New energy power system active splitting method and system based on instability mode

Abstract

The invention relates to the technical field of power system transient stability analysis and control, and discloses a new energy power system active splitting method and system based on an instability mode, and the method comprises the steps: collecting the current system power flow information, and carrying out the short-time simulation to generate fault set data; a convolutional neural network is adopted to extract fault set features to obtain an instability mode of the system; taking reduction of the absolute value of the transmission power of the cut-off line as an optimization target, and searching a system splitting section by adopting a maximum-flow minimum-cut algorithm; new energy frequency limitation is considered, secondary search is carried out on the splitting section, and an active splitting strategy is generated. According to the method, new energy frequency limitation is considered at the same time, active splitting control can be effectively carried out, and the applicability is high. The method can integrate offline data simulation training and system real-time operation data acquisition, can perform splitting control more quickly and effectively, and can be used for online application of stable control of a new energy power system.

Description

Technical Field

[0001] The present invention relates to the technical field of transient stability analysis and control of power systems, and in particular to a method and system for actively decoupling a new energy power system based on an unstable mode. Background Art

[0002] The large-scale connection of renewable energy power generation to the grid has brought new challenges to the safe and stable operation of the power system. As the last link of the three lines of defense of the power system, the decoupling control is an important means to maintain the safe and stable operation of the power system and avoid large-scale power outages. The traditional out-of-step decoupling control formulates corresponding control strategies through a large number of offline simulations, which is increasingly difficult to adapt to the complex and changeable operation modes and numerous accident collections brought about by the connection of renewable energy to the grid. Therefore, the study of the active decoupling method of the renewable energy power system has important theoretical significance and technical value for the stability control and recovery of the power system.

[0003] In the prior art, the power system out-of-step disconnection control determines whether disconnection is needed and the corresponding disconnection plan through changes in system frequency, power and oscillation center, and implements the disconnection operation at the preset section when the out-of-step judgment criterion is activated. However, with the continuous increase in the proportion of new energy and the addition of DC transmission channels, the operation mode of the system has become more complex and changeable, and the characteristics of the system after failure have become more complex. It is difficult to achieve effective control of the system based on out-of-step disconnection based on offline simulation. On the other hand, the active disconnection criterion based on fault triggering may not be able to achieve accurate action under special faults, and needs to be used together with other criteria to improve the accuracy of the judgment. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: how to accurately obtain the advantages of real-time operation data of the system so as to be able to perform decoupling control more quickly and effectively.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for actively decoupling a new energy power system based on an unstable mode, comprising:

[0007] Collect current system power flow information and perform short-term simulation to generate fault set data;

[0008] Convolutional neural network is used to extract fault set features to obtain the system instability mode;

[0009] Taking reducing the absolute value of the transmission power of the cut-off line as the optimization goal, the maximum flow minimum cut algorithm is used to search for the system decoupling section;

[0010] Considering the frequency limitation of new energy, a secondary search is conducted on the disconnection section to generate an active disconnection strategy.

[0011] As a preferred solution of the active disconnection method of the new energy power system based on the instability mode described in the present invention, wherein: the generation of fault set data includes obtaining real-time power flow information in the operation of the system, and modifying the power flow parameters of the system simulation model accordingly according to the real-time power flow information. The system power flow equation in the polar coordinate system is expressed as:

[0012]

[0013]

[0014] Where i and j represent the node numbers in the system, P and Q represent the active power injection and reactive power injection of the node respectively, U represents the node voltage amplitude, θ ij represents the voltage phase difference between nodes i and j, G ij represents the conductance between nodes i and j; B ij represents the susceptance between nodes i and j;

[0015] A short-term simulation traversal of the expected fault is performed in the simulation model containing the real power flow information of the system, the fault set data of the system under the current operation mode is quickly generated, and the N-1 power flow data of the system is generated according to the fault location.

[0016] As a preferred solution of the method for active decoupling of a new energy power system based on an instability mode described in the present invention, the convolutional neural network includes two convolutional neural networks constructed for system stability judgment and power angle instability mode recognition; the output of the neural network is a label required for network training, and the input is relevant parameters of system stability and instability mode; for nodes, voltage, phase angle and frequency deviation are selected as network inputs; for generators, power angle, speed and electromagnetic power are selected as network inputs; and the system instability mode is obtained through network training results.

[0017] As a preferred solution of the active decoupling method of the new energy power system based on the instability mode described in the present invention, wherein: the optimization goal of reducing the absolute value of the power transmission of the cut-off line includes: performing equivalence on the generator sets in the power system, and dividing the leading group and the remaining group according to the power angle instability mode; the generation of equivalent generator nodes mainly includes the system shortest path search; the system shortest path search is performed, firstly obtaining the node and branch information in the system, generating a connectivity matrix L with a dimension of the number of nodes, and the initial value of its elements is 0, by traversing the branch information, setting the elements corresponding to the nodes connected by the branches to 1, and then searching the shortest path between the nodes by the Dijkstra algorithm; by searching and comparing the shortest paths between the generator and the equivalent nodes, the set of nodes passed through is obtained;

[0018] After the power system is disconnected, the power flow of the cut-off line will be transferred to other lines of the disconnected subsystem, which may cause excessive line transmission power, causing line overload, and adversely affecting the stability of the system after disconnection. In order to prevent overload of the line after disconnection and improve the stability of the system, the absolute value of reducing the transmission power of the cut-off line is used as the optimization target. The objective function is as follows:

[0019]

[0020] Among them, R and A represent the node sets of the two subsystems after decomposition respectively; the residual network is introduced, and after finding a feasible flow, while subtracting the capacity of the feasible flow through the line from the size of the feasible flow, a line with a reverse capacity equal to the feasible flow is added, and the generated network flow direction is modified so that the feasible flow finally searched is the maximum flow of the network.

[0021] As a preferred solution of the active decoupling method of the new energy power system based on the instability mode described in the present invention, wherein: the maximum flow minimum cut algorithm includes processing the weighted directed graph problem, defining a directed weighted graph G=(V, E) in graph theory, wherein V represents the set of all nodes in the directed graph, and E represents the set of all edges therein; a two-terminal flow network is a special directed graph, and the weights of all edges in the network should be positive; there is only one source point and one sink point in a two-terminal flow network, the source point s has only outflow lines and no inflow lines, and the sink point t has only inflow lines and no outflow lines;

[0022] For a two-terminal flow network, let s be a source point with only outgoing edges and no incoming edges; let t be a sink point with only incoming edges and no outgoing edges; when there is a combination of flows from source point s to sink point t such that the flow on each edge does not exceed the maximum limit, then the combination is called a feasible flow of the flow network, and its constraints are expressed as:

[0023]

[0024]

[0025] Where e represents the edge of the graph, v represents the node of the graph, f(e) represents the flow on e, and c(e) represents the capacity of e. The value of the source-to-sink flow is expressed as:

[0026]

[0027] The maximum flow problem of the network is to optimize the allocation of the network from the source point s to the sink point t to obtain the maximum feasible flow of the network;

[0028] The minimum cut is a set of edges in all cut sets that minimizes the sum of the capacities of the deleted edges. The minimum cut problem has been proven to be equivalent to the maximum flow problem, and the minimum cut of a flow network can be obtained by finding the maximum flow.

[0029] As a preferred solution of the active disconnection method of the new energy power system based on the instability mode described in the present invention, wherein: the search system disconnection section includes: the new energy unit is sensitive to frequency, and the frequency exceeding the limit will affect the operation of the unit and even cause disconnection from the grid; when the wind farm is lower than 46.5Hz or higher than 51.5Hz, it is directly disconnected from the grid, and when the frequency is between 46.5Hz and 48.5Hz and between 50.5Hz and 51.5Hz, it maintains the grid-connected operation capability for a certain period of time;

[0030] The frequency characteristic formula of the new energy unit under droop control is expressed as

[0031]

[0032] in, Indicates the frequency characteristic coefficient under control; ΔP E represents the active power change of the control part; f0 represents the rated frequency of the system; Δf represents the frequency change; P EN Indicates the rated active power of the control part;

[0033] Frequency static characteristics of synchronous generator:

[0034]

[0035] Among them, K G Indicates the frequency static characteristic coefficient of the synchronous generator; ΔP G represents the change in the active power of the generator; f0 represents the rated frequency of the system; Δf represents the frequency change; P GN Indicates the rated active power of the generator;

[0036] The comprehensive power-frequency characteristic coefficient of the system:

[0037]

[0038] in, Indicates the comprehensive power-frequency characteristic coefficient of the system; Indicates the change in active power of the entire system; Δf* indicates the change in frequency of the entire system;

[0039] The frequency deviation constraint formula of the subsystem after decoupling is expressed as:

[0040]

[0041] Where Δf S Indicates the frequency deviation of the subsystem; Δfmax Indicates the maximum allowable frequency deviation; ΔP max represents the maximum allowable active power change; f0 represents the rated frequency of the system; P N Indicates the rated active power of the system; Represents the comprehensive power-frequency characteristic coefficient of the system.

[0042] As a preferred scheme of the active decoupling method of the new energy power system based on the unstable mode described in the present invention, the secondary search for the decoupling section includes: after decoupling the system according to the preliminary decoupling section, the unbalanced power of the subsystem is set to be the total load power of the system minus the output power of the generator, the unbalanced power of each subsystem is calculated, and the frequency deviation is calculated according to the comprehensive power-frequency characteristic coefficient of the system; if the frequency exceeds the limit, a secondary search for the decoupling section of the system is performed.

[0043] A new energy power system active decoupling system based on an unstable mode using any method described in the present invention comprises:

[0044] The data acquisition module collects the current system power flow information, modifies the parameters of the simulation model, and performs short-term simulation to generate fault set data;

[0045] The neural network model module uses convolutional neural network to extract fault set features and obtain the system instability mode;

[0046] The system decoupling module takes reducing the absolute value of the transmission power of the cut-off line as the optimization goal and uses the maximum flow minimum cut algorithm to search for the system decoupling section;

[0047] The feedback regulation module considers the frequency limitation of new energy sources, conducts a secondary search for the disconnection section, and generates an active disconnection strategy.

[0048] A computer device comprises: a memory and a processor; the memory stores a computer program, comprising: the steps of implementing any one of the methods of the present invention when the processor executes the computer program.

[0049] A computer-readable storage medium stores a computer program, comprising: when the computer program is executed by a processor, the steps of implementing any one of the methods of the present invention are implemented.

[0050] The beneficial effects of the present invention are as follows: (1) It is applicable to active disconnection control of new energy power systems. The present invention first obtains the instability mode of the system through a convolutional neural network, and then searches for the active disconnection section of the system through the maximum flow minimum cut algorithm. At the same time, it takes into account the frequency limit of new energy sources, and can effectively perform active disconnection control, with strong applicability. (2) It can integrate offline data simulation training and system real-time operation data acquisition, and can perform disconnection control more quickly and effectively, which can be used for online applications of stable control of new energy power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0052] Figure 1 An overall flow chart of a method for actively decoupling a new energy power system based on an unstable mode provided by the first embodiment of the present invention;

[0053] Figure 2 A structural diagram of a test system for a new energy power system active decoupling method based on an unstable mode provided in a second embodiment of the present invention;

[0054] Figure 3 A post-fault equivalent generator node diagram of a new energy power system active decoupling method based on an unstable mode provided by a second embodiment of the present invention;

[0055] Figure 4 A typical fault maximum flow minimum cut search after disconnection cross-section diagram of a new energy power system active disconnection method based on an unstable mode provided by the second embodiment of the present invention;

[0056] Figure 5 A typical fault secondary search and disconnection cross-section diagram of a new energy power system active disconnection method based on an unstable mode provided in a second embodiment of the present invention;

[0057] Figure 6 A power angle curve diagram of a subsystem after decoupling of a new energy power system active decoupling method based on an unstable mode provided in a second embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0059] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a method for actively decoupling a new energy power system based on an unstable mode, comprising:

[0060] S1: Collect current system power flow information and perform short-term simulation to generate fault set data.

[0061] Furthermore, the real-time power flow information of the system during operation is obtained, the parameters of the system simulation model are modified according to the real-time power flow information, a short-time simulation traversal of the expected faults is performed in the simulation model containing the real power flow information of the system, the fault set data of the system under the current operation mode is quickly generated, and the N-1 power flow data of the system is generated according to the fault location.

[0062] Furthermore, the generation of fault set data includes obtaining real-time power flow information during system operation and modifying parameters of the system simulation model according to the real-time power flow information; the multi-order model includes detailed generator dynamic characteristics, and the simulation model of the excitation system and the speed regulator system is expressed as:

[0063]

[0064]

[0065]

[0066]

[0067] Among them, E′ q and E′ d is the transient potential of the generator; T′ do and T′ q0 is the transient time constant; x d and x q are the direct-axis and quadrature-axis reactances; I d and I q are the direct-axis and quadrature-axis currents; δ represents the generator power angle; ω is the angular velocity of the generator; ω0 is the synchronous angular velocity; H is the inertia constant of the generator; P m is the mechanical power; P e is the electromagnetic power; D is the damping coefficient;

[0068] Furthermore, the real-time power flow information of the system is obtained, and the power flow parameters of the system simulation model are modified accordingly according to the real-time power flow information. The power flow equation of the system in the polar coordinate system is expressed as:

[0069]

[0070]

[0071] Where i and j represent the node numbers in the system, P and Q represent the active power injection and reactive power injection of the node respectively, U represents the node voltage amplitude, θ ij represents the voltage phase difference between nodes i and j, G ij represents the conductance between nodes i and j; B ij Represents the susceptance between nodes i and j.

[0072] Furthermore, a short-term simulation traversal of anticipated faults is performed in a simulation model containing real power flow information of the system, the fault set data of the system under the current operation mode is quickly generated, and the N-1 power flow data of the system is generated according to the fault location.

[0073] It should be noted that in the power system, the decoupling section refers to the division of the power system into several subsystems by cutting off specific transmission lines. The selection of the decoupling section is crucial to the stability of the system, especially in the event of a fault or emergency. The initial decoupling section can be determined by the maximum flow minimum cut algorithm, but the decoupling subsystem may not be stable due to frequency exceeding the limit. In order to ensure the stability of the system after decoupling, a secondary search is required to optimize the decoupling section, taking into account the unbalanced power and frequency deviation of each subsystem.

[0074] S2: Convolutional neural network is used to extract fault set features to obtain the system instability mode.

[0075] Furthermore, two convolutional neural networks are constructed for system stability judgment and power angle instability mode recognition. The output of the neural network is the label required for network training, that is, the stability and instability mode of the system, and the input is the relevant parameters of the system stability and instability mode.

[0076] Furthermore, the stability discrimination convolutional neural network contains two convolutional layers and two fully connected layers, and the power angle instability pattern recognition neural network contains four convolutional layers and two fully connected layers. The activation functions of the convolutional layer and the fully connected layer both use the elu function:

[0077]

[0078] Among them, α is a hyperparameter and is generally taken as 1.

[0079] Furthermore, the network training uses the mean square error (MSE) function as the loss function to measure the predictive ability of the model. The formula is as follows:

[0080]

[0081] Where n is the number of samples, L j is the actual value, L′ j Predicted value for the neural network.

[0082] Furthermore, for nodes, the voltage will fluctuate greatly after the system becomes unstable, and the phase angle and frequency deviation will also change continuously. Therefore, voltage, phase angle and frequency deviation are selected as network inputs. For generators, the power angle and speed of the generators in the group change rapidly after the power angle becomes unstable, and the electromagnetic power of the generators near the fault point fluctuates greatly. Therefore, the power angle, speed and electromagnetic power are selected as network inputs. The system instability mode is obtained through network training results.

[0083] Furthermore, data collection and preprocessing collects electrical parameter data of the system under different operating states and fault conditions, including voltage, phase angle, frequency deviation of nodes, power angle, speed and electromagnetic power of generators. Preprocess the data, including normalization and denoising, to ensure data quality and consistency. Construct a system stability discrimination convolutional neural network (CNN-Stability). Design the CNN model structure, including convolutional layers, pooling layers and fully connected layers. The specific number of layers and parameters are adjusted according to the size and complexity of the data set.

[0084] Furthermore, the input layer receives the voltage, phase angle and frequency deviation data of the node. The output layer is the stability label of the system, indicating whether the system is stable in the current state. Constructing a convolutional neural network for power angle instability mode recognition (CNN-InstabilityMode): Design another CNN model structure to identify the power angle instability mode. The input layer receives the power angle, speed and electromagnetic power data of the generator. The output layer is the instability mode label, indicating the specific instability mode of the system in the current state.

[0085] It should be noted that by constructing and applying two convolutional neural networks (CNN-Stability and CNN-InstabilityMode), the stability judgment of the power system and the recognition of the power angle instability mode can be effectively realized. This process combines big data and deep learning technology, uses the electrical parameters of nodes and generators as input, extracts features through convolutional neural networks, and outputs the stability and instability mode labels of the system, providing important technical support for the safe operation of the power system.

[0086] S3: Taking reducing the absolute value of the transmission power of the cut-off line as the optimization goal, the maximum flow minimum cut algorithm is used to search for the system decoupling section.

[0087] Furthermore, the maximum flow minimum cut algorithm is an algorithm for dealing with weighted directed graph problems. In graph theory, a directed weighted graph G = (V, E) is defined, where V is the set of all nodes in the directed graph and E is the set of all edges in it. A two-terminal flow network is a special directed graph, and the weights of all edges in the network should be positive. In a two-terminal flow network, there is only one source point and one sink point. The source point s has only outflow routes and no inflow routes, and the sink point t has only inflow routes and no outflow routes. When there is a combination of flows from the source point s to the sink point t so that the flow on each edge does not exceed the maximum limit, then the combination is called a feasible flow of the flow network. The maximum flow problem of the network is to optimize the allocation of the network from the source point s to the sink point t in order to obtain the maximum feasible flow of the network. If a group of edges in the flow network are deleted so that the network cannot flow, and the flow cannot flow from the source point s to the sink point t, then the group of edges is called the cut set of the flow network. The minimum cut is a set of edges in the cut set that minimizes the sum of the capacities of the deleted edges. In graph theory, the minimum cut problem is proved to be equivalent to the maximum flow problem, so finding the maximum flow of a flow network can find its minimum cut.

[0088] Furthermore, when there is a combination of flows from source point s to sink point t such that the flow on each edge does not exceed the maximum limit, then the combination is called a feasible flow of the flow network, and its constraints are expressed as:

[0089]

[0090]

[0091] Where e is the edge of the graph, v is the node of the graph, f(e) is the flow on e, and c(e) is the capacity of e. The value of the source-to-sink flow is:

[0092]

[0093] Furthermore, in the maximum flow minimum cut problem, the flow network has only one source point and one sink point, so it is necessary to make the generator set equivalent in the power system, and divide the leading group and the remaining group according to the power angle instability mode. The generation of equivalent generator nodes mainly includes the system shortest path search. Before searching the system shortest path, first obtain the node and branch information in the system, generate a connectivity matrix L with a dimension of the number of nodes, and its element initial value is 0. By traversing the branch information, set the corresponding elements of the nodes connected by the branch to 1, and then search the shortest path between the nodes through the Dijkstra algorithm. Then, by searching and comparing the shortest paths between the generator and the equivalent node, find the set of nodes that form a network that is interconnected between the generators in the same group.

[0094] Furthermore, after the power system is disconnected, the power flow of the cut-off line will be transferred to other lines of the disconnected subsystem, which may cause excessive line transmission power, causing line overload, and adversely affecting the stability of the system after disconnection. In order to prevent overload of the line after disconnection and improve the stability of the system, the absolute value of reducing the transmission power of the cut-off line is used as the optimization target, and the objective function is as follows:

[0095]

[0096] Among them, R and A represent the node sets of the two subsystems after decomposition respectively; the residual network is introduced, and after finding a feasible flow, while subtracting the capacity of the feasible flow through the line from the size of the feasible flow, a line with a reverse capacity equal to the feasible flow is added, and the generated network flow direction is modified so that the feasible flow finally searched is the maximum flow of the network.

[0097] Furthermore, according to the above optimization objective function, a residual network is introduced. After a feasible flow is found, the capacity of the feasible flow through the route is subtracted from the size of the feasible flow, and a reverse route with a capacity equal to the feasible flow is added. The generated network flow direction can be modified so that the feasible flow finally searched is the maximum flow of the network. This method essentially searches for the augmented chain of the network, that is, the link that can make the flow flow from the source point s to the sink point t under the feasible flow. By introducing the residual network, only the forward arc can be considered in the search process, which simplifies the search process.

[0098] Furthermore, when there is a combination of flows from the source point s to the sink point t such that the flow on each edge does not exceed the maximum limit, the combination is called a feasible flow of the flow network; the flow constraint of the feasible flow is: for each edge (i, j) ∈ E, the flow fij satisfies:

[0099] 0≤f ij ≤wi j

[0100] Furthermore, flow conservation, for each node v∈N\{s,t}, satisfies:

[0101]

[0102] Furthermore, the maximum flow problem of the network is to optimize the allocation of the network from the source point s to the sink point t to obtain the maximum feasible flow of the network; the maximum flow problem is to solve the maximum feasible flow F from the source point s to the sink point t, and the formula is expressed as:

[0103]

[0104] Furthermore, the search system disconnection section includes that the new energy units are sensitive to frequency, and frequency exceeding the limit will affect the operation of the units and even cause them to be disconnected from the grid; then when the frequency of the wind farm is lower than 46.5Hz or higher than 51.5Hz, it will be directly disconnected from the grid, and when the frequency is between 46.5Hz and 48.5Hz and 50.5Hz and 51.5Hz, it will maintain the ability to connect to the grid for a certain period of time.

[0105] Furthermore, the frequency characteristic formula of the new energy unit under droop control is expressed as:

[0106]

[0107] in, Indicates the frequency characteristic coefficient under control; ΔP E represents the active power change of the control part; f0 represents the rated frequency of the system; Δf represents the frequency change; P EN Indicates the rated active power of the control part.

[0108] Furthermore, the frequency static characteristics of the synchronous generator are:

[0109]

[0110] Among them, K G Indicates the frequency static characteristic coefficient of the synchronous generator; ΔP G represents the change in the active power of the generator; f0 represents the rated frequency of the system; Δf represents the frequency change; P GN Indicates the rated active power of the generator.

[0111] Furthermore, the comprehensive power-frequency characteristic coefficient of the system is:

[0112]

[0113] in, Indicates the comprehensive power-frequency characteristic coefficient of the system; It represents the change of active power of the whole system; Δf* represents the change of frequency of the whole system.

[0114] Furthermore, the frequency deviation constraint formula of the decoupled subsystem is expressed as:

[0115]

[0116] Where Δf S Indicates the frequency deviation of the subsystem; Δf max Indicates the maximum allowable frequency deviation; ΔP max represents the maximum allowable active power change; f0 represents the rated frequency of the system; P N Indicates the rated active power of the system; Represents the comprehensive power-frequency characteristic coefficient of the system.

[0117] It should be noted that the minimum cut is a set of edges in all cut sets that minimizes the sum of the capacities of the deleted edges; the minimum cut problem has been proved to be equivalent to the maximum flow problem, and the minimum cut can be obtained by finding the maximum flow of the flow network.

[0118] S4: Considering the frequency limitation of new energy, a secondary search is conducted on the disconnection section to generate an active disconnection strategy.

[0119] Furthermore, the secondary search for the disconnection section includes, after disconnecting the system according to the preliminary disconnection section, setting the unbalanced power of the subsystem to be the total load power of the system minus the generator output power, calculating the unbalanced power of each subsystem, and calculating the frequency deviation according to the comprehensive power-frequency characteristic coefficient of the system; if the frequency exceeds the limit, a secondary search for the disconnection section of the system is performed.

[0120] It should be noted that unbalanced power refers to the difference between the load power in the subsystem and the generator output power. Unbalanced power will cause frequency changes and affect system stability. Therefore, the frequency deviation must be calculated based on the comprehensive power-frequency characteristic coefficient to ensure that the frequency of the subsystem is within the allowable range. If the frequency exceeds the limit, it is necessary to readjust the decoupling section and perform a secondary search to obtain a stable decoupling solution.

[0121] On the other hand, this embodiment also provides a new energy power system active decoupling system based on an unstable mode, which includes:

[0122] The data acquisition module collects the current system power flow information, modifies the parameters of the simulation model, and performs short-term simulation to generate fault set data.

[0123] The neural network model module uses convolutional neural network to extract fault set features to obtain the system's instability mode.

[0124] The system decoupling module takes reducing the absolute value of the transmission power of the cut-off line as the optimization goal, and uses the maximum flow minimum cut algorithm to search for the system decoupling section.

[0125] The feedback regulation module considers the frequency limitation of new energy sources, conducts a secondary search for the disconnection section, and generates an active disconnection strategy.

[0126] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0127] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0128] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0129] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0130] Example 2, reference Figure 2-6 , which is an embodiment of the present invention, provides a method for active disconnection of a new energy power system based on an unstable mode. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0131] First, the power flow information of the system must be obtained and the fault set to be analyzed must be determined. In this embodiment, the IEEE 39-node system is used as the test system. There are 10 generators and 46 lines in the system, including 12 transformer lines and 34 transmission lines. A new energy wind turbine with a capacity of about 200MW is connected at nodes 3, 13 and 24 respectively, and the load at the node where the new energy is connected is increased by about 100MW as the rated operation mode. Figure 2 On the bus16-bus17 line, a three-phase grounding fault lasting 120ms occurred at 10% of the first section of the line, and the line was cut off after the fault was resolved.

[0132] A convolutional neural network is used to extract fault set features to obtain the instability mode of the system. In this embodiment, the stability and instability mode of the system are determined by a trained convolutional neural network. After the fault occurs, the generators are divided into two groups. The power angle difference between the two groups continues to increase, and the system eventually becomes unstable. The instability mode is [0,0,1,1,1,1,1,0,0,0], that is, generators G1, G2, G8, G9, and G10 are in one group, and the remaining generators are in another group. The nodes are equivalent according to the grouping of the generators. The equivalent generator nodes are as follows: Figure 3 As shown in the figure, the blue line is the fault line bus16-bus17, the orange area is the equivalent generator node 1, and the blue area is the equivalent generator node 2.

[0133] Taking the reduction of the absolute value of the power transmission of the cut-off line as the optimization goal, the maximum flow minimum cut algorithm is used to search the system disconnection section. According to the equivalent generator nodes and the remaining free nodes, the disconnection section is searched. The preliminary disconnection section is as follows: Figure 4As shown in the figure, the decoupled sections are bus6-bus11 and bus4-bus14, which are the red lines in the figure. By disconnecting these lines, the system is decoupled into two independent subsystems.

[0134] Considering the frequency limit of new energy, a secondary search is performed on the disconnection section. When the system frequency estimated by unbalanced power is lower than 46.5Hz or higher than 51.5Hz, it is considered that the frequency exceeds the limit, and the secondary search of the disconnection section is started. In order to verify the effectiveness of the secondary search of the disconnection section considering the frequency limit of new energy, the comprehensive power-frequency characteristic coefficient of the system is taken as a smaller value to amplify the impact of unbalanced power on the system frequency. Set the subsystem 1 Subsystem 2 At this time, the frequency of subsystem 1 is 48.30Hz, and the frequency of subsystem 2 is 51.60Hz. The frequency of subsystem 2 is higher than 51.5Hz, exceeding the specified frequency limit. In order to ensure that the new energy unit wendin is connected to the grid, a secondary search is conducted on the disconnection section, and the disconnection sections obtained are bus3-bus4, bus4-bus5, and bus6-bus11. The disconnection sections of the system secondary search are as follows: Figure 5 As shown in the figure, the blue line is the fault line and the red line is the disconnection section line. Through this disconnection method, the system is disconnected into two independent subsystems.

[0135] The decoupling control is started 200ms after the fault of bus16-bus17 is removed. The power angle curves of subsystem 1 and subsystem 2 after decoupling are as follows: Figure 6 As shown in the figure, the power angle swing amplitude of the subsystem gradually decreases and the system remains stable. At this time, the frequency of subsystem 1 is 49.12Hz, and the frequency of subsystem 2 is 50.86Hz, which meets the frequency requirements. After decoupling, the new energy unit can maintain grid-connected operation.

[0136] In summary, the present invention can perform online stability control on power systems containing new energy units. Compared with the existing out-of-step decoupling method, the present invention performs active decoupling control through data-driven instability pattern recognition, which can make full use of massive offline simulation data, improve simulation accuracy, and has strong applicability. At the same time, the algorithm of the present invention is based on the maximum flow minimum cut algorithm and the minimum path search, while taking into account the frequency limit of the new energy unit, and can formulate a decoupling strategy online according to the actual operation mode of the system to ensure the rapidity of decoupling.

[0137] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for actively decoupling a new energy power system based on an unstable mode, characterized in that: include: Collect current system power flow information and perform short-term simulation to generate fault set data; Convolutional neural network is used to extract fault set features to obtain the system instability mode; Taking reducing the absolute value of the transmission power of the cut-off line as the optimization goal, the maximum flow minimum cut algorithm is used to search for the system decoupling section; Considering the frequency limitation of new energy, a secondary search is conducted on the disconnection section to generate an active disconnection strategy.

2. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 1, characterized in that: The generation of fault set data includes obtaining real-time power flow information during system operation, and modifying the power flow parameters of the system simulation model accordingly according to the real-time power flow information. The system power flow equation in the polar coordinate system is expressed as: Where i and j represent the node numbers in the system, P and Q represent the active power injection and reactive power injection of the node respectively, U represents the node voltage amplitude, θ ij represents the voltage phase difference between nodes i and j, G ij represents the conductance between nodes i and j; B ij represents the susceptance between nodes i and j; A short-term simulation traversal of the expected fault is performed in the simulation model containing the real power flow information of the system, the fault set data of the system under the current operation mode is quickly generated, and the N-1 power flow data of the system is generated according to the fault location.

3. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 2, characterized in that: The convolutional neural network includes: constructing two convolutional neural networks for system stability judgment and power angle instability mode recognition; the output of the neural network is the label required for network training, and the input is the relevant parameters of system stability and instability mode; for the node, the voltage, phase angle and frequency deviation are selected as the network input; for the generator, the power angle, speed and electromagnetic power are selected as the network input; the system instability mode is obtained through the network training result.

4. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 3, characterized in that: The optimization goal of reducing the absolute value of the power transmission of the cut-off line includes equivalently dividing the generator sets in the power system into leading generator sets and remaining generator sets according to the power angle instability mode; To search for the shortest path of the system, first obtain the node and branch information in the system, generate a connectivity matrix L with the dimension of the number of nodes, and set the initial value of its elements to 0. By traversing the branch information, set the elements corresponding to the nodes connected by the branches to 1, and then search for the shortest path between the nodes through the Dijkstra algorithm; by searching and comparing the shortest paths between the generator and the equivalent nodes, obtain the set of nodes passed through; After the power system is disconnected, the power flow of the cut-off line will be transferred to other lines of the disconnected subsystem, which may cause excessive power transmission of the line, causing line overload, and adversely affecting the stability of the system after disconnection; In order to prevent overload of the line after disconnection and improve the stability of the system, the transmission power P of the disconnected line is reduced. ij The absolute value of is the optimization target, and the objective function formula is expressed as: Among them, R and A represent the node sets of the two subsystems after decomposition respectively; the residual network is introduced, and after finding a feasible flow, while subtracting the capacity of the feasible flow through the line from the size of the feasible flow, a line with a reverse capacity equal to the feasible flow is added, and the generated network flow direction is modified so that the feasible flow finally searched is the maximum flow of the network.

5. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 4, characterized in that: The maximum flow minimum cut algorithm includes processing a directed graph problem with weights, defining a directed weighted graph G=(V,E) in graph theory, wherein V represents the set of all nodes in the directed graph, and E represents the set of all edges therein; a two-terminal flow network is a special directed graph, and the weights of all edges in the network should be positive; a two-terminal flow network has only one source point and one sink point, the source point s has only outflow lines and no inflow lines, and the sink point t has only inflow lines and no outflow lines; For a two-terminal flow network, let s be a source point with only outgoing edges and no incoming edges; let t be a sink point with only incoming edges and no outgoing edges; when there is a combination of flows from source point s to sink point t such that the flow on each edge does not exceed the maximum limit, then the combination is called a feasible flow of the flow network, and its constraints are expressed as: Where e represents the edge of the graph, v represents the node of the graph, f(e) represents the flow on e, and c(e) represents the capacity of e. The value of the source-to-sink flow is expressed as: The maximum flow problem of the network is to optimize the allocation of the network from the source point s to the sink point t to obtain the maximum feasible flow of the network; The minimum cut is a set of edges in all cut sets that minimizes the sum of the capacities of the deleted edges. The minimum cut problem has been proven to be equivalent to the maximum flow problem, and the minimum cut of a flow network can be obtained by finding the maximum flow.

6. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 5, characterized in that: The search system disconnection section includes: the new energy unit is sensitive to frequency, and the frequency exceeding the limit will affect the operation of the unit and even cause it to be disconnected from the grid; when the frequency of the wind farm is lower than 46.5Hz or higher than 51.5Hz, it will be directly disconnected from the grid, and when the frequency is between 46.5Hz and 48.5Hz and 50.5Hz and 51.5Hz, it will maintain the grid-connected operation capability for a certain period of time; The frequency characteristic formula of the new energy unit under droop control is expressed as in, Indicates the frequency characteristic coefficient under control; ΔP E represents the active power change of the control part; f0 represents the rated frequency of the system; Δf represents the frequency change; P EN Indicates the rated active power of the control part; Frequency static characteristics of synchronous generator: Among them, K G Indicates the frequency static characteristic coefficient of the synchronous generator; ΔP G represents the change in the active power of the generator; f0 represents the rated frequency of the system; Δf represents the frequency change; P GN Indicates the rated active power of the generator; The comprehensive power-frequency characteristic coefficient of the system: in, Indicates the comprehensive power-frequency characteristic coefficient of the system; Indicates the change in active power of the entire system; Δf * Indicates the frequency change of the system as a whole; The frequency deviation constraint formula of the subsystem after decoupling is expressed as: Where Δf S Indicates the frequency deviation of the subsystem; Δf max Indicates the maximum allowable frequency deviation; ΔP max represents the maximum allowable active power change; f0 represents the rated frequency of the system; P N Indicates the rated active power of the system; Represents the comprehensive power-frequency characteristic coefficient of the system.

7. The method for actively disconnecting a new energy power system based on an unstable mode according to claim 6, characterized in that: The secondary search for the disconnection section includes: after disconnecting the system according to the preliminary disconnection section, setting the unbalanced power of the subsystem to be the total load power of the system minus the generator output power, calculating the unbalanced power of each subsystem, and calculating the frequency deviation according to the comprehensive power-frequency characteristic coefficient of the system; if the frequency exceeds the limit, a secondary search for the disconnection section of the system is performed.

8. An active disconnection system for a new energy power system based on an unstable mode using the method according to any one of claims 1 to 7, characterized in that: The data acquisition module collects the current system power flow information, modifies the parameters of the simulation model, and performs short-term simulation to generate fault set data; The neural network model module uses convolutional neural network to extract fault set features and obtain the system instability mode; The system decoupling module takes reducing the absolute value of the transmission power of the cut-off line as the optimization goal and uses the maximum flow minimum cut algorithm to search for the system decoupling section; The feedback regulation module considers the frequency limitation of new energy sources, conducts a secondary search for the disconnection section, and generates an active disconnection strategy.

9. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the method for actively decoupling a new energy power system based on an unstable mode are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for actively decoupling a new energy power system based on an unstable mode are implemented.

Images