CN103605852A - Parallel topology method for electromechanical transient real-time simulation for large-scale power network - Google Patents

Abstract

The invention provides a parallel topology method for electromechanical transient real-time simulation for a large-scale power network. The parallel topology method includes steps of dividing a network on the basis of a main topology structure of the network; layering and partitioning the network; performing boundary system topology and sub-network full topology; performing sub-network variation topology; integrating electromechanical transient real-time simulation; solving the network after topology changes. The parallel topology method for electromechanical transient real-time simulation for the large-scale power network has the advantages that an MPI (message passing interface)-based parallel computation structure is adopted, so that the topology efficiency can be improved, state change elements are modified by the aid of local admittance matrixes, and the computation speed can be increased owing to a local factor analysis technology; only voltage levels of shifting elements in local sub-networks are subjected to topology after the topology changes, so that the searching range can be reduced, and the efficiency can be improved.

Description

A kind of concurrent topology method for the real-time simulation of large scale electric network electromechanical transient

Technical field

The present invention relates to a kind of topological method, specifically relate to a kind of concurrent topology method for the real-time simulation of large scale electric network electromechanical transient.

Background technology

Electrical-Machanical Transient Simulation of Power System refers to the electromechanical transient stability of the method research electric system of using time-domain-simulation, be that electric system is subject to after large interference, each synchronous generator keeps synchronous operation and is transitioned into ability new or that return to original steady-state operation mode.Electrical-Machanical Transient Simulation of Power System is widely used at aspects such as electric power system design, planning, operation and scheduling, is the effective tool of analyzing Operation of Electric Systems stability and dynamic property.

Along with developing rapidly of electric system, the scale expanding day of electrical network.Requirement to electric system simulation scale progressively improves.The growth of the necessary adaptive system scale of electromechanical transient simulation software, possesses direct modeling large scale electric network, carries out the ability of real-time simulation.The simulation process that the computing velocity that so-called electromechanical transient real-time simulation is exactly requirement software can strictly be synchronizeed with actual electric network physics dynamic process.The electromechanical transient simulation software that possesses real-time simulation ability can access actual physical device and control device, therefore verification checking, closed loop test and the parameter tuning etc. of controlling apparatus of electric power system, steady control strategy etc. is had to practical significance.

Improving electromechanical transient simulation speed except upgrading computer hardware equipment, is exactly the fast algorithm that design accuracy meets the demands, and parallel computing is one of means that improve computing velocity.The method of large-scale electrical power system electromechanical transient parallel artificial is exactly large electrical network to be divided into sub-electrical network and the boundary system of several phase compartments, then by no computer CPU, is born the calculation task of each subnet, improves thus simulation calculation speed.

Power System Network Topology Analysis Using is the utility module of the senior applied analysis program of electric system, its basic function is by the physics connection mode generation of electrical network, to calculate bus model according to power network switch state (folding) and network element state (throwing is moved back), and will have the bus set of electrical link to turn to electric island.Power System Network Topology Analysis Using requires must be fast reliable, and be easy to application.

Conventional network topology analytical method can be divided into two large classes: the 1) method based on from multiply matrix and adjacency matrix; 2) method based on graph search, comprises depth-first and breadth First.The former analysis speed is slower, and applying at present more is the latter.

Power System Network Topology Analysis Using based on graph search is divided into two basic steps:

(1) electric pressure bus is analyzed: according to the state that cut-offs of switch and exiting/returning to form of element, by nodal analysis method, form bus model.Function is that the node of analyzing in a certain voltage level of a certain plant stand is connected into how many buses by Closing Switch, consequently certain electric pressure is divided into several buses.

(2) grid analysis: the bus of analyzing whole electrical network is connected into how many sub-electrical networks by closed branch road (as AC line, transformer coil etc.), every sub-electrical network is the set that has the bus of electrical link, also claim electric island, in computation process, form according to this network equation group.The rudimentary algorithm that electric pressure bus is analyzed adopts " storehouse " principle, from a node, utilizes a storehouse to deposit the node of medial fascicle, along certain paths, searches for forward, and the node division that Closing Switch is linked together is a bus; From the node of another unallocated female wire size, adopt the search that uses the same method again, until all nodes distribution bus-bar number all.It is consistent that grid is analyzed the process of analyzing with electric pressure bus, just node is changed to bus, and switch is changed to two ends branch road (transformer, circuit etc.).

Summary of the invention

In order to overcome above-mentioned the deficiencies in the prior art, the invention provides a kind of concurrent topology method for the real-time simulation of large scale electric network electromechanical transient, the parallel computation structure of employing based on MPI, topology efficiency is improved, state variation element adopts local admittance battle array to revise, and local factors decomposition technique improves computing velocity, and variation topology is only carried out topology to the electric pressure of book intranetwork element part displacement, contracting hunting zone, raises the efficiency.

In order to realize foregoing invention object, the present invention takes following technical scheme:

The invention provides a kind of concurrent topology method for the real-time simulation of large scale electric network electromechanical transient, said method comprising the steps of:

Step 1: main topological structure Network Based carries out network to be cut apart;

Step 2: by network hierarchy subregion;

Step 3: boundary system topological sum subnet is topology entirely;

Step 4: subnet changes topology;

Step 5: integrated with electromechanical transient real-time simulation;

Step 6: after change in topology, carry out Solution To The Network.

Described step 1 comprises the following steps:

Step 1-1: main topological structure Network Based is divided into subnet and boundary system by network, completes network and cuts apart;

In the main topological structure of described electrical network, all switches are closed, and all elements are all in the state of putting into operation;

Step 1-2: master control process distribution subnet data are to respective subnet process;

Step 1-3: each subnet process receives subnet process data from master control process.

In described step 2, master control process and each subnet process are divided into a plurality of subregions by network, and each subregion is divided into plant stand, AC line and AC line; Plant stand is divided into transformer and electric pressure, and electric pressure is divided into switch/disconnecting link, generator, electric capacity/reactance, physics bus, load and node.

Described step 3 comprises the following steps:

Step 3-1: boundary system deposit data, in master control process, only has node and contact branch road, after boundary system topology completes, forms the corresponding relation of boundary node and border bus, sends to each subnet process;

Step 3-2: each subnet removes border bus from the female wire size list of the main topological structure of network, is each node group distribution bus-bar number in conjunction with on off state, element connection status; If there is not boundary node in node group, in conjunction with the female wire size list of the main topological structure of network distribution bus-bar number; Otherwise this node group is divided into corresponding border bus;

Step 3-3: if run into the female wire size in different borders in node group ergodic process, automatically be modified to the female wire size in the border running into for the first time, and record node and female wire size corresponding relation in makeover process, after the analysis of electric pressure bus finishes, send to the revise the boundary topological relation of system of master control process.

Described step 4 comprises the following steps:

Step 4-1: each subnet process reads the displacement information of this subnet switch;

Step 4-2: according to the electric pressure formation and modification electric pressure list of switch place;

Step 4-3: carry out topology within the scope of variation electric pressure;

Step 4-4: after variation topology finishes, collect by program message passing model M PI the boundary node and the border bus mapping relations that change in each subnet process and arrive master control process;

Step 4-5: upgrade boundary system topology;

Step 4-6: submit to the change information of element connection status to electromechanical transient real-time simulation subprocess, if boundary system interconnection has change in topology, processed the change in topology of boundary system by master control process.

Described step 5 comprises the following steps:

Step 5-1: read master control process topological data and the full topological data of each subnet;

Step 5-2: read master control process network trend, zero sequence data and transient stability data;

Step 5-3: master control process distribution topological data is given each subnet process;

Step 5-4: the temporarily steady computational data of master control process distribution;

Step 5-5: each subnet is topology entirely, sets up female wire size mapping relations, and upgrades the female wire size in transient state real-time simulation;

Step 5-6: transient stability initialization.

Described step 6 comprises the following steps:

Step 6-1: revise network admittance matrix;

Network topology structure situation of change comprises newly-increased branch road or element in parallel, branch road or element in parallel remove from electrical network, branch road or element generation bus in parallel numbering change; Network topology structure change to the modification of network admittance matrix be presented as that matrix increases ranks, matrix ranks are constant and puncture table in corresponding ranks;

Step 6-2: revise the part of factor matrix;

When network topology structure changes, establishing the part that former network admittance matrix A changes is Δ A ₂₂, and in the lower right corner of former network admittance matrix A, be expressed as:

$\tilde{A}=A+\mathrm{\ΔA}=A+\left[\begin{array}{cc}0& 0\\ 0& \mathrm{\Δ}{A}_{22}\end{array}\right]---\left(1\right)$

Wherein,

for network topology structure change after network admittance matrix;

Network admittance matrix after network topology structure is changed

the form of being write as partitioned matrix, has

$\tilde{A}=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& {\tilde{A}}_{22}\end{array}\right]---\left(2\right)$

${\tilde{A}}_{22}=A_{22}+{\mathrm{\ΔA}}_{22}---\left(3\right)$

Wherein, the Factorization of former network admittance matrix A is write as following form according to block form:

$A=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& A_{22}\end{array}\right]=\left[\begin{array}{cc}{L}_{11}& 0\\ L_{21}& L_{22}\end{array}\right]\left[\begin{array}{cc}{D}_{11}& 0\\ 0& D_{22}\end{array}\right]\left[\begin{array}{cc}{U}_{11}& U_{12}\\ 0& U_{22}\end{array}\right]---\left(4\right)$

So

${\tilde{A}}_{22}=L_{21}{D}_{11}{U}_{12}+{\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}---\left(5\right)$

${\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}=L_{22}{D}_{22}{U}_{22}+{\mathrm{\ΔA}}_{22}---\left(6\right)$

Wherein,

for there is the submatrix of localized variation; with

be respectively

triangle submatrix on lower triangle after decomposition, diagonal sum, L ₁₁, L ₂₁, L ₂₂, D ₁₁, D ₂₂, U ₁₁, U ₁₂and U ₂₂be the corresponding part in the factor table matrix of former network admittance matrix A.

Compared with prior art, beneficial effect of the present invention is:

1, adopt the parallel computation structure based on program message passing model M PI, topological efficiency is improved;

2, divide network method and electromechanical transient to divide network method compatible, therefore can be integrated with electromechanical transient program;

3, preferentially carry out the topology of boundary system, can guarantee that the bus numbering of boundary system is minimum, thereby reduce the size of subnet bus numbering;

4, change topology and only the electric pressure of book intranetwork element part displacement is carried out to topology, dwindle hunting zone, raise the efficiency;

5, automatically process the mapping relations between topological structure of electric after the original topological structure of electrical network and concurrent topology;

6, state variation element adopts local admittance battle array to revise, and local factors decomposition technique improves computing velocity;

7, in the time of can processing Three-phase Power Systems difference, cut-off problem.

Accompanying drawing explanation

Fig. 1 is the concurrent topology method flow diagram for the real-time simulation of large scale electric network electromechanical transient;

Fig. 2 is that cut-off switch cannot be determined subnet wiring situation schematic diagram;

Fig. 3 is the interconnection schematic diagram of dividing by interconnection of optimizing in the table method of border;

Fig. 4 is the interconnection schematic diagram of dividing by I side bus of optimizing in the table method of border;

Fig. 5 is the interconnection schematic diagram of dividing by J side bus of optimizing in the table method of border;

Fig. 6 carries out network dividing mode schematic diagram in the main topological structure lower network of network by a certain side bus;

Fig. 7 carries out network dividing mode schematic diagram in the main topological structure lower network of network by interconnection;

Fig. 8 is that network hierarchy divides cell relation schematic diagram;

Fig. 9 is a complete topological process flow diagram of subnet;

Figure 10 is the complete topological interactive relationship figure of boundary system topological sum subnet;

Figure 11 is that subnet changes topological process flow diagram;

Figure 12 is the transient stability program initialization process flow diagram that contains concurrent topology;

Figure 13 is the temporarily steady parallel artificial basic flow sheet containing concurrent topology function.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail.

The Key Term of using in the present invention is explained as follows:

Network topology is in electric system, the basic function of network topology is by the physics connection mode generation of electrical network, to calculate bus model according to power network switch state and network element state, and by there being the bus set of electrical link to turn to electric island, for other electric system advanced applied software such as trend, state estimation, transient emulation lay the foundation.

MPI(Message Passing Interface) being the representative of a kind of standard or standard, is a kind of program message passing model, and becomes the representative of this programming model, and its object is to serve interprocess communication.It has defined the set of one group of normalized function or process, can in multiple programming language, be called, thereby realizes the information interaction between different processes.MPI is a kind of data exchange ways of realizing parallel computation.

Parallel computation refers to use multiple computational resource as the process of a large amount of computational problems of solution such as CPU, storage simultaneously.For carrying out parallel computation, computational resource should comprise computing machine, a high-speed communicating network of being furnished with many CPU, or both are combined with.The fundamental purpose of parallel computation is the large-scale and complicated computational problem of quick solution, improves the speed of calculating, and the storage restricted problem that solves resource.Along with the expansion of electric system scale, some computational problems are running into speed bottle-neck as electromechanical transient, electro-magnetic transient etc. aspect real-time simulation, so by parallel computing strategy, improve computing velocity.

It is for realizing the spatially task division process for mutually not having related subsystem and getting in touch with subsystem by studied system divides of a kind of spatial parallelism computation process that network is cut apart.In electric system, network is cut apart electrical network is divided into and does not mutually have associated sub-electrical network and a small amount of interconnection, thereby can the dispensed of this little electrical network or interconnection is upper to different CPU, carries out parallel computation, accelerates computing velocity.

The general designation of a kind of folding equipment defining in switch electrical network, comprises isolating switch, isolation switch, grounding switch etc.This kind equipment only has two states: disconnect or closure, also claim logic element.

Common point between the equipment that node links together, is used for characterizing the topological relation of the equipment that connects each other and a virtual binding site.Node is static, and the variation of on off state can not affect node.

The set of calculating the some nodes that link together by Closing Switch of bus, also claims topological node.Have one at least with the bypass elements of this set associative, or have a power injection element at least.Bus is dynamic, and the variation of on off state may be that bus changes.

The set that the calculating bus being communicated with in the electrical network of electric island forms.If there is the equipment such as generator or equivalent power supply in electric island, claim this electric Dao Weihuo island.

Electromechanical transient simulation is the electromechanical transient stability of using the method research electric system of time-domain-simulation, i.e. electric system is subject to after large interference, and each synchronous generator keeps synchronous operation and is transitioned into ability new or that return to original steady-state operation mode.Electromechanical transient real-time simulation refers to the simulation process that simulation velocity can strictly be synchronizeed with actual electric network physics dynamic process.

As Fig. 1, in figure, suspension points represents that native system is except process 0, can also configure a plurality of subprocesss of similar process n.In can carrying out n subnet process of concurrent topology analysis, the Processing tasks of process 0, except carrying out general topology, also will be taken into account the Processing tasks of boundary system and the distribute work of data etc. simultaneously, and therefore, process 0 is also referred to as master control process.In parallel computation task, divide timing, process 0 is effectively to calculate the subnet of bus number minimum, its objective is the calculated amount of balance the whole network parallel task.

Concurrent topology method for the real-time simulation of large scale electric network electromechanical transient, comprises the following steps:

Step 1: main topological structure Network Based carries out network to be cut apart;

Step 2: by network hierarchy subregion;

Step 3: boundary system topological sum subnet is topology entirely;

Step 4: subnet changes topology;

Step 5: integrated with electromechanical transient real-time simulation;

Step 6: after change in topology, carry out Solution To The Network.

It is the prerequisite of carrying out subnetting parallel computation that network is cut apart, and is the preliminary work before parallel computation.In the present invention, optimization border table network dividing method is carried out to some alternative use, make it be adapted to concurrent topology algorithm.

In electromechanical transient parallel artificial, the starting point of emulation is trend equilibrium solution, and the diakoptic algorithm therefore in the past using in parallel algorithm, also based on initial trend, cannot be determined the subnet attaching relation of inner each switch of Substation Bus Arrangement.Because in cut-offfing state, or the switch in calculating bus not, cannot determine their ownership for some.As three hollow cut-off switch in the middle of Fig. 2, their both sides belong to respectively in different calculating bus situation, if bus is calculated in both sides, belong to different subnets, cannot determine which subnet these switches belong to.

For avoiding determining the situation of switch subnet ownership, the present invention implements subnetting algorithm on a kind of main topological structure of electrical network without logicality element states such as consideration switches.In the main topological structure of electrical network, all switches are closed, and all elements are all in the state of putting into operation;

Under this electric network composition, the result that execution topology search obtains is the main topological structure of electrical network.Because all switches are all closed, so there will not be the problem that cannot determine switch subnet ownership.But it is pointed out that due to the electrical network of implementing diakoptic algorithm is changed to the main topological model of electrical network by the initial electric network model of trend, may make calculated amount no longer be uniformly distributed and reduce parallel efficiency.

In concurrent topology, minimum to the processing cost of boundary system, this is because all switches have all been divided in subnet.

Fig. 3, Fig. 4 and Fig. 5 are corresponding respectively optimizes the three types that the interconnection in the table method of border is divided, be respectively by interconnection divide, I side bus is divided and J side bus is divided.In these three kinds of dividing mode, only have by interconnection dividing mode and just only branch road (AC line or transformer coil) is divided into master control (process 0).Element on I side and J side bus is all divided into respective subnet.

In Fig. 6 and Fig. 7, provided a kind of result of dividing in the main topological structure lower network of electrical network, they are corresponding to bus and two kinds of modes of carrying out network division by interconnection respectively.Under this network dividing mode, no longer there is switch in boundary system, only remaining branch road and the some boundary nodes of dividing by bus mode of dividing by interconnection mode.Therefore, the topological cost of boundary system is minimum, only takies very little calculated amount.But the problem that this situation is brought is boundary system topology may bring wrong result, but can in subnet topology, revise concurrently.

Step 1 comprises the following steps:

Step 1-1: main topological structure Network Based is divided into subnet and boundary system by network, completes network and cuts apart;

Step 1-2: master control process distribution subnet data are to respective subnet process;

Step 1-3: each subnet process receives subnet process data from master control process.

In described step 2, master control process and each subnet process are divided into a plurality of subregions by network, and each subregion is divided into plant stand, AC line and AC line; Plant stand is divided into transformer and electric pressure, and electric pressure is divided into switch/disconnecting link, generator, electric capacity/reactance, physics bus, load and node.

In Fig. 8, network is comprised of subregion, under subregion, can Further Division be plant stand, AC line and AC line.Inner in plant stand, can further data be given to layering.Electric pressure layer and mechanical floor.Separated by transformer between electric pressure layer, in mechanical floor except transformer, unique certain electric pressures that belongs to all.The minimal processing unit of network topology change is exactly electric pressure.In the present invention, each subnet is all processed according to above-mentioned Organization of Data relation.Therefore, the scope of its topology search is further dwindled.

Fig. 9 has described the complete topological flow process of subnet that in the present invention, step 3 adopts, and this flow process is applicable to the full topology of all-ones subnet.When full topology starts, first according to the main topology of electrical network, form the female wire size of main topology of node.It doesn't matter for the state of the female wire size of main topology and switch and element, in topological process afterwards, can not change.Then consider on off state and by the branch road connection of Closing Switch UNICOM node set together, the female wire size of the calculating of distribution node.In the process of distribution bus-bar number, if the female wire size of main topology of this node is not used, this main bus-bar number of priority allocation, distributes to this node otherwise can search female wire size numbering minimum and the female wire size of non-main topology from bus list.Female wire size forms bus and element associated relation after dividing.Based on this relation, adopt storehouse search technique, carry out system busbar analysis, determine the element state that puts into operation.

In the present invention, the full topology of boundary system topological sum subnet, in different processes, needs the border of Coordination Treatment boundary system, makes the boundary node of identical boundary system in different sub-network, have the female wire size in identical border.Accompanying drawing 10 is boundary system topology and the complete topological interactive relationship figure of subnet, and step 3 specifically comprises the following steps:

Step 3-1: boundary system deposit data, in master control process, only has node and contact branch road, after boundary system topology completes, forms the corresponding relation of boundary node and border bus, sends to each subnet process;

Step 3-2: each subnet removes border bus from the female wire size list of the main topological structure of network, is each node group distribution bus-bar number in conjunction with on off state, element connection status; If there is not boundary node in node group, in conjunction with the female wire size list of the main topological structure of network distribution bus-bar number; Otherwise this node group is divided into corresponding border bus;

Step 3-3: if run into the female wire size in different borders in node group ergodic process, automatically be modified to the female wire size in the border running into for the first time, and record node and female wire size corresponding relation in makeover process, after the analysis of electric pressure bus finishes, send to the revise the boundary topological relation of system of master control process.

Figure 11 has described subnet and has changed topological flow process, and step 4 specifically comprises the following steps:

Step 4-1: each subnet process reads the displacement information of this subnet switch;

Step 4-2: according to the electric pressure formation and modification electric pressure list of switch place;

Step 4-3: carry out topology within the scope of variation electric pressure;

Step 4-4: after variation topology finishes, collect by program message passing model M PI the boundary node and the border bus mapping relations that change in each subnet process and arrive master control process;

Step 4-5: upgrade boundary system topology;

Step 4-6: submit to the change information of element connection status to electromechanical transient real-time simulation subprocess, if boundary system interconnection has change in topology, processed the change in topology of boundary system by master control process.

Electromechanical transient parallel simulation program comes from electromechanical transient serial program, has passed through the development of two more than ten years, has accumulated a large amount of technology and experience.But this program does not have plant stand topological analysis, only have branch road bus numbering, and adopted in a large number modular design, size of code is huge.Parallel network topology program is integrated with it, mainly need to solve female wire size mapping problems that distinct program brings and the hybrid programming interface problem of different programming languages.

Female wire size mapping will realize the mapping of two aspects, comprises from the initial female wire size of dynamo-electric side trend and is mapped to the female wire size of each subnet and is mapped to the initial female wire size of dynamo-electric side trend from the female wire size of subnet.In the present invention, the trend initial topology of electromechanical transient parallel artificial, concurrent topology based on network cut apart topology and all by the serial implementation of parallel network topology, provided, they have identical bus numbering.So can set up the mapping relations of original female wire size and device name when the initialization of concurrent topology subnet, after concurrent topology completes, again set up the mapping relations of original female wire size and the female wire size of subnet topology according to device name.More than setting up mapping relations process can executed in parallel, and only need during initialization, complete, and efficiency is very high.

Aspect concurrent topology and parallel temporarily steady hybrid programming Interface realization, the present invention has designed an interface module, is responsible for realizing the data in the parallel side access of electromechanics concurrent topology side, or calls its function.The function that interface module realizes has:

1) concurrent topology initialization, exits cleaning;

2) communication domain of concurrent topology in MPI environment is set;

3) the full topological analysis function of subnet;

4) subnet changes topological analysis function;

5) topological data read functions;

6) topological data distribution function;

7) various data access function interfaces, comprise internal storage data table, female wire size mapping etc.

Figure 12 is the temporarily steady program initialization process flow diagram that contains concurrent topology.Wherein the function of black overstriking frame is that concurrent topology is all, and other are temporarily steady parallel artificial function.In this accompanying drawing, mainly listed the relation of the full topological functions of subnet of temporarily steady initialization and concurrent topology.

Figure 13 is the temporarily steady parallel artificial basic flow sheet containing concurrent topology function, and the same figure is the same, and black overstriking frame is that concurrent topology is all.Each subnet is responsible for collecting displacement switching information in this subnet, and then topology in related electric pressure in this subnet, if find that there is boundary system change in topology, collects change information and upgrade boundary system topology in master control.The affected element unification of change in topology is recorded, and carries out the renewal of admittance battle array, factor table and revises.Then enter that temporarily steady fault handling, Injection Current solve, the iterative of dynamic element and network etc.

In the present invention, to the displacement information of switch, to there is a preprocessing process, to adapt to switch a, b, the non-problem of simultaneously cut-offfing of c three-phase.Only have the Switch Three-Phase of working as to disconnect simultaneously, or the state disconnecting before corresponding, there is a phase state closed, such switch changed position information is just submitted to concurrent topology and is processed.Otherwise, be processed into the equipment failure of connecting with switch, submit to transient stability program.

Detailed processing logic is as table 1:

Table 1

Step 5 specifically comprises the following steps:

Step 5-1: read master control process topological data and the full topological data of each subnet;

Step 5-2: read master control process network trend, zero sequence data and transient stability data;

Step 5-3: master control process distribution topological data is given each subnet process;

Step 5-4: the temporarily steady computational data of master control process distribution;

Step 5-5: each subnet is topology entirely, sets up female wire size mapping relations, and upgrades the female wire size in transient state real-time simulation;

Step 5-6: transient stability initialization.

Described step 6 comprises the following steps:

Step 6-1: revise network admittance matrix;

Network topology structure situation of change comprises newly-increased branch road or element in parallel (be presented as newly-increased bus numbering or drop into new element between existing bus numbering in network topology bus numbering link), branch road or element in parallel remove (in network topology bus numbering link, being presented as between existing bus is numbered have element to exit or have bus numbering to disappear) from electrical network, (this situation is to cause the division of bus numbering or merge causing because element state changes for branch road or the variation of element generation bus in parallel numbering, as bus connection switch in transformer station disconnects or closure, a kind of special circumstances in above-mentioned 2 kinds of situations),

The generation of above three kinds of buses numbering situation the modification of admittance matrix is presented as to matrix increases ranks, matrix ranks are constant and puncture table in corresponding three kinds of situations of ranks.In actual implementation procedure, for program realizes conveniently, the operation of puncture table ranks can consider with the constant merging of matrix ranks, and just in corresponding ranks, off-diagonal element is put the diagonal element with null value and put with 1 processing, thus the resource overhead of avoiding matrix element to move bringing.Obviously, the input of any element with exit the admittance matrix element value that all needs to revise corresponding ranks, the input of element also may need for matrix increases ranks, the branch road additional method of available admittance matrix is modified.

Step 6-2: revise the part of factor matrix;

When network topology structure changes, establishing the part that former network admittance matrix A changes is Δ A ₂₂, and in the lower right corner of former network admittance matrix A, be expressed as:

$\tilde{A}=A+\mathrm{\ΔA}=A+\left[\begin{array}{cc}0& 0\\ 0& \mathrm{\Δ}{A}_{22}\end{array}\right]---\left(1\right)$

Wherein, for network topology structure change after network admittance matrix;

Network admittance matrix after network topology structure is changed

the form of being write as partitioned matrix, has

$\tilde{A}=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& {\tilde{A}}_{22}\end{array}\right]---\left(2\right)$

${\tilde{A}}_{22}=A_{22}+{\mathrm{\ΔA}}_{22}---\left(3\right)$

Wherein, the Factorization of former network admittance matrix A is write as following form according to block form:

$A=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& A_{22}\end{array}\right]=\left[\begin{array}{cc}{L}_{11}& 0\\ L_{21}& L_{22}\end{array}\right]\left[\begin{array}{cc}{D}_{11}& 0\\ 0& D_{22}\end{array}\right]\left[\begin{array}{cc}{U}_{11}& U_{12}\\ 0& U_{22}\end{array}\right]---\left(4\right)$

So

${\tilde{A}}_{22}=L_{21}{D}_{11}{U}_{12}+{\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}---\left(5\right)$

${\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}=L_{22}{D}_{22}{U}_{22}+{\mathrm{\ΔA}}_{22}---\left(6\right)$

Wherein,

for there is the submatrix of localized variation;

with be respectively triangle submatrix on lower triangle after decomposition, diagonal sum, L ₁₁, L ₂₁, L ₂₂, D ₁₁, D ₂₂, U ₁₁, U ₁₂and U ₂₂be the corresponding part in the factor table matrix of former network admittance matrix A.The main calculated amount that local factors is decomposed concentrates on two parts: multiplication L ₂₂d ₂₂u ₂₂with right

factorization.When

element sum be less than change after admittance battle array all decompose the half because of subnumber of generation, local factors decomposition method is consuming time is less than full decomposition.Therefore, this algorithm calculated amount is stable, and generally can produce higher efficiency.

Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit, although the present invention is had been described in detail with reference to above-described embodiment, those of ordinary skill in the field are to be understood that: still can modify or be equal to replacement the specific embodiment of the present invention, and do not depart from any modification of spirit and scope of the invention or be equal to replacement, it all should be encompassed in the middle of claim scope of the present invention.

Claims (7)

1. for a concurrent topology method for large scale electric network electromechanical transient real-time simulation, it is characterized in that: said method comprising the steps of:

Step 1: main topological structure Network Based carries out network to be cut apart;

Step 2: by network hierarchy subregion;

Step 3: boundary system topological sum subnet is topology entirely;

Step 4: subnet changes topology;

Step 5: integrated with electromechanical transient real-time simulation;

Step 6: after change in topology, carry out Solution To The Network.

2. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 1, is characterized in that: described step 1 comprises the following steps:

Step 1-1: main topological structure Network Based is divided into subnet and boundary system by network, completes network and cuts apart;

In the main topological structure of described electrical network, all switches are closed, and all elements are all in the state of putting into operation;

Step 1-2: master control process distribution subnet data are to respective subnet process;

Step 1-3: each subnet process receives subnet process data from master control process.

3. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 2, it is characterized in that: in described step 2, master control process and each subnet process are divided into a plurality of subregions by network, and each subregion is divided into plant stand, AC line and AC line; Plant stand is divided into transformer and electric pressure, and electric pressure is divided into switch/disconnecting link, generator, electric capacity/reactance, physics bus, load and node.

4. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 1, is characterized in that: described step 3 comprises the following steps:

Step 3-1: boundary system deposit data, in master control process, only has node and contact branch road, after boundary system topology completes, forms the corresponding relation of boundary node and border bus, sends to each subnet process;

Step 3-2: each subnet removes border bus from the female wire size list of the main topological structure of network, is each node group distribution bus-bar number in conjunction with on off state, element connection status; If there is not boundary node in node group, in conjunction with the female wire size list of the main topological structure of network distribution bus-bar number; Otherwise this node group is divided into corresponding border bus;

Step 3-3: if run into the female wire size in different borders in node group ergodic process, automatically be modified to the female wire size in the border running into for the first time, and record node and female wire size corresponding relation in makeover process, after the analysis of electric pressure bus finishes, send to the revise the boundary topological relation of system of master control process.

5. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 1, is characterized in that: described step 4 comprises the following steps:

Step 4-1: each subnet process reads the displacement information of this subnet switch;

Step 4-2: according to the electric pressure formation and modification electric pressure list of switch place;

Step 4-3: carry out topology within the scope of variation electric pressure;

Step 4-4: after variation topology finishes, collect by program message passing model M PI the boundary node and the border bus mapping relations that change in each subnet process and arrive master control process;

Step 4-5: upgrade boundary system topology;

Step 4-6: submit to the change information of element connection status to electromechanical transient real-time simulation subprocess, if boundary system interconnection has change in topology, processed the change in topology of boundary system by master control process.

6. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 1, is characterized in that: described step 5 comprises the following steps:

Step 5-1: read master control process topological data and the full topological data of each subnet;

Step 5-2: read master control process network trend, zero sequence data and transient stability data;

Step 5-3: master control process distribution topological data is given each subnet process;

Step 5-4: the temporarily steady computational data of master control process distribution;

Step 5-5: each subnet is topology entirely, sets up female wire size mapping relations, and upgrades the female wire size in transient state real-time simulation;

Step 5-6: transient stability initialization.

7. the concurrent topology method for the real-time simulation of large scale electric network electromechanical transient according to claim 1, is characterized in that: described step 6 comprises the following steps:

Step 6-1: revise network admittance matrix;

Network topology structure situation of change comprises newly-increased branch road or element in parallel, branch road or element in parallel remove from electrical network, branch road or element generation bus in parallel numbering change; Network topology structure change to the modification of network admittance matrix be presented as that matrix increases ranks, matrix ranks are constant and puncture table in corresponding ranks;

Step 6-2: revise the part of factor matrix;

When network topology structure changes, establishing the part that former network admittance matrix A changes is Δ A ₂₂, and in the lower right corner of former network admittance matrix A, be expressed as:

$\tilde{A}=A+\mathrm{\ΔA}=A+\left[\begin{array}{cc}0& 0\\ 0& \mathrm{\Δ}{A}_{22}\end{array}\right]---\left(1\right)$

Wherein,

for network topology structure change after network admittance matrix;

Network admittance matrix after network topology structure is changed the form of being write as partitioned matrix, has

$\tilde{A}=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& {\tilde{A}}_{22}\end{array}\right]---\left(2\right)$

${\tilde{A}}_{22}=A_{22}+{\mathrm{\ΔA}}_{22}---\left(3\right)$

Wherein, the Factorization of former network admittance matrix A is write as following form according to block form:

$A=\left[\begin{array}{cc}{A}_{11}& A_{12}\\ A_{21}& A_{22}\end{array}\right]=\left[\begin{array}{cc}{L}_{11}& 0\\ L_{21}& L_{22}\end{array}\right]\left[\begin{array}{cc}{D}_{11}& 0\\ 0& D_{22}\end{array}\right]\left[\begin{array}{cc}{U}_{11}& U_{12}\\ 0& U_{22}\end{array}\right]---\left(4\right)$

So

${\tilde{A}}_{22}=L_{21}{D}_{11}{U}_{12}+{\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}---\left(5\right)$

${\tilde{L}}_{22}{\tilde{D}}_{22}{\tilde{U}}_{22}=L_{22}{D}_{22}{U}_{22}+{\mathrm{\ΔA}}_{22}---\left(6\right)$

Wherein,

for there is the submatrix of localized variation;

with

be respectively

triangle submatrix on lower triangle after decomposition, diagonal sum, L ₁₁, L ₂₁, L ₂₂, D ₁₁, D ₂₂, U ₁₁, U ₁₂and U ₂₂be the corresponding part in the factor table matrix of former network admittance matrix A.

Images