Power system voltage dominant cascading failure weak line identification method, system and device

Abstract

The application provides a power system voltage dominant cascading failure weak line identification method, system and device, which constructs a voltage estimation knowledge graph according to the topological structure, fault history data and simulation data of the power system; a fault scene set is established, each fault scene is selected in turn, a corresponding white-box tree spike network model is determined by using the knowledge graph, and the node voltage under the scene is calculated; whether a voltage response event occurs is judged according to the calculation result, if the condition is met, the event is identified and whether the event involves device locking or off-grid is judged, so that whether the line is a weak line is determined; if the condition is not met, other fault scenes are searched. The application can accurately identify the weak line in the power system, and provides a reliable basis for the design, operation control and protection measure optimization of the power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system protection and fault analysis, and particularly relates to a power system voltage dominant cascading failure weak line identification method, system and device. BACKGROUND

[0002] With the rapid development of renewable energy, large-scale power electronic equipment such as wind power, photovoltaic power generation and direct current transmission is connected to the power system, which promotes the power electronic transformation of the power system. The power electronic equipment has high sensitivity and strong coupling, which makes the fault evolution process in the power electronic power system more complex. After the fault of the power grid, the change of the system voltage will cause the switching of the operation mode and control strategy of the power electronic equipment, or trigger the protection action, and then cause the further change of the power grid voltage. The voltage change will again affect the output characteristics of the power electronic equipment, forming a cyclic feedback process. In the power electronic power system, the influence range of a single fault may rapidly expand, leading to the disconnection of the power electronic equipment, and even causing system stability problems, presenting a new form of voltage dominant cascading failure.

[0003] According to the inducing factors of cascading failure, existing researches mainly divide the cascading failure into three types: overload dominant type, structure dominant type and coordination dominant type. The overload dominant type of cascading failure refers to the transfer of power flow due to the fault, which leads to the overload of other lines and equipment, and triggers cascading trip. The structure dominant type of cascading failure is caused by the failure of key nodes or lines, which changes the topology of the system and then triggers the cascading failure. The coordination dominant type of cascading failure is caused by improper or delayed protection device settings, which triggers the failure of other equipment. In addition, some researches also focus on the cascading reaction caused by abnormal frequency fluctuations. However, the generation mechanism and form of voltage dominant cascading failure are significantly different from the overload, structure, coordination dominant or frequency dominant cascading failure, and the existing researches are difficult to explain the evolution rule of voltage dominant cascading failure.

[0004] With the wide access of power electronic equipment, the characteristics of fault transmission through voltage have gradually attracted attention. Existing researches mainly focus on the influence of voltage change on a single power electronic equipment or the response of multiple power electronic equipments through voltage coupling, and have not fully considered the influence of secondary response of power electronic equipment on voltage and its transmission effect. Therefore, the research on voltage dominant cascading failure is still in its infancy, and the related mechanism and rule have not been clearly defined.

[0005] In order to effectively prevent cascading failures, it is necessary to optimize the topology of the power grid by identifying the key lines that cause failure propagation and expansion. Currently, the identification of weak lines in cascading failures mainly adopts mathematical modeling and simulation, complex network theory, data-driven methods, and reliability evaluation and optimization methods. Mathematical modeling establishes a dynamic model or load flow model of the power system to analyze the dynamic behavior of the system, and identifies weak lines through key evaluation indicators and sensitivity analysis; complex network theory studies the relationship between nodes and edges in the power system using graph theory and network analysis techniques to identify key nodes and weak lines; data-driven methods combine historical and real-time monitoring data to identify weak lines in cascading failures using machine learning and data mining techniques; reliability evaluation methods evaluate the performance of lines under different failure scenarios by calculating the reliability indicators and failure rates of lines, and then identify weak lines that may trigger cascading failures. However, existing methods for identifying weak lines in cascading failures mainly focus on the power flow transfer characteristics of overload-dominated cascading failures, and do not fully consider the voltage change process caused by failures and the changes in the operation mode of power electronic equipment, so they are not suitable for identifying weak lines in voltage-dominated cascading failures.

[0006] Based on the above background, it is necessary to develop a weak line identification method for voltage-dominated cascading failures to address the shortcomings of existing technologies in identifying voltage-dominated cascading failures and improve the stability and reliability of power systems. SUMMARY

[0007] To address the shortcomings of the existing technology, the present application provides a method for identifying weak lines in voltage-dominated cascading failures in a power system. This method uses the topology of the power system, historical failure data, and simulation data to construct a voltage estimation model. For each failure scenario, it calculates the node voltage and determines whether a voltage response event has occurred. Then, based on the voltage response event that has occurred, it determines whether it is a weak line. This application effectively solves the problem of identifying weak lines in voltage-dominated cascading failures in power systems, providing better technical support for the operation and maintenance of power electronic power systems.

[0008] To solve the above technical problems, the present application adopts the following technical solutions:

[0009] A method for identifying weak lines in voltage-dominated cascading failures in a power system, characterized by the following steps:

[0010] S1, a set of failure scenarios to be searched is established for the power system, and a power system voltage estimation knowledge graph is constructed based on the topology of the power system and historical failure data; the historical failure data includes the active power, reactive power output by power electronic equipment, and the voltage of the access node under short-circuit failure of the power system;

[0011] S2, select one fault scenario from the set of fault scenarios, and determine a white-box tree-dendritic network voltage estimation model of the fault scenario by using the power system voltage estimation knowledge graph;

[0012] S3, estimate the power system node voltage of the fault scenario by using the white-box tree-dendritic network voltage estimation model;

[0013] S4, determine whether a voltage response event will occur according to the estimated power system node voltage; if yes, identify the voltage response event that will occur, and perform step S5; otherwise, return to step S2;

[0014] S5, determine whether the voltage response event is power electronic equipment lockout or off-grid according to the active power and reactive power output by the power electronic equipment after the voltage response event occurs; if yes, perform step S7; otherwise, perform step S6;

[0015] S6, re-estimate the power system node voltage of the fault scenario after the voltage response event occurs by using the white-box tree-dendritic network voltage estimation model of the fault scenario according to the active power and reactive power output by the power electronic equipment after the voltage response event occurs, and return to step S4;

[0016] S7, determine that the line on which the fault point in the current fault scenario is located is a weak line; then, determine whether each fault scenario in the set of fault scenarios has been searched; if yes, return to step S2 to select a fault scenario; otherwise, end the search;

[0017] Thus, each fault scenario in the set of fault scenarios is searched, and the weak line identification operation of each fault scenario is completed, and the cascading failure weak line identification result of the power system is obtained.

[0018] Specifically, in step S1, the set of fault scenarios is constructed as follows:

[0019] Select any transmission line in the power system, select a position in each preset designated position on the transmission line to set a short-circuit fault point, the fault condition of the short-circuit fault point is one of three-phase short-circuit, two-phase short-circuit or single-phase short-circuit, and the fault transition impedance value of the short-circuit fault point is set in the preset fault transition impedance value interval range, thereby forming a fault scenario;

[0020] Different transmission lines selected in the power system, or different positions in the preset designated positions on the transmission line to set the short-circuit fault point, or different fault conditions of the short-circuit fault point, or different fault transition impedance values of the short-circuit fault point, are all respectively regarded as forming different fault scenarios;

[0021] This results in several different fault scenarios forming a fault scenario set.

[0022] Specifically, in step S1, the power system voltage estimation knowledge graph is constructed as follows:

[0023] The initial network architecture of the power system is used as the first-level entity of the knowledge graph. Faults on different transmission lines within the power system form a basic fault network, which serves as the second-level entity of the knowledge graph. Faults at different locations on each transmission line form a detailed fault network, which serves as the third-level entity of the knowledge graph. Different fault transition impedance values ​​at each fault location form specific fault scenarios, which serve as the fourth-level entity of the knowledge graph. The white-box dendritic network voltage estimation model under specific fault scenarios serves as the attribute of the fourth-level entity of the knowledge graph. The relationships between first- and second-level entities are based on the fault line, between second- and third-level entities on the fault location, and between third- and fourth-level entities on the fault transition impedance. Thus, a power system voltage estimation knowledge graph is constructed.

[0024] Specifically, in step S2, the voltage estimation model for the white-box dendritic network is determined as follows:

[0025] Using a power system voltage estimation knowledge graph, we determine the corresponding secondary entities based on the faulty lines in the fault scenario; the corresponding tertiary entities based on the fault location in the fault scenario; the corresponding quaternary entities based on the fault transition impedance in the fault scenario; and the white-box dendritic network voltage estimation model for the fault scenario based on the relationship between the quaternary entities and their attributes.

[0026] Specifically, in step S3, the method for estimating the power system node voltages in this fault scenario using the white-box dendritic network voltage estimation model is as follows:

[0027] ;

[0028] , ;

[0029] In the formula, This represents the node voltage of the power system in this fault scenario; and The first is the voltage estimation model of the white box dendritic network. Layer and first The nonlinear output of the layer, , This indicates the number of dendrite layers in the white-box dendrite network voltage estimation model. This represents the nonlinear output of the final layer in the white-box dendritic network voltage estimation model; This represents matrix multiplication. denotes a Hadamard product operation; is the weight parameter of the lth layer of the white-box dendritic network voltage estimation model; is the injection power vector of the nodes of the power system, wherein are the active power and the reactive power injected into the nodes of the power system, respectively, , is the number of nodes in the power system;

[0030] The power of the power electronic equipment at the moment of the fault occurrence is unchanged, and the injection power vector of the nodes of the power system is determined according to the active power and the reactive power output by the power electronic equipment before the fault, so as to estimate the node voltage after the fault. After the voltage response event occurs, the injection power vector of the nodes of the power system is determined according to the active power and the reactive power output by the power electronic equipment after the event, so as to estimate the node voltage after the voltage response event.

[0031] Specifically, in step S4, the specific manner of judging whether the voltage response event will occur is that when the voltage of the access node of any power electronic equipment satisfies the starting voltage range of the voltage response event of the power electronic equipment, it is judged that the voltage response event will occur.

[0032] The specific manner of identifying the voltage response event to be occurred is that if only one starting voltage range of the voltage response event is satisfied, it is determined that the voltage response event is the voltage response event to be occurred; if multiple starting voltage ranges of the voltage response events are satisfied, it is determined that the voltage response event with the shortest remaining action time is the voltage response event to be occurred.

[0033] The remaining action time of the voltage response event is calculated in the following manner:

[0034] ;

[0035] wherein denotes the remaining action time of the voltage response event ; denotes the trigger action time or the end action time of the voltage response event , denotes the time when the voltage satisfies the trigger or end voltage range of the voltage response event , denotes the action time variable of the voltage response event, denotes the trigger action time when = trigger, denotes the end action time when = end; is the current time. ​

[0036] Specifically, in step S5, the active and reactive power output by the power electronic equipment after the voltage response event occurs is determined as follows:

[0037] ;

[0038] wherein, and are the active and reactive power output by the power electronic equipment of the access node after the voltage response event occurs; represents the voltage of the node ; represents the start voltage range or the maintain voltage range of the voltage response event ; represents the duration for which the device grid-connected point voltage satisfies the start voltage range or the maintain voltage range of the voltage response event ; represents the voltage range variable of the voltage response event, = start represents the start voltage range, = maintain represents the maintain voltage range; and respectively represent the mathematical relationship of the active power and the reactive power output by the power electronic equipment of the access node after the voltage response event occurs, with respect to the voltage of the node ;

[0039] If the power electronic equipment is off-grid / locked out, and are zero; if the voltage support control of the power electronic equipment and the high voltage ride-through control are started, and are determined by the control strategy; if the crowbar of the VSC type power electronic equipment is put in or the commutation failure of the LCC type power electronic equipment occurs, and are obtained from the state equation of the power electronic equipment or according to the fitting of the power-voltage external characteristic curve of the power electronic equipment.

[0040] Specifically, in step S6, according to the active and reactive power output by the power electronic equipment after the voltage response event occurs, the specific way of re-estimating the voltage of the power system node of the fault scenario after the above voltage response event occurs by using the white-box tree spike network voltage estimation model of the fault scenario is as follows:

[0041] ;

[0042] , ;

[0043] wherein, represents the node voltage of the power system of the fault scenario; and are the nonlinear outputs of the first layer and the second layer of the white-box dendritic network voltage estimation model, respectively, , represents the number of dendritic layers of the white-box dendritic network voltage estimation model, represents the nonlinear output of the final layer of the white-box dendritic network voltage estimation model; represents the matrix multiplication operation; represents the Hadamard product operation; is the weight parameter of the lth layer of the white-box dendritic network voltage estimation model; is the power vector output by the power electronic equipment after the voltage response event occurs, , and are the active power and the reactive power output by the power electronic equipment connected to the node after the voltage response event occurs, respectively, , is the number of nodes in the power system.

[0044] In a second aspect, the present application further provides a power system voltage dominant cascading failure weak line identification system, comprising a data acquisition module, a display module, a first storage module, a second storage module, a first calculation module, a second calculation module, a third calculation module, a fourth calculation module, a first comparison module, a second comparison module, and a third comparison module.

[0045] The data acquisition module is used to acquire the active power, the reactive power, and the voltage of the node output by the power electronic equipment of the power system.

[0046] The display module is used to display the cascading failure weak line identification result of the power system.

[0047] The first storage module is used to store the power system voltage estimation knowledge graph.

[0048] The second storage module is used to store the starting voltage range of the voltage response event and the action time of the voltage response event.

[0049] The first calculation module estimates and calculates the node voltage of the power system according to the active power and the reactive power output by the electronic equipment collected by the data acquisition module at the moment of the fault, or the active power and the reactive power output by the power electronic equipment calculated by the third calculation module after the voltage response event occurs.

[0050] The second calculation module is used for calculating the remaining action time of the voltage response event.

[0051] The third calculation module is used for calculating the active power and the reactive power output by the power electronic equipment after the voltage response event occurs.

[0052] The fourth calculation module trains the weight parameters of the white-box dendritic network voltage estimation model according to the historical data of the active power, the reactive power and the access node voltage collected by the data acquisition module, and obtains the white-box dendritic network voltage estimation model with the determined weight parameters.

[0053] The first comparison module is used for comparing whether the node voltage estimation value meets the starting voltage range of the voltage response event in the termination state or the maintaining voltage range of the voltage response event in the triggering state, and if yes, the second comparison module is called, otherwise, the first calculation module and the second calculation module are called for recalculation.

[0054] The second comparison module is used for comparing whether the remaining action time of the voltage response event is 0, and if yes, the third comparison module is called, otherwise, the third calculation module is called for recalculation.

[0055] The third comparison module is used for comparing and judging whether the voltage response event occurred is the power electronic equipment lockout or off-grid, and if yes, it is determined that the line on which the fault point in the current fault scenario is located is a weak line, and the display module is called to display the cascading failure weak line identification result of the power system, otherwise, the third calculation module is called for recalculation.

[0056] In a third aspect, the present application further provides a power system voltage dominant cascading failure weak line identification device, comprising a processor and a storage medium; the storage medium is used for storing a computer program; the processor is connected with the storage medium, and is used for executing the computer program stored in the storage medium, so that the power system voltage dominant cascading failure weak line identification device executes the power system voltage dominant cascading failure weak line identification method described above.

[0057] Compared with the prior art, the present application has the following beneficial effects:

[0058] 1. The prior art focuses on overload or topology failure dominant power system cascading failure, which cannot reflect the process of secondary response of power electronic equipment due to voltage disturbance, and cannot be used for voltage dominant cascading failure; the present application identifies the influence of node voltage change on power electronic equipment and the effect of power electronic equipment control or protection response on node voltage, and can identify the problem of voltage dominant weak line.

[0059] 2. The prior art assesses the line most likely to trigger large-scale power outage under various scenarios through failure rate, Monte Carlo simulation and other probability methods; the present application models the dynamic evolution process of cascading failure, describes the voltage response mode of power electronic equipment, and identifies voltage dominant weak line through physical and data driven way, which has the advantages of high efficiency and accuracy.

[0060] 3. The prior art simulates the dynamic process of equipment after line disconnection to identify the key line or failure mode that leads to system collapse, but detailed modeling of power electronic power system is extremely difficult, and voltage dominant cascading failure is an electromagnetic transient process with huge calculation amount, so simulation is not applicable; the present application can quickly identify the weak line of the most voltage dominant cascading failure in a short time by constructing power system voltage estimation knowledge graph and using white box tree spike network for fast reasoning of node voltage. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to make the purpose, technical scheme and advantages of the application clearer, the application will be further described in detail below with reference to the drawings, in which:

[0062] Figure 1 The flow chart of the power system voltage dominant cascading failure weak line identification method of the present application;

[0063] Figure 2 The structure diagram of power system containing power electronic equipment access;

[0064] Figure 3 The simulation waveform of the grid-connected point voltage, active power and reactive power of power electronic equipment during line L 10-11 DETAILED DESCRIPTION

[0065] ​In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present application.

[0066] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present application.

[0067] The present application provides a method for identifying weak lines of voltage dominant cascading failures in a power system, which uses the topological structure of the power system, historical failure data and simulation data to construct a voltage estimation model. For each failure scenario, the node voltage is calculated and it is determined whether a voltage response event occurs. Then, it is determined whether the weak line is a weak line according to the voltage response event. The present application effectively solves the problem of identifying weak lines of voltage dominant cascading failures in a power system, and provides technical support for the operation and maintenance of power electronic power systems.

[0068] Based on the above technical design idea, the present application provides a method for identifying weak lines of voltage dominant cascading failures in a power system, and the specific process is as shown in Figure 1 The method comprises the following steps:

[0069] S1, a set of failure scenarios to be searched is established for a power system, and a voltage estimation knowledge graph of the power system is constructed according to the topological structure of the power system and historical failure data. The historical failure data includes active power, reactive power and voltage of the access node output by power electronic equipment under short-circuit failure of the power system.

[0070] Specifically, the power electronic equipment on the line in the power system includes wind farms, photovoltaic power stations, energy storage power stations, direct current converter stations and the like.

[0071] S2, a failure scenario is selected from the set of failure scenarios, and a white-box tree spike network voltage estimation model of the failure scenario is determined by using the voltage estimation knowledge graph of the power system.

[0072] S3, estimating the node voltage of the power system in the fault scenario by using the white-box tree-branch network voltage estimation model;

[0073] S4, judging whether a voltage response event will occur according to the estimated node voltage of the power system, if yes, identifying the voltage response event to occur, and performing step S5, otherwise, returning to step S2;

[0074] S5, judging whether the voltage response event is power electronic equipment lockout or off-grid according to the active power and reactive power output by the power electronic equipment after the voltage response event occurs, if yes, performing step S7, otherwise, performing step S6;

[0075] S6, re-estimating the node voltage of the power system in the fault scenario after the voltage response event occurs by using the white-box tree-branch network voltage estimation model of the fault scenario according to the active power and reactive power output by the power electronic equipment after the voltage response event occurs, and returning to step S4;

[0076] S7, determining that the line where the fault point in the current fault scenario is located is a weak line, then judging whether each fault scenario in the fault scenario set has been searched, if yes, returning to step S2 to select a fault scenario, otherwise, ending the search;

[0077] Thus, each fault scenario in the fault scenario set is searched, the weak line identification operation of each fault scenario is completed, and the cascading failure weak line identification result of the power system is obtained.

[0078] In specific implementation, in step S1, the fault scenario set is constructed as follows:

[0079] Any transmission line in the power system is selected, a short-circuit fault point is set at a position in each preset designated position on the transmission line, the fault condition of the short-circuit fault point is one of three-phase short-circuit, two-phase short-circuit or single-phase short-circuit, and the fault transient impedance value of the short-circuit fault point is set in the preset fault transient impedance value interval range, thereby forming a fault scenario;

[0080] Different transmission lines are selected in the power system, or different positions in the preset designated positions on the transmission line are selected to set the short-circuit fault point, or different fault conditions are set for the short-circuit fault point, or different fault transient impedance values are set for the short-circuit fault point, which are respectively regarded as forming different fault scenarios;

[0081] Thus, the fault scenario set is formed by a plurality of different fault scenarios.

[0082] For example, if a designated position is preset every 5% length of the transmission line, the impedance value range is 0 ohm to 1000 ohm, and the integer is taken; when the fault scene set is constructed, each transmission line in the power system can be selected in turn, a designated position of 5% length of the transmission line is selected, a short-circuit fault point is selected in turn, three-phase short-circuit, two-phase short-circuit and single-phase short-circuit are set in turn to form three forms of fault conditions, and the fault transition impedance is set once every 1 ohm from 0 ohm to 1000 ohm, and 1000 different fault conditions can be set in each fault condition; thus, any fault condition formed is regarded as a fault scene, and the set of all fault conditions constitutes the fault scene set.

[0083] In specific implementation, in step S1, the power system voltage estimation knowledge graph is constructed in the following manner:

[0084] The initial network framework of the power system is taken as a first-level entity of the knowledge graph; the basic fault network framework formed by different transmission line faults in the power system is taken as a second-level entity of the knowledge graph; the detailed fault network framework formed by different fault positions of each transmission line is taken as a third-level entity of the knowledge graph; the specific fault scene formed by different fault transition impedance values under each fault position is taken as a fourth-level entity of the knowledge graph; the white-box tree-dendritic network voltage estimation model under the specific fault scene is taken as an attribute of the fourth-level entity of the knowledge graph; the fault line is taken as the relationship between the first-level entity and the second-level entity, the fault position is taken as the relationship between the second-level entity and the third-level entity, and the fault transition impedance is taken as the relationship between the third-level entity and the fourth-level entity; thus, the power system voltage estimation knowledge graph is constructed.

[0085] In specific implementation, in step S2, the white-box tree-dendritic network voltage estimation model is determined in the following manner:

[0086] The power system voltage estimation knowledge graph is used to determine the corresponding second-level entity according to the fault line of the fault scene, determine the corresponding third-level entity according to the fault position corresponding to the fault scene, determine the corresponding fourth-level entity according to the fault transition impedance corresponding to the fault scene, and determine the white-box tree-dendritic network voltage estimation model of the fault scene according to the relationship between the fourth-level entity and the attribute.

[0087] In specific implementation, in step S4, the white-box tree-dendritic network voltage estimation model is used to calculate the node voltage of the power system in the fault scene in the following manner:

[0088]

[0089] In the formula, V is the node voltage of the power system in the fault scene, and ​​respectively, are the nonlinear outputs of the first and the second layers of the white-box dendritic network voltage estimation model, , represents the number of layers of the dendritic layer of the white-box dendritic network voltage estimation model, represents the nonlinear output of the final layer of the white-box dendritic network voltage estimation model; represents a matrix multiplication operation; represents a Hadamard product operation; is a weight parameter of the lth layer of the white-box dendritic network voltage estimation model; is an injection power vector of nodes of a power system, wherein are active power and reactive power injected into nodes , , is the number of nodes in the power system;

[0090] The power of the power electronic equipment at the moment of fault occurrence is unchanged, and the injection power vector of the nodes of the power system is determined according to the active power and the reactive power output by the power electronic equipment before the fault, so as to estimate the node voltage after the fault. After the voltage response event occurs, the injection power vector of the nodes of the power system is determined according to the active power and the reactive power output by the power electronic equipment after the event, so as to estimate the node voltage after the voltage response event.

[0091] In a specific implementation, the weight parameters of each layer in the white-box dendritic network voltage estimation model are determined in the following manner:

[0092] The injection power vector of the nodes of the power system and the node voltage form a sample pair, represents the voltage of the node , the voltage of the node of the power system is calculated through a forward operation, and then the weight parameters are updated according to the estimated value and the actual value through a back propagation algorithm; this is repeated until the weight parameters converge, thereby forming the weight parameters of the white-box dendritic network voltage estimation model under the current fault scenario;

[0093] Wherein, the forward operation is implemented in the following manner:

[0094] ;

[0095] The back propagation algorithm updates the weight parameters in the following manner:

[0096] ;

[0097] In the formula, M is the number of samples used for single training.​

[0098] In implementation, in step S5, whether a voltage response event occurs is determined as follows:

[0099] When the voltage at any access node of the power electronic equipment meets the starting voltage range of the voltage response event of the power electronic equipment, it is determined that the voltage response event will occur.

[0100] The voltage response event that occurs is identified as follows:

[0101] If only one starting voltage range of the voltage response event is met, the voltage response event is the voltage response event that will occur; if multiple starting voltage ranges of the voltage response events are met, the voltage response event with the shortest remaining action time is the voltage response event that will occur.

[0102] In implementation, the remaining action time of the voltage response event is calculated as follows:

[0103] ;

[0104] In the formula, represents the remaining action time of the voltage response event ; represents the trigger action time or the end action time of the voltage response event , represents the time when the voltage meets the trigger or end voltage range of the voltage response event , represents the action time variable of the voltage response event, = trigger represents the trigger action time, = end represents the end action time; is the current time.

[0105] In implementation, in step S5, the active and reactive power output by the power electronic equipment after the voltage response event occurs is determined as follows:

[0106] ;

[0107] In the formula, and are the active and reactive power output by the power electronic equipment at the access node after the voltage response event occurs; represents the voltage at the node ; represents the starting voltage range or the holding voltage range of the voltage response event , denotes the voltage range of the voltage response event , the duration of the start voltage range or the maintain voltage range, denotes the voltage range variable of the voltage response event, denotes the start voltage range when = start, denotes the maintain voltage range when = maintain; and denote the active power and the reactive power output by the power electronic equipment after the voltage response event occurs; and respectively; the mathematical relationship of the node voltage;

[0108] if the power electronic equipment is off-grid / locked out, and are zero; if the voltage support control of the power electronic equipment, the high voltage ride-through control is started, and are determined by the control strategy; if the crowbar of the VSC type power electronic equipment is put in, the commutation failure of the LCC type power electronic equipment occurs, and are obtained from the state equation of the power electronic equipment or according to the power-voltage external characteristic curve fitting of the power electronic equipment.

[0109] In a specific implementation, in step S6, according to the active power and the reactive power output by the power electronic equipment after the voltage response event occurs, the specific manner of re-estimating the node voltage of the power system of the fault scenario after the voltage response event is as follows by using the white-box tree network voltage estimation model of the fault scenario:

[0110]

[0111] Similarly, in the formula, denotes the node voltage of the power system of the fault scenario; and are the nonlinear outputs of the first layer and the first layer of the white-box tree network voltage estimation model respectively, , denotes the number of layers of the tree layer of the white-box tree network voltage estimation model, denotes the nonlinear output of the final layer of the white-box tree network voltage estimation model; denotes the matrix multiplication operation; denotes the Hadamard product operation; The weight parameter of the lth layer of the white-box tree network voltage estimation model is The power vector output by the power electronic equipment after the voltage response event occurs, , And The active power and the reactive power output by the power electronic equipment of the access node after the voltage response event occurs, , The number of nodes in the power system.

[0112] Compared with the prior art, the present application has the following technical advantages:

[0113] 1. The prior art focuses on overload or topology failure dominant power system cascading failure, and cannot reflect the process of secondary response of power electronic equipment due to voltage disturbance, and cannot be used for voltage dominant cascading failure. The present application identifies the influence of node voltage change on power electronic equipment and the effect of power electronic equipment control or protection response on node voltage, and can identify the problem of voltage dominant weak line.

[0114] 2. The prior art assesses the line most likely to trigger large-scale power outage under various scenarios through failure rate, Monte Carlo simulation and other probability methods; the present application models the dynamic evolution process of cascading failure, describes the voltage response mode of power electronic equipment, and identifies the voltage dominant weak line through the joint driving of physics and data, with the advantages of high efficiency and accuracy.

[0115] 3. The prior art simulates the dynamic process of equipment after line disconnection to identify the key line or failure mode that causes system collapse, but detailed modeling of power electronic power system is extremely difficult, and voltage dominant cascading failure is an electromagnetic transient process with huge calculation amount, so simulation is not applicable; the present application builds a power system voltage estimation knowledge graph, uses a white-box tree network to quickly reason the node voltage, and can quickly identify the weak line of the most voltage dominant cascading failure in a short time.

[0116] In the second aspect, the present application further provides a power system voltage dominant cascading failure weak line identification system, comprising a data acquisition module, a display module, a first storage module, a second storage module, a first calculation module, a second calculation module, a third calculation module, a fourth calculation module, a first comparison module, a second comparison module and a third comparison module.

[0117] The data acquisition module is used to acquire the active power, the reactive power output by the power electronic equipment of the power system and the voltage of the access node;

[0118] The display module is used to display the cascading failure weak line identification result of the power system.

[0119] The first storage module is used for storing the power system voltage estimation knowledge graph.

[0120] The second storage module is used for storing the starting voltage range of the voltage response event and the action time of the voltage response event.

[0121] The first calculation module estimates and calculates the power system node voltage according to the active power and the reactive power output by the electronic equipment collected by the data acquisition module at the fault moment, or the active power and the reactive power output by the power electronic equipment after the voltage response event occurs calculated by the third calculation module.

[0122] The second calculation module is used for calculating the remaining action time of the voltage response event.

[0123] The third calculation module is used for calculating the active power and the reactive power output by the power electronic equipment after the voltage response event occurs.

[0124] The fourth calculation module trains the weight parameters of the white-box dendritic network voltage estimation model according to the historical data of the active power, the reactive power and the access node voltage collected by the data acquisition module, to obtain the white-box dendritic network voltage estimation model with the determined weight parameters.

[0125] The first comparison module is used for comparing whether the node voltage estimation value meets the starting voltage range of the voltage response event in the termination state or the holding voltage range of the voltage response event in the triggering state; if yes, the second comparison module is called; otherwise, the first calculation module and the second calculation module are called for recalculation.

[0126] The second comparison module is used for comparing whether the remaining action time of the voltage response event is 0; if yes, the third comparison module is called; otherwise, the third calculation module is called for recalculation.

[0127] The third comparison module is used for comparing and judging whether the occurred voltage response event is the power electronic equipment lockout or off-grid; if yes, it is determined that the line on which the fault point in the current fault scene is located is a weak line, and the display module is called to display the cascading failure weak line identification result of the power system; otherwise, the third calculation module is called for recalculation.

[0128] The power system voltage dominant cascading failure weak line identification system is designed to execute the power system voltage dominant cascading failure weak line identification method provided in the present application, and has the corresponding technical advantages of the method.

[0129] ​Thirdly, the present invention also provides a power system voltage-dominated cascading fault weak line identification device, characterized in that it includes a processor and a storage medium; the storage medium is used to store a computer program; the processor is connected to the storage medium and is used to execute the computer program stored in the storage medium, so that the power system voltage-dominated cascading fault weak line identification device performs the power system voltage-dominated cascading fault weak line identification method provided above by the present invention.

[0130] The storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0131] The code for a computer program that performs the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and ++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer.

[0132] Example:

[0133] To verify the effectiveness of the present invention, as follows Figure 2 The following analysis will be based on a schematic diagram of a power system structure including the access of power electronic equipment. Figure 2In the embodiment shown, Bus 30, 31, 34, 36, 37 are synchronous machine nodes, with rated power of 250 MW, 520 MW, 508 MW, 560 MW, and 540 MW respectively; Bus 32, 33, 38 are doubly-fed wind farm AC grid-connected nodes, with rated power of 650 MW, 632 MW, and 830 MW respectively; Bus 39 is a conventional DC transmission system (LCC-HVDC) access point, with rated power of 1000 MW; Bus 35 is a flexible DC transmission system (VSC-HVDC) access point, with rated power of 650 MW; Bus 1, 3, 4, 7-9, 12, 15, 16, 18, 20, 21, 23-29, 39 are load nodes, with total active load of 6254.23 MW and total reactive load of 1387 Mvar.

[0134] The blocking of the wind farm and the VSC-HVDC is determined according to the grid code, that is, the under-voltage (<0.9 p.u.) or over-voltage (>1.1 p.u.) at the wind farm or VSC-HVDC grid-connected point lasts for more than the setting time, and the setting time is referred to in the Technical Regulation for Access of Wind Farms to Power Systems. The LCC-HVDC is blocked when continuous commutation failure or low voltage occurs for more than 3 times or for too long, and the low voltage threshold value is selected as 0.75 p.u. and the maximum allowable low voltage duration is selected as the typical value of 0.5 s.

[0135] In the embodiment, a three-phase metallic ground fault occurs in the power system, the fault transition impedance is 1 Ω, the fault duration is 1 s, the fault position is set at intervals of 10%, and all lines are traversed. Figure 2 The system shown has a total of 46 lines, each line has 9 fault positions, and the fault transition impedance is 1 Ω, thereby generating 414 fault scenarios. In each fault scenario, the power of the PEE injection node is changed, the power and voltage data are collected, the white-box dendritic network voltage estimation model training samples are formed, and the white-box dendritic network voltage estimation model training weight parameters are obtained. Therefore, the white-box dendritic network voltage estimation model can be constructed as the attribute of the power system voltage estimation knowledge graph. Figure 2 The grid structure shown is a first-level entity, which has 46 second-level entities according to the difference in fault lines, each second-level entity has 9 third-level entities according to the difference in fault positions, and each third-level entity has 1 fourth-level entity, and each fourth-level entity has the white-box dendritic network voltage estimation model as the attribute of the power system voltage estimation knowledge graph.

[0136] In the embodiment, the wind farm is disconnected, the VSC-HVDC is blocked, or the LCC-HVDC is blocked as the key event search, and a total of 29 voltage-dominant cascading fault weak lines are obtained, which are L 1-2 , L 1-39 , L 2-3 , L 2-25 , L3-4 L 3-18 L 4-5 L 4-14 L 5-6 L 5-8 L 6-7 L 6-11 L 7-8 L 8-9 L 9-39 L 10-11 L 10-13 L 13-14 L 14-15 L 15-16 L 16-17 L 16-21 L 16-24 L 17-18 L 17-27 L 21-22 L 22-23 L 25-26 L 28-29 The key events of weak lines and the cascading failures they cause are shown in Table 1.

[0137] Table 1. Critical events and vulnerable lines in a cascading failure pattern.

[0138]

[0139] In this embodiment, for Figure 2 Simulations were performed on all lines of the power system shown, and the resulting power electronic equipment disconnection and blocking events caused by line faults, along with their corresponding times, are shown in Table 2. The power electronic equipment disconnection and blocking events caused by the faults correspond well with the search results shown in Table 1. Therefore, the proposed method for searching weak lines in voltage-dominated cascading faults in power electronic systems has good accuracy and can effectively search for weak lines in power systems with voltage-dominated cascading faults.

[0140] Table 2 Power electronic equipment disconnection and blocking events and their corresponding occurrence times

[0141]

[0142] In this embodiment, Figure 3 The simulated waveforms of the grid-connected voltage U, active power P, and reactive power Q of the power electronic equipment during a fault on line L10-11 are shown. It can be seen that after the fault persists for a period of time, the active and reactive power of LCC-HVDC 39 and wind farm 32 change abruptly and stabilize at 0, indicating that LCC-HVDC 39 is blocked and wind farm 32 is disconnected from the grid. Based on the simulated power waveforms, it can be determined whether a line fault will lead to the blocking or disconnection of the power electronic equipment.

[0143] In summary, the application provides a power system voltage dominant cascading failure weak line identification method, system and device, which constructs a voltage estimation knowledge graph according to the topological structure, fault history data and simulation data of the power system; a fault scene set is established, each fault scene is selected in turn, a corresponding white box tree spike network model is determined by using the knowledge graph, and the node voltage under the scene is calculated; whether a voltage response event occurs is judged according to the calculation result, if the condition is met, the event is identified and whether it involves device lockout or off-grid is judged, so whether the line is a weak line is determined; if the condition is not met, other fault scenes are searched. The method can accurately identify the weak line in the power system, and provide a reliable basis for the design, operation control and protection measure optimization of the power system.

[0144] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit the technical solutions. Those of ordinary skill in the art should understand that modifications or equivalent replacements to the technical solutions of the present application without departing from the purpose and scope of the technical solutions should be covered in the scope of the claims of the present application.

Claims

1. A method for identifying weak lines in power systems with voltage-dominated cascading faults, characterized in that, Includes the following steps: S1. Establish a set of fault scenarios to be searched for the power system, and construct a power system voltage estimation knowledge graph based on the power system topology and historical fault data; the historical fault data includes the active power, reactive power output of power electronic equipment and the voltage of the access node under power system short-circuit faults. S2. Select a fault scenario from the set of fault scenarios, and use the power system voltage estimation knowledge graph to determine the white-box dendritic network voltage estimation model for the fault scenario. S3. Using the white-box dendritic network voltage estimation model, estimate the power system node voltage for this fault scenario; S4. Based on the estimated power system node voltage, determine whether a voltage response event will occur; if so, identify the voltage response event to be occurred and proceed to step S5; otherwise, return to step S2. S5. Based on the active and reactive power output by the power electronic equipment after the voltage response event occurs, determine whether the voltage response event is a power electronic equipment blocking or disconnection from the grid; if so, proceed to step S7. Otherwise, proceed to step S6; S6. Based on the active and reactive power output by the power electronic equipment after the voltage response event, use the white-box dendritic network voltage estimation model of the fault scenario to re-estimate the power system node voltage after the above voltage response event, and return to step S4. S7. Determine that the line where the fault point is located in the current fault scenario is a weak line; then, determine whether every fault scenario in the fault scenario set has been traversed and searched; if yes, return to step S2 to select a fault scenario again; otherwise, end the search. Therefore, by traversing and searching each fault scenario in the set of fault scenarios, the weak line identification operation for each fault scenario is completed, and the weak line identification result of the cascading faults in the power system is obtained.

2. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 1, characterized in that, In step S1, the set of fault scenarios is constructed in the following manner: Select any transmission line in the power system, select one of the preset designated locations on the transmission line to set a short-circuit fault point. The fault condition of the short-circuit fault point is one of three-phase short circuit, two-phase short circuit or single-phase short circuit. Set the fault transition impedance value of the short-circuit fault point within the preset fault transition impedance value range, thereby forming a fault scenario. Different transmission lines selected in the power system, or different locations set as short-circuit fault points in the preset designated locations on the transmission lines, or different fault conditions set at the short-circuit fault points, or different fault transition impedance values ​​set at the short-circuit fault points, are all considered to form different fault scenarios. This results in several different fault scenarios forming a fault scenario set.

3. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 2, characterized in that, In step S1, the power system voltage estimation knowledge graph is constructed as follows: The initial network architecture of the power system is used as the first-level entity of the knowledge graph. Faults on different transmission lines within the power system form a basic fault network, which serves as the second-level entity of the knowledge graph. Faults at different locations on each transmission line form a detailed fault network, which serves as the third-level entity of the knowledge graph. Different fault transition impedance values ​​at each fault location form specific fault scenarios, which serve as the fourth-level entity of the knowledge graph. The white-box dendritic network voltage estimation model under specific fault scenarios serves as the attribute of the fourth-level entity of the knowledge graph. The relationships between first- and second-level entities are based on the fault line, between second- and third-level entities on the fault location, and between third- and fourth-level entities on the fault transition impedance. Thus, a power system voltage estimation knowledge graph is constructed.

4. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 3, characterized in that, In step S2, the voltage estimation model for the white-box dendritic network is determined as follows: Using a power system voltage estimation knowledge graph, we determine the corresponding secondary entities based on the faulty lines in the fault scenario; the corresponding tertiary entities based on the fault location in the fault scenario; the corresponding quaternary entities based on the fault transition impedance in the fault scenario; and the white-box dendritic network voltage estimation model for the fault scenario based on the relationship between the quaternary entities and their attributes.

5. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 1, characterized in that, In step S3, the specific method for estimating the power system node voltages for this fault scenario using the white-box dendritic network voltage estimation model is as follows: ； ， ； In the formula, This represents the node voltage of the power system in this fault scenario; and The first is the voltage estimation model of the white box dendritic network. Layer and first The nonlinear output of the layer, , This indicates the number of dendrite layers in the white-box dendrite network voltage estimation model. This represents the nonlinear output of the final layer in the white-box dendritic network voltage estimation model; This represents matrix multiplication. This represents the Hadamard product operation; The weight parameters of the l-th layer in the white-box dendritic network voltage estimation model; This represents the injected power vector for a power system node. ,in Injected into the node respectively Active power and reactive power, , The number of nodes in the power system; At the instant a fault occurs, the power output of the power electronic equipment remains unchanged. The injected power vector of the power system node is determined based on the active and reactive power output of the power electronic equipment before the fault. This allows for the estimation of node voltages after a fault; after a voltage response event occurs, the injected power vector of the power system node is determined based on the active and reactive power output of the power electronic equipment after the event. This allows for the estimation of node voltages after a voltage response event occurs.

6. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 1, characterized in that, In step S4, the specific method for determining whether a voltage response event will occur is as follows: if the voltage of any access node of any power electronic equipment meets the starting voltage range of the voltage response event of that power electronic equipment, then it is determined that a voltage response event will occur. The specific method for identifying the voltage response event to occur is as follows: if the starting voltage range of only one voltage response event is met, then that voltage response event is determined to be the voltage response event to occur; if the starting voltage range of multiple voltage response events is met, then the voltage response event with the shortest remaining action time is determined to be the voltage response event to occur. The remaining action time of the voltage response event is calculated as follows: ； In the formula, Indicates voltage response event The remaining action time; Indicates voltage response event The trigger time or termination time of the action. This indicates that the voltage meets the voltage response event. The moment when the voltage range is triggered or terminated. This represents the action time variable of a voltage response event. = trigger indicates the time when the action is triggered. = end indicates the time to terminate the action; This refers to the current moment.

7. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 1, characterized in that, In step S5, the active and reactive power output by the power electronic equipment after the voltage response event occurs is determined as follows: ； In the formula, and Connecting nodes after voltage response events occur The active and reactive power output of the power electronic equipment; Represents a node The voltage; Indicates voltage response event The start-up voltage range or holding voltage range, This indicates that the voltage at the device's grid connection point meets the voltage response event. The duration of the start-up voltage range or holding voltage range, The voltage range variable represents the voltage response event. = start indicates the starting voltage range. =maintain indicates that the voltage range is maintained; Indicates voltage response event The trigger time or termination time of the action. This represents the action time variable of a voltage response event. = trigger indicates the time when the action is triggered. = end indicates the time to terminate the action; and These represent voltage response events. Access node after occurrence Power electronic equipment output active power and reactive power Each regarding nodes Voltage Mathematical relationships; If the power electronic equipment is disconnected from the grid / locked and The value is zero; if the voltage support control and high voltage ride-through control of the power electronic equipment are activated, and Determined by the control strategy; if the VSC type power electronic equipment crowbar is engaged, or the LCC type power electronic equipment experiences commutation failure, and It is obtained from the state equation of the power electronic equipment or by fitting the power-voltage external characteristic curve of the power electronic equipment.

8. The method for identifying weak lines in power systems with voltage-dominant cascading faults according to claim 1, characterized in that, In step S6, based on the active and reactive power output by the power electronic equipment after the voltage response event, the specific method for re-estimating the power system node voltage after the aforementioned voltage response event using the white-box dendritic network voltage estimation model for this fault scenario is as follows: ； ， ； In the formula, This represents the node voltage of the power system in this fault scenario; and The first is the voltage estimation model of the white box dendritic network. Layer and first The nonlinear output of the layer, , This indicates the number of dendrite layers in the white-box dendrite network voltage estimation model. This represents the nonlinear output of the final layer in the white-box dendritic network voltage estimation model; This represents matrix multiplication. This represents the Hadamard product operation; The weight parameters of the l-th layer in the white-box dendritic network voltage estimation model; This represents the power vector output by the power electronic equipment after a voltage response event occurs. , and Connecting nodes after voltage response events occur The active and reactive power output of the power electronic equipment. , This represents the number of nodes in the power system.

9. A power system voltage-dominant type cascading fault weak line identification system, characterized in that, It includes a data acquisition module, a display module, a first storage module, a second storage module, a first calculation module, a second calculation module, a third calculation module, a fourth calculation module, a first comparison module, a second comparison module, and a third comparison module; The data acquisition module is used to collect the active power, reactive power, and voltage of the access nodes output by the power electronic equipment in the power system. The display module is used to display the identification results of weak lines with cascading faults in the power system. The first storage module is used to store a knowledge graph of power system voltage estimation; The second storage module is used to store the start voltage range and the action time of the voltage response event; The first calculation module estimates and calculates the node voltage of the power system based on the active and reactive power output of the electronic equipment collected by the data acquisition module at the moment of the fault, or the active and reactive power output of the power electronic equipment after the voltage response event is calculated by the third calculation module. The second calculation module is used to calculate the remaining action time of the voltage response event; The third calculation module is used to calculate the active power and reactive power output by the power electronic equipment after the voltage response event occurs. The fourth calculation module trains the weight parameters of the white-box dendritic network voltage estimation model based on the historical data of active power, reactive power and access node voltage collected by the data acquisition module, and obtains a white-box dendritic network voltage estimation model with determined weight parameters. The first comparison module is used to compare whether the estimated node voltage value meets the start voltage range of a voltage response event in the terminated state or the holding voltage range of a voltage response event in the triggered state; if so, the second comparison module is invoked. Otherwise, the first and second calculation modules are invoked for recalculation; The second comparison module is used to compare whether the remaining action time of the voltage response event is 0; if so, the third comparison module is called; otherwise, the third calculation module is called to recalculate. The third comparison module is used to compare and determine whether the voltage response event is a power electronic equipment lockout or disconnection; if so, it determines that the line where the fault point is located in the current fault scenario is a weak line, and calls the display module to display the identification result of the weak line of the cascading fault in the power system. Otherwise, the third calculation module is invoked for recalculation.

10. A device for identifying weak lines in a power system with voltage-dominated cascading faults, characterized in that, It includes a processor and a storage medium; the storage medium is used to store a computer program; the processor is connected to the storage medium and is used to execute the computer program stored in the storage medium, so that the power system voltage-dominated cascading fault weak line identification device performs the power system voltage-dominated cascading fault weak line identification method as described in any one of claims 1 to 8.

Images