CN112765828B - Real-time simulation network subdivision method for complex power system - Google Patents
Real-time simulation network subdivision method for complex power system Download PDFInfo
- Publication number
- CN112765828B CN112765828B CN202110117489.0A CN202110117489A CN112765828B CN 112765828 B CN112765828 B CN 112765828B CN 202110117489 A CN202110117489 A CN 202110117489A CN 112765828 B CN112765828 B CN 112765828B
- Authority
- CN
- China
- Prior art keywords
- switch
- node
- subsystem
- subdivision
- switches
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/04—Power grid distribution networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/20—Design reuse, reusability analysis or reusability optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a real-time simulation network subdivision method of a complex power system, which comprises the following steps of 1) traversing a system switch element by adopting a depth-first search method; 2) subdivision of the subsystem based on switching node transfer; 3) and partitioning the sub-system internal sub-network based on the merging of the adjacent switch nodes. According to the subdivision process of the subsystem and the subdivision process of the sub-network in the single subsystem, different subdivision targets are adopted to carry out hierarchical network subdivision, and the problems that a complex power system is difficult to realize in real-time simulation and a traditional network subdivision method is low in subdivision efficiency, high in blindness and dependent on modeling experience of designers are solved.
Description
Technical Field
The invention relates to the field of real-time simulation of power systems, in particular to a real-time simulation network subdivision method of a complex power system.
Background
The high-efficiency real-time simulation method is the key for realizing the real-time simulation of the large-scale power system. In the real-time simulation process, hundreds of thousands of nodes of power system network topology electromagnetic transient state real-time simulation under microsecond-level calculation step length is realized, and strict requirements are provided for the rationality and the high efficiency of system network subdivision. The existing power system comprises a large number of power electronic devices, and elements of the power electronic devices are unevenly distributed, so that the real-time computing efficiency of the Rt-lab real-time simulation platform which takes the switching elements as main computing efficiency influence factors is low; in the traditional real-time simulation network subdivision process, arbitrary subdivision is often performed by means of modeling experience of designers, the blindness is high, subdivision targets of different subdivision steps are mixed, the subdivision efficiency is low due to the fact that the method is continuously applied to network subdivision of the existing complex power system, the optimality of a generated subdivision scheme cannot be guaranteed, and the difficulty is brought to the realization of real-time simulation of the complex power system with a large number of switching elements. Therefore, a new network subdivision method needs to be developed to meet the requirements of high efficiency and optimality of the network subdivision of the power system. However, when a real-time simulation network partitioning method of a complex power system is researched, the following three difficulties exist:
1) the power electronic components are various in types and are not uniformly distributed. The existing power system comprises a plurality of types of power electronic equipment and irregular distribution of power electronic elements, so that an Rt-lab real-time simulator which takes a switch element as a main simulation calculation efficiency influence factor cannot efficiently process the switch elements, and the real-time simulation is difficult to realize.
2) In the process of the network subdivision of the power system, subdivision targets of subsystems are different from those of subnets in a single subsystem. In the real-time simulation network subdivision process, subdivision targets of the subsystems are in switch average distribution and the number of cutting branches is the minimum, subdivision targets of sub-networks in a single subsystem are in switch average distribution and the number of state space nodes is the minimum, and if the subdivision targets are in a rough way, the optimality of a subdivision scheme cannot be guaranteed.
3) The real-time simulation calculation precision requirement of the power system is high. In order to accurately depict the dynamic characteristics of equipment contained in a power system and release risks for equipment commissioning and engineering construction in advance, high-precision full-system real-time operation needs to be realized, and then more accurate quantitative analysis and design of the system are carried out.
Therefore, it is necessary to provide a network partitioning method, which partitions the computation content of the whole system into different CPU cores for parallel computation reasonably and efficiently, so as to implement high-precision real-time operation of the whole system.
Disclosure of Invention
The invention aims to provide a real-time simulation network subdivision method of a complex electric power system aiming at the defects of the prior art, and solves the problems of low network subdivision efficiency and difficult realization of high-precision real-time simulation of the complex electric power system containing a large number of switching elements by analyzing different subdivision targets to be met when network subdivision in different steps is performed and adopting a hierarchical network subdivision method based on depth-first search result trees.
The invention provides a real-time simulation network subdivision method of a complex power system, which comprises the following steps:
1) traversing the system switch element by adopting a depth-first search method;
2) subdivision of the subsystem based on switching node transfer;
3) and partitioning the sub-system internal sub-network based on the merging of the adjacent switch nodes.
Further, the specific steps of step 1) are as follows:
1.1) selecting a switching node with the number of 1 as an initial state of system traversal for a power system topology with N switching nodes based on system connectivity, giving a mark 1, and unfolding traversal, wherein an integral element cluster is used as a single switching node to participate in traversal;
1.2) judging whether the switch node adjacent to the Mth switch node is given a mark or not for the Mth switch node, wherein M belongs to N; if all the M switch nodes are marked, judging that all the adjacent connection switch nodes of the M switch node are marked; if the switch nodes which are not marked exist, giving the switch node with the minimum number in all the switch nodes which are not marked with the minimum number to the minimum continuous mark which is not used, and continuously searching the adjacent nodes by taking the switch node given the minimum continuous mark as a starting point;
1.3) if the switch node with the mark number of K is judged to be completely marked by all adjacent connected switch nodes, judging whether the value of the mark number meets the condition that K is not equal to N, if so, returning to the parent node of the switch node with the mark number of K along the historical path, and continuing to execute the switch node marking operation in the step 1.2); and if the mark number K is equal to N, ending the traversal, and obtaining a depth-first search switch distribution undirected graph with the mark numbers of 1-K.
Further, the specific steps of step 2) are as follows:
2.1) arbitrarily dividing the object system into P subsystems, setting a switch distribution balance index alpha, and calculating the switch distribution balance h of the subsystems 1-P i ：
Wherein h is i The switch distribution balance degree of the ith subsystem; s i The number of switching elements included in the ith subsystem;
ideally, the number of switching elements included in a single subsystem, i.e. the number of switching elements included in a single subsystem, is such that the switches in the different subsystems are equally distributed
Rounding the result of F/P, wherein F is the total number of switching elements contained in the system, and P is the number of subdivision subsystems;
if the difference between the subsystem with the most switches and the subsystem with the least switches is not more than D, the value of the switch distribution balance index alpha can be obtained according to the formula T 1 Calculating to obtain:
2.2) determining an initial subsystem i, making i equal to 1, and comparing the switch distribution balance degree of a subsystem j adjacent to the subsystem i and the system i, wherein j is more than 1 and less than or equal to P;
if | h i -h j If | is less than or equal to alpha, the switch distribution balance degree of two adjacent subsystems of the subsystem i and the subsystem j is judged to meet the requirementMarking requirements; if | h i -h j |>Alpha, judging that the switch distribution balance degree of two adjacent subsystems of the subsystem i and the subsystem j does not meet the index requirement;
2.3) if the switch distribution balance degree of the two adjacent subsystems in the step 2.2) does not meet the index requirement, taking the subsystem with a large number of switches in the two subsystems, determining the connected switch nodes 1-c on the branch connected with the other subsystem in the subsystem with the large number of switches, and calculating the cutting branch number r of the switch nodes 1-c 1 ～r c R is to 1 ～r c Sorting from small to large, and taking r 1 ～r c The switch node corresponding to the minimum value in the group carries out transfer operation, and r is 1 ～r c The switch node corresponding to the minimum numerical value in the step 2.2) is transferred to another subsystem until the switch distribution balance degree of the subsystems i and j in the subsystem meets the index requirement;
2.4) if the switch distribution balance of the two adjacent subsystems in the step 2.2) meets the index requirement, judging whether the serial number i of the current subsystem meets the condition that i is less than P: if yes, making i equal to i +1, and continuing to execute the switch node transfer operation in the step 2.3); if not, judging the switch unbalancing degree h of the subsystems 1-P 1 ～h P Whether the absolute value of the difference between the switch distribution balance degrees of any two adjacent subsystems is less than a preset index value alpha or not, namely whether the formula T is met or not 2 ：
If not satisfying the formula T 2 If yes, the node transfer operation of the step 2.2) to the step 2.3) is executed again from the initial subsystem; if the formula T is satisfied 2 And ending the subdivision of the subsystem.
Further, the specific steps of step 3) are as follows:
3.1) determining an initial value of the number of switches in the single subnet; calculating complexity and solving complexity of a node equation according to a state space equation, and estimating an initial value of the optimal switch number in a single subnet; total number of switches in known systemThe quantity is F, and assuming that the number of switches in a single subnet is S, the number of subnets is [ F/S [ ]]The maximum order of the node equation is [ F/S ]]-1, the overall time complexity calculation formula T 3 Comprises the following steps:
T 3 ＝2 S +([N/S]-1) 3
let formula T 3 Taking the minimum value, determining the initial value S of the number of switches in a single subnet 0 Determining the root node, and making the initial value of simulation step size step s ＝s r Wherein s is r Selecting an unacceptable real-time simulation step value according to application requirements for starting a subdivision process;
3.2) defining the number of nodes contained in the path from the leaf node to the root node as the length of the leaf node, taking all subtrees with the longest length in the subtree set, wherein the length is L, and judging the number S of switches contained in the leaf node of the subtree L Whether or not S is satisfied L ≥S 0 : if yes, splitting the leaf node of the longest sub-tree into a plurality of switches with the number not exceeding S 0 The leaf nodes of (4) are put into the subnet set R; if not, judging whether the leaf node of the longest subtree is combined with the mother node of the leaf node of the longest subtree and the number of switches is satisfied with S L +S L-1 ≥S 0 If yes, not executing S L And S L-1 The merging operation of (1) directly putting the leaf nodes of the longest subtree into the subnet set R; if not, merging the leaf node of the longest subtree with the mother node of the leaf node of the longest subtree;
3.3) judging whether the length L of the subtree meets L >1, if so, making L equal to L-1, and continuing to perform the adjacent switch node merging operation in the step 3.2) on other subtrees; if not, judging that all the subtrees are completely merged;
3.4) carrying out fine adjustment on the formed subnet set R by taking minimum switch distribution balance degree as a target, carrying out real-time simulation test, and recording the minimum simulation step size which can be achieved under the condition s ；
3.5) judging the minimum simulation step size step which can be reached at the moment s If the increase is not started, let S be S 0 -1 or S ═ S 0 +1，The operation steps of combining the adjacent switch nodes in the steps 3.2) to 3.4) are repeatedly executed from two directions of reducing and increasing the switch number S in the single subnet; and if the sub-network size begins to increase, selecting the subdivision scheme corresponding to the minimum simulation step size which can be reached, determining the subdivision scheme as the optimal subdivision scheme, and finishing subdivision of the sub-network in the single sub-system.
Compared with the prior art, the invention has the following beneficial effects:
1. the real-time simulation network subdivision method of the complex power system effectively solves the problem that the real-time simulation of the power system is difficult to realize due to the fact that power electronic equipment in the system are various in types and uneven in distribution and the subdivision efficiency of the traditional network subdivision method is low, and realizes high-precision real-time operation of the whole system under given simulation resources;
2. in the sub-system dividing step, two dividing targets of switch distribution balance degree and cutting branch number are considered, and a sub-system dividing method with uniform switch distribution and less communication quantity is formed; in the sub-network dividing step in a single subsystem, two dividing targets of switch distribution balance and node number are considered, and a sub-network dividing method with uniform switch distribution and lower SSN node equation order is formed;
3. the method avoids the defect that the traditional method depends on modeling experience of designers, and realizes automatic execution of the whole operation process; a complete real-time simulation network subdivision method system of the complex power system is constructed, and technical support is provided for accurate quantitative analysis and design of the system.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a flow chart of a subdivision method of a subsystem in a hierarchical network subdivision method;
FIG. 3 is a flow chart of the subdivision of sub-networks within a single sub-system in a hierarchical network subdivision method;
FIG. 4 is a diagram of the results of a depth-first search of an example system model of a medium voltage DC integrated power system;
FIG. 5 is a graph of the results of arbitrarily dividing a medium voltage DC integrated power system into 8 subsystems;
FIG. 6 is a diagram of a subdivision result of a subsystem transferred through a switching node;
FIG. 7 is a diagram of subnet partition results merged by neighboring switch nodes inside different subsystems;
FIG. 8 is a simulation diagram of input current of the inverter when the rotating speed command value of the multiphase induction motor is decreased from 200rpm to 20rpm and immediately increased from 20rpm to 200rpm under the condition of no load;
FIG. 9 is a simulation diagram of the variation curve of the rotating speed of the induction motor when the rotating speed command value of the multi-phase induction motor is reduced from 200rpm to 20rpm and is immediately increased from 20rpm to 200rpm under the condition of no load;
FIG. 10 is a graph showing a simulation of the output voltage of the inverter with the load suddenly increasing from no load to 40% load while the multiphase induction motor is maintained at 200 rpm;
FIG. 11 is a graph showing a simulation of the inverter output current with the load suddenly increasing from no load to 40% load while the multiphase induction motor is maintained at 200 rpm;
FIG. 12 is a graph showing a simulation of the output voltage of the inverter when the multiphase induction motor is maintained at a speed of 200rpm and the load on the inverter is suddenly unloaded from full load to no load;
FIG. 13 is a graph showing a simulation of the output current of the inverter when the multiphase induction motor is maintained at a speed of 200rpm and the load on the inverter is suddenly unloaded from full load to no load;
FIG. 14 is a diagram illustrating resource consumption of a real-time simulator after network subdivision is performed according to the present invention.
Detailed Description
The invention will be further described in detail with reference to the following drawings and specific examples, which are not intended to limit the invention, but are for clear understanding.
The real-time simulation network subdivision method of the complex power system shown in fig. 1, 2 and 3 comprises the following steps:
1) traversing system switching elements using depth-first search
1.1) selecting a switch node with the number of 1 as an initial state of system traversal for a power system topology with N switch nodes based on system connectivity, giving a mark number of 1, and unfolding traversal, wherein an integral element cluster is used as a single switch node to participate in traversal;
1.2) judging whether the switch node adjacent to the Mth switch node is given a mark or not for the Mth switch node, wherein M belongs to N; if all the M switch nodes are marked, judging that all the adjacent connection switch nodes of the M switch node are marked; if the unmarked switch nodes exist, giving the switch node with the minimum number in all the unmarked switch nodes to the unused minimum continuous mark number, and continuously searching the adjacent nodes by taking the switch node given the minimum continuous mark number as the starting point;
1.3) if the switch node with the mark number of K is judged to be completely marked by all adjacent connected switch nodes, judging whether the value of the mark number meets the condition that K is not equal to N, if so, returning to the parent node of the switch node with the mark number of K along the historical path, and continuing to execute the switch node marking operation in the step 1.2); if the mark number K is equal to N, the traversal is finished, and a depth-first search switch distribution undirected graph with the mark numbers of 1-K is obtained;
2) subdivision of the subsystem based on switching node transfer;
2.1) arbitrarily dividing the object system into P subsystems, setting a switch distribution balance index alpha, and calculating switch distribution balances h of the subsystems 1-P i ：
Wherein h is i The switch distribution balance degree of the ith subsystem; s i The number of switching elements included in the ith subsystem;
ideally, the number of switching elements included in a single subsystem, i.e. the number of switching elements included in a single subsystem, is equal when the switches in different subsystems are equally distributed
Rounding the result of F/P, where F is included in the systemThe total number of the switching elements is P, and the number of the subdivision subsystems is P;
if the difference between the subsystem with the most switches and the subsystem with the least switches is not more than D, the value of the switch distribution balance index alpha can be obtained according to the formula T 1 Calculating to obtain:
2.2) determining an initial subsystem i, making i equal to 1, and comparing the switch distribution balance degree of a subsystem j adjacent to the subsystem i and the system i, wherein j is more than 1 and less than or equal to P;
if | h i -h j If the | is less than or equal to alpha, judging that the switch distribution balance degrees of two adjacent subsystems of the subsystem i and the subsystem j meet the index requirement; if | h i -h j |>Alpha, judging that the switch distribution balance degree of two adjacent subsystems of the subsystem i and the subsystem j does not meet the index requirement;
2.3) if the switch distribution balance degree of the two adjacent subsystems in the step 2.2) does not meet the index requirement, taking the subsystem with a large number of switches in the two subsystems, determining the connected switch nodes 1-c on the branch connected with the other subsystem in the subsystem with the large number of switches, and calculating the cutting branch number r of the switch nodes 1-c 1 ～r c R is to 1 ～r c Sorting from small to large, and taking r 1 ～r c The switch node corresponding to the minimum value in the data carries out transfer operation, and r is converted into 1 ～r c The switch node corresponding to the minimum numerical value in the step 2.2) is transferred to another subsystem until the switch distribution balance degree of the subsystems i and j in the subsystem meets the index requirement;
2.4) if the switch distribution balance degree of the two adjacent subsystems in the step 2.2) meets the index requirement, judging whether the serial number i of the current subsystem meets the condition that i is less than P: if yes, making i equal to i +1, and continuing to execute the switch node transfer operation in the step 2.3); if not, judging the switch unbalancing degree h of the subsystems 1-P 1 ～h P Whether all of the switch distribution balance degrees of any two adjacent subsystems are satisfiedThe absolute value of the difference is less than the preset index value alpha, namely whether the formula T is met 2 ：
If not satisfying the formula T 2 If yes, the node transfer operation of the step 2.2) to the step 2.3) is executed again from the initial subsystem; if the formula T is satisfied 2 And ending the subdivision of the subsystem.
3) Partitioning the sub-system internal sub-network based on the adjacent switch node combination;
3.1) determining an initial value of the number of switches in the single subnet; calculating complexity and solving complexity of a node equation according to a state space equation, and estimating an initial value of the optimal switch number in a single subnet; given that the number of total switches in a system is F, assuming that the number of switches in a single subnet is S, the number of subnets is [ F/S ]]The maximum order of the node equation is [ F/S ]]-1, the overall time complexity calculation formula T 3 Comprises the following steps:
T 3 ＝2 S +([N/S]-1) 3
let formula T 3 Taking the minimum value, determining the initial value S of the number of switches in a single subnet 0 Determining the root node, and making the initial value of simulation step size step s ＝s r Wherein s is r Selecting an unacceptable real-time simulation step size value according to application requirements for starting a subdivision process;
3.2) defining the number of nodes contained in a path from a leaf node to a root node as the length of the leaf node, taking all subtrees with the longest length in the subtree set, wherein the length is L, and judging the number S of switches contained in the leaf node of each subtree L Whether or not S is satisfied L ≥S 0 : if yes, splitting leaf nodes of the longest subtree into a plurality of switches with the number not exceeding S 0 The leaf nodes of (2) are put into the subnet set R; if not, judging whether the leaf node of the longest subtree is combined with the mother node of the leaf node of the longest subtree and the number of switches is satisfied with S L +S L-1 ≥S 0 If yes, not executing S L And S L-1 The merging operation of (2) directly putting the leaf nodes of the longest sub-tree into the sub-network set R; if not, merging the leaf node of the longest subtree with the mother node of the leaf node of the longest subtree;
3.3) judging whether the length L of the subtree meets L >1, if so, making L equal to L-1, and continuing to perform the adjacent switch node merging operation in the step 3.2) on other subtrees; if not, judging that all the subtrees are completely merged;
3.4) fine-tuning the formed subnet set R by taking the minimum switch distribution balance degree as a target, carrying out real-time simulation test, and recording the minimum simulation step length step which can be achieved under the condition s ；
3.5) judging the minimum simulation step size step which can be reached at the moment s If the increase is not started, let S be S 0 -1 or S ═ S 0 +1, the adjacent switch node merging operation steps in the steps 3.2) to 3.4) are repeatedly executed from two directions of decreasing and increasing the number S of switches in a single sub-network; if the sub-network size begins to increase, selecting a subdivision scheme corresponding to the minimum simulation step length which can be reached, determining the subdivision scheme as the optimal subdivision scheme, and finishing subdivision of the sub-network in a single sub-system;
4) and (5) verifying real-time performance.
Based on the Rt-lab real-time simulation platform, different subsystems formed by the hierarchical network subdivision schemes in the steps 1) to 3) correspond to different CPU cores of the simulator, according to the characteristics of the simulator, the subsystem with more total number of switches corresponds to the secondary subsystem, the subsystem with the least total number of switches corresponds to the main subsystem, and real-time verification is carried out. Therefore, a set of complete automatic hierarchical subdivision method for real-time simulation of large-scale complex power systems is obtained, and the method can be used for real-time simulation network subdivision of power systems containing a large number of switching elements.
The real-time simulation network subdivision method of the complex power system effectively solves the problem that the real-time simulation of the power system is difficult to realize due to the fact that power electronic equipment in the system are various in types and uneven in distribution and the subdivision efficiency of the traditional network subdivision method is low, and realizes high-precision real-time operation of the whole system under given simulation resources; in the sub-system dividing step, two dividing targets of switch distribution balance degree and cutting branch number are considered, and a sub-system dividing method with uniform switch distribution and less communication quantity is formed; in the sub-network dividing step in a single subsystem, two dividing targets of switch distribution balance and node number are considered, and a sub-network dividing method with uniform switch distribution and lower SSN node equation order is formed; meanwhile, the method avoids the defect that the traditional method depends on modeling experience of designers, and realizes automatic execution of the whole operation process; finally, a complete complex power system real-time simulation network subdivision method system is constructed, and technical support is provided for accurate quantitative analysis and design of the system.
As shown in fig. 4, according to the network dividing steps, the dividing results of different dividing steps of the medium-voltage direct-current integrated power system shown in fig. 6 and 7 can be obtained.
Parameters of twelve same-step rectification generators (equivalent to 4 three-phase synchronous generators) are as follows: stator resistance 0.003pu, stator leakage reactance 0.147pu, stator d-axis armature reaction reactance 1.26pu, excitation resistance 0.001pu, excitation leakage reactance 0.182pu, d-axis damping winding resistance 0.01pu, d-axis damping winding reactance 0.057pu, stator q-axis armature reaction reactance 1.262pu, q-axis damping winding resistance 0.003pu, q-axis damping winding leakage reactance 0.218 pu;
fifteen-phase induction motor (equivalent to five-phase induction motor) parameters: the rotor equivalent fundamental wave resistance is 0.104pu, the single five-phase winding self fundamental wave mutual leakage reactance is 0.01pu, the fundamental wave excitation reactance is 0.057pu, and the fundamental wave mutual leakage reactance between two adjacent five-phase windings is-0.001 pu;
DC/DC converter parameters: LCL filter inductance L DC1 250 microhenry, filter capacitance C DC 5 millifarad, filter capacitance L DC2 50 microHenry, voltage outer loop controller proportion link K DCpp 23, voltage outer loop controller integral link K DCpi 600, current inner loop controller proportion link K DCip Integral link K of current inner loop controller (0.01) DCii ＝34；
DC/AC inverter parameters: filter resistor R of LCL filter AC1 0.002 ohm ═ 0.002 ohmFilter inductance L AC1 125 microhenry, filter capacitance C AC 630 microfarads, filter resistance R AC2 0.0005 ohm, filter inductance L AC2 38 microhenry, voltage outer loop controller proportional link K ACpp 0.2, voltage outer loop controller integral link K ACpi 300, current inner loop controller proportion link K ACip 1, current inner loop controller integral link K ACii ＝5；
Example system parameters: the length of the cable from the generator to the distribution board is 50 meters, and the line resistance R L1 0.0004 ohm, line inductance L L1 15 microhenry; the length of the cable of the cross-over screen is 100 meters, and the circuit resistance R L2 0.0014 ohm, line inductance L L2 15 microhenry; the length of the cable from the distribution board to the fifteen-phase induction motor is 50 meters, and the line resistance R L3 0.0004 ohm, line inductance L L3 15 microhenry; the cable length from the distribution board to the DC/DC converter is 50 meters, and the line resistance R L4 0.0003 ohm, line inductance L L4 20 microhenry; the length of the cable from the DC/DC converter to the bus-bar screen is 30 meters, and the line resistance R L5 0.0007 ohm, line inductance L L5 20 microhenry; the length from the bus-coupled screen to the DC/AC inverter cable is 30 meters, and the line resistance R L6 0.0014 ohm, line inductance L L6 15 microhenry; the local distribution part loads are respectively configured into 6 370 kilowatt resistive-inductive loads and 6 630 kilowatt resistive-inductive loads, and the power factors are all 0.8;
setting parameters of a hierarchical network subdivision method: the number P of the subdivision subsystems is 8, the difference between the number of the switches of the subnet with the maximum number of switches and the number of the switches of the subnet with the minimum number of switches is not more than 20, and the step of calculating the target calculation step in real-time simulation t ≤50μs；
Specific examples of the invention are given below:
real-time simulation network subdivision method of complex power system
1) Traversing switch elements in a system by adopting a depth-first search method
1.1) based on system connectivity, the node number of a switch element in the system is 1-148, 148 switch nodes are totally arranged, traversal is started from the switch node with the number of 1, and a mark number of 1 is given;
1.2) checking whether the switch node with the number of 1 has an unmarked switch node, namely the switch node with the number of 2, and giving the minimum continuous mark number 2 to the node which is not used; taking the node with the mark number of 2 as a new starting point, continuously searching the switch nodes with the mark number of 2 for the switch nodes with unmarked switch nodes, namely the switch nodes with the numbers of 3 and 4, giving the minimum continuous mark number 3 of the node 3 with the minimum number value which is not used at the moment, and repeating the operation;
1.3) continuously and repeatedly searching unmarked switch nodes and backspacing operations, finally giving a mark number 148 to the switch node with the number of 12, judging that all adjacent connection nodes of the node with the mark number of 148 are marked, wherein the mark number K is 148, the total number N of the switch nodes in the system is 148, and the two are equal, and traversing is finished to obtain a switch distribution undirected graph with the mark numbers of 1-148;
FIG. 4 is a diagram showing the results of a medium voltage DC integrated power system traversed by a depth-first search method
2) Node transfer based subdivision of subsystems
2.1) the topology of the medium-voltage direct-current integrated power system is arbitrarily divided into 8 subsystems, and the number of switches contained in different subsystems is respectively as follows: s 1 ＝81，S 2 ＝86，S 3 ＝80，S 4 ＝98，S 5 ＝63，S 6 ＝59，S 7 ＝59，S 8 62, 588 switching elements, ideally, the number of switching elements contained in a single subsystem is equal to the number of switching elements contained in a single subsystem when the switches in different subsystems are evenly distributed
Calculating the switch distribution balance degrees of different subsystems as h 1 ＝0.095，h 2 ＝0.162，h 3 ＝0.081，h 4 ＝0.324，h 5 ＝0.149，h 6 ＝0.203，h 7 ＝0.203，h 8 ＝0.162；
According to known conditions, from the formula T 1 And the switch distribution balance index alpha is calculated as follows:
alpha is more than or equal to 0.02.
Fig. 5 shows the result of arbitrarily dividing the medium voltage dc integrated power system into 8 subsystems.
2.2) taking the subdivision of the #1 subsystem and the #2 subsystem as an example
Determining the initial subsystem as numbered subsystem 1, making i equal to 1, comparing the switch distribution balance of the #1 subsystem and the #2 subsystem, and | h 1 -h 2 |＝0.067>Alpha is 0.02, and the switch distribution balance degree of the two adjacent subsystems is judged to be not meeting the index requirement;
2.3) taking the #2 subsystem with a large number of switches, determining a connected node on a connected branch between the #2 subsystem and the #1 subsystem, and calculating the number of cutting branches of the connected node, wherein the connected node in the #2 subsystem is known as nodes marked with 7 and 18, if the node 7 is transferred to the #1 region, the number of the connected branches is reduced by 1, if the node 18 is transferred to the #1 region, the number of the connected branches is increased by 1, so that the number of the cutting branches of the node marked with 7 is smaller than that of the node marked with 18, and at this time, the node transfer operation is executed on the switch node marked with 7. This step is repeated until | h 1 -h 2 |≤α；
2.4) after the above-mentioned switch node transfer operation is performed on the subsystems # 1 to #8, when i is 8, the number of switches in the finally obtained subsystems # 1 to #8 is respectively: s 1 ＝82，S 2 ＝82，S 3 ＝81，S 4 ＝80，S 5 ＝67，S 6 ＝66，S 7 ＝65，S 8 65, the switch distribution balance of each subsystem is respectively: h is 1 ＝0.108，h 2 ＝0.108，h 3 ＝0.095，h 4 ＝0.081，h 5 ＝0.095，h 6 ＝0.108，h 7 ＝0.122，h 8 0.122, the switch distribution balance degree of any two adjacent subsystems meets the index requirement, namely the formula T 2 ：
And finishing the subdivision of the subsystem.
Fig. 6 shows the subdivision results of the subsystem.
3) Subdivision of sub-system internal sub-network based on adjacent node combination
Taking the subnet partition in the subsystem 1 as an example:
3.1) determining the number of switches in a single subnet.
From the formula T 3 Determining an initial value S for the number of switches in a single subnet 0 The total number of switches F in the area 1 is 82, and the formula T is substituted into 3 Obtaining S 0 When the number of the single subnet switches in the area 1 is determined to be not more than 8, the other subsystems are used for solving the same subsystem 1, and meanwhile, the node 1 is selected as a root node, so that step is performed s ＝s r And (5) 100, opening the subdivision process.
3.2) there are 5 leaf nodes in the subsystem, 12, 15, 18, 27 and 6, whose lengths are: 11. 11, 9 and 6, namely 5 subtrees in total, the subtree with the longest length is selected: a sub-tree where a leaf node 12 is located and a sub-tree where a leaf node 15 is located, respectively, perform neighbor node merging with the leaf node 12 and the leaf node 15 as starting points, the leaf node 12 includes 3, 3<6 switches, but the parent node of the leaf node 12 is a node 11, the leaf node 12 and the node 11 include 3+4 to 7>6 switches in total, do not perform merging operation on the node 12 and the node 11, place the leaf node 12 in a subnet set R, and continue the neighbor merging operation performed on the parent node 11; similarly, the leaf node 15 also performs the merge operation to the parent node until the number of switches included in the node exceeds 6, does not perform the merge, and puts the leaf node into the subnet set R.
3.3) after the merging operation of the adjacent switch nodes of the 5 subtrees, when L is equal to 1, judging that all the subtrees are merged completely.
3.4) fine-tuning the formed subnet set R by taking the minimum switch distribution balance degree as a target, carrying out real-time simulation test, and recording the minimum simulation step length step which can be achieved under the condition s 。
3.5) threshold number of switches from S within a single subnet 0 Subtracting one and adding one at each time, executing subdivision of the sub-network in the single sub-system, recording the minimum achievable simulation step length until the minimum achievable simulation step length begins to increase, selecting all subdivision schemes corresponding to the minimum value in the simulation step lengths, and determining the subdivision schemes as the optimal subdivision schemes 0 Determining S as 8 0 And (4) the subdivision scheme is the optimal subdivision scheme when the subdivision scheme is 8, and the subdivision in the single subsystem is finished.
FIGS. 8-13 show the results of the sub-network subdivision in different subsystems.
4) Real-time verification
Considering the characteristics of the Rt-lab real-time simulation platform, the calculation contents in the #1 subsystem and the #8 subsystem are exchanged, and finally the real-time operation of the medium-voltage direct-current integrated power system with the simulation step length of 40 mu s is realized, wherein the real-time operation is carried out for 40 mu s<step t Fig. 14 shows the resource consumption of the simulator in this case, where continuity timeout does not occur in the whole operation process, and the requirement of the preset index is met.
Those not described in detail in this specification are well within the skill of the art.
Claims (3)
1. A real-time simulation network subdivision method of a complex power system is characterized by comprising the following steps:
1) traversing the system switch element by adopting a depth-first search method;
2) partitioning the subsystem based on switch node transfer;
the specific steps of the step 2) are as follows:
2.1) arbitrarily dividing the object system into P subsystems, setting a switch distribution balance index alpha, and calculating the switch distribution balance h of the subsystems 1-P i ：
Wherein h is i Distributing balance degrees for the switches of the ith subsystem; s. the i The number of switching elements included in the ith subsystem;
ideally, the number of switching elements included in a single subsystem, i.e. the number of switching elements included in a single subsystem, is equal when the switches in different subsystems are equally distributed
Rounding the result of F/P to get an integer, wherein F is the total number of switching elements contained in the system, and P is the number of the subdivision subsystems;
if the difference between the subsystem with the most switches and the subsystem with the least switches is not more than D, the value of the switch distribution balance index alpha can be obtained according to the formula T 1 Calculating to obtain:
2.2) determining an initial subsystem i, setting i to be 1, comparing the switch distribution balance degree of a subsystem j adjacent to the subsystem i and the system i, and setting j to be more than 1 and less than or equal to P;
if | h i -h j If the | is less than or equal to alpha, judging that the switch distribution balance degrees of two adjacent subsystems of the subsystem i and the subsystem j meet the index requirement; if | h i -h j |>Alpha, judging that the switch distribution balance degree of two adjacent subsystems of the subsystem i and the subsystem j does not meet the index requirement;
2.3) if the switch distribution balance degree of the two adjacent subsystems in the step 2.2) does not meet the index requirement, taking the subsystem with a large number of switches in the two subsystems, determining the connected switch nodes 1-c on the branch connected with the other subsystem in the subsystem with the large number of switches, and calculating the cutting branch number r of the switch nodes 1-c 1 ～r c R is to 1 ～r c Sorting from small to large, and taking r 1 ～r c The switch node corresponding to the minimum value in the group carries out transfer operation, and r is 1 ～r c Transferring the switch node corresponding to the minimum numerical value to another subsystem until the switch distribution balance degree of the subsystems i and j in the step 2.2) meets the index requirement;
2.4) if the switch distribution balance of the two adjacent subsystems in the step 2.2) meets the index requirement, judging whether the serial number i of the current subsystem meets the condition that i is less than P: if so, making i equal to i +1, and continuing to execute the switch node transfer operation in the step 2.3); if not, judging the switch unbalancing degree h of the subsystems 1-P 1 ～h P Whether the absolute value of the difference between the switch distribution balance degrees of any two adjacent subsystems is less than a preset index value alpha or not, namely whether the formula T is met or not 2 ：
If not satisfying the formula T 2 If yes, the node transfer operation of the step 2.2) -the step 2.3) is executed again from the initial subsystem; if the formula T is satisfied 2 If yes, ending the subdivision of the subsystem;
3) and partitioning the sub-system internal sub-network based on the merging of the adjacent switch nodes.
2. The real-time simulation network subdivision method for the complex power system according to claim 1, characterized in that: the specific steps of the step 1) are as follows:
1.1) selecting a switching node with the number of 1 as an initial state of system traversal for a power system topology with N switching nodes based on system connectivity, giving a mark 1, and unfolding traversal, wherein an integral element cluster is used as a single switching node to participate in traversal;
1.2) judging whether the switch node adjacent to the Mth switch node is given a mark or not for the Mth switch node, wherein M belongs to N; if all the M switch nodes are marked, judging that all the adjacent connection switch nodes of the M switch node are marked; if the switch nodes which are not marked exist, giving the switch node with the minimum number in all the switch nodes which are not marked with the minimum number to the minimum continuous mark which is not used, and continuously searching the adjacent nodes by taking the switch node given the minimum continuous mark as a starting point;
1.3) if the switch node with the mark number of K is judged to be completely marked by all adjacent connected switch nodes, judging whether the value of the mark number meets the condition that K is not equal to N, if so, returning to the parent node of the switch node with the mark number of K along the historical path, and continuing to execute the switch node marking operation in the step 1.2); and if the mark number K is equal to N, ending the traversal, and obtaining a depth-first search switch distribution undirected graph with the mark numbers of 1-K.
3. The real-time simulation network subdivision method of the complex power system according to claim 1, characterized in that: the specific steps of the step 3) are as follows:
3.1) determining an initial value of the number of switches in the single subnet; calculating complexity and solving complexity of a node equation according to a state space equation, and estimating an initial value of the optimal switch number in a single subnet; given that the number of total switches in a system is F, assuming that the number of switches in a single subnet is S, the number of subnets is [ F/S ]]The maximum order of the node equation is [ F/S ]]-1, the overall time complexity calculation formula T 3 Comprises the following steps:
T 3 ＝2 S +([N/S]-1) 3
let formula T 3 Taking the minimum value, determining the initial value S of the number of switches in a single subnet 0 Determining the root node, and making the initial value of simulation step size step s ＝s r Wherein s is r Selecting an unacceptable real-time simulation step size value according to application requirements for starting a subdivision process;
3.2) defining the number of nodes contained in the path from the leaf node to the root node as the length of the leaf node, taking all subtrees with the longest length in the subtree set, wherein the length is L, and judging the number S of switches contained in the leaf node of the subtree L Whether or not S is satisfied L ≥S 0 : if yes, splitting leaf nodes of the longest subtree into a plurality of switch numbersNot more than S 0 The leaf nodes of (2) are put into the subnet set R; if not, judging whether the leaf node of the longest subtree is combined with the mother node of the leaf node of the longest subtree and the number of switches is satisfied with S L +S L-1 ≥S 0 If yes, not executing S L And S L-1 The merging operation of (1) directly putting the leaf nodes of the longest subtree into the subnet set R; if not, merging the leaf node of the longest subtree with the mother node of the leaf node of the longest subtree;
3.3) judging whether the length L of the subtree meets L >1, if so, making L equal to L-1, and continuing to perform the adjacent switch node merging operation in the step 3.2) on other subtrees; if not, judging that all the subtrees are completely merged;
3.4) fine-tuning the formed subnet set R by taking the minimum switch distribution balance degree as a target, carrying out real-time simulation test, and recording the minimum simulation step length step which can be achieved under the condition s ；
3.5) judging the minimum simulation step size step which can be reached at the moment s If the increase is not started, let S be S 0 -1 or S ═ S 0 +1, the adjacent switch node merging operation steps in the steps 3.2) to 3.4) are repeatedly executed from two directions of decreasing and increasing the number S of switches in a single sub-network; and if the sub-network size begins to increase, selecting the subdivision scheme corresponding to the minimum simulation step size which can be reached, determining the subdivision scheme as the optimal subdivision scheme, and finishing subdivision of the sub-network in the single sub-system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110117489.0A CN112765828B (en) | 2021-01-28 | 2021-01-28 | Real-time simulation network subdivision method for complex power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110117489.0A CN112765828B (en) | 2021-01-28 | 2021-01-28 | Real-time simulation network subdivision method for complex power system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112765828A CN112765828A (en) | 2021-05-07 |
| CN112765828B true CN112765828B (en) | 2022-09-27 |
Family
ID=75706383
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110117489.0A Active CN112765828B (en) | 2021-01-28 | 2021-01-28 | Real-time simulation network subdivision method for complex power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112765828B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113884039B (en) * | 2021-09-17 | 2023-08-08 | 沪东中华造船(集团)有限公司 | Multi-element analysis-based container ship cable length estimation method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104578055A (en) * | 2015-01-12 | 2015-04-29 | 浙江大学 | Power system transient stability simulation method based on network node numbering optimization |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102280885B (en) * | 2011-08-29 | 2013-07-31 | 天津大学 | Method for islanding electric power system into a plurality of subsystems |
| CN106374453B (en) * | 2016-09-05 | 2022-03-18 | 中国电力科学研究院 | Power system reconstruction method |
| CN108429259B (en) * | 2018-03-29 | 2019-10-18 | 山东大学 | A kind of online dynamic decision method and system of unit recovery |
| CN110059094A (en) * | 2019-04-26 | 2019-07-26 | 山东大学 | Partition method, medium and the equipment of active distribution network distributions estimation |
| CN111709107B (en) * | 2020-06-16 | 2023-03-24 | 全球能源互联网研究院有限公司 | Power network simulation method and system containing large-scale power electronic equipment |
-
2021
- 2021-01-28 CN CN202110117489.0A patent/CN112765828B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104578055A (en) * | 2015-01-12 | 2015-04-29 | 浙江大学 | Power system transient stability simulation method based on network node numbering optimization |
Non-Patent Citations (1)
| Title |
|---|
| 考虑环网检测的配电网拓扑重构遗传算法;孙孔明等;《电力***自动化》;20180206(第11期);70-77 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112765828A (en) | 2021-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xu et al. | Selection methods of main circuit parameters for modular multilevel converters | |
| Di Gerlando et al. | Comprehensive steady‐state analytical model of a three‐phase diode rectifier connected to a constant DC voltage source | |
| Fallahzadeh‐Abarghouei et al. | Online hierarchical and distributed method for voltage control in distribution smart grids | |
| CN114070115A (en) | Multi-alternating-current-port high-voltage direct-hanging energy storage power conversion system and control method thereof | |
| CN112765828B (en) | Real-time simulation network subdivision method for complex power system | |
| CN110135039A (en) | Wind-powered electricity generation collects regional non-equilibrium among three phase voltages and determines method and device | |
| CN105680442B (en) | Consider that the expectation of trend and sensitivity consistency equivalence lacks power supply volume appraisal procedure | |
| Munir et al. | Residential distribution system harmonic compensation using priority driven droop controller | |
| CN105449674A (en) | Generalized state space averaging modeling method for aviation 12-pulse inverter | |
| Yu et al. | Efficient and accurate hybrid model of modular multilevel converters for large MTDC systems | |
| CN111062145B (en) | Electromagnetic transient simulation equivalence method based on three-layer partition and electrical distance | |
| CN112564094A (en) | Machine room power control method, device and system | |
| Moghbel et al. | Optimal placement and sizing of multiple STATCOM in distribution system to improve voltage profile | |
| CN113162056B (en) | Automatic voltage control method and device for extra-high voltage power grid | |
| Tiwari et al. | A computer package for multi-contingency constrained reactive power planning | |
| Grégoire et al. | Fpga-based real-time simulation of multilevel modular converter hvdc systems | |
| Cao et al. | Real-time electromagnetic transient simulation for regional power grid based on cloudpss | |
| Pinto et al. | Needs and Improvements in Power Flow Analysis | |
| CN112670992A (en) | Method and system for analyzing stability and correcting instability of power distribution network containing energy router | |
| Girinon et al. | Analytical input filter design in DC distributed power systems approach taking stability and quality criteria into account | |
| Hao et al. | Hierarchical network partition of DC power systems containing a large number of switches | |
| Xu et al. | Research on Electromechanical-electromagnetic Hybrid Simulation Algorithm for Large Scale Power System based on Boundary Nodes Grouping and Decoupling | |
| CN112765920B (en) | Direct-current short-circuit current calculation method and system based on differential-common mode transformation | |
| Camargos et al. | Time-varying harmonic distortions in AC drives | |
| CN117526393B (en) | Method for determining oscillation risk of flexible direct current transmission system under different operation modes |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |