination inspection results. The theoretical protection settings of the protection equipment are obtained based on the real-time distribution network diagram. The actual protection settings of the protection equipment are compared with the theoretical protection settings to obtain the setting inspection results. The inspection status of the protection equipment is obtained by combining the setting inspection results with the cooperative inspection results.

2. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The identification of suspected change nodes and suspected change branches based on real-time power grid measurement data includes: Based on the data representation status attributes, real-time power grid measurement data is divided into electrical connectivity data and electrical operation data. The electrical connectivity data is compared with the equipment connection data in the power grid ledger to identify the first suspected change node; The electrical operation data of the branch where the power grid node is located is combined with the electrical connectivity data of the power grid node for effective verification. The power grid node that fails the effective verification is designated as the second suspected change node. The first suspected change node and the second suspected change node constitute the suspected change node. Based on the electrical operation data of the power grid branch, determine whether the power grid branch has undergone a step change. The power grid branch that has undergone a step change and whose two ends are not suspected change nodes is regarded as a suspected change branch.

3. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The process of obtaining the electrical coupling observation matrix and state transition matrix based on historical electrical operation data and historical power grid topology includes: Obtain the mean and covariance matrix of historical electrical operation data corresponding to the topology state in the historical power grid topology; Conditional independence tests are performed on historical electrical operation data corresponding to topology states in the historical topology of the power grid to determine causal relationships. Based on the historical electrical operation data corresponding to topology states in the historical topology of the power grid, several historical operating conditions are identified, and the absolute deviation value of historical electrical operation data corresponding to each historical operating condition under the same topology state is obtained. Based on the absolute deviation value, the disturbance deviation corresponding to each historical operating condition under the same topology state is obtained, and the disturbance matrix of historical electrical operation data corresponding to topology states is obtained through the disturbance deviation. The final covariance matrix is ​​obtained based on the perturbation matrix, causal relationship and covariance matrix. The joint Gaussian distribution function is obtained based on the final covariance matrix and the mean. Typical historical electrical operation data in the historical electrical operation data are input into the joint Gaussian distribution function to obtain the conditional likelihood probability, and then the electrical coupling observation matrix is ​​obtained. State change records are extracted from the historical topology of the power grid and assigned to the corresponding operating conditions. The frequency of transitions between topology states in each operating condition is counted to obtain the state transition matrix corresponding to each operating condition.

4. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 3, characterized in that, The process of constructing a hidden state vector using suspected changing node states and suspected changing branch states, and obtaining effective changing nodes and effective changing branches using a hidden Markov model through a state transition matrix and an electrical coupling observation matrix, includes: The current operating condition is determined based on real-time power grid measurement data. The target state transition matrix is ​​then selected from the state transition matrix based on the current operating condition. The real-time power grid measurement data is matched with typical historical electrical operation data to locate the corresponding area of ​​the electrical coupling observation matrix and obtain the observation condition probability. The optimal hidden state vector is derived using a hidden Markov model through the hidden state vector, the observation condition probability, and the target state transition matrix, thereby obtaining the effective change nodes and effective change branches.

5. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The reconstructing of the power grid topology based on effective changing nodes and effective changing branches includes: Based on the equipment connection relationships in the power grid ledger data, a basic topology framework is established. If the effective change node is a closing node, a connection between the effective change node and the corresponding electrical associated node is established in the basic topology framework. If the effective change node is a opening node, the connection between the effective change node and the corresponding electrical associated node is disconnected in the basic topology framework. If the effective change branch is an elimination branch, a connection between the nodes at both ends of the effective change branch is established in the basic topology framework. If the effective change branch is an addition branch, a node is added between the nodes at both ends of the effective change branch in the basic topology framework, and a connection is established between the nodes at both ends and the added node, thereby obtaining the power grid topology.

6. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The process of obtaining power grid equipment parameters and line parameters based on real-time power grid measurement data and power grid topology includes: Starting from the power generation nodes of the power grid, power flow calculations are performed along the power grid topology to obtain predicted electrical operation data. The predicted electrical operation data is then matched with the actual electrical operation data. If the match is successful, the initial parameters of the power grid equipment and the initial parameters of the lines are used as the parameters of the power grid equipment and the lines, respectively. If the match is unsuccessful, anomaly areas are located based on the data deviation distribution characteristics between the predicted electrical operation data and the actual electrical operation data. The initial parameters of the power grid equipment and the initial parameters of the lines within the anomaly areas are iteratively corrected to obtain the parameters of the power grid equipment and the lines.

7. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 6, characterized in that, The step of locating abnormal areas based on the data deviation distribution characteristics between predicted electrical operation data and actual electrical operation data includes: The system obtains the data deviation between the predicted electrical operation data and the actual electrical operation data corresponding to the power grid equipment. The power grid equipment with the largest data deviation is identified as the core abnormal equipment. The power grid topology is traversed with the location of the core abnormal equipment as the center, and the abnormal region is formed with the power grid equipment whose data deviation is greater than the preset deviation as the boundary.

8. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The method involves obtaining the theoretical cross-device coordination protection constraint values ​​of protection devices in the power grid based on the real-time distribution network diagram model, and comparing the actual cross-device coordination protection constraint values ​​of the protection devices with the theoretical cross-device coordination protection constraint values ​​to obtain the coordination inspection results, including: The final protection coordination pair is obtained based on the upstream and downstream connection relationship of the protection equipment in the real-time distribution network diagram, and the initial protection coordination pair is obtained through the equipment connection relationship in the power grid ledger data. The initial protection coordination pair and the final protection coordination pair are then matched across the entire domain. Perform single-domain matching between the initial protection pair that failed to match and the final protection pair to obtain the number of the final protection pair that were matched and the number of the initial protection pair; The theoretical coordination protection constraint value across devices is obtained based on the final number of protection coordination pairs and the preset time limit difference. If the theoretical coordination protection constraint value across devices is greater than or equal to the actual coordination protection constraint value across devices of the matched initial protection coordination pair, it indicates that the coordination inspection is qualified; otherwise, it indicates that it is unqualified.

9. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 1, characterized in that, The theoretical protection settings of the protection equipment obtained based on the real-time distribution network diagram include: Based on the real-time distribution network model, the fault conditions of the power grid under the maximum and minimum operating modes are simulated respectively, so as to obtain the maximum fault current and minimum fault current of the fault point. The theoretical protection settings of the protection equipment for the main protection section are obtained based on the maximum and minimum fault currents. The theoretical protection settings of the protection equipment for the backup section of the entire line are obtained based on the setting matching relationship between the protection equipment and the minimum fault current. The maximum load current of the power grid is obtained based on the real-time distribution network diagram. The theoretical protection settings of the protection equipment for the remote backup section are obtained based on the maximum load current and the minimum fault current of the power grid.

10. The method for autonomous equipment protection inspection based on distribution network diagram model according to claim 9, characterized in that, The process of comparing the actual protection settings of the protection equipment with the theoretical protection settings to obtain the setting inspection results, and combining the setting inspection results with the coordinated inspection results to obtain the protection equipment inspection status, includes: The theoretical protection settings of the protection equipment for the main protection section, the theoretical protection settings of the protection equipment for the entire backup section, and the theoretical protection settings of the protection equipment for the remote backup section are obtained respectively. The first, second, and third differences between these values ​​and the corresponding actual protection settings are then obtained. If the first, second, and third differences are all greater than the preset difference, the setting inspection is considered qualified; otherwise, the setting inspection is considered unqualified. When both the setpoint inspection result and the coordinated inspection result are qualified, it means that the protection equipment inspection is qualified; otherwise, it means that the protection equipment inspection is unqualified.

Images