A method for parallel optimization of stochastic electromagnetic transients in DC power distribution systems

Abstract

The embodiment of the present invention discloses a method for parallel optimization of random electromagnetic transients in a DC power distribution system, and relates to the technical field of DC power distribution. The present invention selectively implements voltage correction and accelerates through optimal node sorting and local contraction of subsystem networks, so as to realize efficient parallel simulation of random electromagnetic transients in a DC power distribution system under limited resource conditions, and is suitable for a parallel processing scheme in which associated branches are set between subsystems and delayed transmission through boundary voltages. In order to improve the simulation accuracy of random electromagnetic transients under the parallel simulation scheme, a voltage selective correction step is designed, and an acceleration technology including optimal node sorting and local contraction of subsystem networks is developed to reduce the computational burden of the correction step, which can fully guarantee the efficiency of parallel analysis of random electromagnetic transients in complex DC power distribution systems.

Description

Technical Field

[0001] The present invention relates to the technical field of direct current power distribution, and in particular to a method for random electromagnetic transient parallel optimization of a direct current power distribution system. Background Art

[0002] DC distribution systems are widely used in modern power systems, especially in urban power grids and electrified transportation. However, with the interaction between converters and passive components in DC distribution systems, instability may occur in different frequency ranges. In order to explore the impact of configuration control and protection strategies on system operation, electromagnetic transient simulation analysis is usually required. Existing electromagnetic transient analysis tools have an insurmountable contradiction between the expansion of system scale and the reduction of analysis step size. Therefore, in order to ensure the analysis efficiency and accuracy of DC distribution systems, it is necessary to configure corresponding methods and solutions.

[0003] Considering the uncertainty and diversity of operation scenarios, multiple analysis under multi-parameter settings will bring significant computational overhead. Therefore, random electromagnetic transient analysis based on stochastic differential equation theory has become an important means to reveal system dynamics and test control strategies. However, due to the combined influence of factors such as random process simulation, numerical algorithm convergence and switch device trigger correction mechanism, the numerical stability of random electromagnetic transient analysis is generally weaker than that of deterministic analysis. In random electromagnetic transient analysis, the need to process a large number of nodes and complex computing tasks has led to the application of parallel processing technology to improve computing speed and efficiency. However, new problems have emerged. Parallel processing methods may cause error accumulation due to the combined effect of factors such as time step, associated branch impedance and numerical algorithm convergence, especially when the switch device trigger correction mechanism is not accurate enough. In addition, as the scale of the system expands, the allocation and management of computing resources have become more complicated, further exacerbating the problem of computing efficiency.

[0004] Therefore, in view of the problems of large random calculation amount and poor numerical stability in the parallel processing scheme of setting associated branches between subsystems and transmitting through boundary voltage delay, how to optimize and improve the efficiency of parallel analysis of random electromagnetic transients in DC distribution systems has become a topic that needs to be studied. Summary of the invention

[0005] The embodiments of the present invention provide a method for parallel optimization of random electromagnetic transients in a DC power distribution system, which can improve the efficiency of parallel analysis of random electromagnetic transients in a DC power distribution system.

[0006] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0007] A method for parallel optimization of random electromagnetic transients in DC power distribution systems, such as Figure 3 As shown, including:

[0008] S1, decomposing the DC power distribution system to obtain various subsystems, and then reconfiguring associated branches;

[0009] S2, initialize the nodes in the subsystem and then sort the nodes;

[0010] S3. Filter the sorted nodes, where the screening results are related to the degree of influence of the nodes on the voltage correction. After that, the correction steps can be set to set a unified voltage result accuracy standard. e v ;

[0011] S4, obtaining the node voltage at the current moment;

[0012] S5. According to the set voltage result accuracy standard, detect whether the node voltage at the current moment meets the correction condition. If it does, perform voltage correction; otherwise, repeat S4 to S5; wherein, according to the solution result of S4 and the set voltage result accuracy standard, e v , determine whether the current voltage result meets the calibration conditions, if yes, go to step 6; otherwise, t = t +△ t , repeat step 4, △ t is the set simulation time step;

[0013] S6. Synchronize the data of each subsystem.

[0014] In this embodiment, S1 includes: dividing the DC power distribution system into at least two subsystems according to a preset subsystem boundary, and recording the electrical parameters and connection relationships of the components in each subsystem; wherein the subsystem boundary includes: physical structure and functional partitioning. In practical applications, the entire system can be divided into several subsystems according to different functional areas (such as converter area, load area, etc.). Establish each subsystem model according to the electrical parameters and connection relationships of the components in the subsystem. The re-association branch includes: replacing the node voltage of a subsystem at the previous moment with the node voltage of another subsystem associated with the subsystem at the current moment.

[0015] In this embodiment, S2 includes: classifying the nodes in the system according to the type of connected components, the classification results include: dynamic nodes, static nodes and switch nodes; numbering the classified nodes, and then generating storage vectors for each type of node, wherein the numbering priority is from high to low: static nodes, switch nodes and dynamic nodes, and the length of the storage vector is equal to the number of nodes of the same type; generating a correlation matrix according to the obtained storage vectors and the maximum sorting length obtained by sorting the nodes.

[0016] In this embodiment, S3 includes: determining the local shrinkage range of the node, and then removing the nodes with little influence on the voltage correction from the sorted nodes, wherein the influence of the node on the voltage correction is quantified by standardized sensitivity. The standardized sensitivity includes: ,in, S Ak It corresponds to k The sensitivity matrix of the working state is expressed as S A , S A Each element in S A ( i , j ) indicates that when j The voltage deviation of the associated branch has an impact on the i The influence of voltage correction on each node is represented by j, i represents the number of the associated branch, and j represents the number of the node. j L is the total number of associated branches accessing the subsystem, k S is the number of possible working states; the elimination of nodes with little influence on voltage correction includes: according to s i , nodes with lower normalized sensitivity are removed from S A Removed.

[0017] In this embodiment, the single-step simulation stage is processed in parallel, and the node voltage at the current time t is obtained by linear static iteration. S4 includes: establishing a node voltage model for subsystem A and subsystem B ,in, G AB,n and G BA,n is the node admittance submatrix corresponding to the associated branch between subsystems A and B, G A,n and G B,n are the node admittance matrices corresponding to the reconstruction of the companion circuits of subsystem A and subsystem B, I A,n and I B,n are the injected currents of subsystem A and subsystem B respectively, V A,n and V B,n are the node voltages of subsystem A and subsystem B respectively. The subscript n indicates the nth step of the simulation, i.e., t n=nDt,t n represents the time, Dt represents the simulation step; the node voltage model is processed by linear static iteration until convergence.

[0018] Specifically, the node voltage model is processed by linear static iteration until convergence, including: establishing a node admittance matrix according to the node voltage model: , where Mn, Vn, Nn, and In represent the block diagonal matrix formed by the node admittance matrix corresponding to the reconstruction of the companion circuits of subsystems A and subsystem B, the node voltage vector formed by the node voltages of subsystems A and subsystem B, the zero-complement matrix formed by the node admittance submatrices corresponding to the associated branches between subsystems A and B, and the injected current vector formed by the injected currents of subsystems A and subsystem B, respectively;

[0019] The node admittance matrix is ​​processed by a linear static iterative model until convergence, the linear static iterative model comprising: , where the superscript ( k ) is the k The corresponding result of the iteration; the superscript -1 indicates the matrix inversion operation.

[0020] In this embodiment, in S5, the correction condition includes: condition 1 and condition 2, wherein condition 1 is used to detect whether at the current time t n Execute the correction, the condition 2 is used to detect whether at the next moment t n+1 Enable the correction step; the condition 1 includes: , e v To unify the accuracy standard value of the set voltage result, express The part corresponding to the boundary node, , is the voltage of the boundary node in the first solution of the subsystem at the current moment, V O,n-1 represents the boundary node voltage obtained by the remaining subsystem at the last moment; the condition 2 includes: , yes The part corresponding to the boundary node; , and V A,n They represent the node voltage vector of the companion circuit of subsystem A at the current moment, which has taken the error correction step into account, and the node voltage vector of the companion circuit of subsystem A at the current moment, respectively. and V A,n satisfy: , G Link,n is the node admittance submatrix of all link branches at the current moment, V O,n-1 is the voltage result of the boundary node obtained by other subsystems at the previous moment, is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject the current vector into the accompanying circuit in the subsystem at the current moment, is the voltage at the boundary node of the first solution at the current moment.

[0021] In this embodiment, in S6, it includes: updating the current injection vector of the subsystem at the current moment; updating the voltage value of the node according to the updated current injection vector; synchronizing the current injection vector and voltage value of all subsystems. Wherein, the updating of the current injection vector of the subsystem at the current moment includes: ，G Link,n-1 is the node admittance submatrix of all associated branches at the previous moment, V O,n-2 is the voltage result of the boundary nodes of the remaining subsystem obtained at the first two moments, is the inverse matrix of the node admittance matrix corresponding to the current subsystem at the previous moment, I A,n-1 Injecting a current vector into the accompanying circuit in the current subsystem at the last moment; and updating the voltage value of the node according to the updated current injection vector, including: , is the node voltage value to be corrected, V A,n is the node voltage value after correction, G Link,n is the node admittance submatrix of all associated branches at the current moment; V O,n-1 is the voltage result of the boundary nodes of the remaining subsystems obtained at the previous moment; is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject current vector into the accompanying circuit in the subsystem at the current moment; is the voltage of the boundary node of the first solution at the current moment.

[0022] The method for parallel optimization of random electromagnetic transients in a DC distribution system provided in an embodiment of the present invention selectively implements voltage correction and accelerates through optimal node sorting and local contraction of subsystem networks, thereby realizing efficient parallel simulation of random electromagnetic transients in a DC distribution system under limited resource conditions. The method is suitable for a parallel processing scheme in which associated branches are set between subsystems and delayed transmission is performed through boundary voltages. In order to improve the simulation accuracy of random electromagnetic transients under the parallel simulation scheme, a voltage selective correction step is designed, and an acceleration technology including optimal node sorting and local contraction of subsystem networks is developed to reduce the computational burden of the correction step, thereby fully ensuring the efficiency of parallel analysis of random electromagnetic transients in complex DC distribution systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flow chart of a random electromagnetic transient parallel simulation optimization method for a DC power distribution system provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of a FPGA system solution for random electromagnetic transient parallel simulation of a DC power distribution system provided by an embodiment of the present invention;

[0026] Figure 3 A schematic diagram of the main steps provided for an embodiment of the present invention;

[0027] Figure 4 A schematic diagram of a specific example provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention will be described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention. It can be understood by those skilled in the art that, unless specifically stated, the singular forms "one", "one", "said" and "the" used herein may also include plural forms. It should be further understood that the term "including" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used here may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items. It can be understood by those skilled in the art that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as generally understood by those skilled in the art in the field to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0029] The general design idea of ​​this embodiment is as follows: First, initialize and sort the nodes of the entire DC distribution system. Specifically, it is necessary to sort the static nodes, switch nodes and dynamic nodes, and generate corresponding storage vectors, so as to ensure the orderly progress of subsequent calculations. Secondly, according to the sensitivity matrix of voltage correction, determine the set of nodes that need to be locally shrunk in each subsystem in the voltage correction step. Then, set a unified accuracy standard, and confirm whether to enable the correction step based on the parallel single-step simulation results and the set correction conditions. If enabled, calculate the corresponding voltage correction amount by solving the linear equations, and update the related matrix to reflect the latest system status. Finally, calculation modules including initialization, parameter update, associated branch update and local network contraction are written. Combining all the calculation modules, an FPGA-based random electromagnetic transient parallel simulation implementation scheme is designed.

[0030] The advantages of the scheme of this embodiment are that it selectively implements voltage correction and accelerates it through optimal node sorting and local contraction of subsystem networks, thereby achieving efficient parallel simulation of random electromagnetic transients in DC distribution systems under limited computing resources. It is suitable for setting associated branches between subsystems and parallel processing schemes that transmit voltage delays through boundaries, and is easy to implement and easy to promote.

[0031] The embodiment of the present invention provides a method for parallel optimization of random electromagnetic transients in a DC power distribution system, such as Figure 1 As shown, including:

[0032] Step 1: Subsystem decomposition and associated branch settings.

[0033] Step 2: Node initialization and optimal sorting.

[0034] Step 3: Determine the local shrinkage range and eliminate nodes that have little impact on voltage correction.

[0035] Step 4: Calibration step setting, set a unified voltage result accuracy standard e v。

[0036] Step 5: Parallel single-step simulation stage, use the linear static iteration method to solve the current moment t The node voltage.

[0037] Step 6: Solve the result according to step 5, and according to the set voltage result accuracy standard e v , determine whether the current voltage result meets the calibration conditions, if yes, go to step 7; otherwise, t = t +△ t , repeat step 5, △ t is the set simulation time step; repeat step 4.

[0038] Step 7: Calculate voltage correction value and synchronize data.

[0039] Preferably, the subsystem decomposition and associated branch setting in step 1 include the following steps:

[0040] Step 1.1: Determine the subsystem boundaries, clarify the physical structure and functional divisions of the DC distribution system, and divide the entire system into several subsystems according to different functional areas (such as converter area, load area, etc.). Each subsystem contains its related components and connection relationships.

[0041] Step 1.2: Establish subsystem models. These models should include the electrical parameters and connection methods of all components in the subsystem (e.g., for the converter area, a high-frequency switch model can be used to simulate its fast switching characteristics; for the load area, a constant power or current model can be used to describe its dynamic response).

[0042] Step 1.3: Define association vectors, which are determined by the access position and are used to describe the access mode of the associated branches.

[0043] Step 1.4: In order to simplify the coupling problem in parallel simulation, use delayed voltage for approximation. Take two subsystems as an example: when it is necessary to solve the single-step simulation of subsystem A, a delay can be introduced in advance, that is, the node voltage at the previous moment from subsystem B is used instead of the node voltage at the current moment. This approximation method can also be used for subsystem B.

[0044] Step 1.5: Through the above replacement, the companion circuits of the two subsystems will be reconfigured. If there are multiple subsystems, the multiple subsystems can be regarded as a whole and the above steps can be repeated. Preferably, the companion circuits reconfigured in step 1.5 are as follows: Figure 4 shown.

[0045] Preferably, the node initialization and optimal sorting in step 2 includes the following steps:

[0046] Step 2.1: Classify the nodes in the system according to the type of connected components. Nodes connected to components with parameter changes are called dynamic nodes, nodes connected only to ordinary concentrated components are called static nodes, and the remaining nodes connected to switching devices are called switching nodes.

[0047] Step 2.2: Number the static nodes, switch nodes, and dynamic nodes. Static nodes are numbered first, followed by switch nodes, and finally dynamic nodes.

[0048] Step 2.3: Generate storage vectors for static nodes, switch nodes, and dynamic nodes. The lengths of these vectors are m f , m s and m d ,in m f represents the number of static nodes, m s represents the number of switch nodes, m d Indicates the number of dynamic nodes.

[0049] Step 2.4: Calculate the sort lengths separately m 1= mf , m 2= m f + m s , m 3= m f + m s + m d ,These three lengths determine the maximum sorting length of the node admittance matrices of the three blocks (static nodes, switching nodes and dynamic nodes).

[0050] Step 2.5: Generate the correlation matrix using the above stored vectors and maximum sort length.

[0051] Step 2.6: Sort each block internally to form the desired lower triangular matrix and upper triangular matrix. By decoupling, the fixed part can be reused to avoid re-performing the inverse operation each time. In the specific implementation, it is necessary to count and decompose the non-zero elements in each block to construct a new correlation matrix. Specifically, the internal sorting is formed by counting the node degrees of each block. In the specific implementation, it is necessary to count and decompose the non-zero elements in each block to construct a new correlation matrix. This method ensures that the fixed part can be reused each time it is updated, avoiding frequent calculation of the inverse operation. For example, step 2.6 is coded and executed.

[0052] function[ x i , S =Block-ordering( x i , m i , S ):

[0053] m = length( x i ) / / Total number of nodes to be sorted

[0054] for k =1: m do

[0055] z = x i ( k : m ); / / x i The number of nodes that need to be sorted

[0056] p = [0, …, 0] m-k+1 ; / / Store the number of non-zero elements injected

[0057] get p by decomposing S : Will z The node in is set to k + m i - m Nodes, for S Decompose, count the number of non-zero elements, and use that number as p value;

[0058] j =min( p );

[0059] get the node to be indexed: if length( j )>1, S The node with the smallest degree is x i ( j );otherwise j = j ;

[0060] update S by latest ordering: Exchange S No. k Row and x i ( j ) line, and k Column and x i ( j )List;

[0061] x i ( k )↔ x i ( j );

[0062] end for

[0063] return x i , S 。

[0064] Preferably, determining the local shrinkage range and eliminating nodes with little influence on voltage correction in step 3 includes the following steps:

[0065] Step 3.1: Define the sensitivity matrix S A , which represents the influence of the boundary node voltage deviation of other subsystems on the node voltage correction of the current subsystem. Each element in the sensitivity matrix S A ( i , j ) indicates that when j The voltage deviation of the associated branch has an impact on the i The impact of node voltage correction;

[0066] Step 3.2: To simplify the analysis, reduce the amount of calculation by normalization and standardize the sensitivity s i The expression is Where: S Ak It corresponds to k The sensitivity matrix of each working state; j L is the total number of branches associated with the access subsystem, k S is the number of possible working states.

[0067] Step 3.3: Based on the calculated s i , nodes with lower normalized sensitivity are removed from S A ; where nodes with low normalized sensitivity are removed from S A Remove, determine s i Whether the value is low can be judged by setting a threshold, which needs to be determined based on the specific sensitivity matrix S A Setting, specifically: 1) Empirical method: A preliminary threshold range can be set based on historical data and experience. For example, for a subsystem A in a DC power distribution system, the voltage and current changes under different fault conditions can be simulated to calculate s i 1) Dynamic adjustment method: In actual operation, the threshold can be adjusted dynamically according to the real-time operation situation. For example, in the transient process caused by the action of the switch device, the default setting is a high threshold to improve accuracy; in the steady state stage, the threshold can be appropriately lowered to save computing resources; 2) Segmented setting method: Different thresholds can be set for different nodes and subsystems. For example, for nodes and subsystems with greater impact, a higher threshold can be set; for nodes and subsystems with less impact, a lower threshold can be set.

[0068] Step 3.4: Reconstruct the corrected sensitivity matrix S A ,The new matrix only contains the remaining high sensitivity nodes, ensuring the accuracy of voltage correction and the improvement of computational efficiency.

[0069] Preferably, the linear static iteration method for solving the node voltage in step 5 includes the following steps:

[0070] Step 5.1: Taking two subsystems as an example, the node voltage equation can be expressed as: , where: G AB,n and G BA,n is the node admittance submatrix corresponding to the associated branch between subsystems A and B; G A,n and G B,n are the node admittance matrices corresponding to the reconstruction of the companion circuits of subsystem A and subsystem B respectively; I A,n and I B,n are the injected currents of subsystem A and subsystem B respectively; V A,n and V B,n are the node voltages of subsystem A and subsystem B respectively.

[0071] Step 5.2: By decomposing the node admittance matrix in step 5.1, we can get: , where: G AB,n and G BA,n is the node admittance submatrix corresponding to the associated branch between subsystems A and B; G A,n and G B,n are the node admittance matrices corresponding to the reconstruction of the companion circuits of subsystem A and subsystem B respectively; I A,n and I B,n are the injected currents of subsystem A and subsystem B respectively; V A,n and V B,n are the node voltages of subsystem A and subsystem B respectively.

[0072] Step 5.3: The linear equation in step 5.2 can be solved by linear static iteration method, and the node voltage equation can be calculated as follows until convergence: , where: G AB,n andG BA,n is the node admittance submatrix corresponding to the associated branch between subsystems A and B; G A,n and G B,n are the node admittance matrices corresponding to the reconstruction of the companion circuits of subsystem A and subsystem B respectively; I A,n and I B,n are the injected currents of subsystem A and subsystem B respectively; V A,n and V B,n are the node voltages of subsystem A and subsystem B respectively; the superscript ( k ) is the k The corresponding result of the iteration; the superscript -1 indicates the matrix inversion operation.

[0073] Preferably, the correction conditions in step 6 include the following conditions: Condition 6.1: used to determine whether it is necessary to t n Execute the correction step. This condition focuses on the convergence performance of the remaining subsystems (i.e., the subsystems except the ones to be corrected). The specific formula is: , where: e v To set a uniform accuracy standard; yes The part corresponding to the boundary node; ; is the voltage of the boundary node in the first solution of the subsystem being sought at the current moment; V O,n-1 Represents the boundary node voltage of the remaining subsystem obtained at the last moment.

[0074] Condition 6.2: Used to determine whether at the next moment t n+1 Enable the correction step, this condition depends on the current time t n The correction effect mainly focuses on the convergence performance of the current subsystem (i.e., the subsystem to be corrected) after correction. The specific formula is: Where: yes The part corresponding to the boundary node; ; and V A,n satisfy:

[0075] , where: G Link,nis the node admittance submatrix of all link branches at the current moment; V O,n-1 is the voltage result of the boundary node of other subsystems obtained at the previous moment; is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject current vector into the accompanying circuit in the subsystem at the current moment; is the voltage of the boundary node of the first solution at the current moment.

[0076] Preferably, determining whether the correction condition is met in step 6 comprises the following steps:

[0077] Step 6.1: Determine whether condition 6.1 is satisfied. If so, perform the correction step; if not, proceed to step 6.2;

[0078] Step 6.2: Determine whether condition 6.2 is satisfied. If so, execute the correction step; if not, determine that the correction step does not need to be executed.

[0079] Preferably, calculating the voltage correction value and synchronizing the data in step 7 includes the following steps: Step 7.1: calculating the current injection vector according to the previous moment, V A,n-1 , and , update the current injection vector of each subsystem at the current moment. Among them, yes The part corresponding to the boundary node; is the voltage of the boundary node in the first solution at the previous moment; Satisfy the following formula: , where: G Link,n-1 is the node admittance submatrix of all associated branches at the previous moment; V O,n-2 is the voltage result of the boundary nodes of the remaining subsystems obtained at the first two moments; is the inverse matrix of the node admittance matrix corresponding to the current subsystem at the previous moment, I A,n-1 Inject the current vector into the companion circuit in the current subsystem at the previous moment.

[0080] Step 7.2: Recalculate the voltage value of each node based on the updated current injection vector. The specific formula is: , where: is the node voltage value to be corrected, V A,n is the node voltage value after correction, G Link,nis the node admittance submatrix of all associated branches at the current moment; V O,n-1 is the voltage result of the boundary nodes of the remaining subsystems obtained at the previous moment; is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject current vector into the accompanying circuit in the subsystem at the current moment; is the voltage of the boundary node of the first solution at the current moment.

[0081] Step 7.3: Synchronize and exchange the data of each subsystem, such as voltage correction value, injection current vector, admittance matrix information, boundary node voltage, equivalent current deviation, etc., including the voltage correction value, current injection information and other relevant data of each subsystem.

[0082] The present invention discloses a random electromagnetic transient parallel optimization system for a DC power distribution system, wherein the system is an FPGA system, such as Figure 2 As shown, the following calculation modules are included:

[0083] Calculation module ①: used to call and generate the lower triangular matrix of the complete admittance matrix and the truncated admittance matrix.

[0084] Computation module ②: used to generate and initialize the sub-matrices needed in the correction step so that these matrices can be quickly accessed when executing the correction step to improve the calculation efficiency.

[0085] Calculation module ③: used for parallel single-step simulation process, updating the current injection value according to the voltage results and historical current data of the boundary nodes.

[0086] Calculation module ④: used for parallel single-step simulation process, updates the upper triangular matrix and lower triangular matrix of the admittance matrix according to the voltage and current data of the current time step.

[0087] Calculation module ⑤: used for parallel single-step simulation process, updates the sub-matrices required for correction according to the latest voltage and current data. These updated matrices will be used for subsequent voltage correction calculations to synchronize the latest boundary voltage information of each parallel subsystem.

[0088] Calculation module ⑥: used for parallel single-step simulation process, using the updated admittance matrix and current data to solve the voltage value of each node.

[0089] Calculation module ⑦: used for parallel single-step simulation process to determine whether to execute the correction step within the current time step.

[0090] Calculation module ⑧: used for parallel single-step simulation process, used in conjunction with calculation module ⑦. If correction is required, the voltage correction value is calculated and synchronized based on the updated current data and admittance matrix.

[0091] Calculation module ⑨: used for parallel single-step simulation process, calculates the current value of each branch according to the node voltage and the associated branch admittance matrix.

[0092] The DC power distribution system random electromagnetic transient parallel simulation optimization method and system designed in this embodiment selectively implements voltage correction and accelerates through optimal node sorting and local contraction of subsystem networks, thereby realizing efficient parallel simulation of random electromagnetic transients in DC power distribution systems under limited resource conditions. It is suitable for setting associated branches between subsystems and parallel processing solutions that transmit voltage delays through boundaries. The designed FPGA system implementation plan ensures the efficiency and accuracy of parallel simulation of complex DC power distribution systems.

[0093] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The above is only a specific implementation method of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or replacements that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for parallel optimization of random electromagnetic transients in a DC power distribution system, characterized in that: include: S1, decomposing the DC power distribution system to obtain various subsystems, and then reconfiguring associated branches; S2, initialize the nodes in the subsystem and then sort the nodes; S3, screening the sorted nodes, wherein the screening result is related to the influence degree of the nodes on the voltage correction; S4, obtaining the node voltage at the current moment; S5. According to the set voltage result accuracy standard, detect whether the node voltage at the current moment meets the correction condition. If so, perform voltage correction; otherwise, enter the next moment and repeat S4 to S5; S6, synchronize the data of each subsystem; S1, including: Dividing the DC power distribution system into at least two subsystems according to preset subsystem boundaries, and recording electrical parameters and connection relationships of components in each subsystem; Establish each subsystem model according to the electrical parameters and connection relationships of the components in the subsystem; The reconfiguration of the associated branches includes: replacing the node voltage of a subsystem at the previous moment with the node voltage of another subsystem associated with the subsystem at the current moment; S2, including: Classify the nodes in the system according to the type of connected components, and the classification results include: dynamic nodes, static nodes and switch nodes; The classified nodes are numbered, and then a storage vector of each type of node is generated. The priority of the numbering is from high to low: static node, switch node, and dynamic node. The length of the storage vector is equal to the number of nodes of the same type. Generate a correlation matrix according to the obtained storage vector and the maximum sorting length obtained by sorting the nodes; S3, comprising: determining a local contraction range of the node, and then removing nodes with a small influence on the voltage correction from the sorted nodes, wherein the influence of the node on the voltage correction is quantified by standardized sensitivity; The method of determining the local shrinkage range and eliminating nodes with less influence on voltage correction includes: defining a sensitivity matrix, each element in the sensitivity matrix represents the influence of the voltage deviation of the associated branch on the node voltage correction, wherein the amount of calculation is reduced by standardization processing; according to the calculated standardized sensitivity, the nodes with lower standardized sensitivity are removed from the sensitivity matrix; and then, the corrected sensitivity matrix is ​​reconstructed; S4 includes: For subsystem A and subsystem B, establish node voltage models ,in, G AB,n and G BA,n is the node admittance submatrix corresponding to the associated branch between subsystems A and B, G A,n and G B,n are the node admittance matrices corresponding to the reconstruction of the companion circuits of subsystem A and subsystem B, I A,n and I B,n are the injected currents of subsystem A and subsystem B respectively, V A,n and V B,n are the node voltages of subsystem A and subsystem B respectively. The subscript n indicates the nth step of the simulation, i.e., t n =nDt,t n represents the time at that time, and Dt represents the simulation step length; Processing the node voltage model by linear static iteration until convergence; The node voltage model is processed by linear static iteration until convergence, including: establishing a node admittance matrix according to the node voltage model: , where Mn, Vn, Nn, and In represent the block diagonal matrix formed by the node admittance matrix corresponding to the reconstruction of the companion circuits of subsystems A and subsystem B, the node voltage vector formed by the node voltages of subsystems A and subsystem B, the zero-complement matrix formed by the node admittance submatrices corresponding to the associated branches between subsystems A and B, and the injected current vector formed by the injected currents of subsystems A and subsystem B, respectively; The node admittance matrix is ​​processed by a linear static iterative model until convergence, the linear static iterative model comprising: , where the superscript ( k ) is the k The corresponding result of the iteration; the superscript -1 indicates the matrix inversion operation.

2. The method according to claim 1, characterized in that The standardized sensitivity includes: ,in, S Ak It corresponds to k The sensitivity matrix of each working state is expressed as S A , S A Each element in S A ( i , j ) indicates that when j The voltage deviation of the associated branch has an impact on the i The influence of voltage correction on each node is represented by j, i represents the number of the associated branch, and j represents the number of the node. j L is the total number of associated branches accessing the subsystem, k S is the number of possible working states; the elimination of nodes with little influence on voltage correction includes: according to s i The value of , will move nodes with low normalized sensitivity from S A Removed.

3. The method according to claim 1, characterized in that In S5, the correction conditions include: condition 1 and condition 2, wherein condition 1 is used to detect whether at the current time t n Execute the correction, the condition 2 is used to detect whether at the next moment t n+1 Enable the correction step; The condition 1 includes: , e v To unify the accuracy standard value of the set voltage result, express The part corresponding to the boundary node, , is the voltage of the boundary node in the first solution of the subsystem at the current moment, V O,n-1 represents the boundary node voltage obtained by the remaining subsystem at the last moment; The condition 2 includes: , yes The part corresponding to the boundary node; , and V A,n They represent the node voltage vector of the companion circuit of subsystem A at the current moment, which has taken the error correction step into account, and the node voltage vector of the companion circuit of subsystem A at the current moment, respectively. and V A,n satisfy: , G Link,n is the node admittance submatrix of all link branches at the current moment, V O,n-1 is the voltage result of the boundary node obtained by other subsystems at the previous moment, is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject the current vector into the accompanying circuit in the subsystem at the current moment, is the voltage at the boundary node of the first solution at the current moment.

4. The method according to claim 3, characterized in that In S6, it includes: Update the current injection vector of the subsystem at the current moment; Update the voltage value of the node according to the updated current injection vector; Synchronize the current injection vectors and voltage values ​​of all subsystems.

5. The method according to claim 4, characterized in that The updating of the current injection vector of the subsystem at the current moment includes: ，G Link,n-1 is the node admittance submatrix of all associated branches at the previous moment, V O,n-2 is the voltage result of the boundary nodes of the remaining subsystem obtained at the first two moments, is the inverse matrix of the node admittance matrix corresponding to the current subsystem at the previous moment, I A,n-1 Inject current vector into the accompanying circuit in the current subsystem at the previous moment; The updating of the voltage value of the node according to the updated current injection vector includes: , is the node voltage value to be corrected, V A,n is the node voltage value after correction, G Link,n is the node admittance submatrix of all associated branches at the current moment; V O,n-1 is the voltage result of the boundary nodes of the remaining subsystems obtained at the previous moment; is the inverse matrix of the node admittance matrix corresponding to the subsystem at the current moment, I A,n Inject current vector into the accompanying circuit in the subsystem at the current moment; is the voltage at the boundary node of the first solution at the current moment.

Images