CN116706886A - Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system - Google Patents
Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system Download PDFInfo
- Publication number
- CN116706886A CN116706886A CN202310625703.2A CN202310625703A CN116706886A CN 116706886 A CN116706886 A CN 116706886A CN 202310625703 A CN202310625703 A CN 202310625703A CN 116706886 A CN116706886 A CN 116706886A
- Authority
- CN
- China
- Prior art keywords
- model
- splitting
- simulation
- vertex
- power system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 94
- 238000000034 method Methods 0.000 title claims abstract description 67
- 230000001052 transient effect Effects 0.000 title claims abstract description 30
- 238000013507 mapping Methods 0.000 claims abstract description 10
- 230000003595 spectral effect Effects 0.000 claims abstract description 6
- 230000005540 biological transmission Effects 0.000 claims description 27
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000007670 refining Methods 0.000 claims description 6
- 230000006854 communication Effects 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 230000001360 synchronised effect Effects 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 235000008733 Citrus aurantifolia Nutrition 0.000 claims 1
- 235000011941 Tilia x europaea Nutrition 0.000 claims 1
- 239000004571 lime Substances 0.000 claims 1
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 101100499229 Mus musculus Dhrsx gene Proteins 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
An electromagnetic transient parallel simulation self-adaptive model splitting method of an electric power system comprises the following steps: step 1, a parallel simulation interface, namely introducing a Berhelson model as an interface, and decoupling the split sub-model so as to realize parallel simulation; step 2, a model splitting process and a parallel simulation process; step 3, calculating amount and principle of model splitting; step 4, mapping between the original data and the graph structure; step 5, simplifying pretreatment of the complex graph structure; and 6, splitting the spectral clustering self-adaptive model. And provides an electromagnetic transient parallel simulation self-adaptive model splitting device for the power system. The invention has large treatment scale and high resolution speed; the splitting is flexible; the split model is uniform in height; the expandability is good; the adaptability is strong.
Description
Technical Field
The invention belongs to the field of power system detection, and relates to a method and a device for splitting an electromagnetic transient parallel simulation self-adaptive model of a power system.
Background
Electromagnetic transient simulation of a large-scale power system generally breaks the whole power system model into a plurality of sub-models to perform parallel simulation so as to accelerate the simulation speed. The existing power system generally uses a splitting method based on geographic information to split the model, and the method is easy to execute, but does not have self-adaptive capacity, so that the number of sub-models is severely limited, the scale is uneven, the associated lines are more, the calculated amount of the split model is uneven in simulation, and the difficulty and the workload of splitting can be increased due to the change of the power system.
As the structure of the power system becomes more and more complex, the related power electronic devices become more and more, and the time of electromagnetic transient simulation of the large-scale power system increases dramatically. Due to the high sparsity of the power system, the parallel simulation method can effectively reduce simulation time, and a power system model splitting algorithm for parallel simulation and a parallel segmentation strategy based on multi-layer and multi-region characteristics of the power system are paid attention to. However, the most commonly used method still carries out model splitting based on geographic information, and is easy to execute, but the result is uneven, and the method cannot adapt to the change of the topological structure of the power system.
For example Cao, wang Lijiang, zhao Yongfei, zhang Xiuqi. Large-scale New energy grid-connected electromagnetic transient parallel simulation based on PSCAD/EMTDC [ J ]. Chinese electric power, 2020,53 (11): 154-161. The method for evaluating the node weight of the bus is not suitable for power electronic equipment elements, and has the problems that the single node number is used for partitioning, the consideration of the interface communication efficiency is insufficient, the simulation calculation amount of the diversified model is not accurately represented, and the like. The method can be applied to electromagnetic transient simulation parallel calculation of the power system, and has the problems of limited efficiency improvement of the large-scale power system, large adaptability difference to different systems and the like.
For example, chinese patent application No. 2009101358605 discloses an electromagnetic transient off-line non-real-time parallel simulation system and a simulation method, in which the network of the power system is divided into a plurality of subnets, each of the subnets respectively forms an electromagnetic transient simulation system, and an interface control system for performing data exchange with each of the electromagnetic transient simulation systems is provided between each of the electromagnetic transient simulation systems, so as to realize off-line non-real-time parallel simulation of each of the electromagnetic transient simulation systems. The simulation method of the invention is that the electromagnetic transient simulation system carries out electromagnetic transient calculation on each subnet, and the calculation variable of the subnet is transmitted to other subnets connected with the interface system through the interface control system, thus realizing off-line non-real-time parallel simulation of the whole electromagnetic transient simulation system. The simulation speed of the method is generally slower than that of digital real-time simulation, and a large-scale power system model of any topological structure cannot be adaptively split.
Disclosure of Invention
In order to overcome the defects of the existing method for splitting the electromagnetic transient parallel simulation model of the power system, the invention provides a method and a device for splitting the electromagnetic transient parallel simulation self-adaptive model of the power system, and the speed of electromagnetic transient parallel simulation of the power system is improved.
The technical scheme adopted for solving the technical problems is as follows:
an electromagnetic transient parallel simulation self-adaptive model splitting method of an electric power system comprises the following steps:
step 1, a parallel simulation interface, namely introducing a Berhelson model as an interface, and decoupling the split sub-model, so as to realize parallel simulation, wherein the process is as follows:
for a uniform lossless transmission line, let its wave impedance be Z, length be l, wave speed be v, then the transmission delay beThe inductance and capacitance of the line unit length are constants L which are not affected by frequency 0 And C 0 ，u k 、u m Representing node k and node respectivelyVoltage of m, i k ，i m The current of the node k and the current of the node m are respectively represented, and t is the current simulation time;
then obtaining equivalent current sources I of node k and node m k 、I m Is a recursive formula of (a):
step 2, a model splitting process and a parallel simulation process, wherein the steps are as follows:
2.1 Model splitting process
Preprocessing the original data of the power system, splitting the graph to obtain a split model, and simulating;
2.2 Parallel simulation process
Starting the threads with the same number to perform parallel simulation according to the number of the submodels, wherein in the simulation process, the data are required to be synchronized and serially operated after each step;
step 3, calculation amount and principle of model splitting
The influence of the elements on the simulation time is obtained through testing the simulation time of the common elements in the power system, and the total calculated amount of the simulation is defined as L:
L node l is the calculated amount of the node gen L is the calculated amount of the power supply trans Calculated amount of transformer, L line The calculated amount of the transmission line;
the simulation time of the slowest simulation model is t slow Simulation time t of the fastest simulation model fast ，t wait Representing a time difference between the simulated slowest model and the fastest model;
t wait ＝t slow -t fast (3)
t com is the time consumption of the communication process, which represents all of the stepsThe time consumption of the model for communication after the simulation is completed, and the following two principles are required to be followed when the model is split:
a) The transmission lines among the models are as few as possible, so that t can be effectively reduced wait ；
b) The calculated amount of each model should be equal as much as possible, and t can be reduced com ；
Step 4, mapping between the original data and the graph structure
Due to the complexity of the power system, the original data of the power system is selectively mapped into a graph structure so as to simplify the splitting process;
the mapping relationship is as follows:
bus-vertex;
transmission line, transformer-side;
other elements are added to the bus, are represented by the weights of the vertexes, and the topology structure of the graph is not changed when the elements are transferred by using the weights of the vertexes;
W i is the vertex v i The weight of the transformer is equal to the sum of the calculated amounts of the vertex elements, and the coefficient is 1/2 because the calculated amounts of the transmission line and the transformer are shared by two ends;
step 5, simplifying pretreatment of a complex graph structure, wherein the process is as follows:
5.1 Transformer clustering
Centralizing all vertexes connected by the transformer into one vertex, namely a T vertex;
5.2 Transmission line clustering
For the berlong model, the delay τ of the transmission line should be greater than a simulation step t step ，
τ≥t step (5)
The vertexes connected by the short lines are concentrated into one vertex C, which is called a complex vertex;
5.3 Multi-level clustering
Before formal graph splitting, performing multi-level clustering on the graph g=g (V, E) to further simplify the structure, wherein V represents a vertex and E represents an edge;
step 6, splitting a spectral clustering self-adaptive model, wherein the process is as follows:
6.1 Selecting an initial cluster center
Selecting an initial cluster center by adopting a DPC algorithm, wherein the DPC algorithm selects the cluster center by the following criteria: a. vertices with higher local densities; b. the distance between different cluster centers is relatively far;
6.2 Initial clustering)
Placing all the directly connected vertexes of the center into corresponding subgraphs to expand the initial clustering center into a small network, and then clustering by using an LDG algorithm to obtain complete K subgraphs, wherein the functions are as follows:
V k is a collection of vertices, ω (V k ) Is V k Is calculated as c= |v 0 I/K, where I V 0 I is the sum of all vertex calculations, adj (v) i ) Is the vertex v i Is defined in the constraint function f (v i ) In, |V k ∩Adj(v i ) I represents vertex v i And set V k Correlation between 1-omega (V) k ) C is used to balance the calculated amount of all subgraphs;
6.3 Restoring the graph structure and refining the graph structure, and optimizing the splitting result by using multi-level refinement and restoration, wherein the steps are as follows:
a. for the original clustering result, mapping the subgraph to the upper layer to restore the graph structure;
b. refining the restored subgraph, and improving the resolution quality;
c. the two steps are repeated until the influence of multi-level clustering is completely counteracted.
The LDG function is optimized as follows:
n average is the average calculated amount of the vertexes in the whole graph, sigma n Is the standard deviation of the number of vertex calculations,is the vertex v i Is>Is->A set of directly connected adjacent vertices, w isWeights of adjacent vertexes are intermediately connected.
Further, the multi-level clustering of 5.3) greatly reduces the number of vertexes before splitting, is beneficial to accelerating the splitting process, enlarges the equivalent distance between vertexes and is beneficial to the selection of the original clustering center;
limitation of multi-level clustering: number of vertices Size Limit Not greater than the limit of equation (5);
N total for the total number of nodes of the power system, K is the number of sub-graphs, E represents the minimum number of vertices in each sub-graph, and each vertex is only allowed to be clustered once.
Still further, in said step 6.3), the formula is introducedTo enlarge ginsengThe test point set is helpful to avoid useless operations when processing huge power systems; in a power system of less than 10000 nodes, a set of reference vertices greater than a preset threshold will interfere with normal distribution, thus setting the weight w of indirect neighboring vertices to 0.
An electromagnetic transient parallel simulation self-adaptive model splitting device of an electric power system has the processing process of the splitting method.
The technical conception of the invention is as follows: by utilizing the high sparsity of the power system, a spectral clustering self-adaptive model splitting algorithm is provided, wherein the algorithm maps the original data of the power system to a graph structure, combines a simplified preprocessing technology of a complex graph structure with an LDG algorithm, can adaptively split a large-scale power system model with any topological structure, and obtains high-quality sub-models, and the number of the sub-models is not limited. The method overcomes the defects of the existing model splitting method, and further improves the speed of electromagnetic transient parallel simulation of the power system.
The beneficial effects of the invention are mainly shown in the following steps: and (1) the treatment scale is large, and the resolution speed is high. And (2) flexible splitting. (3) the split model is highly uniform. (4) scalability is good. And (5) the adaptability is strong.
Drawings
Fig. 1 is a schematic diagram of a single lossless transmission line and a berlong model thereof.
Fig. 2 is a parallel simulation flow diagram.
FIG. 3 is a diagram of a multi-level clustering process.
FIG. 4 is a flowchart of an adaptive model splitting method based on spectral clustering.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 to 4, a method for splitting an electromagnetic transient parallel simulation self-adaptive model of a power system comprises the following steps:
step 1, a parallel simulation interface, namely introducing a Berhelson model as an interface, and decoupling the split sub-model, so as to realize parallel simulation, wherein the process is as follows:
for a uniform lossless transmissionThe line is provided with the wave impedance Z, the length l and the wave speed v, and the transmission delay isThe inductance and capacitance of the line unit length are constants L which are not affected by frequency 0 And C 0 ，u k 、u m Representing the voltages at nodes k and m, i, respectively k ，i m The current of the node k and the current of the node m are respectively represented, and t is the current simulation time;
then obtaining equivalent current sources I of node k and node m k 、I m Is a recursive formula of (a):
the Berlong model converts the wave process of the distributed parameter transmission line into a lumped parameter circuit with only resistors and historical current sources, and delay of the wave process decouples two ends of the transmission line, so that possibility is provided for electromagnetic transient parallel simulation;
step 2, a model splitting process and a parallel simulation process, wherein the steps are as follows:
2.1 Model splitting process
Preprocessing the original data of the power system, splitting the graph to obtain a split model, and simulating;
2.2 Parallel simulation process
Starting the threads with the same number to perform parallel simulation according to the number of the submodels, wherein in the simulation process, the data are required to be synchronized and serially operated after each step;
step 3, calculation amount and principle of model splitting
The influence of the elements on the simulation time is obtained through testing the simulation time of the common elements in the power system, and the total calculated amount of the simulation is defined as L:
L node l is the calculated amount of the node gen L is the calculated amount of the power supply trans Calculated amount of transformer, L line The calculated amount of the transmission line;
the simulation time of the slowest simulation model is t slow Simulation time t of the fastest simulation model fast ，t wait Representing a time difference between the simulated slowest model and the fastest model;
t wait ＝t slow -t fast (11)
t com is the time consumption of the communication process, which represents the time consumption of all models in each step to communicate after the simulation is completed, and the following two principles are required to be followed when the models are split:
a) The transmission lines among the models are as few as possible, so that t can be effectively reduced wait ；
b) The calculated amount of each model should be equal as much as possible, and t can be reduced com ；
Step 4, mapping between the original data and the graph structure
Due to the complexity of the power system, the original data of the power system is selectively mapped into a graph structure so as to simplify the splitting process;
the mapping relationship is as follows:
bus-vertex;
transmission line, transformer-side;
other elements are added to the bus, are represented by the weights of the vertexes, and the topology structure of the graph is not changed when the elements are transferred by using the weights of the vertexes;
W i is the vertex v i The weight of the transformer is equal to the sum of the calculated amounts of the vertex elements, and the coefficient is 1/2 because the calculated amounts of the transmission line and the transformer are shared by two ends;
step 5, simplifying pretreatment of a complex graph structure, wherein the process is as follows:
5.1 Transformer clustering
Because the berzuron model decoupling submodel of the transmission line is used, the transformer cannot be a crossing element between submodels, and all vertexes connected by the transformer are concentrated into one vertex, namely a T vertex, so that the disassembly at the transformer is avoided, and the graph structure is simplified to a certain extent;
5.2 Transmission line clustering
For the berlong model, the delay τ of the transmission line should be greater than a simulation step t step ，
τ≥t step (13)
The vertexes connected by the short lines are concentrated into one vertex C, which is called a complex vertex;
5.3 Multi-level clustering
Before formal graph splitting, performing multi-level clustering on the graph g=g (V, E) to further simplify the structure, wherein V represents a vertex and E represents an edge;
the multi-level clustering of 5.3) greatly reduces the number of vertexes before splitting, is beneficial to accelerating the splitting process, enlarges the equivalent distance between vertexes and is beneficial to the selection of the original clustering center.
Limitation of multi-level clustering
Number of vertices Size Limit Not greater than the limit of equation (5);
N total for the total number of nodes of the power system, K is the number of sub-graphs, E represents the minimum number of vertices in each sub-graph, and each vertex is only allowed to be clustered once.
Step 6, splitting a spectral clustering self-adaptive model, wherein the process is as follows:
6.1 Selecting an initial cluster center
Selecting an initial cluster center by adopting a DPC algorithm, wherein the DPC algorithm selects the cluster center by the following criteria: a. vertices with higher local densities; b. the distance between different cluster centers is relatively far;
6.2 Initial clustering)
Placing all the directly connected vertexes of the center into corresponding subgraphs to expand the initial clustering center into a small network, and then clustering by using an LDG algorithm to obtain complete K subgraphs, wherein the functions are as follows:
V k is a collection of vertices, ω (V k ) Is V k Is calculated as c= |v 0 I/K, where I V 0 I is the sum of all vertex calculations, adj (v i ) Is the vertex v i Is defined in the constraint function f (v i ) In, |V k ∩Adj(v i ) I represents vertex v i And set V k Correlation between 1-omega (V) k ) the/C can balance the calculated amount of all sub-graphs;
6.3 Restoring the graph structure and refining the graph structure, and optimizing the splitting result by using multi-level refinement and restoration, wherein the steps are as follows:
a. for the original clustering result, mapping the subgraph to the upper layer to restore the graph structure;
b. refining the restored subgraph, and improving the resolution quality;
c. repeating the steps a and b until the influence of multi-level clustering is completely counteracted;
the LDG function is optimized as follows:
n average is the average calculated amount of the vertexes in the whole graph, sigma n Is the standard deviation of the number of vertex calculations,is the vertex v i Is>Is->A set of directly connected adjacent vertices, w isThe weight of the adjacent vertex is 0.00005; formula introduction +.>To expand the reference vertex set, helping to avoid useless operations when handling large power systems; in a low power system (e.g., a power system of less than 10000 nodes), a larger (greater than a preset threshold) set of reference vertices may interfere with normal distribution, thus setting the weight w of indirect neighboring vertices to 0.
An electromagnetic transient parallel simulation self-adaptive model splitting device of an electric power system has the processing process of the splitting method.
In this embodiment, the LDG (Linear Deterministic Greedy) algorithm places adjacent vertices together at the time of splitting to reduce splitting of the line. A greedy algorithm is adopted to place one vertex in the sub-graph with the most adjacent vertices, and meanwhile vertex load balance of each sub-graph is guaranteed.
The DPC (Density Peaks Clustering) algorithm further divides the data points of each graph into two parts, core and edge points, and detects noise points. Wherein, the core point is the core part of the graph, and the density value is larger; the edge points are located in the border area of the graph and have smaller density values, and the distinction between the two is defined by means of the average local density of the border area.
The embodiments described in this specification are merely illustrative of the manner in which the inventive concepts may be implemented. The scope of the present invention should not be construed as being limited to the specific forms set forth in the embodiments, but the scope of the present invention and the equivalents thereof as would occur to one skilled in the art based on the inventive concept.
Claims (4)
1. The utility model provides a power system electromagnetism transient state parallel simulation self-adaptation model split method which is characterized in that the method includes the following steps:
step 1, a parallel simulation interface, namely introducing a Berhelson model as an interface, and decoupling the split sub-model, so as to realize parallel simulation, wherein the process is as follows:
for a uniform lossless transmission line, let its wave impedance be Z, length be l, wave speed be v, then the transmission delay beThe inductance and capacitance of the line unit length are constants L which are not affected by frequency 0 And C 0 ，u k 、u m Representing the voltages at nodes k and m, i, respectively k ，i m The current of the node k and the current of the node m are respectively represented, and t is the current simulation time;
then obtaining equivalent current sources I of node k and node m k 、I m Is a recursive formula of (a):
step 2, a model splitting process and a parallel simulation process, wherein the steps are as follows:
2.1 Model splitting process
Preprocessing the original data of the power system, splitting the graph to obtain a split model, and simulating;
2.2 Parallel simulation process
Starting the threads with the same number to perform parallel simulation according to the number of the submodels, wherein in the simulation process, the data are required to be synchronized and serially operated after each step;
step 3, calculation amount and principle of model splitting
The influence of the elements on the simulation time is obtained through testing the simulation time of the common elements in the power system, and the total calculated amount of the simulation is defined as L:
L node l is the calculated amount of the node gen L is the calculated amount of the power supply trans Calculated amount of transformer, L lime The calculated amount of the transmission line;
the simulation time of the slowest simulation model is t slow Simulation time t of the fastest simulation model fast ，t wait Representing a time difference between the simulated slowest model and the fastest model;
t wait ＝t slow -t fast (3)
t com is the time consumption of the communication process, which represents the time consumption of all models in each step to communicate after the simulation is completed, and the following two principles are required to be followed when the models are split:
a) The transmission lines among the models are as few as possible, so that t can be effectively reduced wait ；
b) The calculated amount of each model should be equal as much as possible, and t can be reduced com ；
Step 4, mapping between the original data and the graph structure
Due to the complexity of the power system, the original data of the power system is selectively mapped into a graph structure so as to simplify the splitting process;
the mapping relationship is as follows:
bus-vertex;
transmission line, transformer-side;
other elements are added to the bus, are represented by the weights of the vertexes, and the topology structure of the graph is not changed when the elements are transferred by using the weights of the vertexes;
W i is the vertex v i Weights of (1), itThe sum of the calculated amounts of the vertex elements is equal, and the coefficient is 1/2 because the calculated amounts of the transmission line and the transformer are shared by two ends;
step 5, simplifying pretreatment of a complex graph structure, wherein the process is as follows:
5.1 Transformer clustering
Centralizing all vertexes connected by the transformer into one vertex, namely a T vertex;
5.2 Transmission line clustering
For the berlong model, the delay τ of the transmission line should be greater than a simulation step t step ，
τ≥t step (5)
The vertexes connected by the short lines are concentrated into one vertex C, which is called a complex vertex;
5.3 Multi-level clustering
Before formal graph splitting, performing multi-level clustering on the graph g=g (V, E) to further simplify the structure, wherein V represents a vertex and E represents an edge;
step 6, splitting a spectral clustering self-adaptive model, wherein the process is as follows:
6.1 Selecting an initial cluster center
Selecting an initial cluster center by adopting a DPC algorithm, wherein the DPC algorithm selects the cluster center by the following criteria: a. vertices with higher local densities; b. the distance between different cluster centers is relatively far;
6.2 Initial clustering)
Placing all the directly connected vertexes of the center into corresponding subgraphs to expand the initial clustering center into a small network, and then clustering by using an LDG algorithm to obtain complete K subgraphs, wherein the functions are as follows:
V k is a collection of vertices, ω (V k ) Is V k Is calculated as c= |v 0 I/K, where I V 0 I is the sum of all vertex calculations, adj (v) i ) Is the vertex v i Is defined by a set of adjacent vertices of the model,in the constraint function f (v i ) In, |V k ∩Adj(v i ) I represents vertex v i And set V k Correlation between 1-omega (V) k ) C is used to balance the calculated amount of all subgraphs;
6.3 Restoring the graph structure and refining the graph structure, and optimizing the splitting result by using multi-level refinement and restoration, wherein the steps are as follows:
a. for the original clustering result, mapping the subgraph to the upper layer to restore the graph structure;
b. refining the restored subgraph, and improving the resolution quality;
c. repeating the two steps until the influence of multi-level clustering is completely counteracted;
the LDG function is optimized as follows:
n average is the average calculated amount of the vertexes in the whole graph, sigma n Is the standard deviation of the number of vertex calculations,is the vertex v i Is>Is->A set of directly connected adjacent vertices, w isWeights of adjacent vertexes are intermediately connected.
2. The method for splitting the electromagnetic transient parallel simulation self-adaptive model of the power system according to claim 1, wherein the multi-level clustering of 5.3) is characterized in that the number of vertexes is greatly reduced before splitting, so that the splitting process is accelerated, the equivalent distance between vertexes is enlarged, and the selection of an original clustering center is facilitated;
limitation of multi-level clustering: number of vertices Size Limit Not greater than the limit of equation (5);
N total for the total number of nodes of the power system, K is the number of sub-graphs, E represents the minimum number of vertices in each sub-graph, and each vertex is only allowed to be clustered once.
3. The method for splitting the electromagnetic transient parallel simulation adaptive model of the power system according to claim 1 or 2, wherein in the step 6.3), a formula is introducedTo expand the reference vertex set, helping to avoid useless operations when handling large power systems; in a power system of less than 10000 nodes, a set of reference vertices greater than a preset threshold will interfere with normal distribution, thus setting the weight w of indirect neighboring vertices to 0.
4. A splitting device realized by the splitting method of the electromagnetic transient parallel simulation self-adaptive model of the power system according to claim 1, wherein the processing procedure of the device is the splitting method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310625703.2A CN116706886A (en) | 2023-05-30 | 2023-05-30 | Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310625703.2A CN116706886A (en) | 2023-05-30 | 2023-05-30 | Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116706886A true CN116706886A (en) | 2023-09-05 |
Family
ID=87844410
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310625703.2A Pending CN116706886A (en) | 2023-05-30 | 2023-05-30 | Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116706886A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103729502A (en) * | 2013-12-19 | 2014-04-16 | 南京南瑞继保电气有限公司 | Method for increasing electromagnetic transient simulation speed of power system |
| CN115186891A (en) * | 2022-07-08 | 2022-10-14 | 广西电网有限责任公司电力科学研究院 | Transient simulation analysis decision method, device and equipment for power grid harmonic oscillation |
| CN115203940A (en) * | 2022-07-15 | 2022-10-18 | 国家电网公司华中分部 | Self-adaptive grid division method for transient simulation of large power grid |
-
2023
- 2023-05-30 CN CN202310625703.2A patent/CN116706886A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103729502A (en) * | 2013-12-19 | 2014-04-16 | 南京南瑞继保电气有限公司 | Method for increasing electromagnetic transient simulation speed of power system |
| CN115186891A (en) * | 2022-07-08 | 2022-10-14 | 广西电网有限责任公司电力科学研究院 | Transient simulation analysis decision method, device and equipment for power grid harmonic oscillation |
| CN115203940A (en) * | 2022-07-15 | 2022-10-18 | 国家电网公司华中分部 | Self-adaptive grid division method for transient simulation of large power grid |
Non-Patent Citations (1)
| Title |
|---|
| LUWEI SHI: "Research_on_Self-adaptive_Network_Partitioning_Algorithm_for_Electromagnetic_Transient_Parallel_Simulation_of_Power_System", 《2022 IEEE 5TH INTERNATIONAL ELECTRICAL AND ENERGY CONFERENCE (CIEEC)》, pages 1238 - 1243 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11669085B2 (en) | Method and system for determining system settings for an industrial system | |
| CN110366193B (en) | Slice deployment method and device for network arrangement layer bearing of smart grid | |
| CN114896937A (en) | Integrated circuit layout optimization method based on reinforcement learning | |
| CN110336270B (en) | Updating method of transient stability prediction model of power system | |
| CN105186578A (en) | Distributed automatic dispatching method for power system with accurate network loss calculation capability | |
| WO2024037393A1 (en) | Topology recognition model compression method and apparatus, electronic device and medium | |
| CN110264012A (en) | Renewable energy power combination prediction technique and system based on empirical mode decomposition | |
| CN110290542A (en) | Marine UAV Communication coverage optimization method and system | |
| CN110610019A (en) | Dynamics analysis method of Markov jump system with partially unknown transition probability | |
| CN111723516A (en) | Multi-target seawater intrusion management model based on adaptive DNN (deep dynamic network) substitution model | |
| CN109391511A (en) | It is a kind of based on can outward bound network intelligence communication resource allocation policy | |
| Xu et al. | Online topology‐based voltage regulation: A computational performance enhanced algorithm based on deep reinforcement learning | |
| CN116706886A (en) | Method and device for splitting electromagnetic transient parallel simulation self-adaptive model of power system | |
| CN115953651A (en) | Model training method, device, equipment and medium based on cross-domain equipment | |
| CN105893667A (en) | Heterogeneous simulation data unified integrated visual method | |
| CN115203940A (en) | Self-adaptive grid division method for transient simulation of large power grid | |
| Zhang et al. | Distributed parallel computing technique for EM modeling | |
| CN111146815A (en) | Intelligent power distribution network distributed power generation planning configuration method | |
| CN109783845B (en) | Electromechanical-electromagnetic hybrid simulation network separation method and device for boundary point grouping decoupling | |
| CN112018757A (en) | General example generation method for morphological evolution of power distribution network | |
| CN117494816B (en) | Model reasoning method, device, equipment and medium based on computing unit deployment | |
| CN116187723B (en) | Resource scheduling method and device applied to distribution line loss reduction scene | |
| CN113055111B (en) | Channel modeling method and system based on Bayesian optimization | |
| CN115563543B (en) | Channel quality multi-terminal electric quantity protection principle selection method, device and system | |
| CN109862574B (en) | Dynamic collaborative topology planning method and system for wireless backhaul problem |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |