CN102904256A - Method and system for rapid self-healing of power grid - Google Patents

Abstract

The invention discloses a method for rapid self-healing of a power grid. The method comprises the following steps of: obtaining operation data of all machine units in the power grid, and establishing mathematic models of all the machine units; determining a starting sequence of non-started machine units and searching for power supply routes from started machine units to the non-started machine units according to the mathematic models; calculating capacities and schedulable loads of the started machine units by using an optimal trend algorithm; and starting the non-started machine units in sequence according to the starting sequence, the power supply routes, and the capacities and the schedulable loads of a generator unit. Correspondingly, the invention further provides a system for the rapid self-healing of the power grid. According to the method and the system, the rapid self-healing of the power grid is realized by considering a plurality of machine units with black-start functions, and the steady trend constraint and the machine unit operation constraint are met.

Description

A kind of method and system of electrical network rapidly self-healing

Technical field

The present invention relates to electric network fault self Healing Technology field in the electric power system, particularly relate to a kind of method of electrical network rapidly self-healing, and a kind of system of electrical network rapidly self-healing.

Background technology

Modern interconnected network scale constantly enlarges, and has improved significantly the runnability of electric power system, has satisfied the develop rapidly of national economy.Simultaneously, because the expansion of power supply scale, the loss that electric network fault causes will be more huge.Electric power system is the most complicated man-made system of generally acknowledging, as a network configuration, local fault might be brought out chain reaction, causes large-area power-cuts, even causes the collapse of whole electrical network.How to prevent large-area power-cuts, realize the pre-control of system's potential risk, the rapid recovery after the fault is the key subjects that modern power systems operation and planning face.

The electrical network Self healing Strategy be structured in the difference that depends on to a great extent each electrical network self structure.To an actual electric network, need to formulate recovery policy according to its specific features.In general, in recovery process, different electrical networks has some and similarly recovers feature, usually the self-healing process can be divided into general three phases: power up stage, rack recover stage and load restoration stage.

High efficiency electrical network Self healing Strategy can reduce system recovery time, improves the reliability of operation of power networks.Research experience shows, recovery capability is directly related after the safe operation ability of electric power system and the fault.Statistics shows: the system that reduces to depend on to a great extent fast of breakdown loss is recovered, and along with the fast development of electrical network scale and the raising of control device, this trend is more obvious.

The scientific and reasonable configuration that starts power supply is the basic guarantee of realizing the electrical network rapidly self-healing.Water power and small fuel oil unit can provide the startup power supply.In addition, some unit that possesses quick extrusion pressure function (fast cut back is hereinafter to be referred as FCB) also can be born the task that power system restoration starts power supply.The unit that possesses black start-up performance of arranging some can significantly improve rapidly self-healing ability under the electric network fault as starting power supply.But the modeling of the parameters such as the control characteristic of present dissimilar unit, unit capacity, climbing time in electrical network self-healing process, the position in network all still lack the model and method of qualitative and qualitative assessment to the impact that the system self-healing scheme generates.Now need the electrical network self-healing scheme of angle research polytype unit during preventing large-area power-cuts from dispatching of power netwoks badly, and consider a plurality of electrical network rapidly self-healing methods of black start-up performance unit after large-area power-cuts that possess.

Summary of the invention

Based on this, the invention provides a kind of method and system of electrical network rapidly self-healing, can consider a plurality of rapidly self-healings that possess the units realization electrical network of black start-up performance, can satisfy steady-state load flow constraint and unit operation constraint simultaneously.

A kind of method of electrical network rapidly self-healing comprises the steps:

Obtain the service data of all units in the electrical network, set up the Mathematical Modeling of described unit;

According to described Mathematical Modeling, determine that the boot sequence sequence and the search that do not start unit have started unit to the described supply path that does not start unit;

Utilize the calculating of optimal load flow algorithm to start exerting oneself of unit and can dispatch load;

According to exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set, start successively the described unit that do not start.

Accordingly, the present invention also provides a kind of system of electrical network rapidly self-healing, comprises model building module, sequence and path determination module, the computing module and start module of exerting oneself and load;

Described model building module is used for obtaining the service data of all units in the electrical network, sets up the Mathematical Modeling of described unit;

Described sequence and path determination module are used for according to described Mathematical Modeling, determine not start the boot sequence sequence of unit and search to have started unit to the described supply path that does not start unit;

The described computing module of exerting oneself and load is used for utilizing the calculating of optimal load flow algorithm to start exerting oneself of unit and can dispatch load;

Described startup module is used for starting successively the described unit that do not start according to the exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set.

The method and system of electrical network rapidly self-healing of the present invention, boot sequence sequence and the supply path of at first definite unit, calculating has started exerting oneself of unit and can dispatch load again, can realize the rapidly self-healing of electrical network.By these steps, generator and circuit in definite generator boot sequence and circuit recovery order, have adopted the optimal load flow algorithm respectively according to sequence starting and recovery after optimizing.Based on efficient optimal load flow algorithm, can realize faster the process that whole Self healing Strategy makes up.The result who finally obtains is one group of starter-generator and recovers the shortest scheduling operation sequence of key duration of load application.

Owing to existing a plurality of black startup power supplys (to comprise Hydropower Unit in the network system, the pumped storage unit, possess the unit of FCB function etc.), utilize the solution of the present invention, in neighborhood search, Path selection, all can support a plurality of each other disjunct electric islands to carry out simultaneously self-healing during optimal load flow calculates, can realize in electrical network self-healing process that therefore a plurality of electric islands carry out rapidly self-healing simultaneously, the present invention is fast automatic, system recovery time is very fast, and the reliability of electrical network self-healing is very high.

Description of drawings

Fig. 1 is the method schematic flow sheet in one embodiment of electrical network rapidly self-healing of the present invention.

Fig. 2 is the schematic diagram of Mathematical Modeling among Fig. 1.

Fig. 3 is the schematic diagram calculation of boot sequence among the step S12 among Fig. 1.

Fig. 4 is the schematic diagram that makes up undirected adjacency matrix among Fig. 1 among the step S13.

Fig. 5 is system's structural representation in one embodiment of electrical network rapidly self-healing of the present invention.

Embodiment

Below in conjunction with embodiment and accompanying drawing the present invention is described in further detail, but embodiments of the present invention are not limited to this.

As shown in Figure 1, be the method schematic flow sheet in one embodiment of electrical network rapidly self-healing of the present invention, comprise the steps:

S11, obtain the service data of all units in the electrical network, set up the Mathematical Modeling of described unit;

In this step, because the machine set type of the connection in the electrical network is more, comprise Hydropower Unit, common fired power generating unit, pumped storage unit, have the unit of FCB ability, the black unit etc. that starts, the various units of model are in the Mathematical Modeling of the unified stable state in electrical network self-healing stage in this step, as shown in Figure 2, it is the schematic diagram of this Mathematical Modeling, the parameter of this Mathematical Modeling is: C is unit capacity, a% is the minimum technology percentage of exerting oneself, R is the startup desired volume (be 0 for the black unit that starts unit and possess the FCB function) of non-black startup unit, t ₀The time (this variable is the output variable of electrical network Self healing Strategy) that non-black startup is accepted external electric energy, recovered station service, unit t ₁-t ₀That non-black startup unit recovers the required time of electrical production (for the black unit that starts unit and possess the FCB function, t ₁=t ₀), k is unit climbing rate.

Based on the physical characteristic of generator, except will providing the generator Common Parameters such as capacity, climbing rate, minimum output, the peculiar parameter of some other generator also needs to provide.Such as, need to provide starting capacity and the maximum crash time of unit for the thermal power generation unit; Start for black, such as the hydraulic turbine, starting capacity and the time of being incorporated into the power networks all are zero.For simplified model, key load is regarded as the generator with negative starting capacity, its climbing rate is zero.Therefore, for a generating set, be t start-up time ₀, condition below needing to satisfy:

t ₀＜ _T2 （1）

t ₀＞T ₃ （2）

t ₁-t ₀＝T ₁ （3）

In the following formula, T ₁Be generator connecting in parallel with system required time, T ₂For generator the latest must recovery time, T ₃For generator allows recovery time the earliest;

When generator is in charging when returning to form, i.e. t ₀＜t＜t ₁, it is exerted oneself and is-R.As t 〉=t ₁, generator begins to increase exerts oneself, and its recruitment of exerting oneself is k * (t-t ₁) (work as t ₀＜t＜t ₁The time this is 0), maximum is no more than C.Generator output curve shown in Figure 3 is the linear superposition of both of these case, can be expressed as:

P(t)＝min{k.max[t-(t ₀+T ₁),0],C}-R.U(t-t ₀) （4）

U (t-t wherein ₀) be the step function of unit, be defined as:

$U\left(X\right)=\left\{\begin{array}{cc}1& X\&GreaterEqual;0\\ 0& X<0\end{array}\right.---\left(5\right).$

S12, according to described Mathematical Modeling, determine that startup does not start the boot sequence sequence of unit and searches and started unit to the described supply path that does not start unit;

When electrical network breaks down, the purpose of electrical network rapidly self-healing is to start as early as possible blackly to start power supply and without the unit of black start-up ability, calculate the black unit that starts to the feasible supply path of non-black startup unit, and supply path is charged;

The optimization aim in this stage is the power up shortest time.This is the multistage optimization problems of classics.In this process, a stage is regarded in the startup of each unit as.In the end a stage, should recover the unit that all satisfy condition.In this process, utilize dynamic programming problems to form all possible recovery sequence tree, wherein mainly comprise two key issues:

Searching can recover the recovery order of unit and critical load.This is an associativity problem, therefore must solve the problem of multiple shot array.For make whole decision tree within the acceptable range, just the generating set when near the generating set the network that restores electricity can be used as next step startup.

Seek the shortest supply path of generating set and key load.At this one-phase, system is in light condition.In order to obtain the generator powered path, utilize the charging current of circuit to analyze whether overvoltage of supply path.In the situation that satisfy condition, select the circuit of minimum charge current, can avoid like this not satisfying the overvoltage constraint.

In a preferred embodiment, this step S12 specifically can comprise:

Steps A, according to described Mathematical Modeling, utilize heuritic approach to make up the described boot sequence sequence that does not start unit;

Utilizing in the present embodiment heuritic approach to determine the boot sequence of generator, such as containing two non-black starter-generator x and y, as shown in Figure 3, is the schematic diagram calculation of both boot sequences, and their black startups require power to be respectively S _xAnd S _y, the climbing rate of two generators is respectively K _xAnd K _y, and satisfy between them:

S _x＞S _y,K _x＜K _y (6)

So generator y should start first than generator x.If but

S _x＞S _y,K _x＞K _y (7)

After need analyzing, the order that starts obtains.In order to recover as early as possible without the black start-up ability unit, when under condition as shown in Figure 3, and work as

t′ _y＜t _x (8)

At this moment, generator x should start first.Formula 3 can be obtained by following derivation, according to Fig. 3 (a) as can be known

$t_x-{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">t</figure-callout>}}_0=\frac{S_x}{k_0+k_y}---\left(9\right)$

k ₀For recovering the average climbing rate of unit.

Can obtain according to Fig. 3 (b):

$t_x^{\&prime;}-{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">t</figure-callout>}}_0=\frac{S_x-S_y}{k_0}--\left(10\right)$

$t_y^{\&prime;}-t_x^{\&prime;}=\frac{S_y}{k_0+k_x}---\left(11\right)$

Replace inequality (8) with formula (9) (10) (11), can obtain:

$\frac{{\mathrm{<figure-callout\; id="0"\; label="S\; y\; k"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">S</figure-callout>}}_y}{k_{}}+\frac{S_x-{\mathrm{<figure-callout\; id="0"\; label="S\; y\; k"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">S</figure-callout>}}_y}{k_{}}<\frac{{\mathrm{<figure-callout\; id="0"\; label="S\; y\; k"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">S</figure-callout>}}_y}{k_{}}---\left(12\right)$

(12) are carried out conversion can be obtained:

$S_x(1+\frac{k_0}{k_x})<S_y(1+\frac{k_0}{k_y})---\left(13\right)$

Formula (13) has represented the priority criterion that starts, if starting power is less and creep speed is larger, this does not start unit and has higher priority.

Step B, utilize connection matrix mapping algorithm search and the not startup unit that starts unit and is connected, definitely started the access path that unit does not extremely start unit;

In this step, the generator that satisfies the breaker operator requirement just can be used as the startup unit in stage.In practical engineering application, in order to guarantee the highly effective and safe of electrical network self-healing, be activated first near the generator that recovers sub-block, then start generating set far away.

The key component of this step is to determine contiguous recovery sub-block.Common direct search is very consuming time, especially for the network system that is on a grand scale.Utilize in the present embodiment the algorithm based on the connection matrix conversion to come detection range to restore electricity sub-block apart from all nodes for D bar branch road.

In a preferred embodiment, this step specifically can comprise the steps:

Step B1, with all units as node, set up the topological structure of described electrical network; Wherein, described topological structure is undirected adjacency matrix;

Step B2, according to described undirected adjacency matrix, make up corresponding transformation matrix;

Step B3, described undirected adjacency matrix and described transformation matrix are multiplied each other, obtain described access path

At first construct one with network system in the adjacency matrix CM of all nodes same dimension the topological structure of whole network is described.

So-called adjacency matrix refers to represent whether there is adjacent relation between the summit with the form of matrix.Adjacency matrix can be divided into two kinds of Digraph adjacent matrix and undirected adjacency matrix.Be used in the present embodiment asking the adjacency matrix of shortest path to belong to undirected adjacency matrix category, so the undirected adjacency matrix of the present embodiment emphasis on analyzing.A figure at this hypothesis G=(V, E), V={v1 wherein, v2 ..., vn}.The adjacency matrix of G then is a n rank square formation.The topological structure of network system as shown in Figure 4 can represent by the undirected adjacency matrix shown in the following formula:

$\mathrm{CM}=\left(\begin{array}{ccccc}0& 1& 1& 1& 0\\ 1& 0& 1& 1& 1\\ 1& 1& 0& 1& 0\\ 1& 1& 1& 0& 1\\ 0& 1& 0& 1& 0\end{array}\right)---\left(14\right)$

Construct one with the transformation matrix TM of dimension, its element value is as follows:

Diagonal element is set to 1, and when i to j connected branch road in the band electric network, the capable j of i row and the capable i column element of j are set to 1;

With CM * TM ^D-1Calculate in the matrix of gained, take out the row with the radio network node place, the non-zero entry in each row is at most by what D bar circuit can be linked and is not with electrical nodes, has namely determined to start the access path that unit does not extremely start unit.

Step C, according to the branch road weight of described boot sequence sequence, described access path and default described access path, utilize Di Jiesitela (Dijkstra) algorithm to determine to have started unit to the described supply path that does not start unit busbar charging electric capacity minimum;

When the calculating generator supply path, need charging current with each bar circuit as the measurement factor.In order to calculate the charging capacitor of supply path, compose with circuit and transformer weight, utilize dijkstra's algorithm to come the minimum supply path of total weight (being total charging capacitor) on the searching route, the method is the graph search algorithm that solves the single supply shortest path.Utilize the method can find the supply path of total charging current minimum, the out-of-limit of overvoltage constraints can be avoided in this path.

Usually the charging current of transformer is less, and the circuit that therefore contains transformer has higher priority.But the transformer increase can cause the transformer fe magnetic resonance.Therefore for fear of this problem, best solution is to consider two factors: the operating time of short-circuiting device and charging current.In this step, the circuit that contains transformer has lower priority.

But classical dijkstra's algorithm can only find the shortest path between two buses, adopt in this step improved dijkstra's algorithm, the algorithm after the improvement can be determined to have recovered to arrive arbitrary non-black supply path that starts the total charging capacitor minimum between power plant's high voltage bus with electric network.

The step of improved dijkstra's algorithm is: all are made as " 0 " with the branch road weight in the electric network, all do not arrange according to the branch road charging capacitor with the branch road weight in the electric network (is the large circuit of line charging electric capacity, when calculating shortest path, its " distance " is just large), transformer has unified weight numerical value " 10 ".Then calculate according to the dijkstra's algorithm of classics that arbitrary node begins from the band electric network, to " the shortest " path of specifying bus, namely obtain from electric network to the path of specifying generator bus, charging capacitor minimum.

S13, utilize the optimal load flow algorithm to calculate to have started exerting oneself and can dispatching load of unit;

The purpose of this step is to determine the strong situation of generator and the load dispatched that guarantees the network system constraint; Utilize the optimal load flow method to find the solution the situation of exerting oneself and the load level of generator, all generator regulated quantity minimums are as target function.Utilize the method, consider generator climbing rate, can within the shortest time, obtain more excellent running status, can specifically comprise the steps:

Step D, with the regulated quantity minimum of all units as target function, utilize the optimal load flow algorithm to calculate described the exerting oneself of unit that started;

In order to satisfy under the constraints, minimize generator start-up time, regard this problem as optimal power flow problems.Generator model as shown in Figure 2, regarding important load as the climbing rate is zero generator model.Utilize least regulating amount to be the optimal load flow model (shown in the formula 15) of target function, the shortest time that in subject matter, can obtain operating.Preferentially adopt in this step interior point method to ask optimal power flow problems.

$\min \underset{i}{\mathrm{\&Sigma;}}{(P_{G.i}-{\mathrm{<figure-callout\; id="0"\; label="p\; G"\; filenames="HDA00002292415700012.png"\; state="\{\{state\}\}">p</figure-callout>}}_G)}^2$

$s.t.\left\{\begin{array}{c}{P}_{G.i}-P_{L.i}-U_i\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{U}_j(G_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}}+B_{\mathrm{ij}}{\sin \mathrm{\&delta;}}_{\mathrm{ij}})=0\\ Q_{G.i}-Q_{L.i}-U_i\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{U}_j(G_{\mathrm{ij}}\sin {\mathrm{\&delta;}}_{\mathrm{ij}}-B_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}})=0\end{array}\right.---\left(15\right)$ $\left(15\right)$

${\stackrel{\&OverBar;}{P}}_{G.i}\&le;P_{G.i}\&le;{\underset{\&OverBar;}{P}}_{G.i}$

${\stackrel{\&OverBar;}{Q}}_{G.i}\&le;Q_{G.i}\&le;{\underset{\&OverBar;}{Q}}_{G.i}$

${\stackrel{\&OverBar;}{U}}_i\&le;U_i\&le;{\underset{\&OverBar;}{U}}_i$

In the formula, P _G.iExert oneself P for the upper generator of bus i _G0.iFor the upper generator of bus i is exerted oneself P at Last status _L.iAnd Q _L.iBe respectively meritorious demand and the reactive requirement of the upper load of bus i, U _iBe the voltage magnitude of bus i, δ _IjBe the phase difference of voltage between bus i and the bus j, G _IjAnd B _IjBe respectively real part and the imaginary part of admittance matrix element.Symbol

The lower limit and the upper limit that represent respectively variable.

Step e, set up load model, utilize the optimal load flow algorithm to calculate the described load dispatched that starts unit; Wherein, described load model comprises continuous meritorious and load or burden without work model, and discrete meritorious and idle model;

In order to guarantee power grid operation in the self-healing process, need sub-load in the recovery system.These loads are used for guaranteeing that system operates in the admissible constraint scope.Since in the recovery process, the unsteadiness of system running state, and the load of recovery will have flexibility.Need to consider two kinds of situations of continuous and discontinuous model of load in this step;

Continuous meritorious and load or burden without work model, this kind load can simply utilize inequality constraints to represent:

Meritorious: P _Max〉=P _t〉=P _Min

Idle: Q _Max〉= _QT 〉=Q _Min

In the formula, P _MaxAnd Q _MaxBe respectively meritorious and the idle upper limit, P _MinAnd Q _MinBe respectively meritorious and idle lower limit, P _tAnd Q _tBe the meritorious and idle size of load when the time t.To this model, load can be set in any point in the restriction range;

Discrete meritorious and idle model, this model is based on a kind of complex model in the actual industrial.The meritorious and idle of load all disperses.At any time, this load can only be adjusted with discrete magnitude.

In actual applications, all operation of power networks constraints all will be satisfied, and major parameter comprises: the constraint of described unit, overvoltage steady state stability, switch temporary overvoltage, voltage stability and capacity of trunk constraint.

Load can be dispatched continuously and classical optimal load flow computational methods can be adopted.But for discontinuous load, need to improve classical optimal load flow.In this step, can be divided into for two steps: the first step, at first utilize optimal load flow to find the solution continuous duty.According to the result of the first step, regulate discrete load at second one, and then call the optimal load flow program.In this step, load remains unchanged, and only adjusts exerting oneself of generator.Owing to need to call optimal load flow program twice, will be to greater than the continuous duty model in the time of calculating discrete load, and need to consider more constraints.Therefore for same network system, algorithm may be sought the scheme of recovering continuous duty.

In the optimal load flow program, only need to comprise and to dispatch load.But to some network system, do not have to dispatch load in the generator powered path, therefore need to determine to guarantee in the starting stage load dispatched of system stability.Utilize the contiguous method of having recovered sub-block of search among the step B, in limited step search, find and to dispatch load.These can be dispatched load and will progressively be resumed.If but system can not to all restoration schedule load power supplies, then should set deletion from decision-making.

S14, according to the exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set, start successively the described unit that do not start;

By above-mentioned steps S11 ~ S13, can finish the structure of the electrical network Self healing Strategy of recovering generator in this step, start successively the described unit that do not start, finish the rapidly self-healing of fault electrical network.

The present invention also provides a kind of system of electrical network rapidly self-healing, as shown in Figure 5, comprises model building module 51, sequence and path determination module 52, the computing module 53 and start module 54 of exerting oneself and load;

Described model building module 51 is used for obtaining the service data of all units in the electrical network, sets up the Mathematical Modeling of described unit; Wherein, the parameter of described Mathematical Modeling comprises: unit capacity, minimum technology are exerted oneself, and the startup desired volume of percentage, non-black startup unit, non-black startup are accepted external electric energy, are recovered the time of station service, non-black startup unit recovers required time, the unit climbing rate of electrical production;

In this module, because the machine set type of the connection in the electrical network is more, comprise Hydropower Unit, common fired power generating unit, pumped storage unit, have the unit of FCB ability, the black unit etc. that starts, the various units of model are in the Mathematical Modeling of the unified stable state in electrical network self-healing stage in this module, as shown in Figure 2, it is the schematic diagram of this Mathematical Modeling, C is unit capacity, a% is the minimum technology percentage of exerting oneself, R is the startup desired volume (be 0 for the black unit that starts unit and possess the FCB function) of non-black startup unit, t ₀The time (this variable is the output variable of electrical network Self healing Strategy) that non-black startup is accepted external electric energy, recovered station service, unit t ₁-t ₀That non-black startup unit recovers the required time of electrical production (for the black unit that starts unit and possess the FCB function, t ₁=t ₀), k is unit climbing rate.

Described sequence and path determination module 52 is used for according to described Mathematical Modeling, determines that startup does not start the boot sequence sequence of unit and searches to have started unit to the described supply path that does not start unit;

When electrical network breaks down, the purpose of electrical network rapidly self-healing is to start as early as possible blackly to start power supply and without the unit of black start-up ability, calculate the black unit that starts to the feasible supply path of non-black startup unit, and supply path is charged;

The optimization aim in this stage is the power up shortest time.This is the multistage optimization problems of classics.In this process, a stage is regarded in the startup of each unit as.In the end a stage, should recover the unit that all satisfy condition.In this process, utilize dynamic programming problems to form all possible recovery sequence tree, wherein mainly comprise two key issues:

Searching can recover the recovery order of unit and critical load.This is an associativity problem, therefore must solve the problem of multiple shot array.For make whole decision tree within the acceptable range, just the generating set when near the generating set the network that restores electricity can be used as next step startup.

Seek the shortest supply path of generating set and key load.At this one-phase, system is in light condition.In order to obtain the generator powered path, utilize the charging current of circuit to analyze whether overvoltage of supply path.In the situation that satisfy condition, select the circuit of minimum charge current, can avoid like this not satisfying the overvoltage constraint.

In a preferred embodiment, this sequence of modules and path determination module 52 specifically comprise structure block, path determination module and minimal path determination module;

Described structure block is used for according to described Mathematical Modeling, utilizes heuritic approach to make up the described boot sequence sequence that does not start unit;

Utilize in the present embodiment heuritic approach to determine the boot sequence of generator, such as containing two non-black starter-generator x and y, formula (13)

Represented the priority criterion that starts, if starting power is less and creep speed is larger, this does not start unit and has higher priority.

Described path determination module is used for utilizing the search of connection matrix mapping algorithm and the not startup unit that starts unit and is connected, has definitely started the access path that unit does not extremely start unit;

In this module, the generator that satisfies the breaker operator requirement just can be used as the startup unit in stage.In practical engineering application, in order to guarantee the highly effective and safe of electrical network self-healing, be activated first near the generator that recovers sub-block, then start generating set far away.

The key component of this module is to determine contiguous recovery sub-block.Common direct search is very consuming time, especially for the network system that is on a grand scale.Utilize in the present embodiment the algorithm based on the connection matrix conversion to come detection range to restore electricity sub-block apart from all nodes for D bar branch road.

In a preferred embodiment, described path module can comprise that specifically topological structure is set up module, matrix makes up module and the module that multiplies each other;

Described topological structure is set up module and is used for all units setting up the topological structure of described electrical network as node; Wherein, described topological structure is undirected adjacency matrix;

Described matrix makes up module and is used for according to described undirected adjacency matrix, makes up corresponding transformation matrix;

The described module that multiplies each other is used for described undirected adjacency matrix and described transformation matrix are multiplied each other, and obtains described access path.

At first construct one with network system in the adjacency matrix CM of all nodes same dimension the topological structure of whole network is described;

So-called adjacency matrix refers to represent whether there is adjacent relation between the summit with the form of matrix.Adjacency matrix can be divided into two kinds of Digraph adjacent matrix and undirected adjacency matrix.Be used in the present embodiment asking the adjacency matrix of shortest path to belong to undirected adjacency matrix category, so the undirected adjacency matrix of the present embodiment emphasis on analyzing.A figure at this hypothesis G=(V, E), V={v1 wherein, v2 ..., vn}.The adjacency matrix of G then is a n rank square formation.

Construct one with the transformation matrix TM of dimension, its element value is as follows:

Diagonal element is set to 1, and when i to j connected branch road in the band electric network, the capable j of i row and the capable i column element of j are set to 1;

With CM * TM ^D-1Calculate in the matrix of gained, take out the row with the radio network node place, the non-zero entry in each row is at most by what D bar circuit can be linked and is not with electrical nodes, has namely determined to start the access path that unit does not extremely start unit.

Described minimal path determination module is used for the branch road weight according to described boot sequence sequence, described access path and default described access path, utilizes the Di Jiesitela algorithm to determine to have started unit to the described supply path that does not start unit busbar charging electric capacity minimum.

When the calculating generator supply path, need charging current with each bar circuit as the measurement factor.In order to calculate the charging capacitor of supply path, compose with circuit and transformer weight, utilize dijkstra's algorithm to come the minimum supply path of total weight (being total charging capacitor) on the searching route, the method is the graph search algorithm that solves the single supply shortest path.Utilize the method can find the supply path of total charging current minimum, the out-of-limit of overvoltage constraints can be avoided in this path.

Usually the charging current of transformer is less, and the circuit that therefore contains transformer has higher priority.But the transformer increase can cause the transformer fe magnetic resonance.Therefore for fear of this problem, best solution is to consider two factors: the operating time of short-circuiting device and charging current.In this module, the circuit that contains transformer has lower priority.

But classical dijkstra's algorithm can only find the shortest path between two buses, adopt improved dijkstra's algorithm in this module, the algorithm after the improvement can be determined to have recovered to arrive arbitrary non-black supply path that starts the total charging capacitor minimum between power plant's high voltage bus with electric network.

The thinking of improved dijkstra's algorithm is: all are made as " 0 " with the branch road weight in the electric network, all do not arrange according to the branch road charging capacitor with the branch road weight in the electric network (is the large circuit of line charging electric capacity, when calculating shortest path, its " distance " is just large), transformer has unified weight numerical value " 10 ".Then calculate according to the dijkstra's algorithm of classics that arbitrary node begins from the band electric network, to " the shortest " path of specifying bus, namely obtain from electric network to the path of specifying generator bus, charging capacitor minimum.

The described computing module 53 of exerting oneself and load is used for utilizing the calculating of optimal load flow algorithm to start exerting oneself of unit and can dispatch load;

The purpose of this module is to determine the strong situation of generator and the load dispatched that guarantees the network system constraint; Utilize the optimal load flow method to find the solution the situation of exerting oneself and the load level of generator, all generator regulated quantity minimums are as target function.Utilize the method, consider generator climbing rate, can obtain more excellent running status within the shortest time, the described computing module 53 of exerting oneself and load specifically can comprise output calculation module and load computing module:

Described output calculation module is used for regulated quantity minimum with all units as target function, utilizes the optimal load flow algorithm to calculate described the exerting oneself of unit that started;

In order to satisfy under the constraints, minimize generator start-up time, regard this problem as optimal power flow problems.Generator model as shown in Figure 2, regarding important load as the climbing rate is zero generator model.Utilize least regulating amount to be the optimal load flow model (shown in the formula 15) of target function, the shortest time that in subject matter, can obtain operating.The preferential interior point method that adopts is asked optimal power flow problems in this module.

Described load computing module is used for setting up load model, utilizes the optimal load flow algorithm to calculate the described load dispatched that starts unit; Wherein, described load model comprises continuous meritorious and load or burden without work model, and discrete meritorious and idle model;

In order to guarantee power grid operation in the self-healing process, need sub-load in the recovery system.These loads are used for guaranteeing that system operates in the admissible constraint scope.Since in the recovery process, the unsteadiness of system running state, and the load of recovery will have flexibility.Need to consider two kinds of situations of continuous and discontinuous model of load in this module;

Continuous meritorious and load or burden without work model, this kind load can simply utilize inequality constraints to represent:

Meritorious: P _Max〉=P _t〉=P _Min

Idle: Q _Max〉=Q _t〉=Q _Min

In the formula, P _MaxAnd Q _MaxBe respectively meritorious and the idle upper limit, P _MinAnd Q _MinBe respectively meritorious and idle lower limit, P _tAnd Q _tBe the meritorious and idle size of load when the time t.To this model, load can be set in any point in the restriction range;

Discrete meritorious and idle model, this model is based on a kind of complex model in the actual industrial.The meritorious and idle of load all disperses.At any time, this load can only be adjusted with discrete magnitude.

In actual applications, all operation of power networks constraints all will be satisfied, and major parameter comprises: the constraint of described unit, overvoltage steady state stability, switch temporary overvoltage, voltage stability and capacity of trunk constraint.

Load can be dispatched continuously and classical optimal load flow computational methods can be adopted.But for discontinuous load, need to improve classical optimal load flow.In this module, can be divided into for two steps: the first step, at first utilize optimal load flow to find the solution continuous duty.According to the result of the first step, regulate discrete load at second one, and then call the optimal load flow program.In this step, load remains unchanged, and only adjusts exerting oneself of generator.Owing to need to call optimal load flow program twice, will be to greater than the continuous duty model in the time of calculating discrete load, and need to consider more constraints.Therefore for same network system, algorithm may be sought the scheme of recovering continuous duty.

In the optimal load flow program, only need to comprise and to dispatch load.But to some network system, do not have to dispatch load in the generator powered path, therefore need to determine to guarantee in the starting stage load dispatched of system stability.Utilize the contiguous method of having recovered sub-block of search in the determination module of path, in limited step search, find and to dispatch load.These can be dispatched load and will progressively be resumed.If but system can not to all restoration schedule load power supplies, then should set deletion from decision-making.

Described startup module 54 is used for starting successively the described unit that do not start according to the exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set;

According to above-mentioned module 51 ~ 53, in this module 54, can finish the structure of the electrical network Self healing Strategy of recovering generator, start successively the described unit that do not start, finish the rapidly self-healing of fault electrical network.

The method and system of electrical network rapidly self-healing of the present invention, boot sequence sequence and the supply path of at first definite unit, calculating has started exerting oneself of unit and can dispatch load again, can realize the rapidly self-healing of electrical network.By these steps, generator and circuit in definite generator boot sequence and circuit recovery order, have adopted the optimal load flow algorithm respectively according to sequence starting and recovery after optimizing.Based on efficient optimal load flow algorithm, can realize faster the process that whole Self healing Strategy makes up.The result who finally obtains is one group of starter-generator and recovers the shortest scheduling operation sequence of key duration of load application.

Owing to existing a plurality of black startup power supplys (to comprise Hydropower Unit in the network system, the pumped storage unit, possess the unit of FCB function etc.), utilize the solution of the present invention, in neighborhood search, Path selection, all can support a plurality of each other disjunct electric islands to carry out while self-healing during optimal load flow calculates, so can realize that in electrical network self-healing process a plurality of electric islands carry out rapidly self-healing simultaneously, the present invention is fast automatic, system recovery time is very fast, and the reliability of electrical network self-healing is very high.

The above embodiment has only expressed several execution mode of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection range of patent of the present invention should be as the criterion with claims.

Claims (10)

1. the method for an electrical network rapidly self-healing is characterized in that, comprises the steps:

Obtain the service data of all units in the electrical network, set up the Mathematical Modeling of described unit;

According to described Mathematical Modeling, determine that the boot sequence sequence and the search that do not start unit have started unit to the described supply path that does not start unit;

Utilize the calculating of optimal load flow algorithm to start exerting oneself of unit and can dispatch load;

According to exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set, start successively the described unit that do not start.

2. the method for electrical network rapidly self-healing according to claim 1, it is characterized in that, described according to described Mathematical Modeling, determine not start the boot sequence sequence of unit and search and started unit to the described step that does not start the supply path of unit and specifically comprise the steps:

According to described Mathematical Modeling, utilize heuritic approach to make up the described boot sequence sequence that does not start unit;

Utilize connection matrix mapping algorithm search and the not startup unit that starts unit and is connected, definitely started the access path that unit does not extremely start unit;

According to the branch road weight of described boot sequence sequence, described access path and default described access path, utilize the Di Jiesitela algorithm to determine to have started unit to the described supply path that does not start unit busbar charging electric capacity minimum.

3. the method for electrical network rapidly self-healing according to claim 2, it is characterized in that, described connection matrix mapping algorithm search and the not startup unit that starts unit and is connected of utilizing definitely started the step that unit extremely do not start the access path of unit and specifically comprised the steps:

All units as node, are set up the topological structure of described electrical network; Wherein, described topological structure is undirected adjacency matrix;

According to described undirected adjacency matrix, make up corresponding transformation matrix;

Described undirected adjacency matrix and described transformation matrix are multiplied each other, obtain described access path.

4. the method for electrical network rapidly self-healing according to claim 1 is characterized in that, describedly utilizes the optimal load flow algorithm to calculate to have started the step exerting oneself and can dispatch load of unit specifically to comprise the steps:

As target function, utilize the optimal load flow algorithm to calculate described the exerting oneself of unit that started the regulated quantity minimum of all units;

Set up load model, utilize the optimal load flow algorithm to calculate the described load dispatched that starts unit; Wherein, described load model comprises continuous meritorious and load or burden without work model, and discrete meritorious and idle model.

5. the method for electrical network rapidly self-healing according to claim 4 is characterized in that,

Described continuous meritorious model and load or burden without work model are: $\left\{\begin{array}{c}{P}_{\max }\&GreaterEqual;P_t\&GreaterEqual;P_{\min }\\ Q_{\max }\&GreaterEqual;Q_t\&GreaterEqual;Q_{\min }\end{array}\right.;$ Wherein, P _MaxAnd Q _MaxBe respectively meritorious and the idle upper limit, P _MinAnd Q _MinBe respectively meritorious and idle lower limit, P _tAnd Q _tBe the described meritorious and idle size of load when the time t of dispatching;

Described discrete parameter meritorious and idle model comprises: the constraint of described unit, overvoltage steady state stability, switch temporary overvoltage, voltage stability and capacity of trunk constraint.

6. the system of an electrical network rapidly self-healing is characterized in that, comprises model building module, sequence and path determination module, the computing module and start module of exerting oneself and load;

Described model building module is used for obtaining the service data of all units in the electrical network, sets up the Mathematical Modeling of described unit;

Described sequence and path determination module are used for according to described Mathematical Modeling, determine not start the boot sequence sequence of unit and search to have started unit to the described supply path that does not start unit;

The described computing module of exerting oneself and load is used for utilizing the calculating of optimal load flow algorithm to start exerting oneself of unit and can dispatch load;

Described startup module is used for starting successively the described unit that do not start according to the exerting oneself and the described load of dispatching of described boot sequence sequence, described supply path, described generating set.

7. the method for electrical network rapidly self-healing according to claim 6 is characterized in that, described sequence and path determination module specifically comprise structure block, path determination module and minimal path determination module:

Described structure block is used for according to described Mathematical Modeling, utilizes heuritic approach to make up the described boot sequence sequence that does not start unit;

Described path determination module is used for utilizing the search of connection matrix mapping algorithm and the not startup unit that starts unit and is connected, has definitely started the access path that unit does not extremely start unit;

Described minimal path determination module is used for the branch road weight according to described boot sequence sequence, described access path and default described access path, utilizes the Di Jiesitela algorithm to determine to have started unit to the described supply path that does not start unit busbar charging electric capacity minimum.

8. the method for electrical network rapidly self-healing according to claim 7 is characterized in that, described path module comprises that specifically topological structure is set up module, matrix makes up module and the module that multiplies each other:

Described topological structure is set up module and is used for all units setting up the topological structure of described electrical network as node; Wherein, described topological structure is undirected adjacency matrix;

Described matrix makes up module and is used for according to described undirected adjacency matrix, makes up corresponding transformation matrix;

The described module that multiplies each other is used for described undirected adjacency matrix and described transformation matrix are multiplied each other, and obtains described access path.

9. the method for electrical network rapidly self-healing according to claim 1 is characterized in that, the described computing module of exerting oneself and load specifically comprises output calculation module and load computing module:

Described output calculation module is used for regulated quantity minimum with all units as target function, utilizes the optimal load flow algorithm to calculate described the exerting oneself of unit that started;

Described load computing module is used for setting up load model, utilizes the optimal load flow algorithm to calculate the described load dispatched that starts unit; Wherein, described load model comprises continuous meritorious and load or burden without work model, and discrete meritorious and idle model.

10. the method for electrical network rapidly self-healing according to claim 9 is characterized in that,

Described continuous meritorious model and load or burden without work model in the described load computing module are:

$\left\{\begin{array}{c}{P}_{\max }\&GreaterEqual;P_t\&GreaterEqual;P_{\min }\\ Q_{\max }\&GreaterEqual;Q_t\&GreaterEqual;Q_{\min }\end{array}\right.;$ Wherein, P _MaxAnd Q _MaxBe respectively meritorious and the idle upper limit, P _MinAnd Q _MinBe respectively meritorious and idle lower limit, P _tAnd Q _tBe the described meritorious and idle size of load when the time t of dispatching;

Described discrete parameter meritorious and idle model comprises: the constraint of described unit, overvoltage steady state stability, switch temporary overvoltage, voltage stability and capacity of trunk constraint.

Images