Fine-grained parallel electromagnetic transient simulation method, system, terminal and medium for distribution network

Abstract

The present invention provides a method, system, terminal and medium for fine-grained parallel electromagnetic transient simulation of a complex distribution network with distributed power access. For a complex distribution network with distributed power access, a fine-grained parallel solution method for the main components of the complex distribution network is proposed, and the parallelism of the solution method is increased to a single branch and a single node. Combined with the massive computing resources of the GPU, parallel solution within the system can be achieved; in addition, a network interface based on dual-port Thevenin equivalence is further introduced to decouple the entire system into a transmission network part, a synchronous motor part and a power electronics part. The solution of the three subsystems only depends on the value of the interface capacitor. Combined with the multi-threading characteristics of the GPU, parallel solution between subsystems can be achieved. As a result, the system simulation time will not change significantly with the increase of the system simulation scale, thereby greatly improving the simulation efficiency of the complex distribution network with distributed power access.

Description

Technical Field

[0001] The present invention relates to the technical field of power system simulation, and in particular to a method, system, terminal and medium for fine-grained parallel electromagnetic transient simulation of a complex distribution network with distributed power supply access. Background Art

[0002] Due to the uneven distribution of power resources and load centers in my country, large-scale distributed power access to the distribution network has become an important supplementary means for the development of the power system. Therefore, on the one hand, the structure of the distribution network is becoming more and more complex, and the simulation scale is growing rapidly. On the other hand, the distribution network presents a high degree of power electronic characteristics, and its strong nonlinear characteristics are becoming more and more obvious, which leads to a large number of complex electromagnetic transient problems, which seriously threaten the safe and stable operation of the distribution network. Electromagnetic transient simulation has gradually become an effective method to accurately simulate the operating status of future power systems.

[0003] Traditional electromagnetic transient simulation is limited by the computing power of simulation hardware and the parallel degree of simulation algorithms. In recent years, with the emergence of high-performance computing platforms, electromagnetic transient simulation platforms have developed from single machines to computer clusters, and from multi-core CPUs to heterogeneous computing platforms based on GPUs and FPGAs. The performance improvement of the simulation platform has greatly improved the calculation speed of the model and the simulation efficiency of the power system. In terms of algorithm structure, some scholars have obtained simplified models with high efficiency and good numerical stability through dimensionality reduction methods such as spatial projection transformation and polynomial expansion; other scholars focus on using partitioning methods to model large power systems. The long transmission line decoupling method is the earliest method to achieve parallel computing, but its requirements for network separation are relatively strict, and it is difficult to achieve flexible network separation. Then, some new network separation methods are proposed, mainly including state space node method, multi-region Thevenin equivalent method, ideal transformer model method and node segmentation method, which can improve the efficiency of electromagnetic transient simulation from the perspective of the algorithm. Therefore, the development of the simulation platform and the improvement of the algorithm structure have laid the foundation for the application of electromagnetic transient simulation in large-scale power systems. However, the methods mentioned above are essentially highly serial algorithms. As the scale of the system further expands, it will gradually become difficult to meet the requirements of simulation speed.

[0004] Since Professor Dommel proposed the EMTP simulation framework in the 1960s, the augmented node analysis method of EMTP-RV and the state space node method of RT-LAB have been further designed based on the node analysis framework. The node analysis method has a high degree of serialization and its numerical complexity is O(n 2), which is more suitable for solving on the CPU. However, with the failure of Moore's Law, the clock frequency of the CPU has reached saturation. Obviously, with the continuous expansion of the scale of the power system, the limited improvement of CPU performance alone can no longer meet the needs of simulation. Therefore, the demand for simulation platforms based on other hardware architectures still exists. With the emergence of GPUs and FPGAs, many scholars have begun to study parallel electromagnetic transient simulation algorithms based on GPUs and FPGAs. Most of the current research focuses on parallel solution of node admittance matrices or decoupled subsystems, which basically does not change the simulation algorithm framework based on the node analysis method. They still have quadratic complexity. In addition, the existing network segmentation methods are low-density system-level parallel methods, which cannot give full play to the advantages of parallel hardware such as GPUs and FPGAs. Summary of the invention

[0005] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method, system, terminal and medium for fine-grained parallel electromagnetic transient simulation of a complex distribution network with distributed power supply access.

[0006] According to one aspect of the present invention, a fine-grained parallel electromagnetic transient simulation method for a distribution network is provided, comprising:

[0007] Constructing a simulation solution model for a complex distribution network with distributed power access, the simulation solution model for a complex distribution network with distributed power access includes: a transmission network model, a synchronous motor model, a power electronic equipment model and a network interface model, wherein the transmission network model, the synchronous motor model and the power electronic equipment model can realize independent fine-grained parallel solution only by relying on the interface capacitor voltage of the network interface model, and the network interface model is solved according to the current update results of the transmission network model, the synchronous motor model and the power electronic equipment model;

[0008] For the simulation solution model of the complex distribution network with distributed power access, the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology are obtained, and then the corresponding interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix are formed;

[0009] Solving the system power flow of the complex distribution network simulation solution model with distributed power generation access according to the transmission network parameter matrix, and obtaining the power flow data of each node and each branch in the complex distribution network system;

[0010] According to the power flow data, the constant impedance model parameters of the load nodes in the simulation solution model of the complex distribution network with distributed power generation access are solved, and the constant impedance model parameters of the load nodes are added to the transmission network parameter matrix to form a new transmission network parameter matrix; at the same time, according to the power flow data, the voltage and current vector of the synchronous motor model are initialized;

[0011] According to the interface capacitance parameter matrix and its node voltage solution equation, the data required by the network interface model is initialized to obtain the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation, and pre-store them;

[0012] According to the new transmission network parameter matrix and its node voltage and branch current solution equations, the data required by the transmission network model are initialized to obtain the coefficient matrix required for solving the node voltage and branch current of the transmission network model and the initial data required for the main loop simulation, and pre-store them;

[0013] According to the synchronous motor parameter matrix and its electrical equation and rotor motion equation, the data required for the synchronous motor model is initialized to obtain the voltage and current of the synchronous motor model motor and the coefficient matrix required for the rotor dynamic solution of the motor and the initial data required for the main loop simulation, and pre-store them;

[0014] According to the distributed power system parameter matrix and its main circuit solution equation and control system solution equation, the data required by the power electronic equipment model is initialized to obtain the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model and the initial data required for the main cycle simulation, and pre-store them;

[0015] The transmission network model, synchronous motor model, power electronic equipment model and network interface model are respectively updated with the initial data required for the corresponding main loop simulation for the next step of voltage and current; wherein: the parallelism of the transmission network model is increased to a single node and a single branch, and the GPU is used to realize the parallel solution of each node and each branch in the transmission network model to obtain the next step of voltage and current update; the parallelism of the synchronous motor model is increased to a single motor, and the solution of each motor is performed based on the GPU using the direct LU decomposition method, thereby realizing the parallel solution between and within the synchronous motors to obtain The next step is voltage and current update; the parallelism of the main circuit of the power electronic equipment model is improved to be consistent with the transmission network model, and each node and each branch of the main circuit are solved in parallel. The parallelism of the control circuit of the power electronic equipment model is improved to a single device, in which each converter control system allocates a thread for parallel solution to obtain the next step of voltage and current update; the parallelism of the network interface model is improved to a single node, and the GPU is used to allocate a corresponding number of threads to achieve parallel solution of each node to obtain the next step of voltage and current update; thereby, fine-grained parallel solution between and within each model is achieved;

[0016] This cycle continues until the entire simulation process is completed.

[0017] According to another aspect of the present invention, a distribution network fine-grained parallel electromagnetic transient simulation system is provided, comprising:

[0018] A modeling module for a complex distribution network with distributed power access, which constructs a simulation solution model for a complex distribution network with distributed power access, wherein the simulation solution model for a complex distribution network with distributed power access includes: a transmission network model, a synchronous motor model, a power electronic device model, and a network interface model, wherein the transmission network model, the synchronous motor model, and the power electronic device model can realize independent fine-grained parallel solution only by relying on the voltage of the interface capacitor of the network interface model, and the network interface model is solved according to the current update results of the transmission network model, the synchronous motor model, and the power electronic device model;

[0019] A parameter matrix construction module, which obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology for the complex distribution network simulation solution model with distributed power access, and then forms the corresponding interface capacitance parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix; according to the transmission network parameter matrix, solves the system flow of the complex distribution network simulation solution model with distributed power access, and obtains the flow data of each node and each branch in the complex distribution network system; according to the flow data, solves the constant impedance model parameters of the load nodes in the complex distribution network simulation solution model with distributed power access, and adds the constant impedance model parameters of the load nodes to the transmission network parameter matrix to form a new transmission network parameter matrix; at the same time, initializes the voltage and current vector of the synchronous motor model according to the flow data;

[0020] A coefficient matrix and simulation initial data acquisition module, which performs initialization operations on the data required for the transmission network model, synchronous motor model, power electronic equipment model and network interface model according to the new transmission network parameter matrix, synchronous motor parameter matrix, distributed power supply parameter matrix and their solution equations, obtains the coefficient matrix required for solving the node voltage and branch current of the transmission network model, the coefficient matrix required for solving the voltage and current of the motor of the synchronous motor model and the rotor dynamics of the motor, the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model, and the initial data required for the main cycle simulation of each model, and pre-stores them;

[0021] A fine-grained parallel simulation module, which uses the initial data to perform the next step of voltage and current update on the transmission network model, the synchronous motor model, the power electronic equipment model and the network interface model; the parallelism of the transmission network model is increased to a single node and a single branch, and after being combined with the GPU massive computing unit, the parallel solution of each node and each branch in the transmission network model is realized; the parallelism of the synchronous motor model is increased to a single motor, wherein the solution of each motor is carried out by direct LU decomposition method, thereby realizing parallel solution between synchronous motors and within synchronous motors; the solution of the main circuit of the power electronic equipment model is consistent with the transmission network model, and each node and each branch of the main circuit is executed in parallel, and the parallelism of the control circuit of the power electronic equipment model is a single device, wherein the control system of each converter is allocated a thread for parallel solution; the parallelism of the network interface model is a single node, and the parallel solution of each node is realized after the GPU is allocated a corresponding number of threads; thereby, fine-grained parallel solution between and within the models is realized.

[0022] According to a third aspect of the present invention, a terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can be used to execute any of the above-described methods when executing the program.

[0023] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can be used to execute any of the methods described above.

[0024] Due to the adoption of the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0025] The fine-grained parallel electromagnetic transient simulation method, system, terminal and medium provided by the present invention relate to a fine-grained parallel electromagnetic transient simulation technology for a complex distribution network with distributed power access, which uses a delayed insertion method and a dual-port Thevenin equivalent to decouple the complex distribution network with distributed power access into multiple independently solved subsystems (models), improve the algorithm parallelism to a single branch, a single node and a single device, and combine the massive computing resources of the GPU to achieve fine-grained parallel solutions between and within each decoupled subsystem. As a result, the system simulation time will not change significantly with the increase of the system simulation scale, thereby greatly improving the simulation efficiency of the complex distribution network with distributed power access. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0027] Figure 1 It is a flow chart of a fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power access in one embodiment of the present invention;

[0028] Figure 2 It is a flowchart of a method for fine-grained parallel electromagnetic transient simulation of a complex distribution network with distributed power access in a preferred embodiment of the present invention;

[0029] Figure 3 A schematic diagram of fine-grained modeling of a power transmission network model in a preferred embodiment of the present invention;

[0030] Figure 4 A schematic diagram of modeling a network interface between a power transmission network model and a synchronous motor part and a power electronic device part in a preferred embodiment of the present invention;

[0031] Figure 5 A schematic diagram of a control system of a power electronic equipment model in a preferred embodiment of the present invention;

[0032] Figure 6A schematic diagram of main circuit modeling of a power electronic device model in a preferred embodiment of the present invention;

[0033] Figure 7 It is a schematic diagram of a timing diagram for solving each model after decoupling in a preferred embodiment of the present invention;

[0034] Figure 8 An improved 118-node topology diagram with 20 distributed power sources connected in a preferred embodiment of the present invention;

[0035] Fig. 9 Schematic diagram of the simulation accuracy verification of the algorithm proposed in a preferred embodiment of the present invention; wherein NAM represents the solution using the node analysis method, SC-LIM represents the solution using the proposed algorithm, (a) is the simulation accuracy verification of the power transmission network part, V5 a 、V5 b 、V5 c a, b, c represent the three-phase values ​​of node 5 voltage respectively; (b) Partial simulation accuracy verification of synchronous motor, ISG1 a 、ISG1 b , ISG1. Represent the three-phase values ​​of a, b, and c of the motor 1 current respectively; (c) Simulation accuracy verification of power electronic equipment, VCV1 a 、VCV1 b 、VCV1 c Respectively represent the three-phase values ​​of a, b, and c of the output voltage of converter 1;

[0036] Fig.10 It is a schematic diagram of simulation efficiency verification of the algorithm proposed in a preferred embodiment of the present invention, which represents the comparison of single-step simulation time of systems of different scales, wherein NAM+CPU represents the solution efficiency of the node analysis method on the CPU, SC-LIM-CPU represents the solution efficiency of the proposed algorithm on the CPU, and SC-LIM-GPU represents the solution efficiency of the proposed algorithm on the CPU-GPU joint platform;

[0037] Fig.11 The present invention is a schematic diagram of the composition structure of a fine-grained parallel electromagnetic transient simulation system for a complex distribution network with distributed power generation access in one embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following is a detailed description of the embodiments of the present invention: This embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and a specific operation process are given. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

[0039] Figure 1A flow chart of a fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power generation access provided by an embodiment of the present invention.

[0040] like Figure 1 As shown, the fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power supply access provided in this embodiment may include the following steps:

[0041] S100, constructing a simulation solution model for a complex distribution network with distributed power generation access, including the following subsystems: a transmission network part, a synchronous motor part, a power electronic equipment part, and a network interface part, wherein the transmission network part, the synchronous motor part, and the power electronic equipment part can be solved independently and in parallel, and the solution of each part only depends on the interface capacitor voltage of the network interface part to achieve independent fine-grained parallel solution, and the network interface part is solved according to the current update results of the above three parts;

[0042] S200, for a complex distribution network simulation solution model with distributed power access, obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters related to the required simulation topology, and then forms the corresponding interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix;

[0043] S300, solving the distribution network system flow of the simulation solution model of the complex distribution network with distributed generation access according to the transmission network parameter matrix, and obtaining the flow data of each node and each branch of the complex distribution network with distributed generation access;

[0044] S400, based on the power flow data, initialize the voltage and current vector of the synchronous motor, and solve the constant impedance model parameters of the load node in the complex distribution network with distributed power generation access, and then add the constant impedance parameters of the load node to the corresponding position of the transmission network parameter matrix to form a new transmission network parameter matrix;

[0045] S500, according to the interface capacitance parameter matrix, the new transmission network parameter matrix, the synchronous motor parameter matrix and the distributed power system parameter matrix and the solution equations thereof, solve and pre-store the coefficient matrix and initial data required for solving each subsystem; specifically:

[0046] According to the interface capacitance parameter matrix and its node voltage solution equation, the data required by the network interface model is initialized to obtain the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation, and pre-store them;

[0047] According to the new transmission network parameter matrix and its node voltage and branch current solution equations, the data required for the transmission network part is initialized, the coefficient matrix required for solving the node voltage and branch current of the transmission network part and the initial data required for the main loop simulation are obtained, and pre-stored;

[0048] According to the synchronous motor parameter matrix and its electrical equation and rotor motion equation, the data required for the synchronous motor part is initialized to obtain the voltage and current of the synchronous motor part and the coefficient matrix required for the rotor dynamic solution of the motor and the initial data required for the main loop simulation, and pre-store them;

[0049] According to the distributed power system parameter matrix and its main circuit solution equations and control system solution equations, the data required by the power electronic equipment part is initialized, the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment part and the initial data required for the main cycle simulation are obtained, and pre-stored;

[0050] S600 uses the initial simulation data and reasonably allocates GPU threads according to the complexity of each subsystem solution to achieve fine-grained parallel solution between and within each subsystem, and repeats this cycle until the simulation process ends; specifically:

[0051] The initial data required for the corresponding main loop simulation are used to update the next step of voltage and current for the transmission network model, synchronous motor model, power electronic equipment model and network interface model respectively; wherein: the parallelism of the transmission network model is increased to a single node and a single branch, and the massive computing resources of the GPU are used to realize the parallel solution of each node and each branch in the transmission network model to obtain the next step of voltage and current update; the parallelism of the synchronous motor model is increased to a single motor, wherein the solution of each motor is based on the GPU using the direct LU decomposition method, thereby realizing the parallel solution between and within the synchronous motors to obtain the next step of voltage and current update; the parallelism of the main circuit of the power electronic equipment model is improved in accordance with the transmission network model, and each node and each branch of the main circuit are solved in parallel, and the parallelism of the control circuit of the power electronic equipment model is increased to a single device, wherein the control system of each converter is allocated a thread for parallel solution to obtain the next step of voltage and current update; the parallelism of the network interface model is increased to a single node, and the GPU is used to allocate the corresponding number of threads to realize the parallel solution of each node to obtain the next step of voltage and current update; thereby, fine-grained parallel solution between and within the models is realized;

[0052] This cycle continues until the entire simulation process is completed.

[0053] In S100 of this embodiment, as a preferred embodiment, constructing a complex distribution network simulation model including distributed power sources may include the following steps:

[0054] The complex distribution network is partitioned and reorganized using the dual-port Thevenin equivalent. The capacitor branch is selected as the interface between the transmission network part and the synchronous motor part and the power electronic equipment part. The network interface model is as follows: Figure 4 As shown. For the transmission network, synchronous motor and power electronic equipment, the ground capacitance connected to them is equivalent to the voltage source on the branch where they are located. At the same time, for the network interface, the subsystems connected to them are equivalent to the ground current source on the node where they are located. The expression for solving each subsystem after decoupling is:

[0055]

[0056] in, is the interface voltage of the next simulation step, f C is the solution function for updating the node voltage of the network interface model, is the branch current of the interface capacitor in the branch where the synchronous motor is located in the current simulation step, is the branch current of the interface capacitor in the transmission network under the current simulation step, is the branch current of the interface capacitor in the branch where the power electronic equipment part is located in the current simulation step, is the branch current of the interface capacitor in the branch where the synchronous motor is located in the next simulation step, f G is the solution function for updating the current of some branches of the synchronous motor, is the branch current of the interface capacitor in the transmission network in the next simulation step, f NET is the solution function for updating the current of some branches of the transmission network, is the branch current of the interface capacitor in the branch where the power electronic equipment part is located in the next simulation step, f CV A solution function for updating the current of some branches of power electronic equipment.

[0057] In S200 of this embodiment, as a preferred embodiment, for a complex distribution network simulation solution model with distributed power access, the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology are obtained, and then the interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix of the system are formed, including:

[0058] Obtain the interface capacitance parameters of the complex distribution network simulation solution model with distributed power access, including the interface node number and the capacitance value corresponding to the interface node, and add the interface node number and the capacitance value corresponding to the interface node to the corresponding position of the interface capacitance parameter matrix to obtain the interface capacitance parameter matrix of the complex distribution network simulation solution model with distributed power access;

[0059] The transmission network parameters of the simulation and solution model of the complex distribution network with distributed generation access are obtained, including node parameters and branch parameters, wherein the node parameters include node number, voltage amplitude, voltage phase angle, motor active power, motor reactive power, load active power, load reactive power, conductance G, susceptance B and node type number. The node parameters required for model flow solution are obtained according to the three different node types of balance node, PV node and PQ node, and they are supplemented to the corresponding positions according to the basic format of node parameter matrix to obtain the node parameter matrix of the model; the branch parameters include the branch start node number, branch end node number, branch resistance value, branch reactance value, branch to ground susceptance value, transformer ratio and phase angle. According to the branch-related parameters of the model, they are supplemented to the corresponding positions according to the basic format of branch parameter matrix to obtain the branch parameter matrix of the model; the node parameter matrix and the branch parameter matrix together constitute the transmission network parameter matrix of the simulation and solution model of the complex distribution network with distributed generation access;

[0060] Obtain the synchronous motor parameters of the simulation solution model of the complex distribution network with distributed power access, including the synchronous motor stator winding resistance, rotor F winding resistance, rotor D winding resistance, rotor H winding resistance, rotor Q winding resistance, stator d-axis reactance, stator d-axis armature reaction inductance, stator q-axis reactance, stator q-axis armature reaction inductance, rotor F winding reactance, rotor F winding and D winding mutual inductance reactance, rotor D winding reactance, rotor H winding reactance, rotor Q winding reactance, synchronous speed of the synchronous motor, according to the synchronous motor parameters of the model, according to the basic format of the synchronous motor parameter matrix to add to the corresponding position, to obtain the synchronous motor parameter matrix of the model;

[0061] The distributed power system parameters of the complex distribution network simulation solution model with distributed power access are obtained, including the main circuit parameters and control system parameters of the distributed power system. The main circuit parameters include DC source, DC side capacitor, filter inductor and filter capacitor. The control system parameters include phase-locked loop PI parameters and traditional PQ control PI parameters. According to the relevant parameters of the distributed power system of the model, they are supplemented to the corresponding positions according to the basic format of the distributed power system parameter matrix to obtain the distributed power system parameter matrix of the model.

[0062] In S300 of this embodiment, as a preferred embodiment, the power flow of the complex distribution network simulation solution model with distributed power generation access is solved according to the transmission network parameter matrix to obtain the power flow data of each node and each branch of the distribution network, including:

[0063] Based on the node parameter matrix and branch parameter matrix in the obtained transmission network parameter matrix, the node admittance matrix and the number of nodes of each type of the simulation model of the complex distribution network with distributed generation access are solved;

[0064] According to the number of nodes, node type, initial value of voltage amplitude, initial value of voltage phase angle, node admittance matrix, generator active power, generator reactive power, load active power, load reactive power and power flow calculation accuracy, calculate the change of node active power and reactive power, and judge the power flow convergence of complex distribution network simulation solution model;

[0065] According to the input voltage amplitude, voltage phase angle, node admittance matrix, and node number of the voltage amplitude and phase angle to be calculated, the Jacobian matrix of the complex distribution network simulation solution model with distributed generation access is obtained;

[0066] According to the Jacobian matrix, the number of iterations and the accuracy of power flow calculation, the node power flow data and branch power flow data of the complex distribution network simulation model with distributed generation access are obtained.

[0067] In S400 of this embodiment, as a preferred embodiment, based on the flow data, the constant impedance model parameters of the load nodes in the complex distribution network with distributed power access are constructed, and the constant impedance parameters of the load nodes are added to the corresponding positions of the node parameter matrix and the branch parameter matrix to form a new transmission network parameter matrix; in addition, the voltage and current vectors of the synchronous motors in the complex distribution network need to be initialized based on the flow data. Including:

[0068] Construct a constant impedance model for the load node:

[0069] Based on the flow data of each node and each branch of the complex distribution network simulation solution model with distributed power generation access, the node voltage amplitude and phase angle parameters of the node where the load is located are obtained, and then the load active power and load reactive power are used to construct a constant impedance model of the load node, and the constant impedance model parameters corresponding to the load node are obtained;

[0070] According to the basic format of the node parameter matrix and the branch parameter matrix, the constant impedance model parameters of the constructed load nodes are added to the corresponding positions of the node parameter matrix and the branch parameter model, and the nodes of the entire complex distribution network simulation model with distributed generation access are renumbered to obtain a new transmission network parameter matrix.

[0071] Initialize the voltage and current vector of the synchronous motor:

[0072] According to the power flow data in the xy0 coordinate system, the voltage and current results of the synchronous motor are solved:

[0073]

[0074]

[0075]

[0076] [U Gd U Gq ] T =T F [U Gx U Gy ] T

[0077] [I Gd I Gq ] T =T F [I Gx I Gy ] T

[0078]

[0079] Where P G , Q G They are respectively the active power and reactive power output by the generator; U Gx , U Gy are the x-axis component and y-axis component of the generator terminal voltage respectively; I Gx ,I Gy are the x-axis component and y-axis component of the generator output current respectively; δ F is the phase angle difference between the xy axis and the dq axis; r is the resistance of the stator winding; x d 、x q are the reactances of the stator d-axis and q-axis respectively; T F is the coordinate transformation matrix from the xy axis to the dq axis; U Gd , U Gq are the d-axis component and q-axis component of the generator terminal voltage respectively; I Gd ,I Gq are the d-axis component and q-axis component of the generator output current respectively; r F is the resistance of the field winding; x ad is the stator d-axis armature reaction inductance; u F is the excitation voltage of the synchronous motor; ω is the synchronous speed.

[0080] In S500 of this embodiment, as a preferred embodiment, according to the interface capacitance parameter matrix and its node voltage solution equation, the data required by the network interface model is initialized, the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation are obtained, and pre-stored, including:

[0081] The solution equation of the network interface model in the simulation solution model of the complex distribution network with distributed generation access is listed, and the node voltage update equation of the network interface model is obtained as follows:

[0082]

[0083] in, is the interface voltage of the next simulation step, is the interface voltage of the current simulation step, h is the simulation step length, C is the interface capacitance value, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the current simulation step, is the branch current of the interface capacitor in the branch of the transmission network model in the current simulation step, It is the branch current of the interface capacitor in the branch where the power electronic device model is located in the current simulation step;

[0084] According to the node voltage update equation of the network interface model, the data required by the network interface model are initialized, thereby obtaining the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation.

[0085] In S500 of this embodiment, as a preferred embodiment, according to the new transmission network parameter matrix, synchronous motor parameter matrix, distributed power system parameter matrix and their solution equations, the coefficient matrix required for solving each subsystem is solved and pre-stored, and the initial data required for the main loop simulation solution is obtained, including:

[0086] According to the network interface parameter matrix and its node voltage solution equation, the coefficient matrix of the network interface part solution is solved and pre-stored: the solution equation of the network interface model in the complex distribution network simulation solution model with distributed power access is listed, and the voltage update equation of the network interface model is obtained as follows:

[0087]

[0088] in, is the interface voltage of the next simulation step, is the interface voltage of the current simulation step, h is the simulation step length, C is the interface capacitance value, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the current simulation step, is the branch current of the interface capacitor in the branch of the transmission network model in the current simulation step, It is the branch current of the interface capacitor in the branch where the power electronic device model is located in the current simulation step.

[0089] According to the new transmission network parameter matrix and its solution equation, solve and pre-store the coefficient matrix of the transmission network part solution: Figure 3 The KCL equation of the node topology structure and the KVL equation of the branch topology structure in the transmission network model are shown, and the voltage and current update equations of the transmission network part are obtained as follows:

[0090]

[0091]

[0092] Among them, M a is the set of all nodes connected to node a. is the current flowing out of the branch connected to node a, is the current value of the ground current source of node a in the current simulation step, G a is the admittance of node a to the ground, is the voltage of node a in the next simulation step, C a is the capacitance of node a to ground, is the voltage of node a in the current simulation step, is the voltage of node b in the next simulation step, is the voltage value of the voltage source on branch ab in the next simulation step, R ab is the branch resistance of branch ab, is the branch current of branch ab in the current simulation step, L ab is the branch inductance of branch ab, is the branch current of branch ab in the next simulation step, and h is the simulation step length.

[0093] According to the voltage and current update equations of the transmission network model, the data required by the transmission network model are initialized, thereby obtaining the coefficient matrix required for solving the transmission network part and the initial data required for the main loop simulation. It can be found that the node voltage of the current simulation step is only related to the value of the previous simulation step, and its update does not depend on any variable of the current simulation step, so the update of the voltage of each node in the entire system is independent of each other; in addition, the branch current of the current simulation step is also only related to the value of the previous simulation step, so the update of the current of each branch of the system is also independent of each other. Therefore, the calculation process of different nodes and branches of the system is decoupled from each other, and the fine-grained decoupling with a single node and a single branch as the minimum unit has very good parallelism.

[0094] In S500 of this embodiment, as a preferred embodiment, according to the synchronous motor parameter matrix and its solution equation, the coefficient matrix of the synchronous motor part solution is solved and pre-stored: considering the stator d and q windings of the synchronous motor, the excitation winding F and the damping winding D of the rotor d axis, and the damping windings H and Q of the rotor q axis, the electrical part equation of the synchronous motor is described as:

[0095] u=pψ+ωψ+ri,ψ=xi

[0096] Among them, u is the motor voltage vector, p is the differential operator, i is the motor current vector, ψ is the motor flux vector, r is the resistance matrix, x is the reactance matrix, and ω is the speed matrix;

[0097] Substitute the flux equation ψ=xi into the voltage equation u=pψ+ωψ+ri, eliminate the flux variable ψ, and get the voltage equation:

[0098] u=xpi+(ωx+r)i=xpi+yi

[0099] Among them, y is the transformed impedance matrix;

[0100] The trapezoidal integration method is used for differential modeling, and then the zero element in the motor voltage vector u is ignored, and the excitation winding voltage is set to remain constant, then the voltage equation is transformed into:

[0101] Ci(t)=D1u dq (t)+D1u dq (th)+D2u F +Ei(th)

[0102] Among them, C, D1, D2, and E are coefficient matrices generated during the equation transformation process, and u dq is the vector formed by the stator d and q axis voltages, u F is the rotor d-axis excitation voltage, and t is the simulation time.

[0103] D2=[0 0 h 0 0 0] T ;

[0104] Further sorting can be obtained:

[0105] i(t)=A G u dq (t)+B G [i(th) u dq (th) u F ] T

[0106] Among them, A G , B G are the coefficient matrices generated during the equation transformation process, A G =C -1 D1, B G =[C -1 EC -1 D1 C - 1 D2];

[0107] Assume that the rotor motion equation of the synchronous motor is:

[0108]

[0109]

[0110] Among them, Tj , D are the inertia time constant and damping coefficient of the synchronous motor, ω, δ, T m 、T e are the synchronous motor speed, rotor angle, mechanical torque and electromagnetic torque respectively, ω0 is the synchronous speed; using the trapezoidal method for differential processing, the differential form shown below can be obtained:

[0111]

[0112]

[0113] Where ω(t) is the speed of the synchronous motor in the current simulation step, and δ(t) is the electrical angle between the synchronous motor rotor q axis and the system reference axis x rotating at the synchronous speed in the current simulation step;

[0114] The electrical part equations and rotor motion equations of the synchronous motor are transformed into the abc coordinate system by using coordinate transformation, and the transformation matrix from the abc coordinate system to the dq coordinate system is defined as:

[0115]

[0116]

[0117] Among them, T1 is the Park transformation matrix, T2 is the Park inverse transformation matrix, θ is the angle that the synchronous motor rotor d axis leads the stator a axis, and θ=θ0+ωt, θ0 is the initial angle, and u dq =T1u abc ,i dq =T1i abc ,u abc =T2u dq ,i abc =T2i dq ;

[0118] Transform the dq coordinate system locally to the abc coordinate system, and we get:

[0119] i′(t)=T ex A G T1u abc (t)+T ex B G T ex2 [i′(th) u abc (th) u F ] T

[0120] Where i′(t) is the synchronous motor current vector in the abc coordinate system, T ex 、T ex2To maintain the same matrix dimension, the augmented coefficient matrix; i′(t) = [i a (t) i b (t) i c (t) i F (t) i D (t) i H (t) i Q (t)] T ,

[0121] i a (t), i b (t), i c (t) is the three-phase current of the synchronous motor stator, i F (t) is the current of the synchronous motor rotor F winding, i D (t) is the current of synchronous motor rotor D winding, i H (t) is the current of the synchronous motor rotor H winding, i Q (t) is the current of the synchronous motor rotor Q winding, I 4x4 is the 4th-order identity matrix, and I is the 1st-order identity matrix;

[0122] The electrical part of the synchronous motor is estimated based on θ(th) at the last simulation time, and the following is obtained:

[0123]

[0124] in, is the estimated value of the angle that the synchronous motor rotor d-axis leads the stator a-axis, θ(th) is the angle that the synchronous motor rotor d-axis leads the stator a-axis in the previous simulation step, and ω(th) is the synchronous motor speed in the previous simulation step;

[0125] T ex2 Directly use the previous moment value θ(th); T1 and T ex Use the estimated value at the current simulation time

[0126] Separate the electrical equations in the abc coordinate system and solve them simultaneously:

[0127] i abc (t) = A Gabc u abc (t)+B Gabc [i′(th) u abc (th) u F ] T @A Gabc u abc (t)+i hGabc

[0128] i rest (t) = A Grest u abc (t)+B Grest [i′(th) u abc (th) u F ] T @A Grest u abc (t)+i hGrest

[0129] Among them, A′ G =T ex A G T1, A Gabc =A′ G (1∶3, :), A Grest =A′ G (4∶7, :), B′ G =T ex B G T ex2 , B Gabc =B′ G (1∶3, :), B Grest =B′ G (4∶7, :); A Gabc is the equivalent network admittance matrix of the motor part viewed from the stator port, i hGabc is the historical current term of the synchronous motor, A Grest is the equivalent network admittance matrix of the remaining motor part, i hGrest is the remaining historical current term of the synchronous motor;

[0130] According to the above calculation process, the solution process of the synchronous motor part can be obtained, including:

[0131] Based on the voltage and current vector of the synchronous motor in the previous simulation step, T ex , A G And the corresponding amount of T1 to obtain A Gabc and i hGabc , and modify the network admittance matrix and network node injection current accordingly in combination with the stator port number;

[0132] Solve the electrical part equations of each motor branch of the motor part to obtain the current synchronous motor port voltage u abc (t);

[0133] Calculate and update the current step synchronous motor stator three-phase current i abc (t) and other winding current vectors i rest(t), then update the rotor dynamics according to the synchronous motor's rotor motion equation, and store the solution of the synchronous motor's stator three-phase current in the global memory;

[0134] According to the electrical part equations and rotor motion equations of the synchronous motor model, the data required for the synchronous motor model are initialized, thereby obtaining the coefficient matrix required for the partial solution of the synchronous motor and the initial data required for the main loop simulation. It can be found that the solution of the synchronous motor parameters in the form of a block diagonal matrix has a higher sparsity and clear block position, which can be easily reduced in order and allocated to parallel computing resources, thereby realizing parallel solution between synchronous motors. In addition, the direct LU decomposition method is used to solve the internal part of a single synchronous motor, further realizing the parallel solution of its internal solution.

[0135] In S500 of this embodiment, as a preferred embodiment, according to the distributed power system parameter matrix and its solution equations, the coefficient matrix solved by the power electronic equipment part is solved and pre-stored: in a specific application example of this preferred embodiment, a grid-connected converter is used as a structure for connecting a distributed power source to a distribution network to illustrate, and the solution equations on the AC side of the converter are consistent with those on the transmission network. Here, the solution equations on the DC side of the converter and the interface update equations on the AC and DC sides of the converter are emphasized.

[0136] The differential equation for the DC side of the converter is as follows:

[0137]

[0138]

[0139] in, and They are the DC side capacitor C in the current simulation step. s1 and C s2 The voltage on crown They are the DC side capacitor C in the next simulation step. s1 and C s2 The voltage on dc is the DC voltage value, R s is the internal resistance of the DC source, and They are the DC side capacitor C in the current simulation step. s1 and C s2 The value of the injected controlled current source at the node.

[0140] according to Figure 6 The converter main circuit modeling method is shown in the figure. The interface update equations on the AC and DC sides of the converter are as follows:

[0141]

[0142]

[0143]

[0144] Among them, S1, S2, S3, S4, S5, and S6 are switch control signals. The value is 1 when the switch is on, and the value is 0 when the switch is off. They are the three-phase values ​​of the controlled voltage source on the AC side in the next simulation step.

[0145]

[0146]

[0147] in, They are the three-phase values ​​of the current on the AC side filter inductor in the next simulation step, They are the DC side capacitor C in the next simulation step. s1 and C s2 The value of the injected controlled current source at the node.

[0148] The control system of the converter adopts a double closed-loop control structure, the outer loop is the power control loop, and the inner loop is the current control loop. Figure 5 Measure the grid-side voltage and current and calculate the output power P g , Q g , and then with the given value of output power P gref , Q gref By comparison, the given value i of the grid-side current dq axis is generated after PI regulation * gd and i * gq , and then compared with the actual measured i gd and i gq After comparison, it is sent to the PI regulator and the voltage compensation term is considered, so that the reference voltage on the dq axis can be obtained, which is used to generate the PWM signal to control the converter. Different from the main circuit solution, the serial degree of the control link is high, so it is considered to solve the control systems of multiple power electronic devices in parallel.

[0149] According to the solution equations of the AC and DC sides of the converter and the interface update equations of the AC and DC sides, the data required for the solution equations of the AC and DC sides of the converter and the interface update equations of the AC and DC sides are initialized, thereby obtaining the coefficient matrix and initial data required for solving the power electronic equipment part, and it can be found that the solution of the main circuit of the power electronic equipment part is consistent with the transmission network, with the parallelism of a single node and a single branch, and the parallelism of the control system solution is a single device.

[0150] In S600 of this embodiment, as a preferred embodiment, fine-grained parallel solving between models and within models includes:

[0151] Firstly, the branch current of the branch where the interface capacitor is located in the synchronous motor part is solved according to the electrical part equation of the synchronous motor part, and the branch current vector of the current simulation step of the synchronous motor part is obtained and stored;

[0152] Then, according to the solution of the AC side and the DC side of the power electronic equipment part, the branch current of the branch where the interface capacitor is located in the power electronic equipment part is solved, and the branch current vector of the current simulation step of the power electronic equipment part is obtained and stored;

[0153] Then, according to the solution equation of the transmission network part, the branch current of the branch where the interface capacitor is located in the transmission network part is solved, and the branch current vector of the current simulation step of the transmission network part is obtained and stored;

[0154] The three branch current vectors obtained above are called, and the node voltage update equation of the network interface model is used to solve the voltage vector of the interface capacitor to obtain the interface node voltage of the current simulation step and store it;

[0155] At this point, the calculation of each subsystem in a simulation cycle is completed after decoupling. In the next simulation step, the transmission network part, the synchronous motor part, and the power electronic equipment part call the stored interface capacitor voltage vector to perform their respective simulation calculations, and then send their respective branch current vectors back to the network interface part to facilitate the network interface part to update the interface voltage vector in the next simulation step. The timing diagram of the single-step simulation solution is shown in the figure below. Figure 7 shown.

[0156] The fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power access provided in this embodiment can combine the characteristics of the GPU parallel computing architecture, make full use of the advantages of the GPU's massive computing resources, and realize fine-grained parallel solution based on a single branch, a single node, and a single device as the minimum solution unit, thereby ensuring that the system simulation efficiency will not be significantly reduced as the scale of the system simulation increases. In this embodiment, fine-grained parallelism is relative to the current mainstream methods that all implement coarse-grained parallel computing for small systems (containing more branches and nodes).

[0157] The fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power access provided in this embodiment adopts a fast transient simulation method - delayed insertion method, and the coefficient matrix formed is a diagonal matrix, which has the advantage of natural component-level parallelism. Its numerical complexity changes linearly with the increase of simulation scale, which is more suitable for processing large-scale complex distribution networks and improves efficiency.

[0158] Figure 2 A flow chart of a fine-grained parallel electromagnetic transient simulation algorithm for a complex distribution network with distributed power generation access provided by a preferred embodiment of the present invention.

[0159] like Figure 2 As shown, the fine-grained parallel electromagnetic transient simulation method for a complex distribution network with distributed power access provided by the preferred embodiment may include the following steps:

[0160] Step 1, constructing an electromagnetic transient simulation model of a complex distribution network with distributed generation access after decoupling;

[0161] Step 2, for the decoupled simulation model, obtaining the topological transmission network parameter matrix, the synchronous motor parameter matrix and the distributed power system parameter matrix;

[0162] Step 3, solving the distribution network power flow according to the transmission network parameter matrix;

[0163] Step 4: Initialize the voltage and current vector of the system synchronous motor according to the power flow results, solve the constant impedance model parameters of the load node in the complex distribution network with distributed power generation access, and then add the load node to the corresponding position of the transmission network parameter matrix in the form of constant impedance;

[0164] Step 5, according to the new transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix and their solution equations, solve and pre-store the coefficient matrix and initial data required for solving each subsystem;

[0165] Step 6: Using the initial simulation data, reasonably allocate GPU threads according to the complexity of each subsystem solution to achieve fine-grained parallel solution between and within subsystems, and repeat this cycle until the simulation process ends.

[0166] According to the above simulation process, it is implemented in combination with the existing high-performance hardware GPU.

[0167] In the preferred embodiment:

[0168] The electromagnetic transient simulation model of a complex distribution network with distributed generation access is constructed using Figure 4 The two-port Thevenin equivalent shown in the figure is used to partition and reorganize the system, and the capacitor branch is selected as the interface between the transmission network part and the synchronous motor part and the power electronic equipment part. For the transmission network part, the synchronous motor part and the power electronic equipment part, the ground capacitance connected to them is equivalent to the voltage source on the branch where they are located. At the same time, for the network interface, the subsystems connected to them are equivalent to the ground current source on the node where they are located.

[0169] Simulate each decoupled subsystem, including:

[0170] The network interface involves the voltage update of a single capacitor, and its differential model can be written as:

[0171]

[0172] in, is the interface voltage of the next simulation step, is the interface voltage of the current simulation step, h is the simulation step length, C is the interface capacitance value, is the branch current of the interface capacitor in the branch where the synchronous motor is located in the current simulation step, is the branch current of the interface capacitor in the transmission network under the current simulation step, It is the branch current of the interface capacitor in the branch where the power electronic equipment part is located in the current simulation step.

[0173] The solution equation of the network interface model in the simulation solution model of the complex distribution network with distributed generation access is listed, and the node voltage update equation of the network interface model is obtained as follows:

[0174]

[0175] in, is the interface voltage of the next simulation step, is the interface voltage of the current simulation step, h is the simulation step length, C is the interface capacitance value, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the current simulation step, is the branch current of the interface capacitor in the branch of the transmission network model in the current simulation step, It is the branch current of the interface capacitor in the branch where the power electronic device model is located in the current simulation step.

[0176] List Figure 3 The voltage and current solving equations in the transmission network model shown are used to solve the initial simulation data and coefficient matrix of the transmission network part, including: the KCL equation of the node topology structure of the transmission network part and the KVL equation of the branch topology structure, and the voltage and current update equation of the transmission network part is obtained as follows:

[0177]

[0178]

[0179] Among them, M a is the set of all nodes connected to node a. is the current flowing out of the branch connected to node a, is the current value of the ground current source of node a in the current simulation step, G a is the admittance of node a to the ground, is the voltage of node a in the next simulation step, C a is the capacitance of node a to ground, is the voltage of node a in the current simulation step, is the voltage of node b in the next simulation step, is the voltage value of the voltage source on branch ab in the next simulation step, R ab is the branch resistance of branch ab, is the branch current of branch ab in the current simulation step, L ab is the branch inductance of branch ab, is the branch current of branch ab in the next simulation step, and h is the simulation step length.

[0180] Write down the solution equations of the synchronous motor part, solve the initial simulation data and coefficient matrix of the synchronous motor part, including: considering the stator d, q windings of the synchronous motor, the excitation winding F and damping winding D of the rotor d axis, and the damping windings H and Q of the rotor q axis, the electrical part equation of the synchronous motor is described as:

[0181] u=pψ+ωψ+ri,ψ=xi

[0182] Among them, u is the motor voltage vector, p is the differential operator, i is the motor current vector, ψ is the motor flux vector, r is the resistance matrix, x is the reactance matrix, and ω is the speed matrix;

[0183] Substitute the flux equation ψ=xi into the voltage equation u=pψ+ωψ+ri, eliminate the flux variable ψ, and get the voltage equation:

[0184] u=xpi+(ωx+r)i=xpi+yi

[0185] Among them, y is the transformed impedance matrix;

[0186] The trapezoidal integration method is used for differential modeling, and then the zero element in the motor voltage vector u is ignored, and the excitation winding voltage is set to remain constant, then the voltage equation is transformed into:

[0187] Ci(t)=D1u dq (t)+D1u dq (th)+D2u F +Ei(th)

[0188] Among them, C, D1, D2, and E are coefficient matrices generated during the equation transformation process, and u dq is the vector formed by the stator d and q axis voltages, u F is the rotor d-axis excitation voltage, and t is the simulation time.

[0189] D2=[0 0 h 0 0 0]T ;

[0190] Further sorting can be obtained:

[0191] i(t)=A G u dq (t)+B G [i(th) u dq (th) u F ] T

[0192] Among them, A G , B G are the coefficient matrices generated during the equation transformation process, A G =C -1 D1, B G =[C -1 EC -1 D1 C - 1 D2];

[0193] Assume that the rotor motion equation of the synchronous motor is:

[0194]

[0195]

[0196] Among them, T j , D are the inertia time constant and damping coefficient of the synchronous motor, ω, δ, T m , T e are the synchronous motor speed, rotor angle, mechanical torque and electromagnetic torque respectively, ω0 is the synchronous speed; using the trapezoidal method for differential processing, the differential form shown below can be obtained:

[0197]

[0198]

[0199] Where ω(t) is the speed of the synchronous motor in the current simulation step, and δ(t) is the electrical angle between the synchronous motor rotor q axis and the system reference axis x rotating at the synchronous speed in the current simulation step;

[0200] The electrical part equations and rotor motion equations of the synchronous motor are transformed into the abc coordinate system by using coordinate transformation, and the transformation matrix from the abc coordinate system to the dq coordinate system is defined as:

[0201]

[0202]

[0203] Among them, T1 is the Park transformation matrix, T2 is the Park inverse transformation matrix, θ is the angle that the synchronous motor rotor d axis leads the stator a axis, and θ=θ0+ωt, θ0 is the initial angle, and u dq =T1u abc ,i dq =T1i abc ,u abc =T2u dq ,i abc =T2i dq ;

[0204] Transform the dq coordinate system locally to the abc coordinate system, and we get:

[0205] i′(t)=T ex A G T1u abc (t)+T ex B G T ex2 [i′(th) u abc (th) u F ] T

[0206] Where i′(t) is the synchronous motor current vector in the abc coordinate system, T ex 、T ex2 To maintain the same matrix dimension, the augmented coefficient matrix; i′(t) = [i a (t) i b (t) i c (t) i F (t) i D (t) i H (t) i Q (t)] T ,

[0207] i a (t), i b (t), i c (t) is the three-phase current of the synchronous motor stator, i F (t) is the current of the synchronous motor rotor F winding, i D (t) is the current of the synchronous motor rotor D winding, i H (t) is the current of the synchronous motor rotor H winding, i Q (t) is the current of the synchronous motor rotor Q winding, I 4x4 is the 4th-order identity matrix, and I is the 1st-order identity matrix;

[0208] The electrical part of the synchronous motor is estimated based on θ(th) at the last simulation time, and the following is obtained:

[0209]

[0210] in, is the estimated value of the angle that the synchronous motor rotor d-axis leads the stator a-axis, θ(th) is the angle that the synchronous motor rotor d-axis leads the stator a-axis in the previous simulation step, and ω(th) is the synchronous motor speed in the previous simulation step;

[0211] T ex2 Directly use the previous moment value θ(th); T1 and T ex Use the estimated value at the current simulation time

[0212] Separate the electrical equations in the abc coordinate system and solve them simultaneously:

[0213] i abc (t) = A Gabc u abc (t)+B Gabc [i′(th) u abc (th) u F ] T @A Gabc u abc (t)+i hGabc

[0214] i rest (t) = A Grest u abc (t)+B Grest [i′(th) u abc (th) u F ] T @A Grest u abc (t)+i hGrest

[0215] Among them, A′ G =T ex A G T1, A Gabc =A′ G (1∶3, :), A Grest =A′ G (4∶7, :), B′ G =T ex B G T ex2 , B Gabc =B′ G (1∶3, :), B Grest =B′G (4∶7, :); A Gabc is the equivalent network admittance matrix of the motor part viewed from the stator port, i hGabc is the historical current term of the synchronous motor, A Grest is the equivalent network admittance matrix of the remaining motor part, i hGrest is the remaining historical current term of the synchronous motor.

[0216] List the solution equations of the power electronic equipment part, and solve the simulation initial data and coefficient matrix of the power electronic equipment part, including: taking the grid-connected converter as an example to explain the access of distributed power sources to the distribution network, the solution equations on the AC side of the converter are consistent with those on the transmission network part, and here we focus on the solution equations on the DC side and the interface update equations on the AC and DC sides.

[0217] according to Figure 6 The converter main circuit modeling method shown in the figure, the solution equation of the converter DC side after differentiation is as follows:

[0218]

[0219]

[0220] in, and They are the DC side capacitor C in the current simulation step. s1 and C s2 The voltage on and They are the DC side capacitor C in the next simulation step. s1 and C s2 The voltage on dc is the DC voltage value, R s is the internal resistance of the DC source, H″1 and H″2 are the DC side capacitance C in the current simulation step. s1 and C s2 The value of the injected controlled current source at the node.

[0221] The interface update equations on the AC and DC sides of the converter are as follows:

[0222]

[0223]

[0224]

[0225] Among them, S1, S2, S3, S4, S5, and S6 are switch control signals. The value is 1 when the switch is on, and the value is 0 when the switch is off. They are the three-phase values ​​of the controlled voltage source on the AC side in the next simulation step.

[0226]

[0227]

[0228] in, They are the three-phase values ​​of the current on the AC side filter inductor in the next simulation step, They are the DC side capacitor C in the next simulation step. s1 and C s2 The value of the injected controlled current source at the node.

[0229] The control system of the converter adopts a double closed-loop control structure, the outer loop is the power control loop, and the inner loop is the current control loop. Figure 5 Measure the grid-side voltage and current and calculate the output power P g , Q g , and then with the given value of output power P gref , Q gref By comparison, the given value i of the grid-side current dq axis is generated after PI regulation * gd and i * gq , and then compared with the actual measured i gd and i gq After comparison, it is sent to the PI regulator and the voltage compensation term is considered, so that the reference voltage on the dq axis can be obtained, which is used to generate the PWM signal to control the converter. Different from the main circuit solution, the serial degree of the control link is high, so it is considered to solve the control systems of multiple power electronic devices in parallel.

[0230] Based on solving equations for the above subsystems, the entire simulation process includes:

[0231] Firstly, the branch current of the branch where the interface capacitor is located in the synchronous motor part is solved according to the electrical part equation of the synchronous motor part, and the branch current vector of the current simulation step of the synchronous motor part is obtained and stored;

[0232] Then, the branch current of the branch where the interface capacitor is located in the power electronic equipment part is solved according to the electrical part equation of the power electronic equipment part, and the branch current vector of the current simulation step of the power electronic equipment part is obtained and stored;

[0233] Then, according to the solution equation of the transmission network part, the branch current of the branch where the interface capacitor is located in the transmission network part is solved, and the branch current vector of the current simulation step of the transmission network part is obtained and stored;

[0234] The three branch current vectors obtained above are called, and the voltage vector of the interface capacitor is solved using the node voltage solution equation mentioned above to obtain the interface node voltage of the current simulation step and store it;

[0235] At this point, the calculation of each subsystem in a simulation cycle is completed after decoupling. In the next simulation step, the transmission network part, the synchronous motor part, and the power electronic equipment part call the stored interface capacitor voltage vector to perform their respective simulation calculations, and then send their respective branch current vectors back to the network interface part to facilitate the network interface part to update the interface voltage vector in the next simulation step, and this cycle continues until the simulation ends. The timing diagram of the single-step simulation solution is shown in Figure 7 shown.

[0236] At this point, the simulation cycle process of the entire decoupling system has been explained.

[0237] For ease of understanding, the following will be Figure 8 The examples shown are used as examples to illustrate the present invention, but should not be used to limit the protection scope of the present invention.

[0238] Table 1 is the node parameter matrix in the transmission network parameter matrix, Table 2 is the branch parameter matrix in the transmission network parameter matrix, Table 3 is the synchronous motor parameter matrix, and Table 4 is the distributed power system parameter matrix. Import the transmission network, synchronous motor and power electronic equipment parameter information of the distribution network example:

[0239] Table 1

[0240]

[0241] Table 2

[0242]

[0243] Table 3

[0244]

[0245] Table 4

[0246]

[0247] The power flow results of the system are shown in Table 5:

[0248] Table 5

[0249]

[0250] The system load nodes are modeled as constant impedance using the power flow results, and the load nodes are added to the system input matrix in the form of constant impedance, as shown in Table 6:

[0251]

[0252] Then, the coefficient matrix required for solving each subsystem is preprocessed. After the relevant parameter matrix is ​​preprocessed, the main loop can be entered to update the output current according to the solution equations of the transmission network part, the synchronous motor part and the power electronic equipment part. The interface capacitor updates the interface node voltage according to the solution results of the corresponding branch current of each subsystem, and then transmits the voltage update results to the corresponding branches of each subsystem in the form of a controlled voltage source. This cycle continues until the simulation process ends. Among them, the pre-storage of the coefficient matrix of each subsystem, the initialization of the synchronous motor and the storage of the results are completed on the CPU. In addition, before using the GPU to solve, each subsystem needs to process the results of the related subsystems to obtain the required information, which is also completed by the CPU. The updates of the transmission network node voltage and branch current, the output current of the generator, the output voltage and current of the power electronic equipment, and the interface capacitor voltage are all completed on the GPU.

[0253] by Figure 8 The example shown is a basic topology to verify the accuracy of distribution network simulation, and different scale systems are used as basic topologies to verify the efficiency of system simulation. The relevant simulation steps are the same as before.

[0254] See the accuracy verification results for Fig. 9 For the efficiency verification results in (a), (b) and (c), please refer to Fig.10 . Fig. 9 (a), (b) and (c) are respectively comparisons of the simulation accuracy of the transmission network part, the synchronous motor part and the power electronic equipment part in the embodiment of the present invention and the accuracy of the results solved by the traditional node analysis method. It can be seen that the fine-grained parallel simulation method in the embodiment of the present invention has a basically consistent response curve with the node analysis method regardless of whether it is in steady state or transient conditions, which verifies the correctness of the fine-grained parallel simulation method as an electromagnetic transient simulation algorithm. In order to verify the superiority of the fine-grained characteristics of the proposed simulation algorithm in efficiency, simulation efficiency tests are considered for systems of different scales. The changes in system scale are mainly reflected in two aspects, namely the number of connected distributed power sources, the number of nodes in the grid-side system, and the number of motors. The simulation efficiency of the system is tested by changing the indicators of these aspects of the system. The relevant efficiency test results are shown in Figure 2. Fig.10 shown.

[0255] Fig.11 A schematic diagram of the structure of a fine-grained parallel electromagnetic transient simulation system for a complex distribution network with distributed power access provided by an embodiment of the present invention. The fine-grained parallel electromagnetic transient simulation system for a complex distribution network with distributed power access provided by this embodiment may include: a complex distribution network modeling module with distributed power access, a parameter matrix construction module, a coefficient matrix and simulation initial data acquisition module, and a fine-grained parallel simulation module. Among them:

[0256] A modeling module for complex distribution networks with distributed power access, which constructs a simulation and solution model for complex distribution networks with distributed power access. The simulation and solution model for complex distribution networks with distributed power access includes: a transmission network model, a synchronous motor model, a power electronic equipment model, and a network interface model. The transmission network model, the synchronous motor model, and the power electronic equipment model can achieve independent fine-grained parallel solutions only by relying on the voltage of the interface capacitor of the network interface model. The network interface model is solved based on the current update results of the transmission network model, the synchronous motor model, and the power electronic equipment model.

[0257] A parameter matrix construction module, which targets the complex distribution network simulation solution model with distributed power generation access, obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology, and then forms the corresponding interface capacitance parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix; according to the transmission network parameter matrix, solves the system flow of the complex distribution network simulation solution model with distributed power generation access, and obtains the flow data of each node and each branch in the complex distribution network system; according to the flow data, solves the constant impedance model parameters of the load nodes in the complex distribution network simulation solution model with distributed power generation access, and adds the constant impedance model parameters of the load nodes to the transmission network parameter matrix to form a new transmission network parameter matrix; at the same time, initializes the voltage and current vector of the synchronous motor model according to the flow data;

[0258] A coefficient matrix and simulation initial data acquisition module, which initializes the data required for the transmission network model, the synchronous motor model, the power electronic equipment model and the network interface model according to the new transmission network parameter matrix, the synchronous motor parameter matrix, the distributed power supply parameter matrix and their solution equations, obtains the coefficient matrix required for solving the node voltage and branch current of the transmission network model, the coefficient matrix required for solving the voltage and current of the motor of the synchronous motor model and the rotor dynamics of the motor, the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model, and the initial data required for the main cycle simulation of each model, and pre-stores them;

[0259] Fine-grained parallel simulation module, which uses the initial data to update the voltage and current of the transmission network model, synchronous motor model, power electronic equipment model and network interface model respectively; the parallelism of the transmission network model is improved to a single node and a single branch, and after combining with the GPU's massive computing units, the parallel solution of each node and each branch in the transmission network model is realized; the parallelism of the synchronous motor model is improved to a single motor, where the solution of each motor is carried out using the direct LU decomposition method, thereby realizing parallel solution between and within synchronous motors; the solution of the main circuit of the power electronic equipment model is consistent with the transmission network model, and each node and each branch of the main circuit is executed in parallel. The parallelism of the control circuit of the power electronic equipment model is a single device, where the control system of each converter is allocated a thread for parallel solution; the parallelism of the network interface model is a single node, and the parallel solution of each node is realized after the GPU is allocated with the corresponding number of threads; thereby, fine-grained parallel solution between and within the models is realized.

[0260] This cycle continues until the simulation ends.

[0261] It should be noted that the steps in the method provided by the present invention can be implemented by using corresponding modules, devices, units, etc. in the system. Those skilled in the art can refer to the technical solution of the method to realize the composition of the system, that is, the embodiments in the method can be understood as preferred examples for constructing the system, which will not be elaborated here.

[0262] An embodiment of the present invention provides a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it can be used to execute any one of the methods in the above embodiments of the present invention.

[0263] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, it can be used to execute any one of the methods in the above embodiments of the present invention.

[0264] Optionally, the memory is used to store programs; the memory may include volatile memory (English: volatile memory), such as random-access memory (English: random-access memory, abbreviated: RAM), such as static random-access memory (English: static random-access memory, abbreviated: SRAM), double data rate synchronous dynamic random access memory (English: Double Data Rate Synchronous Dynamic Random Access Memory, abbreviated: DDR SDRAM), etc.; the memory may also include non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory). The memory is used to store computer programs (such as applications, functional modules, etc. that implement the above method), computer instructions, etc., and the above computer programs, computer instructions, etc. can be partitioned and stored in one or more memories. And the above computer programs, computer instructions, data, etc. can be called by the processor.

[0265] The above-mentioned computer programs, computer instructions, etc. may be stored in one or more memories in partitions, and the above-mentioned computer programs, computer instructions, data, etc. may be called by a processor.

[0266] The processor is used to execute the computer program stored in the memory to implement the various steps in the method involved in the above embodiment. For details, please refer to the relevant description in the above method embodiment.

[0267] The processor and the memory may be independent structures or integrated structures. When the processor and the memory are independent structures, the memory and the processor may be coupled and connected via a bus.

[0268] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art may make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A fine-grained parallel electromagnetic transient simulation method for a distribution network, characterized in that: include: Constructing a simulation solution model for a complex distribution network with distributed power access, the simulation solution model for a complex distribution network with distributed power access includes: a transmission network model, a synchronous motor model, a power electronic equipment model and a network interface model, wherein the transmission network model, the synchronous motor model and the power electronic equipment model can realize independent fine-grained parallel solution only by relying on the interface capacitor voltage of the network interface model, and the network interface model is solved according to the current update results of the transmission network model, the synchronous motor model and the power electronic equipment model; For the simulation solution model of the complex distribution network with distributed power access, the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology are obtained, and then the corresponding interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix are formed; Solving the system power flow of the complex distribution network simulation solution model with distributed power generation access according to the transmission network parameter matrix, and obtaining the power flow data of each node and each branch in the complex distribution network system; According to the power flow data, the constant impedance model parameters of the load nodes in the simulation solution model of the complex distribution network with distributed power generation access are solved, and the constant impedance model parameters of the load nodes are added to the transmission network parameter matrix to form a new transmission network parameter matrix; at the same time, according to the power flow data, the voltage and current vector of the synchronous motor model are initialized; According to the interface capacitance parameter matrix and its node voltage solution equation, the data required by the network interface model is initialized to obtain the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation, and pre-store them; According to the new transmission network parameter matrix and its node voltage and branch current solution equations, the data required by the transmission network model are initialized to obtain the coefficient matrix required for solving the node voltage and branch current of the transmission network model and the initial data required for the main loop simulation, and pre-store them; According to the synchronous motor parameter matrix and its electrical equation and rotor motion equation, the data required for the synchronous motor model is initialized to obtain the voltage and current of the synchronous motor model motor and the coefficient matrix required for the rotor dynamic solution of the motor and the initial data required for the main loop simulation, and pre-store them; According to the distributed power system parameter matrix and its main circuit solution equation and control system solution equation, the data required by the power electronic equipment model is initialized to obtain the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model and the initial data required for the main cycle simulation, and pre-store them; The transmission network model, synchronous motor model, power electronic equipment model and network interface model are respectively updated with the initial data required for the corresponding main loop simulation for the next step of voltage and current; wherein: the parallelism of the transmission network model is increased to a single node and a single branch, and the GPU is used to realize the parallel solution of each node and each branch in the transmission network model to obtain the next step of voltage and current update; the parallelism of the synchronous motor model is increased to a single motor, and the solution of each motor is performed based on the GPU using the direct LU decomposition method, thereby realizing the parallel solution between and within the synchronous motors to obtain The next step is voltage and current update; the parallelism of the main circuit of the power electronic equipment model is improved to be consistent with the transmission network model, and each node and each branch of the main circuit are solved in parallel. The parallelism of the control circuit of the power electronic equipment model is improved to a single device, in which each converter control system allocates a thread for parallel solution to obtain the next step of voltage and current update; the parallelism of the network interface model is improved to a single node, and the GPU is used to allocate a corresponding number of threads to achieve parallel solution of each node to obtain the next step of voltage and current update; thereby, fine-grained parallel solution between and within each model is achieved; This cycle continues until the entire simulation process is completed; The simulation solution model for the complex distribution network with distributed power access obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology, and then forms the corresponding interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix, including: Obtaining interface capacitance parameters of the complex distribution network simulation solution model with distributed power access, including interface node numbers and capacitance values ​​corresponding to the interface nodes, and adding the interface node numbers and capacitance values ​​corresponding to the interface nodes to corresponding positions of the interface capacitance parameter matrix to obtain the interface capacitance parameter matrix of the complex distribution network simulation solution model with distributed power access; The transmission network parameters of the complex distribution network simulation solution model with distributed power generation access are obtained, including node parameters and branch parameters, wherein the node parameters include node number, voltage amplitude, voltage phase angle, motor active power, motor reactive power, load active power, load reactive power, conductance G, susceptance B and node type number. The node parameters required for the power flow solution of the complex distribution network simulation solution model with distributed power generation access are obtained according to the three different node types of balance node, PV node and PQ node, and supplemented to the corresponding positions according to the basic format of the node parameter matrix to obtain the complex distribution network simulation solution model with distributed power generation access. The node parameter matrix of the distribution network simulation solution model; the branch parameters include the branch start node number, the branch end node number, the branch resistance value, the branch reactance value, the branch ground susceptance value, the transformer ratio and the phase angle, and according to the branch-related parameters of the complex distribution network simulation solution model with distributed power access, the basic format of the branch parameter matrix is ​​supplemented to the corresponding position to obtain the branch parameter matrix of the complex distribution network simulation solution model with distributed power access; the node parameter matrix and the branch parameter matrix together constitute the transmission network parameter matrix of the complex distribution network simulation solution model with distributed power access; Acquire the synchronous motor parameters of the complex distribution network simulation solution model with distributed power access, including the synchronous motor stator winding resistance, rotor F winding resistance, rotor D winding resistance, rotor H winding resistance, rotor Q winding resistance, stator d-axis reactance, stator d-axis armature reaction inductance, stator q-axis reactance, stator q-axis armature reaction inductance, rotor F winding reactance, rotor F winding and D winding mutual inductance reactance, rotor D winding reactance, rotor H winding reactance, rotor Q winding reactance and synchronous speed of the synchronous motor, according to the synchronous motor parameters of the complex distribution network simulation solution model with distributed power access, according to the basic format of the synchronous motor parameter matrix, supplement to the corresponding position, and obtain the synchronous motor parameter matrix of the complex distribution network simulation solution model with distributed power access; Obtain the distributed power system parameters of the complex distribution network simulation solution model with distributed power access, including the main circuit parameters and control system parameters of the distributed power system, wherein the main circuit parameters include a DC source, a DC side capacitor, a filter inductor and a filter capacitor; the control system parameters include phase-locked loop PI parameters and PI parameters of traditional PQ control; and according to the distributed power system related parameters of the complex distribution network simulation solution model with distributed power access, supplement them to the corresponding positions in accordance with the basic format of the distributed power system parameter matrix to obtain the distributed power system parameter matrix of the complex distribution network simulation solution model with distributed power access.

2. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The construction of a complex distribution network simulation solution model with distributed power generation access includes: The distribution network system is partitioned and reorganized using the dual-port Thevenin equivalent, and the capacitor branch is selected as the interface between the transmission network model and the synchronous motor model and the power electronic equipment model; for the transmission network model, the synchronous motor model and the power electronic equipment model, the ground capacitance connected to them is equivalent to the voltage source on the branch where they are located, and for the network interface model, each model connected to it is equivalent to the ground current source on the node where it is located; the expressions solved by each model after decoupling are: in, is the interface voltage of the next simulation step, f C is the solution function for updating the node voltage of the network interface model, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the current simulation step, is the branch current of the interface capacitor in the branch of the transmission network model in the current simulation step, is the branch current of the interface capacitor in the branch where the power electronic device model is located in the current simulation step, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the next simulation step, f G is the solution function for updating the branch current of the synchronous motor model, is the branch current of the interface capacitor in the branch of the transmission network model in the next simulation step, f NET is the solution function for updating the branch current of the transmission network model, is the branch current of the interface capacitor in the branch where the power electronic device model is located in the next simulation step, f CV Solver function for branch current updates of power electronic equipment models.

3. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The method of solving the system flow of the complex distribution network simulation solution model with distributed power generation access according to the transmission network parameter matrix to obtain the flow data of each node and each branch in the complex distribution network system includes: Solving the node admittance matrix and the number of nodes of each type of the simulation model of the complex distribution network with distributed power generation access according to the node parameter matrix and the branch parameter matrix in the obtained transmission network parameter matrix; According to the number of nodes, node type, initial value of voltage amplitude, initial value of voltage phase angle, node admittance matrix, motor active power, motor reactive power, load active power, load reactive power and power flow calculation accuracy, calculate the change of node active power and reactive power, and judge the power flow convergence of the simulation solution model of the complex distribution network with distributed power generation access; According to the input voltage amplitude, voltage phase angle, node admittance matrix, and node number of the voltage amplitude and phase angle to be calculated, the Jacobian matrix of the complex distribution network simulation solution model containing distributed power generation access is calculated; According to the Jacobian matrix, the number of iterations and the power flow calculation accuracy, the node power flow data and the branch power flow data of the complex distribution network simulation solution model containing distributed power generation access are obtained.

4. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The constant impedance model parameters of the load nodes in the complex distribution network simulation solution model with distributed power generation access are solved according to the power flow data, and the constant impedance model parameters of the load nodes are added to the power transmission network parameter matrix to form a new power transmission network parameter matrix; At the same time, according to the power flow data, the voltage and current vector of the synchronous motor model is initialized, including: Construct a constant impedance model for the load node: According to the flow data of each node and each branch of the complex distribution network simulation solution model with distributed power generation access, the node voltage amplitude and phase angle parameters of the node where the load is located are obtained, and then the load active power and load reactive power are used to construct a constant impedance model of the load node, and the constant impedance model parameters corresponding to the load node are obtained; According to the basic format of the node parameter matrix and the branch parameter matrix in the transmission network parameter matrix, the constant impedance model parameters of the constructed load nodes are added to the corresponding positions of the node parameter matrix and the branch parameter model, and the nodes of the entire complex distribution network simulation model with distributed power generation access are renumbered to obtain a new transmission network parameter matrix; Initialize the voltage and current vector of the synchronous motor model: According to the power flow data in the xy0 coordinate system, the voltage and current results of the synchronous motor model are solved: [U Gd U Gq ] T =T F [U Gx U Gy ] T [I Gd I Gq ] T =T F [I Gx I Gy ] T Among them, P G , Q G They are respectively the active power and reactive power output by the generator; U Gx , U Gy are the x-axis component and y-axis component of the generator terminal voltage respectively; I Gx ,I Gy are the x-axis component and y-axis component of the generator output current respectively; δ F is the phase angle difference between the xy axis and the dq axis; r is the resistance of the stator winding; x d 、x q are the reactances of the stator d-axis and q-axis respectively; T F is the coordinate transformation matrix from the xy axis to the dq axis; U Gd , U Gq are the d-axis component and q-axis component of the generator terminal voltage respectively; I Gd ,I Gq are the d-axis component and q-axis component of the generator output current respectively; r F is the resistance of the field winding; x ad is the stator d-axis armature reaction inductance; u F is the excitation voltage of the synchronous motor; ω is the synchronous speed.

5. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The method of solving the equation for the interface capacitance parameter matrix and the node voltage thereof, performing an initialization operation on the data required by the network interface model, obtaining the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation, and pre-storing the data, includes: The solution equation of the network interface model in the simulation solution model of the complex distribution network with distributed power access is listed, and the node voltage update equation of the network interface model is obtained as follows: in, is the interface voltage of the next simulation step, is the interface voltage of the current simulation step, h is the simulation step length, C is the interface capacitance value, is the branch current of the interface capacitor in the branch where the synchronous motor model is located in the current simulation step, is the branch current of the interface capacitor in the branch of the transmission network model in the current simulation step, It is the branch current of the interface capacitor in the branch where the power electronic device model is located in the current simulation step; According to the node voltage update equation of the network interface model, the data required by the network interface model is initialized, thereby obtaining the coefficient matrix required for solving the node voltage of the network interface model and the initial data required for the main loop simulation.

6. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: According to the new transmission network parameter matrix and its node voltage and branch current solution equations, the data required by the transmission network model are initialized to obtain the coefficient matrix required for solving the node voltage and branch current of the transmission network model and the initial data required for the main loop simulation, and pre-store them, including: The KCL equations of the node topology structure and the KVL equations of the branch topology structure of the transmission network model in the simulation solution model of the complex distribution network with distributed generation access are listed to obtain the voltage and current update equations of the transmission network model: Among them, M a is the set of all nodes connected to node a. is the current flowing out of the branch connected to node a, is the current value of the ground current source of node a in the current simulation step, G a is the admittance of node a to the ground, is the voltage of node a in the next simulation step, C a is the capacitance of node a to ground, is the voltage of node a in the current simulation step, is the voltage of node b in the next simulation step, is the voltage value of the voltage source on branch ab in the next simulation step, R ab is the branch resistance of branch ab, is the branch current of branch ab in the current simulation step, L ab is the branch inductance of branch ab, is the branch current of branch ab in the next simulation step, and h is the simulation step length; According to the voltage and current update equations of the power transmission network model, the data required by the power transmission network model are initialized, thereby obtaining the coefficient matrix required for solving the power transmission network model and the initial data required for the main loop simulation.

7. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The method of performing an initialization operation on the data required for the synchronous motor model according to the synchronous motor parameter matrix and its electrical equation and rotor motion equation, obtaining the voltage and current of the synchronous motor model motor and the coefficient matrix required for the rotor dynamic solution of the motor and the initial data required for the main loop simulation, and pre-storing the data, includes: Considering the stator d and q windings of the synchronous motor model, the excitation winding F and damping winding D of the rotor d axis, and the damping windings H and Q of the rotor q axis, the electrical part equation of the synchronous motor model is described as: u=pψ+ωψ+ri,ψ=xi Among them, u is the motor voltage vector, p is the differential operator, i is the motor current vector, ψ is the motor flux vector, r is the resistance matrix, x is the reactance matrix, and ω is the speed matrix; Substitute the flux equation ψ=xi into the voltage equation u=pψ+ωψ+ri, eliminate the flux variable ψ, and get the voltage equation: u=xpi+(ωx+r)i=xpi+yi Among them, y is the transformed impedance matrix; The trapezoidal integration method is used for differential modeling, and then the zero element in the motor voltage vector u is ignored, and the excitation winding voltage is set to remain constant, then the voltage equation is transformed into: Ci(t)=D1u dq (t)+D1u dq (t-h)+D2u F +Ei(t-h) Among them, C, D1, D2, and E are coefficient matrices generated during the equation transformation process, and u dq is the vector formed by the stator d and q axis voltages, u F is the rotor d-axis excitation voltage, t is the simulation time; D2=[00h 000] T ; Further sorting can be obtained: i(t)=A G u dq (t)+B G [i(t-h)u dq (t-h)u F ] T Among them, A G , B G are the coefficient matrices generated during the equation transformation process, A G =C -1 D1, B G =[C -1 E C -1 D1 C -1 D2]; Assume that the rotor motion equation of the synchronous motor is: Among them, T j , D are the inertia time constant and damping coefficient of the synchronous motor, ω, δ, T m , T e are the synchronous motor speed, rotor angle, mechanical torque and electromagnetic torque respectively, ω0 is the synchronous speed; the trapezoidal method is used for differential processing to obtain the differential form shown below: Where ω(t) is the speed of the synchronous motor in the current simulation step, and δ(t) is the electrical angle between the synchronous motor rotor q axis and the system reference axis x rotating at the synchronous speed in the current simulation step; The electrical part equations and rotor motion equations of the synchronous motor model are converted to the abc coordinate system by using coordinate transformation, and the transformation matrix from the abc coordinate system to the dq coordinate system is defined as: Among them, T1 is the Park transformation matrix, T2 is the Park inverse transformation matrix, θ is the angle that the synchronous motor rotor d axis leads the stator a axis, and θ=θ0+ωt, θ0 is the initial angle, and u dq =T1u abc ,i dq =T1i abc ,u abc =T2u dq ,i abc =T2i dq ; Transform the dq coordinate system locally to the abc coordinate system, and we get: i'(t)=T ex A G T1u abc (t)+T ex B G T ex2 [i'(t-h)u abc (t-h)u F ] T Where i'(t) is the synchronous motor current vector in the abc coordinate system, T ex , T ex2 To maintain the same matrix dimension, the augmented coefficient matrix; i'(t) = [i a (t) i b (t) i c (t) i F (t) i D (t) i H (t) i Q (t)] T , i a (t), i b (t), i c (t) is the three-phase current of the synchronous motor stator, i F (t) is the current of the synchronous motor rotor F winding, i D (t) is the current of synchronous motor rotor D winding, i H (t) is the current of the synchronous motor rotor H winding, i Q (t) is the current of the synchronous motor rotor Q winding, I 4x4 is the 4th-order identity matrix, and I is the 1st-order identity matrix; The electrical part of the synchronous motor is estimated based on θ(th) at the last simulation time, and the following is obtained: in, is the estimated value of the angle that the synchronous motor rotor d-axis leads the stator a-axis, θ(th) is the angle that the synchronous motor rotor d-axis leads the stator a-axis in the previous simulation step, and ω(th) is the synchronous motor speed in the previous simulation step; T ex2 Directly use the previous moment value θ(th); T1 and T ex Use the estimated value at the current simulation time Separate the electrical equations in the abc coordinate system and solve them simultaneously: i abc (t)=A Gabc u abc (t)+B Gabc [i'(t-h) u abc (t-h) u F ] T @A Gabc u abc (t)+i hGabc i rest (t)=A Grest u abc (t)+B Grest [i'(t-h) u abc (t-h) u F ] T @A Grest u abc (t)+i hGrest Among them, A' G =T ex A G T1, A Gabc =A' G (1:3,:), A Grest =A' G (4:7,:), B' G =T ex B G T ex2 , B Gabc =B' G (1:3,:), B Grest =B' G (4:7,:); A Gabc is the equivalent network admittance matrix of the motor part viewed from the stator port, i hGabc is the historical current term of the synchronous motor, A Grest is the equivalent network admittance matrix of the remaining motor part, i hGrest is the remaining historical current term of the synchronous motor; According to the above calculation process, the solution process of the synchronous motor part can be obtained, including: Based on the voltage and current vector of the synchronous motor in the previous simulation step, T ex , A G And the corresponding amount of T1 to obtain A Gabc and i hGabc , and modify the network admittance matrix and network node injection current accordingly in combination with the stator port number; Solve the electrical part equations of each motor branch of the motor part to obtain the current synchronous motor port voltage u abc (t); Calculate and update the current step synchronous motor stator three-phase current i abc (t) and other winding current vectors i rest (t), then update the rotor dynamics according to the synchronous motor's rotor motion equation, and store the solution of the synchronous motor's stator three-phase current in the global memory; According to the electrical part equations and rotor motion equations of the synchronous motor model, the data required for the synchronous motor model are initialized, thereby obtaining the voltage and current of the synchronous motor model and the coefficient matrix required for the rotor dynamic solution of the motor and the initial data required for the main loop simulation.

8. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The method of initializing the data required by the power electronic equipment model according to the distributed power system parameter matrix and the main circuit solution equation and the control system solution equation, obtaining the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model and the initial data required for the main cycle simulation, and pre-storing the coefficient matrix, includes: The grid-connected converter is used as the structure of the distributed power source accessing the distribution network system; wherein: The solution equation of the AC side of the converter is consistent with the transmission network model; the solution equation of the DC side of the converter and the interface update equation of the AC and DC sides of the converter include: The solution equation of the converter DC side after differentiation is as follows: in, and They are the DC side capacitor C in the current simulation step. s1 and C s2 The voltage on and They are the DC side capacitor C in the next simulation step. s1 and C s2 The voltage on dc is the DC voltage value, R s is the internal resistance of the DC source, and They are the DC side capacitor C in the current simulation step. s1 and C s2 The value of the injected controlled current source at the node; The interface update equations of the converter AC and DC sides are as follows: Among them, S1, S2, S3, S4, S5, and S6 are switch control signals. The value is 1 when the switch is on, and the value is 0 when the switch is off. are the three-phase values ​​of the controlled voltage source on the AC side in the next simulation step; in, They are the three-phase values ​​of the current on the AC side filter inductor in the next simulation step, They are the DC side capacitor C in the next simulation step. s1 and C s2 The value of the injected controlled current source at the node; The control system of the converter adopts a double closed-loop control structure, the outer loop is a power control loop, and the inner loop is a current control loop; the grid-side voltage and current are measured and the output power P is calculated. g , Q g , and then with the given value of output power P gref , Q gref By comparison, the given value i of the grid-side current dq axis is generated after PI regulation * gd and i * gq , and then compared with the actual measured i gd and i gq After comparison, it is sent to the PI regulator, and the voltage compensation term is taken into account to obtain the reference voltage on the dq axis, which is used to generate a PWM signal to control the converter; According to the solution equations of the AC side and the DC side of the converter and the interface update equations of the AC and DC sides, the data required for the solution equations of the AC side and the DC side of the converter and the interface update equations of the AC and DC sides are initialized, thereby obtaining the coefficient matrix required for solving the power electronic equipment model and the initial data required for the main loop simulation.

9. The fine-grained parallel electromagnetic transient simulation method for distribution network according to claim 1, characterized in that: The fine-grained parallel solving between and within the models includes: Solving the branch current of the branch where the interface capacitor is located in the synchronous motor model according to the electrical part equation of the synchronous motor model, obtaining the branch current vector of the current simulation step of the synchronous motor model and storing it; Solving the branch current of the branch where the interface capacitor is located in the power electronic equipment model according to the solution equations of the AC side and the DC side of the power electronic equipment model, obtaining the branch current vector of the current simulation step of the power electronic equipment model and storing it; Solving the branch current of the branch where the interface capacitor is located in the transmission network model according to the solution equation of the transmission network model, obtaining the branch current vector of the current simulation step of the transmission network model and storing it; The three branch current vectors are called, and the node voltage update equation of the network interface model is used to solve the voltage vector of the interface capacitor to obtain the interface node voltage of the current simulation step and store it; At this point, the calculations of each model within a simulation cycle are completed after decoupling. In the next simulation step, the transmission network model, synchronous motor model, and power electronic equipment model call the stored interface capacitor voltage vector to perform their respective simulation calculations, and then send their respective branch current vectors back to the network interface model part, so that the network interface model can update the interface voltage vector in the next simulation step.

10. A fine-grained parallel electromagnetic transient simulation system for a distribution network, characterized in that: include: A modeling module for a complex distribution network with distributed power access, which constructs a simulation solution model for a complex distribution network with distributed power access, wherein the simulation solution model for a complex distribution network with distributed power access includes: a transmission network model, a synchronous motor model, a power electronic device model, and a network interface model, wherein the transmission network model, the synchronous motor model, and the power electronic device model can realize independent fine-grained parallel solution only by relying on the voltage of the interface capacitor of the network interface model, and the network interface model is solved according to the current update results of the transmission network model, the synchronous motor model, and the power electronic device model; A parameter matrix construction module, which obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology for the complex distribution network simulation solution model with distributed power access, and then forms the corresponding interface capacitance parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix; according to the transmission network parameter matrix, solves the system flow of the complex distribution network simulation solution model with distributed power access, and obtains the flow data of each node and each branch in the complex distribution network system; according to the flow data, solves the constant impedance model parameters of the load nodes in the complex distribution network simulation solution model with distributed power access, and adds the constant impedance model parameters of the load nodes to the transmission network parameter matrix to form a new transmission network parameter matrix; at the same time, initializes the voltage and current vector of the synchronous motor model according to the flow data; A coefficient matrix and simulation initial data acquisition module, which performs initialization operations on the data required for the transmission network model, synchronous motor model, power electronic equipment model and network interface model according to the new transmission network parameter matrix, synchronous motor parameter matrix, distributed power supply parameter matrix and their solution equations, obtains the coefficient matrix required for solving the node voltage and branch current of the transmission network model, the coefficient matrix required for solving the voltage and current of the motor of the synchronous motor model and the rotor dynamics of the motor, the coefficient matrix required for solving the main circuit and control circuit of the power electronic equipment model, and the initial data required for the main cycle simulation of each model, and pre-stores them; A fine-grained parallel simulation module, which uses the initial data to perform the next step of voltage and current update on the transmission network model, synchronous motor model, power electronic equipment model and network interface model respectively; the parallelism of the transmission network model is increased to a single node and a single branch, and after being combined with the GPU massive computing unit, the parallel solution of each node and each branch in the transmission network model is realized; the parallelism of the synchronous motor model is increased to a single motor, wherein the solution of each motor is performed by direct LU decomposition method, thereby realizing parallel solution between synchronous motors and within synchronous motors; the solution of the main circuit of the power electronic equipment model is consistent with the transmission network model, and each node and each branch of the main circuit is executed in parallel, and the parallelism of the control circuit of the power electronic equipment model is a single device, wherein the control system of each converter is allocated a thread for parallel solution; the parallelism of the network interface model is a single node, and the parallel solution of each node is realized after the GPU is allocated a corresponding number of threads; thereby, fine-grained parallel solution between and within each model is realized; The simulation solution model for the complex distribution network with distributed power access obtains the transmission network parameters, synchronous motor parameters and distributed power system parameters of the required simulation topology, and then forms the corresponding interface capacitor parameter matrix, transmission network parameter matrix, synchronous motor parameter matrix and distributed power system parameter matrix, including: Obtaining interface capacitance parameters of the complex distribution network simulation solution model with distributed power access, including interface node numbers and capacitance values ​​corresponding to the interface nodes, and adding the interface node numbers and capacitance values ​​corresponding to the interface nodes to corresponding positions of the interface capacitance parameter matrix to obtain the interface capacitance parameter matrix of the complex distribution network simulation solution model with distributed power access; The transmission network parameters of the complex distribution network simulation solution model with distributed power generation access are obtained, including node parameters and branch parameters, wherein the node parameters include node number, voltage amplitude, voltage phase angle, motor active power, motor reactive power, load active power, load reactive power, conductance G, susceptance B and node type number. The node parameters required for the power flow solution of the complex distribution network simulation solution model with distributed power generation access are obtained according to the three different node types of balance node, PV node and PQ node, and supplemented to the corresponding positions according to the basic format of the node parameter matrix to obtain the complex distribution network simulation solution model with distributed power generation access. The node parameter matrix of the distribution network simulation solution model; the branch parameters include the branch start node number, the branch end node number, the branch resistance value, the branch reactance value, the branch ground susceptance value, the transformer ratio and the phase angle, and according to the branch-related parameters of the complex distribution network simulation solution model with distributed power access, the basic format of the branch parameter matrix is ​​supplemented to the corresponding position to obtain the branch parameter matrix of the complex distribution network simulation solution model with distributed power access; the node parameter matrix and the branch parameter matrix together constitute the transmission network parameter matrix of the complex distribution network simulation solution model with distributed power access; Acquire the synchronous motor parameters of the complex distribution network simulation solution model with distributed power access, including the synchronous motor stator winding resistance, rotor F winding resistance, rotor D winding resistance, rotor H winding resistance, rotor Q winding resistance, stator d-axis reactance, stator d-axis armature reaction inductance, stator q-axis reactance, stator q-axis armature reaction inductance, rotor F winding reactance, rotor F winding and D winding mutual inductance reactance, rotor D winding reactance, rotor H winding reactance, rotor Q winding reactance and synchronous speed of the synchronous motor, according to the synchronous motor parameters of the complex distribution network simulation solution model with distributed power access, according to the basic format of the synchronous motor parameter matrix, supplement to the corresponding position, and obtain the synchronous motor parameter matrix of the complex distribution network simulation solution model with distributed power access; Obtain the distributed power system parameters of the complex distribution network simulation solution model with distributed power access, including the main circuit parameters and control system parameters of the distributed power system, wherein the main circuit parameters include a DC source, a DC side capacitor, a filter inductor and a filter capacitor; the control system parameters include phase-locked loop PI parameters and PI parameters of traditional PQ control; and according to the distributed power system related parameters of the complex distribution network simulation solution model with distributed power access, supplement them to the corresponding positions in accordance with the basic format of the distributed power system parameter matrix to obtain the distributed power system parameter matrix of the complex distribution network simulation solution model with distributed power access.

11. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it can be used to perform the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it can be used to execute the method according to any one of claims 1 to 9.

Images