CN106374453A - Electric power system reconstruction method - Google Patents

Abstract

The invention relates to an electric power system reconstruction method. The electric power system reconstruction method comprises the steps of dividing a to-be-reconstructed system into x sub systems according to a sub-system division rule, wherein x is a positive integer; a backbone channel of the sub systems is recovered by a backbone channel recovery model and a constraint condition of the sub systems; a partial power grid of the sub systems is recovered through an objective function and a constraint condition which are optimized by a partial power grid recovery path of the sub systems; and inter-system networking recovery for the sub systems is realized based on a regional networking recovery function and a constraint condition thereof. According to the electric power system reconstruction method provided by the invention, serial recovery is combined with parallel recovery while coordinate matching of the backbone channel, the partial power grid and the region networking is taken into consideration, so that the electric power system recovery speed is improved.

Description

A kind of power system reconstructing method

Technical field

The present invention relates to field of power system control is and in particular to a kind of power system reconstructing method.

Background technology

After there is massive blackout in system, system recovery procedure last longer, short, need several hours, long then even several Its time, and the recovery emphasis of each time period is different.From time angle, according to system recovery procedure in different time sections Feature recovers the different of object from main, generally whole process is divided into three phases: the black starting-up stage, rack Restoration stage and The load restoration stage.

(1) the black starting-up stage

Generally last for 30～60 points.In this stage, first crash time limit is had to trip respectively from startup power supply The power supply of system provides startup power supply so as to recover generating capacity, is again connected to the grid, and forms the subsystem of isolated operation one by one System.The startup power supply of system can be hydrogenerator, gas turbine generator, retain electromotor in systems after accident (such as Electromotor with itself station service after tripping operation) or off-the-line after isolated subsystem and adjacent system support.Relate in this stage And subject matter have: the startup of unit and operation characteristic, charge the self-excitation causing and excessively electric to nonloaded line and transformator The parallel resonance problem that pressure problem, transformator saturation cause, high capacity motor start, isolate frequency modulation and the pressure regulation problem of mini system Deng.This stage is to quasi-stationary recovery process from electromagnetic transient, electromechanical transient process.

(2) rack Restoration stage

Generally last 3～4 hours.This stage is by by starting the large-scale unit with basic load and putting into main transmission of electricity Circuit progressively recovers the rack of major network, on the one hand strengthens contacting to improve the power supply reliability to station service between power plant, On the other hand some subsystems being carried out side by side, thus setting up a stable rack, being to recover load next stage to beat comprehensively Lower basis.Certainly, for the longer interconnection between some regional systems, can postpone putting into and they being placed on load restoration Carry out later, to avoid stability problem, and reduce the intense strain of dispatcher.In addition, to some to outlying district not The circuit that important load is powered is it is also possible to temporarily need not put into.The subject matter that this stage is related to is to avoid electromotor to inhale That receives idle exceedes its under-excitation ability and a large amount of reactive power flows through voltage produced by nonloaded line and raises.Sometimes for absorption Reactive power produced by line capacitance and the noload over voltage reducing circuit, generally require to put into a number of load.

(3) the load restoration stage

When fired power generating unit has been started up and have certain generating capacity, and set up relatively stable rack with Afterwards, due to system can be supplied to active and idle greatly increase, just can gradually recover load.This stage major problem is that So that system frequency and voltage is maintained within allowed band, and make circuit nonoverload.Load due to fired power generating unit increases Rate of acceleration has certain restriction, and therefore load restoration is limited with maximum factor is that system frequency decline should be too much (if do not surpassed Cross 0.5hz), UFLS action can not be caused.

Content of the invention

The present invention provides a kind of power system reconstructing method, its objective is using build down, parallel recovery algorithm phase knot Close, and count and key passage, partial electric grid and area networking cooperation, improve power system recovery speed.

The purpose of the present invention is to be realized using following technical proposals:

A kind of power system reconstructing method, it thes improvement is that, comprising:

According to system subdivision rule, will treat that reconfiguration system is divided into x subsystem, wherein, x is positive integer；

Using the key routing restoration model of described subsystem and its key passage of the constraints described subsystem of recovery；

Using the partial electric grid restoration path of described subsystem, the object function optimizing and its constraints recover described son The partial electric grid of system；

Carry out networking between system using area networking reconstruction and its constraints sub-system to recover.

Preferably, described subsystem includes: at least one black starting-up power supply, be activated power supply, transmission line of electricity, transformer station and Load bus.

Preferably, described system subdivision rule includes:

The number x of subsystem treats black starting-up power supply number in reconfiguration system less than described；

The difference of the generating set total installation of generating capacity between subsystem or maximum carrying capacity is more than subregion size threshold y；

Each subsystem internal loading with go out dynamic balance；

Interconnection quantity between each subsystem is less than or equal to 4, and the electric pressure of described interconnection is 110～220kv；

Subsystem is centered on black starting-up power supply；

Can be grid-connected between subsystem.

Preferably, the described key routing restoration model using described subsystem and its constraints recover described subsystem Key passage, comprising:

Define rack coverage rate and the rack dispersion ratio of subsystem backbone's passage；

Build key routing restoration model and its constraint bar of subsystem using described rack coverage rate and rack dispersion ratio Part；

Destination node, optimization aim node recovery order are chosen according to the key routing restoration model optimization of described subsystem And restoration path, wherein, the key passage that described destination node is constituted need to meet described constraints.

Further, the described rack coverage rate defining subsystem backbone's passage and rack dispersion ratio include:

(1) defines rack coverage rate c of the trunk rack k of subsystem backbone's passage as the following formula_kAnd node j is to trunk rack Shortest path minima d of each node in k_j-k:

$\left\{\begin{array}{c}{c}_k=\underset{i\&element;{\mathrm{\ω}}_k}{\σ}{\mathrm{\α}}_i+\underset{j\&element;{\mathrm{\ω}}_{\stackrel{\&overbar;}{k}}}{\σ}\frac{1}{2^{2_{j-k}}}{\mathrm{\α}}_j\\ d_{j-k}=\min \left(d_{i-i}\right),i\&element;{\mathrm{\ω}}_k\end{array}\right.---\left(1\right)$

In formula (1), ω_kFor the node set in the trunk rack k of described subsystem,Backbone network for described subsystem Node set outside frame k, α_iThe comprehensive importance degree of node i, α_jComprehensive importance degree for node j, d_j-iFor node j to node i Distance, that is, node j to node i shortest path pass through circuit number；

As the following formula (2) define subsystem backbone passage rack dispersion ratio d and recover i-th destination node when network from Divergence d_i:

$\left\{\begin{array}{cc}d=\max \left(d_i\right)& i=1,2,\mathrm{...},n_k\\ d_i=\frac{\max \left(d_{j-k}\right)}{n_{\stackrel{\&overbar;}{g}}}& j\&element;{\mathrm{\ω}}_{ig},k\&element;{\mathrm{\ω}}_{i\stackrel{\&overbar;}{g}}\end{array}\right.---\left(2\right)$

In formula (2), n_kFor destination node sum, ω_igFor recovering to have recovered the electricity in trunk rack during i-th destination node Source node set,For recovering to have recovered the non-power node set in trunk rack during i-th destination node,For recovering Non-power destination node number.

Further, the described key routing restoration mould building subsystem using described rack coverage rate and rack dispersion ratio Type and its constraints, comprising:

The key routing restoration model of (3) structure subsystem as the following formula:

$\begin{array}{cc}\max & f=\frac{{\mathrm{\τc}}_k-\underset{i\&element;l_k}{\σ}{\mathrm{\ω}}_i}{td}\end{array}---\left(3\right)$

In formula (3), f is the key routing restoration efficiency function of subsystem, and τ is the trunk rack coverage rate power of subsystem Weight, c_kFor the rack coverage rate of the trunk rack k of subsystem, l_kFor recovering the line set that trunk rack comprises, ω_iIt is extensive The weight on multiple line road, t is system recovery required time；

(4) determine that the trend of subsystem constrains as the following formula:

$\left\{\begin{array}{cc}{p}_{gi}^{\min }\≤p_{gi}\≤p_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ q_{gi}^{\min }\≤q_{gi}\≤q_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ v_j^{\min }\≤v_j\≤v_j^{\max }& j=1,2,\mathrm{...},n_k\\ i_k\≤i_k^{\max }& k=1,2,\mathrm{...},n_{ki}\end{array}\right.---\left(4\right)$

In formula (4), n_gFor having recurred motor nodes, n in subsystem_kCount for knot cluster in subsystem, n_kiFor subsystem Multiple line way in system,For the active minima of exerting oneself of i-th electromotor node in subsystem,For in subsystem i-th The active maximum of exerting oneself of individual electromotor node, p_giExert oneself for the active of i-th electromotor node in subsystem,For subsystem The idle minima of exerting oneself of i-th electromotor node in system,Exert oneself for the idle of i-th electromotor node in subsystem Big value, q_giExert oneself for the idle of i-th electromotor node in subsystem,Minimum for the voltage of j-th node in subsystem Value, v_jFor the voltage of j-th node in subsystem,For the voltage max of j-th node in subsystem, i_kFor in subsystem The electric current of k-th circuit,Current maxima for k-th circuit in subsystem；

As the following formula (5) determine subsystem non-power nodal parallel recover constraint function:

$l^{*}\≤\sqrt{(1-{f_{\min }^{*}}^2)r_{\σ}^{*}{t}_j}---\left(5\right)$

In formula (5), l^*For the total load of subsystem parallel recovery node,Allow the low-limit frequency occurring for subsystem, (6) determine as the following formulaAnd t_j:

$\left\{\begin{array}{c}{r}_{\σ}^{*}=\underset{i=1}{\overset{n_g}{\σ}}{r}_{gi}^{*}\\ t_j=\underset{i=1}{\overset{n_g}{\σ}}{s}_{ni}^{*}{t}_{ji}\end{array}\right.---\left(6\right)$

In formula (6),For the climbing rate of i-th electromotor in subsystem, t_jiInertia for i-th electromotor in subsystem Time constant,Rated capacity for i-th electromotor in subsystem.

Further, the described key routing restoration model optimization according to described subsystem chooses destination node, comprising:

(7) determine the Reduce function of the key routing restoration model of described subsystem as the following formula:

$\begin{array}{cc}\max & f^{\′}=\frac{c_k^{\′}}{t^{\′}}\end{array}---\left(7\right)$

In formula (7), f ' is the key routing restoration efficiency function abbreviation value of subsystem, and t ' is that the key passage of subsystem is extensive Recovery required time after multiple efficiency function abbreviation, c '_kFor backbone network after the key routing restoration efficiency function abbreviation of subsystem The rack coverage rate of frame k；

Wherein, as the following formula (8) determine the recovery required time t ' after the key routing restoration efficiency function abbreviation of subsystem:

T '=∑ t '_ii∈ω_d(8)

In formula (8), t '_iFor i-th section in destination node set after the key routing restoration efficiency function abbreviation of subsystem Point recovers required time, ω_dFor destination node set；

(9) determine the rack coverage rate of trunk rack k after the key routing restoration efficiency function abbreviation of subsystem as the following formula c′_k:

c_k'=∑ α '_jj∈ω_d(9)

In formula (9), α '_jFor j-th section in destination node set after the key routing restoration efficiency function abbreviation of subsystem The comprehensive importance degree of point；

Further, as the following formula (10) determine destination node set after the key routing restoration efficiency function abbreviation of subsystem In i-th node recover required time t ':

$t_i^{\′}\≈\frac{1}{2}(\underset{j\&element;{\mathrm{\φ}}_{\min 1}^i}{\σ}{t}_j^l+\underset{j\&element;{\mathrm{\φ}}_{\min 2}^i}{\σ}{t}_j^l)---\left(10\right)$

In formula (10),For the set of minimal paths in other destination node paths for the node i,Arrive it for node i Second shortest path set in his destination node path,Recovery time needed for circuit j；

The step of optimum option destination node specifically includes:

A. set all nodes of subsystem and be destination node, obtain functional value by described formula (7)；

B. the minimum node of importance degree in described destination node is classified as non-targeted node；

C. the comprehensive importance degree of non-targeted node is shifted, wherein, if the comprehensive importance degree α of subsystem interior joint i_i Minimum, then make described node i be non-targeted node, change its comprehensive importance degree α '_i=0, by described comprehensive importance degree α_iIt is transferred to The nearest destination node j away from node i, the then comprehensive importance degree α of (11) switch target node j as the following formula_j:

$\begin{array}{cc}{\mathrm{\α}}_j^{\′}={\mathrm{\α}}_j+{\mathrm{\α}}_i/n_d^i{2}^{d_{i-d}}& j\&element;{\mathrm{\ω}}_d^i\end{array}---\left(11\right)$

In formula (11), α '_jFor conversion after destination node j comprehensive importance degree,It is nearest away from i-th non-targeted node Node set, d_i-dFor node i extremelyDistance,ForMiddle destination node number；

D. press described formula (7) and obtain functional value, if this functional value is bigger than functional value in described step a, return to step b, If this functional value is less than functional value in described step a, go to step e；

E. in described step d, whether functional value is that continuous second wheel reduces, and if so, then goes to step f, if it is not, then preserving Current target node set x, goes to step b；

F. export current target node set；

Wherein, the destination node collection that described step f obtains is combined into optimal objective node set.

Further, the described key routing restoration model optimization destination node recovery order according to described subsystem, bag Include:

Obtain the optimum recovery order of destination node using Crossover Particle Swarm Optimization；

Wherein, in described Crossover Particle Swarm Optimization implementation procedure, destination node is pressed its black starting-up in subsystem The distance of power supply is divided into two classes, and by initialization principle, destination node is initialized, and described initialization principle includes: a. belongs to In inhomogeneous destination node, its sequencing is constant；B. can be randomly ordered with apoplexy due to endogenous wind destination node；

In described Crossover Particle Swarm Optimization implementation procedure, object function is the key routing restoration mould of described subsystem Type.

Further, the described key routing restoration model optimization destination node restoration path according to described subsystem, bag Include:

A. the recovery order according to described destination node, recovers to destination node one by one；

B. judge whether current target node is non-electrical source node, be to go to step c；Otherwise go to step f；

Whether the subsequent node c. judging current target node is non-electrical source node, if so, then determines whether that this is follow-up Whether node can be recovered with described current target node simultaneously, if then going to step d；If otherwise going to step e；

D. obtain the restoration path of destination node according to the recovery order of described destination node, and be possible to recover simultaneously Circuit is set to recover circuit simultaneously, goes to step g；

E. obtain the restoration path of destination node according to the recovery order of described destination node, go to step g；

F. obtain the restoration path of destination node according to the recovery order of described destination node, if this node cannot start Recover in time-constrain, then make this pitch point importance be 0；

G. judge whether that all destination nodes all recover, if so, then algorithm terminates；Otherwise, according to described destination node Recovery order recovery next node, go to step b.

Preferably, the described partial electric grid restoration path using described subsystem optimizes object function and its constraints Recover the partial electric grid of described subsystem, comprising:

Define the main electrical path length of network of partial electric grid, average path length and the dynamic discrete rate of subsystem；

Build object function and its constraints that the partial electric grid restoration path of described subsystem optimizes；

The partial electric grid of the object function recovery subsystem that the partial electric grid restoration path according to described subsystem optimizes.

Further, described define the main electrical path length of network of partial electric grid of subsystem, average path length and dynamic from Scattered rate, comprising:

Main electrical path length t of network of the partial electric grid of (12) definition subsystem as the following formula:

$t=\underset{i<j}{\max }{d}_{ij}---\left(12\right)$

In formula (12), d_ijFor the beeline between partial electric grid interior joint i and node j, represented with the number of lines；

(13) define the average path length of the partial electric grid of subsystem: f as the following formula

$f=\frac{2}{n(n+1)}\underset{i<j}{\&sigma;}{d}_{ij}---\left(13\right)$

In formula (13), n is the quantity of partial electric grid interior joint；

(14) define dynamic discrete rate e of the partial electric grid of subsystem as the following formula_c:

$e_c=\frac{t_{new}}{t_{\max }}\&centerdot;\frac{f_{new}}{f_{\max }}---\left(14\right)$

In formula (14), t_newAnd f_newIt is respectively the main electrical path length of network and the net being formed new rack after subsystem changes Network average path length, t_maxThe main footpath of network for whole system before having a power failure on a large scale；f_maxOccurred for during system reconfiguration Macroreticular average path length.

Further, the described partial electric grid restoration path building described subsystem optimizes object function and its constraint bar Part, comprising:

(15) determine the comprehensive weight m of restoration path in partial electric grid as the following formula_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c---\left(15\right)$

In formula (15), μ is the number of times in restoration path through transformator, c_ijFor the circuit under conversion to same electric pressure Charging capacitor, c_maxFor the maximum of single line charging capacitor in system, ξ₁=1 or 0, represent network dispersion consideration with No, e_cFor the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1 and two addend item dimensions are all 1；

The object function that the partial electric grid restoration path of the described subsystem of (16) structure optimizes as the following formula:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(16\right)$

In formula (16), f is partial electric grid path recovery and optimization function, and β is the importance degree sum of each node in restoration path；

As the following formula (17) determine described subsystem partial electric grid restoration path optimize bound for objective function:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(17\right)$

In formula (17), g is to have recovered power supply node, and l is the set of load bus, the burden with power that pj recovers for node Amount, q_jThe load or burden without work amount recovered for node, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Active for the minimum of node Injecting power, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

Further, the described partial electric grid restoration path according to described subsystem optimizes object function and its constraint bar The partial electric grid of part recovery subsystem, comprising:

A. obtain the direct-to-ground capacitance c of amount, loading and every circuit of partial electric grid to be restored_ij ^*, with extensive The key passage of multiple subsystem is the starting point of searching route, and in subsystem, remainder is not recover rack in area, and defines Initial value k=1；

B. recovered rack and be equivalent to a node o_k, with o_kCentered on to obtain radius be all to belong to inextensive in area in r The power supply node of multiple rack and load bus, described power supply node and load bus are put into destination node collection i_k；

C. the weights making circuit in path are direct-to-ground capacitance, by i_kMiddle destination node is to Centroid o_kShortest path put Enter shortest path collection s_k；

D. determine s respectively_kIn each paths the object function that optimizes of partial electric grid restoration path, and delete and be unsatisfactory for about The path of bundle condition；

E. from s_kThe path l of the target function value minimum that the middle partial electric grid restoration path selecting subsystem optimizes, is incorporated to Send a telegram in reply in network；

F. trend verification is carried out to network of sending a telegram in reply, if trend convergence, go to step g；Trend does not restrain, in s_kSend a telegram in reply Delete path l in network, go to step e；

If g. load restoration rate r of this partial electric grid_k＞ 90%, algorithm stops；Otherwise go to step b.

Preferably, described carry out networking between system using area networking reconstruction and its constraints sub-system extensive It is multiple, comprising:

Build area networking Restoration model and its constraints；

Carry out networking between system using area networking reconstruction and its constraints sub-system to recover.

Further, described structure area networking reconstruction and its constraints, comprising:

(18) determine the comprehensive weight m of restoration path between subsystem in area networking as the following formula_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c+{\mathrm{\&xi;}}_2(1-\mathrm{\&delta;}d)---\left(18\right)$

In formula (18), μ is the number of times in restoration path through transformator, c_ijFor the circuit under conversion to same electric pressure Charging capacitor, c_maxFor the maximum of single line charging capacitor in system, δ d is subregion range-sensitivity, ξ₁=1 or 0, table Whether show the consideration of network dispersion, e_cFor the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1,0 ＜ δ d≤1, and Three addend item dimensions are all 1；

(19) structure area networking reconstruction as the following formula:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(19\right)$

In formula (19), f is area networking reconstruction, and β is the importance degree sum of each node in restoration path；

(20) determine the constraints of described area networking reconstruction as the following formula:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(20\right)$

In formula (20), g is to have recovered power supply node, and l is the set of load bus, p_jThe burden with power recovering for node Amount, q_jThe load or burden without work amount recovered for node, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Active for the minimum of node Injecting power, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

Further, described carry out networking between system using area networking reconstruction and its constraints sub-system extensive It is multiple, comprising:

A. obtain the direct-to-ground capacitance c of the amount, loading and every circuit in n subsystem_ij ^*, with each subsystem Black starting-up power supply is to have recovered rack in area, and as the starting point of searching route, remainder is not recover rack in area, and initially Change k=1；Initialize the ξ in each subsystem₁=1, ξ₂=0；

B. it is equivalent to a node o by having recovered rack in the area of subsystem k_kIf, ξ₂=0, then with o_kCentered on obtain half Footpath is r_kInterior all belong to the power supply node not recovering rack in area and load bus, and by described power supply node and load bus Put into destination node collection i_kIf, ξ₂=1, then with o_kCentered on obtain radius be r_kInterior all still unrecovered power supply nodes and negative Lotus node, and described power supply node and load bus are put into destination node collection i_k；

C. the weights of circuit in direct-to-ground capacitance as path, call dijkstra algorithm to obtain i_kIn each destination node Short path, and described shortest path is put into shortest path collection s_k；

If d. ξ₁=1, then obtain the main footpath t of subsystem k network_kWith load restoration rate r_1kIf, t_k＞ 2/3t_max, then by ξ₁Put 0, if r_1k＞ 50%, then by ξ₂Put 1；

E. determine s respectively_kIn each path area networking reconstruction value, and delete the path being unsatisfactory for constraints；

F. from s_kThe minimum path l of middle selection region networking reconstruction value, is incorporated in subsystem k；

G. sub-system k carries out trend verification, if trend convergence, goes to step h；Trend does not restrain, in s_kIn subsystem k Delete path l, go to step f；

H. whether the subsystem k and adjacent subsystems m difference on the frequency in point both sides arranged side by side, voltage amplitude value difference and phase angle difference are verified Meeting and require side by side, if meeting, subsystem k and m being merged, n=n-1, k=1；If being unsatisfactory for, k=(k+1) %n；

If i. load restoration rate r of subsystem k_k＞ 70%, and subsystem number n=1, algorithm stops；Otherwise go to step b.

Beneficial effects of the present invention:

A kind of power system reconstructing method that the present invention provides, based on the black starting-up number of power sources being resumed in system, really Surely can be with the subsystem of parallel recovery；In conjunction with the indexs such as the network coverage, network dispersion ratio and recovery time, optimization aim node After selection, recovery order and restoration path, done subsystem internal skeleton dry passage recovers；After rack coverage rate reaches condition, in conjunction with Network dynamic dispersion ratio index and substep optimizing algorithm, after load restoration rate reaches condition, done subsystem partial electric grid is extensive Multiple；Consider the subregion range index between subsystem, in conjunction with the resume speed of each subsystem, after reaching condition arranged side by side, complete region connection The recovery of net, reconstructing method is applied to the system reconfiguration problem after large-scale electrical power system occurrence of large-area has a power failure, and the present invention carries For technical scheme propose the significance level dynamically considering system interior joint first, application dynamic discrete rate index weighs rack Optimization progress, also contemplates the matching problem of restructuring procedure between each subsystem in addition.The meter that proposes first and key passage, locally The power system reconstructing method of electrical network and area networking cooperation, is conducive to being rapidly completed key passage, partial electric grid and area The recovery of domain networking, powers to power system restoration significant.

Brief description

Fig. 1 is a kind of flow chart of present invention power system reconstructing method.

Specific embodiment

Below in conjunction with the accompanying drawings the specific embodiment of the present invention is elaborated.

Purpose, technical scheme and advantage for making the embodiment of the present invention are clearer, below in conjunction with the embodiment of the present invention In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described it is clear that described embodiment is The a part of embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art The all other embodiment being obtained under the premise of not making creative work, broadly falls into the scope of protection of the invention.

A kind of power system reconstructing method that the present invention provides, needs the background recovered as early as possible based on system after having a power failure on a large scale, Combined using build down, parallel recovery algorithm, and count and key passage, partial electric grid and area networking cooperation, To reach the purpose of recovery system as early as possible.

Generally, build down refers to after system occurs brown-outs accident or full cut-off accident, using in power supply interrupted district The black starting-up power supply in portion or outside support, are charged to core network and possible secondary primary network station, first recover major network, more extensive Multiple little net, the unit for non self starting provides startup power supply, is layered afterwards under conditions of keeping power supply and balancing the load Secondary progressively recovery system normally runs.

After parallel recovery refers to large-scale blackout, original system is decomposed into several and independent there is black start-up ability Subsystem recovers respectively, afterwards by each subsystem between interconnection and main grid structure reconstruction, progressively expand electrical network scale and reinforcement Network structure, until comprehensive recovery system normally runs.

The technical scheme that the present invention provides, as shown in figure 1, comprise the steps:

101., according to system subdivision rule, will treat that reconfiguration system is divided into x subsystem, and wherein, x is positive integer；

Layering and zoning management is had been carried out at present, the division of each regional power grid is more perfect in domestic actual electric network.Region electricity Net is internal to be typically equipped with black starting-up power supply, and the contact point between regional power grid is also very clear and definite with the setting of interconnection.Therefore, On the basis of meeting above-mentioned division principle, the division of China's electrical network subsystem is more ripe.

The 102. key routing restoration models utilizing described subsystem and its backbone of the constraints described subsystem of recovery Passage；

103. object functions of partial electric grid restoration path optimization utilizing described subsystem and its constraints recover institute State the partial electric grid of subsystem；

104. carry out networking between system using area networking reconstruction and its constraints sub-system recovers.

Wherein, described subsystem includes: at least one black starting-up power supply, is activated power supply, transmission line of electricity, transformer station and negative Lotus node.

Described system subdivision rule includes:

The number x of subsystem treats black starting-up power supply number in reconfiguration system less than described；

The difference of the generating set total installation of generating capacity between subsystem or maximum carrying capacity is more than subregion size threshold y；

Each subsystem internal loading with go out dynamic balance；

Interconnection quantity between each subsystem is less than or equal to 4, and the electric pressure of described interconnection is 110～220kv；

Subsystem is centered on black starting-up power supply；

Can be grid-connected between subsystem.

Specifically, described step 102, comprising:

Define rack coverage rate and the rack dispersion ratio of subsystem backbone's passage；

Build key routing restoration model and its constraint bar of subsystem using described rack coverage rate and rack dispersion ratio Part；

Destination node, optimization aim node recovery order are chosen according to the key routing restoration model optimization of described subsystem And restoration path, wherein, the key passage that described destination node is constituted need to meet described constraints.

Further, the described rack coverage rate defining subsystem backbone's passage and rack dispersion ratio include:

(1) defines rack coverage rate c of the trunk rack k of subsystem backbone's passage as the following formula_kAnd node j is to trunk rack Shortest path minima d of each node in k_j-k:

$\left\{\begin{array}{c}{c}_k=\underset{i\&element;{\mathrm{\&omega;}}_k}{\&sigma;}{\mathrm{\&alpha;}}_i+\underset{j\&element;{\mathrm{\&omega;}}_{\stackrel{\&overbar;}{k}}}{\&sigma;}\frac{1}{2^{d_{j-k}}}{\mathrm{\&alpha;}}_j\\ d_{j-k}=\min \left(d_{j-i}\right),i\&element;{\mathrm{\&omega;}}_k\end{array}\right.---\left(1\right)$

In formula (1), ω_kFor the node set in the trunk rack k of described subsystem,Backbone network for described subsystem Node set outside frame k, α_iThe comprehensive importance degree of node i, α_jComprehensive importance degree for node j, d_j-iFor node j to node i Distance, that is, node j to node i shortest path pass through circuit number；

As the following formula (2) define subsystem backbone passage rack dispersion ratio d and recover i-th destination node when network from Divergence d_i:

$\left\{\begin{array}{cc}d=\max \left(d_i\right)& i=1,2,\mathrm{...},n_k\\ d_i=\frac{\max \left(d_{j-k}\right)}{n_{\stackrel{\&overbar;}{g}}}& j\&element;{\mathrm{\&omega;}}_{ig},k\&element;{\mathrm{\&omega;}}_{i\stackrel{\&overbar;}{g}}\end{array}\right.---\left(2\right)$

In formula (2), n_kFor destination node sum, ω_igFor recovering to have recovered the electricity in trunk rack during i-th destination node Source node set,For recovering to have recovered the non-power node set in trunk rack during i-th destination node,For recovering Non-power destination node number.

The described key routing restoration model and its about building subsystem using described rack coverage rate and rack dispersion ratio Bundle condition, comprising:

The key routing restoration model of (3) structure subsystem as the following formula:

$\begin{array}{cc}\max & f=\frac{{\mathrm{\&tau;c}}_k-\underset{i\&element;l_k}{\&sigma;}{\mathrm{\&omega;}}_i}{td}\end{array}---\left(3\right)$

In formula (3), f is the key routing restoration efficiency function of subsystem, and τ is the trunk rack coverage rate power of subsystem Weight, c_kFor the rack coverage rate of the trunk rack k of subsystem, l_kFor recovering the line set that trunk rack comprises, ω_iIt is extensive The weight on multiple line road, t is system recovery required time；

(4) determine that the trend of subsystem constrains as the following formula:

$\left\{\begin{array}{cc}{p}_{gi}^{\min }\&le;p_{gi}\&le;p_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ q_{gi}^{\min }\&le;q_{gi}\&le;q_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ v_j^{\min }\&le;v_j\&le;v_j^{\max }& j=1,2,\mathrm{...},n_k\\ i_k\&le;i_k^{\max }& k=1,2,\mathrm{...},n_{ki}\end{array}\right.---\left(4\right)$

In formula (4), n_gFor having recurred motor nodes, n in subsystem_kCount for knot cluster in subsystem, n_klFor subsystem Multiple line way in system,For the active minima of exerting oneself of i-th electromotor node in subsystem,For in subsystem i-th The active maximum of exerting oneself of individual electromotor node, p_giExert oneself for the active of i-th electromotor node in subsystem,For subsystem The idle minima of exerting oneself of i-th electromotor node in system,Exert oneself for the idle of i-th electromotor node in subsystem Big value, q_giExert oneself for the idle of i-th electromotor node in subsystem,Minimum for the voltage of j-th node in subsystem Value, v_jFor the voltage of j-th node in subsystem,For the voltage max of j-th node in subsystem, i_kFor in subsystem The electric current of k-th circuit,Current maxima for k-th circuit in subsystem；

As the following formula (5) determine subsystem non-power nodal parallel recover constraint function:

$l^{*}\&le;\sqrt{(1-{f_{\min }^{*}}^2)r_{\&sigma;}^{*}{t}_j}---\left(5\right)$

In formula (5), l^*For the total load of subsystem parallel recovery node,Allow the low-limit frequency occurring for subsystem, (6) determine as the following formulaAnd t_j:

$\left\{\begin{array}{c}{r}_{\&sigma;}^{*}=\underset{i=1}{\overset{n_g}{\&sigma;}}{r}_{gi}^{*}\\ t_j=\underset{i=1}{\overset{n_g}{\&sigma;}}{s}_{ni}^{*}{t}_{ji}\end{array}\right.---\left(6\right)$

In formula (6),For the climbing rate of i-th electromotor in subsystem, t_jiInertia for i-th electromotor in subsystem Time constant,Rated capacity for i-th electromotor in subsystem.

The described key routing restoration model optimization according to described subsystem chooses destination node, comprising:

Because rack coverage rate c in the key routing restoration model of formula (3)_kWith circuit weights omega only with bulk transmission grid End form state is relevant；During the final scale of network reconfiguration used time t and bulk transmission grid and reconstruct, nodal parallel recovery rate is all relevant System；Rack dispersion ratio d is only relevant with the recovery order of bulk transmission grid building process interior joint.So, this stage can temporarily be ignored The impact of rack dispersion ratio, simultaneously takes account of that in formula (3), rack coverage rate weight τ value is larger, and circuit weights omega is to target letter The impact very little of number, so object function can temporarily simplify in this stage, (7) determine the key passage of described subsystem as the following formula The Reduce function of Restoration model:

$\begin{array}{cc}\max & f^{\&prime;}=\frac{c_k^{\&prime;}}{t^{\&prime;}}\end{array}---\left(7\right)$

In formula (7), f ' is the key routing restoration efficiency function abbreviation value of subsystem, and t ' is that the key passage of subsystem is extensive Recovery required time after multiple efficiency function abbreviation, c '_kFor backbone network after the key routing restoration efficiency function abbreviation of subsystem The rack coverage rate of frame k；

Wherein, because the impact that the synchronization of non-electrical source node recovers to reduce for system recovery time is linear, and For in the optimization of object function shown in formula (7), the linear minimizing of time t can be ignored, therefore, (8) determine subsystem as the following formula Recovery required time t ' after the key routing restoration efficiency function abbreviation of system:

T '=∑ t '_ii∈ω_d(8)

In formula (8), t '_iFor i-th section in destination node set after the key routing restoration efficiency function abbreviation of subsystem Point recovers required time, ω_dFor destination node set；

(9) determine the rack coverage rate of trunk rack k after the key routing restoration efficiency function abbreviation of subsystem as the following formula c′_k:

c_k'=∑ α '_jj∈ω_d(9)

In formula (9), α '_jFor j-th section in destination node set after the key routing restoration efficiency function abbreviation of subsystem The comprehensive importance degree of point；

Further, as the following formula (10) determine destination node set after the key routing restoration efficiency function abbreviation of subsystem In i-th node recover required time t ':

$t_i^{\&prime;}\&ap;\frac{1}{2}(\underset{j\&element;{\mathrm{\&phi;}}_{\min 1}^i}{\&sigma;}{t}_j^l+\underset{j\&element;{\mathrm{\&phi;}}_{\min 2}^i}{\&sigma;}{t}_j^l)---\left(10\right)$

In formula (10),For the set of minimal paths in other destination node paths for the node i,Arrive it for node i Second shortest path set in his destination node path,Recovery time needed for circuit j；

By above-mentioned formula it can be found that, when destination node is very few, rack coverage rate c_k' too low, target function value is relatively low； When destination node is excessive, system recovery time t ' will be long, and target function value still will not be preferable.It is contemplated that, object function When maximum, destination node number must be in an intermediate value, therefore, optimum option target section in the technical scheme that the present invention provides Point, assumes initially that all nodes are destination node, then progressively removes the minimum node of importance degree in destination node, when initial Object function will be gradually increased, and after continuing for some time, object function will appear from a peak value, reduces afterwards and progressively, target letter Count corresponding destination node during existing peak value and be optimal objective node, concrete steps include:

A. set all nodes of subsystem and be destination node, obtain functional value by described formula (7)；

B. the minimum node of importance degree in described destination node is classified as non-targeted node；

C. the comprehensive importance degree of non-targeted node is shifted, wherein, if the comprehensive importance degree α of subsystem interior joint i_i Minimum, then make described node i be non-targeted node, change its comprehensive importance degree α '_i=0, by described comprehensive importance degree α_iIt is transferred to The nearest destination node j away from node i, the then comprehensive importance degree α of (11) switch target node j as the following formula_j:

$\begin{array}{cc}{\mathrm{\&alpha;}}_j^{\&prime;}={\mathrm{\&alpha;}}_j+{\mathrm{\&alpha;}}_i/n_d^i{2}^{d_{i-d}}& j\&element;{\mathrm{\&omega;}}_d^i\end{array}---\left(11\right)$

In formula (11), α '_jFor conversion after destination node j comprehensive importance degree,It is nearest away from i-th non-targeted node Node set, d_i-dFor node i extremelyDistance,ForMiddle destination node number；

D. press described formula (7) and obtain functional value, if this functional value is bigger than functional value in described step a, return to step b, If this functional value is less than functional value in described step a, go to step e；

E. in described step d, whether functional value is that continuous second wheel reduces, and if so, then goes to step f, if it is not, then preserving Current target node set x, goes to step b；

F. export current target node set；

Wherein, the destination node collection that described step f obtains is combined into optimal objective node set.

After completing the selection of destination node, if can determine the recovery order of destination node, you can with according to this order Search for the optimal recovery path of destination node one by one, recover all destination nodes, thus completion system recovers, obtain system recovery Scheme.The optimization problem of destination node recovery order, is the optimization problem of a generic sequence, for problems, can adopt Document [noble, Yang Jingyu. swarm intelligence algorithm and its application [m]. Beijing: Chinese Water Conservancy water power publishing house, 2006:89-91] in The intersection particle cluster algorithm proposing.

In addition, intersecting particle cluster algorithm efficiency for improving, can manage to optimize primary quality.Due to target function type (3) presence of rack dispersion ratio d in, so rack reconstruct optimal case is necessarily first recovered away from the nearer target section of black starting-up point Point, recovers afterwards apart from remote.So for improving primary quality, first destination node can be pressed it remote with black starting-up point distance Nearly classification, and following principle is observed when initializing particle: 1) belong to inhomogeneous destination node, its sequencing is constant；2) Can be randomly ordered with apoplexy due to endogenous wind destination node.The described key routing restoration model optimization destination node according to described subsystem recovers Sequentially, comprising:

The concrete steps being obtained the optimum recovery order of destination node using Crossover Particle Swarm Optimization are included:

A) destination node is pressed classification far and near away from black starting-up point, and press above-mentioned primary quality optimization method, carry out grain Son initialization.

B) calculate the object function of each particle correspondence system recovery scheme, and find out the individual history optimal sequence of each particle Pxbest and global optimum's sequence gbest.

C) each particle carries out crossover operation respectively at pxbest, gbest.

D) whether parser may be absorbed in locally optimal solution.In this way, then particle is taken certainly to flee from and go to step after measure e)；Otherwise directly go to step e).

E) whether evaluation algorithm meets the condition of convergence, in this way, then exports gbest and terminates to calculate；Otherwise, go to step b).

It is only necessary to find the optimum of destination node in the order one by one to recover road after obtaining destination node optimum recovery order Footpath, you can form system recovery scheme.Find destination node optimum restoration path when should be noted fired power generating unit recovery when Between restricted problem, and non-electrical source node synchronization recover problem, the described key routing restoration model according to described subsystem The concrete steps of optimization aim node restoration path include:

A. the recovery order according to described destination node, recovers to destination node one by one；

B. judge whether current target node is non-electrical source node, be to go to step c；Otherwise go to step f；

Whether the subsequent node c. judging current target node is non-electrical source node, if so, then determines whether that this is follow-up Whether node can be recovered with described current target node simultaneously, if then going to step d；If otherwise going to step e；

D. obtain the restoration path of destination node according to the recovery order of described destination node, and be possible to recover simultaneously Circuit is set to recover circuit simultaneously, goes to step g；

E. obtain the restoration path of destination node according to the recovery order of described destination node, go to step g；

F. obtain the restoration path of destination node according to the recovery order of described destination node, if this node cannot start Recover in time-constrain, then make this pitch point importance be 0；

G. judge whether that all destination nodes all recover, if so, then algorithm terminates；Otherwise, according to described destination node Recovery order recovery next node, go to step b.

In addition it is also necessary to the partial electric grid of sub-system is recovered after the node recovery of done subsystem backbone's passage, described Step 103, comprising:

Define the main electrical path length of network of partial electric grid, average path length and the dynamic discrete rate of subsystem；Wherein, network Main electrical path length: refer to the maximum of any two node shortest distance in network.The main footpath of network: refer to the main electrical path length of network corresponding two Shortest path between node.Network average path length: refer to the arithmetic mean of instantaneous value of any two node shortest distance in network.

Build object function and its constraints that the partial electric grid restoration path of described subsystem optimizes；

The partial electric grid of the object function recovery subsystem that the partial electric grid restoration path according to described subsystem optimizes.

The described main electrical path length of network of partial electric grid defining subsystem, average path length and dynamic discrete rate, comprising:

Main electrical path length t of network of the partial electric grid of (12) definition subsystem as the following formula:

$t=\underset{i<j}{\max }{d}_{ij}---\left(12\right)$

In formula (12), d_ijFor the beeline between partial electric grid interior joint i and node j, represented with the number of lines；

(13) define the average path length of the partial electric grid of subsystem: f as the following formula

$f=\frac{2}{n(n+1)}\underset{i<j}{\&sigma;}{d}_{ij}---\left(13\right)$

In formula (13), n is the quantity of partial electric grid interior joint；

(14) define dynamic discrete rate e of the partial electric grid of subsystem as the following formula_c:

$e_c=\frac{t_{new}}{t_{\max }}\&centerdot;\frac{f_{new}}{f_{\max }}---\left(14\right)$

In formula (14), t_newAnd f_newIt is respectively the main electrical path length of network and the net being formed new rack after subsystem changes Network average path length, t_maxThe main footpath of network for whole system before having a power failure on a large scale；f_maxOccurred for during system reconfiguration Macroreticular average path length.

The described object function of partial electric grid restoration path optimization building described subsystem and its constraints, comprising:

System recovery is carried out using the method for expanded type substep optimizing, wherein the most key is extensive to each stage optimum The selection in multiple path, therefore, (15) determine the comprehensive weight m of restoration path in partial electric grid as the following formula_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c---\left(15\right)$

In formula (15), μ is the number of times in restoration path through transformator, c_ijFor the circuit under conversion to same electric pressure Charging capacitor, c_maxFor the maximum of single line charging capacitor in system, ξ₁=1 or 0, represent network dispersion consideration with No, e_cFor the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1 and two addend item dimensions are all 1；

If restoration path is made up of dry contact and circuit.The recovery of node depends on the power transmission of circuit；Circuit The too busy to get away node that puts into operation voltage support.Exactly the two inseparable cooperation relation can make system reconfiguration smoothly enter OK, in view of this, the object function that as the following formula the partial electric grid restoration path of the described subsystem of (16) structure optimizes:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(16\right)$

In formula (16), f is partial electric grid path recovery and optimization function, and β is the importance degree sum of each node in restoration path, Wherein, the formula of the comprehensive importance degree of node j is: β_j=cr_{K, l}+(1-c)r_{K, s}0 ＜ c ＜ 1, in formula: r_{K, 1}And r_{K, s}Table respectively Show load restoration rate and the power backup rate of subsystem k after recovery nodes j；C is regulatory factor.Pitch point importance is bigger, table It is bright that it is more obvious to the raising effect of load restoration rate or power backup rate.And the size of pitch point importance not only with node itself Characteristic institute charged, loading size relevant, also the recovering process of homologous ray is related.Synthesis by node j is important The formula of degree understands, even if when regulatory factor c is constant, same node also can in the different times of system recovery, significance level Change, this dynamic change reflects the difference for power supply and workload demand degree for the system, and passes through to change the big of c Little, this demand difference suitably can be zoomed in or out, to reach the purpose of screening requirements node.

It should be noted that ensureing the dynamic equilibrium of power supply and load, as the following formula (17) while finding optimum restoration path Determine described subsystem partial electric grid restoration path optimize bound for objective function:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(17\right)$

In formula (17), g is to have recovered power supply node, and l is the set of load bus, p_jThe burden with power recovering for node Amount, the load or burden without work amount that qj recovers for node, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Minimum for node has Work(injecting power, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

For key passage, partial electric grid cooperation problem, based on partial electric grid recovery and optimization algorithm, complete local electricity The recovery of net, the object function that the described partial electric grid restoration path according to described subsystem optimizes and its constraints recover son The partial electric grid of system, comprising:

A. obtain the direct-to-ground capacitance c of amount, loading and every circuit of partial electric grid to be restored_ij ^*, with extensive The key passage of multiple subsystem is the starting point of searching route, and in subsystem, remainder is not recover rack in area, and defines Initial value k=1；

B. recovered rack and be equivalent to a node o_k, with o_kCentered on to obtain radius be all to belong to inextensive in area in r The power supply node of multiple rack and load bus, described power supply node and load bus are put into destination node collection i_k；

C. the weights making circuit in path are direct-to-ground capacitance, by i_kMiddle destination node is to Centroid o_kShortest path put Enter shortest path collection s_k；

D. determine s respectively_kIn each paths the object function that optimizes of partial electric grid restoration path, and delete and be unsatisfactory for about The path of bundle condition；

E. from s_kThe path l of the target function value minimum that the middle partial electric grid restoration path selecting subsystem optimizes, is incorporated to Send a telegram in reply in network；

F. trend verification is carried out to network of sending a telegram in reply, if trend convergence, go to step g；Trend does not restrain, in s_kSend a telegram in reply Delete path l in network, go to step e；

If g. load restoration rate r of this partial electric grid_k＞ 90%, algorithm stops；Otherwise go to step b.

Equally area networking recovery is carried out using the method for expanded type substep optimizing, the comprehensive weight of circuit in restoration path Introduce network dispersion and subregion range-sensitivity dual indexes, its restoration path will be to adjacent subsystems with strong complementarity Direction offsets, and the rack of adjacent subsystems also has identical and recovers trend.Coordinate between each subregion, to recover the cooperation of progress, And the selection on arranged side by side opportunity, described step 104, comprising:

Build area networking Restoration model and its constraints；

Carry out networking between system using area networking reconstruction and its constraints sub-system to recover.

Described structure area networking reconstruction and its constraints, comprising:

Method using expanded type substep optimizing carries out system recovery, and (18) determine in area networking between subsystem as the following formula The comprehensive weight m of restoration path_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c+{\mathrm{\&xi;}}_2(1-\mathrm{\&delta;}d)---\left(18\right)$

In formula (18), μ is the number of times in restoration path through transformator, c_ijFor the circuit under conversion to same electric pressure Charging capacitor, c_maxFor the maximum of single line charging capacitor in system, δ d is subregion range-sensitivity, ξ₁=1 or 0, table Whether show the consideration of network dispersion, e_cFor the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1,0 ＜ δ d≤1, and Three addend item dimensions are all 1；

(19) structure area networking reconstruction as the following formula:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(19\right)$

In formula (19), f is area networking reconstruction, and β is the importance degree sum of each node in restoration path；

It should be noted that ensureing the dynamic equilibrium of power supply and load, as the following formula (20) while finding optimum restoration path Determine the constraints of described area networking reconstruction:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(20\right)$

In formula (20), g is to have recovered power supply node, and l is the set of load bus, p_jThe burden with power recovering for node Amount, q_jThe load or burden without work amount recovered for node, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Active for the minimum of node Injecting power, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

For partial electric grid, area networking cooperation problem, based on area networking recovery and optimization algorithm, complete region connection The recovery of net, described carry out between system networking using area networking reconstruction and its constraints sub-system and recovers, comprising:

A. obtain the direct-to-ground capacitance c of the amount, loading and every circuit in n subsystem_ij ^*, with each subsystem Black starting-up power supply is to have recovered rack in area, and as the starting point of searching route, remainder is not recover rack in area, and initially Change k=1；Initialize the ξ in each subsystem₁=1, ξ₂=0；

B. it is equivalent to a node o by having recovered rack in the area of subsystem k_kIf, ξ₂=0, then with o_kCentered on obtain half Footpath is r_kInterior all belong to the power supply node not recovering rack in area and load bus, and by described power supply node and load bus Put into destination node collection i_kIf, ξ₂=1, then with o_kCentered on obtain radius be r_kInterior all still unrecovered power supply nodes and negative Lotus node, and described power supply node and load bus are put into destination node collection i_k；

C. the weights of circuit in direct-to-ground capacitance as path, call dijkstra algorithm to obtain i_kIn each destination node Short path, and described shortest path is put into shortest path collection s_k；

If d. ξ₁=1, then obtain the main footpath t of subsystem k network_kWith load restoration rate r_1kIf, t_k＞ 2/3t_max, then by ξ₁Put 0, if r_1k＞ 50%, then by ξ₂Put 1；

E. determine s respectively_kIn each path area networking reconstruction value, and delete the path being unsatisfactory for constraints；

F. from s_kThe minimum path l of middle selection region networking reconstruction value, is incorporated in subsystem k；

G. sub-system k carries out trend verification, if trend convergence, goes to step h；Trend does not restrain, in s_kIn subsystem k Delete path l, go to step f；

H. whether the subsystem k and adjacent subsystems m difference on the frequency in point both sides arranged side by side, voltage amplitude value difference and phase angle difference are verified Meeting and require side by side, if meeting, subsystem k and m being merged, n=n-1, k=1；If being unsatisfactory for, k=(k+1) %n；

If i. load restoration rate r of subsystem k_k＞ 70%, and subsystem number n=1, algorithm stops；Otherwise go to step b.

Wherein, to be often referred to phase sequence, phase place, frequency identical with voltage for described requirement arranged side by side, but different electrical network has been also possible to The particular provisions of oneself.

Finally it should be noted that: above example is only not intended to limit in order to technical scheme to be described, to the greatest extent Pipe has been described in detail to the present invention with reference to above-described embodiment, and those of ordinary skill in the art are it is understood that still The specific embodiment of the present invention can be modified or equivalent, and any without departing from spirit and scope of the invention Modification or equivalent, it all should be covered within the claims of the present invention.

Claims (16)

1. a kind of power system reconstructing method is it is characterised in that methods described includes:

According to system subdivision rule, will treat that reconfiguration system is divided into x subsystem, wherein, x is positive integer；

Using the key routing restoration model of described subsystem and its key passage of the constraints described subsystem of recovery；

Using the partial electric grid restoration path of described subsystem, the object function optimizing and its constraints recover described subsystem Partial electric grid；

Carry out networking between system using area networking reconstruction and its constraints sub-system to recover.

2. the method for claim 1 is it is characterised in that described subsystem includes: at least one black starting-up power supply, is opened Galvanic electricity source, transmission line of electricity, transformer station and load bus.

3. the method for claim 1 is it is characterised in that described system subdivision rule includes:

The number x of subsystem treats black starting-up power supply number in reconfiguration system less than described；

The difference of the generating set total installation of generating capacity between subsystem or maximum carrying capacity is more than subregion size threshold y；

Each subsystem internal loading with go out dynamic balance；

Interconnection quantity between each subsystem is less than or equal to 4, and the electric pressure of described interconnection is 110～220kv；

Subsystem is centered on black starting-up power supply；

Can be grid-connected between subsystem.

4. the method for claim 1 it is characterised in that the described key routing restoration model using described subsystem and Its constraints recovers the key passage of described subsystem, comprising:

Define rack coverage rate and the rack dispersion ratio of subsystem backbone's passage；

Build key routing restoration model and its constraints of subsystem using described rack coverage rate and rack dispersion ratio；

Sequentially and extensive according to the key routing restoration model optimization selection destination node of described subsystem, optimization aim node recovery Multiple path, wherein, the key passage that described destination node is constituted need to meet described constraints.

5. method as claimed in claim 4 is it is characterised in that described definition subsystem backbone's rack coverage rate of passage and net Frame dispersion ratio includes:

(1) defines rack coverage rate c of the trunk rack k of subsystem backbone's passage as the following formula_kAnd it is each in node j to trunk rack k Shortest path minima d of node_j-k:

$\left\{\begin{array}{c}{c}_k=\underset{i\&element;{\mathrm{\&omega;}}_k}{\&sigma;}{\mathrm{\&alpha;}}_i+\underset{j\&element;{\mathrm{\&omega;}}_{\stackrel{\&overbar;}{k}}}{\&sigma;}\frac{1}{2^{d_{j-k}}}{\mathrm{\&alpha;}}_j\\ d_{j-k}=\min \left(d_{j-i}\right),i\&element;{\mathrm{\&omega;}}_k\end{array}\right.---\left(1\right)$

In formula (1), ω_kFor the node set in the trunk rack k of described subsystem,Trunk rack k for described subsystem Outside node set, α_iThe comprehensive importance degree of node i, α_jComprehensive importance degree for node j, d_j-iFor node j to node i away from From that is, node j is to the circuit number of the shortest path process of node i；

(2) define the rack dispersion ratio d of subsystem backbone's passage and recover network dispersion during i-th destination node as the following formula d_i:

$\left\{\begin{array}{cc}d=max\left(d_i\right)& i=1,2,...,n_k\\ d_i=\frac{max\left(d_{j-k}\right)}{n_{\stackrel{\&overbar;}{g}}}& j\&element;{\mathrm{\&omega;}}_{ig},k\&element;{\mathrm{\&omega;}}_{i\stackrel{\&overbar;}{g}}\end{array}\right.---\left(2\right)$

In formula (2), n_kFor destination node sum, ω_igFor recovering to have recovered the power supply section in trunk rack during i-th destination node Point set,For recovering to have recovered the non-power node set in trunk rack during i-th destination node,Non- for recovery Power supply destination node number.

6. method as claimed in claim 4 is it is characterised in that described built using described rack coverage rate and rack dispersion ratio The key routing restoration model of subsystem and its constraints, comprising:

The key routing restoration model of (3) structure subsystem as the following formula:

$\begin{array}{cc}\max & f=\frac{{\mathrm{\&tau;c}}_k-\underset{i\&element;l_k}{\&sigma;}{\mathrm{\&omega;}}_i}{td}\end{array}---\left(3\right)$

In formula (3), f is the key routing restoration efficiency function of subsystem, and τ is the trunk rack coverage rate weight of subsystem, c_kFor The rack coverage rate of the trunk rack k of subsystem, l_kFor recovering the line set that trunk rack comprises, ω_iFor being resumed circuit Weight, t is system recovery required time；

(4) determine that the trend of subsystem constrains as the following formula:

$\left\{\begin{array}{cc}{p}_{gi}^{\min }\&le;p_{gi}\&le;p_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ q_{gi}^{\min }\&le;q_{gi}\&le;q_{gi}^{\max }& i=1,2,\mathrm{...},n_g\\ v_j^{\min }\&le;v_j\&le;v_j^{\max }& j=1,2,\mathrm{...},n_k\\ i_k\&le;i_k^{\max }& k=1,2,\mathrm{...},n_{kl}\end{array}\right.---\left(4\right)$

In formula (4), n_gFor having recurred motor nodes, n in subsystem_kCount for knot cluster in subsystem, n_klFor in subsystem Multiple line way,For the active minima of exerting oneself of i-th electromotor node in subsystem,For i-th generating in subsystem The active maximum of exerting oneself of machine node, p_giExert oneself for the active of i-th electromotor node in subsystem,For in subsystem i-th The idle minima of exerting oneself of individual electromotor node,For the idle maximum of exerting oneself of i-th electromotor node in subsystem, q_gi Exert oneself for the idle of i-th electromotor node in subsystem,For the voltage minimum of j-th node in subsystem, v_jFor son The voltage of j-th node in system,For the voltage max of j-th node in subsystem, i_kFor k-th line in subsystem The electric current on road,Current maxima for k-th circuit in subsystem；

As the following formula (5) determine subsystem non-power nodal parallel recover constraint function:

$l^{*}\&le;\sqrt{(1-{f_{\min }^{*}}^2)r_{\&sigma;}^{*}{t}_j}---\left(5\right)$

In formula (5), l^*For the total load of subsystem parallel recovery node,Allow the low-limit frequency occurring for subsystem, press Formula (6) determinesAnd t_j:

In formula (6),For the climbing rate of i-th electromotor in subsystem, t_jiInertia time for i-th electromotor in subsystem Constant,Rated capacity for i-th electromotor in subsystem.

7. method as claimed in claim 4 is it is characterised in that the described key routing restoration model according to described subsystem is excellent Change and choose destination node, comprising:

(7) determine the Reduce function of the key routing restoration model of described subsystem as the following formula:

$\begin{array}{cc}\max & f^{\&prime;}=\frac{c_k^{\&prime;}}{t^{\&prime;}}\end{array}---\left(7\right)$

In formula (7), f ' is the key routing restoration efficiency function abbreviation value of subsystem, and t ' is the key routing restoration effect of subsystem Recovery required time after rate function abbreviation,For trunk rack k after the key routing restoration efficiency function abbreviation of subsystem Rack coverage rate；

Wherein, as the following formula (8) determine the recovery required time t ' after the key routing restoration efficiency function abbreviation of subsystem:

T '=∑ t '_ii∈ω_d(8)

In formula (8), t '_iRecover for i-th node in destination node set after the key routing restoration efficiency function abbreviation of subsystem Required time, ω_dFor destination node set；

(9) determine the rack coverage rate c ' of trunk rack k after the key routing restoration efficiency function abbreviation of subsystem as the following formula_k:

c_k'=∑ α '_jj∈ω_d(9)

In formula (9), α '_jFor in destination node set after the key routing restoration efficiency function abbreviation of subsystem j-th node comprehensive Close importance degree；

Further, as the following formula (10) determine in destination node set after the key routing restoration efficiency function abbreviation of subsystem I node recovery required time t ':

$t_i^{\&prime;}\&ap;\frac{1}{2}\left(\underset{{\mathrm{j\&phi;}}_{\min 1}^i}{\&sigma;}{t}_j^l+\underset{{\mathrm{j\&phi;}}_{\min 2}^i}{\&sigma;}{t}_j^l\right)---\left(10\right)$

In formula (10),For the set of minimal paths in other destination node paths for the node i,For node i to other mesh Second shortest path set in mark node path,Recovery time needed for circuit j；

The step of optimum option destination node specifically includes:

A. set all nodes of subsystem and be destination node, obtain functional value by described formula (7)；

B. the minimum node of importance degree in described destination node is classified as non-targeted node；

C. the comprehensive importance degree of non-targeted node is shifted, wherein, if the comprehensive importance degree α of subsystem interior joint i_iMinimum, Then make described node i be non-targeted node, change its comprehensive importance degree α '_i=0, by described comprehensive importance degree α_iIt is transferred to away from section The nearest destination node j of point i, the then comprehensive importance degree α of (11) switch target node j as the following formula_j:

${\mathrm{\&alpha;}}_j^{\&prime;}={\mathrm{\&alpha;}}_j+{\mathrm{\&alpha;}}_i/n_d^i{2}^{d_{i-d}},j\&element;{\mathrm{\&omega;}}_d^i---\left(11\right)$

In formula (11), α '_jFor conversion after destination node j comprehensive importance degree,It is the node nearest away from i-th non-targeted node Set, d_i-dFor node i extremelyDistance,ForMiddle destination node number；

D. press described formula (7) and obtain functional value, if this functional value is bigger than functional value in described step a, return to step b, if should Functional value is less than functional value in described step a, then go to step e；

E. in described step d, whether functional value is that continuous second wheel reduces, and if so, then goes to step f, if it is not, then preserving current Destination node set x, goes to step b；

F. export current target node set；

Wherein, the destination node collection that described step f obtains is combined into optimal objective node set.

8. method as claimed in claim 4 is it is characterised in that the described key routing restoration model according to described subsystem is excellent Change destination node recovery order, comprising:

Obtain the optimum recovery order of destination node using Crossover Particle Swarm Optimization；

Wherein, in described Crossover Particle Swarm Optimization implementation procedure, destination node is pressed its black starting-up power supply in subsystem Distance be divided into two classes, and by initialization principle destination node is initialized, described initialization principle include: a. belongs to not Similar destination node, its sequencing is constant；B. can be randomly ordered with apoplexy due to endogenous wind destination node；

In described Crossover Particle Swarm Optimization implementation procedure, object function is the key routing restoration model of described subsystem.

9. method as claimed in claim 4 is it is characterised in that the described key routing restoration model according to described subsystem is excellent Change destination node restoration path, comprising:

A. the recovery order according to described destination node, recovers to destination node one by one；

B. judge whether current target node is non-electrical source node, be to go to step c；Otherwise go to step f；

Whether the subsequent node c. judging current target node is non-electrical source node, if so, then determines whether this subsequent node Whether can recover with described current target node, if then going to step d simultaneously；If otherwise going to step e；

D. obtain the restoration path of destination node according to the recovery order of described destination node, and be possible to the circuit simultaneously recovering It is set to recover circuit simultaneously, go to step g；

E. obtain the restoration path of destination node according to the recovery order of described destination node, go to step g；

F. according to described destination node recovery order obtain destination node restoration path, if this node cannot on startup between Recover in constraint, then make this pitch point importance be 0；

G. judge whether that all destination nodes all recover, if so, then algorithm terminates；Otherwise, extensive according to described destination node Multiple order recovery next node, goes to step b.

10. the method for claim 1 is it is characterised in that the described partial electric grid restoration path using described subsystem The object function optimizing and its partial electric grid of the constraints described subsystem of recovery, comprising:

Define the main electrical path length of network of partial electric grid, average path length and the dynamic discrete rate of subsystem；

Build object function and its constraints that the partial electric grid restoration path of described subsystem optimizes；

The partial electric grid of the object function recovery subsystem that the partial electric grid restoration path according to described subsystem optimizes.

11. methods as claimed in claim 10 are it is characterised in that the main path length of network of the partial electric grid of described definition subsystem Degree, average path length and dynamic discrete rate, comprising:

Main electrical path length t of network of the partial electric grid of (12) definition subsystem as the following formula:

$t=\underset{i<j}{\max }{d}_{ij}---\left(12\right)$

In formula (12), d_ijFor the beeline between partial electric grid interior joint i and node j, represented with the number of lines；

(13) define the average path length of the partial electric grid of subsystem: f as the following formula

$f=\frac{2}{n(n+1)}\underset{i<j}{\&sigma;}{d}_{ij}---\left(13\right)$

In formula (13), n is the quantity of partial electric grid interior joint；

(14) define dynamic discrete rate e of the partial electric grid of subsystem as the following formula_c:

$e_c=\frac{t_{new}}{t_{\max }}\&centerdot;\frac{f_{new}}{f_{\max }}---\left(14\right)$

In formula (14), t_newAnd f_newIt is respectively the main electrical path length of network being formed new rack after subsystem changes and network is flat All paths, t_maxThe main footpath of network for whole system before having a power failure on a large scale；f_maxFor the maximum web occurring during system reconfiguration Network average path length.

12. methods as claimed in claim 10 are it is characterised in that the partial electric grid restoration path of the described subsystem of described structure The object function optimizing and its constraints, comprising:

(15) determine the comprehensive weight m of restoration path in partial electric grid as the following formula_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c---\left(15\right)$

In formula (15), μ is the number of times in restoration path through transformator, c_ijFor the line charging under conversion to same electric pressure Electric capacity, c_maxFor the maximum of single line charging capacitor in system, ξ₁=1 or 0, whether the consideration of expression network dispersion, e_c For the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1 and two addend item dimensions are all 1；

The object function that the partial electric grid restoration path of the described subsystem of (16) structure optimizes as the following formula:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(16\right)$

In formula (16), f is partial electric grid path recovery and optimization function, and β is the importance degree sum of each node in restoration path；

As the following formula (17) determine described subsystem partial electric grid restoration path optimize bound for objective function:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(17\right)$

In formula (17), g is to have recovered power supply node, and l is the set of load bus, p_jThe burden with power amount recovered for node, q_jFor The load or burden without work amount that node recovers, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Minimum active injection work(for node Rate, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

13. methods as claimed in claim 10 are it is characterised in that the described partial electric grid restoration path according to described subsystem The object function optimizing and its partial electric grid of constraints recovery subsystem, comprising:

A. obtain the direct-to-ground capacitance c of amount, loading and every circuit of partial electric grid to be restored_ij ^*, with recovery The key passage of subsystem is the starting point of searching route, and in subsystem, remainder is not recover rack in area, and defines initial Value k=1；

B. recovered rack and be equivalent to a node o_k, with o_kCentered on to obtain radius be all to belong to non-instauration net in area in r The power supply node of frame and load bus, described power supply node and load bus are put into destination node collection i_k；

C. the weights making circuit in path are direct-to-ground capacitance, by i_kMiddle destination node is to Centroid o_kShortest path put into Short path collection s_k；

D. determine s respectively_kIn each paths partial electric grid restoration path optimize object function, and delete be unsatisfactory for constrain bar The path of part；

E. from s_kThe path l of the target function value minimum that the middle partial electric grid restoration path selecting subsystem optimizes, is incorporated to and sends a telegram in reply In network；

F. trend verification is carried out to network of sending a telegram in reply, if trend convergence, go to step g；Trend does not restrain, in s_kSend a telegram in reply network Middle deletion path l, goes to step e；

If g. load restoration rate r of this partial electric grid_k＞ 90%, algorithm stops；Otherwise go to step b.

14. the method for claim 1 are it is characterised in that described utilization area networking reconstruction and its constraints Sub-system carries out networking between system and recovers, comprising:

Build area networking Restoration model and its constraints；

Carry out networking between system using area networking reconstruction and its constraints sub-system to recover.

15. methods as claimed in claim 14 it is characterised in that described structure area networking reconstruction and its constraint bar Part, comprising:

(18) determine the comprehensive weight m of restoration path between subsystem in area networking as the following formula_ij:

$m_{ij}=2^{\mathrm{\&mu;}}\&centerdot;\frac{c_{ij}}{c_{\max }}+{\mathrm{\&xi;}}_1{e}_c+{\mathrm{\&xi;}}_2(1-\mathrm{\&delta;}d)---\left(18\right)$

In formula (18), μ is the number of times in restoration path through transformator, c_ijFor the line charging under conversion to same electric pressure Electric capacity, c_maxFor the maximum of single line charging capacitor in system, δ d is subregion range-sensitivity, ξ₁=1 or 0, represent net Whether the consideration of network dispersion, e_cFor the dynamic discrete rate of the partial electric grid of subsystem, 0 ＜ ec≤1,0 ＜ δ d≤1, and three Addend item dimension is all 1；

(19) structure area networking reconstruction as the following formula:

$\begin{array}{cc}\min & f=\frac{m_{ij}}{\mathrm{\&beta;}}\end{array}---\left(19\right)$

In formula (19), f is area networking reconstruction, and β is the importance degree sum of each node in restoration path；

(20) determine the constraints of described area networking reconstruction as the following formula:

$\left\{\begin{array}{c}\underset{i\&element;g}{\&sigma;}{p}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{p}_j\&le;\underset{i\&element;g}{\&sigma;}{p}_{i,\max }\\ \underset{i\&element;g}{\&sigma;}{q}_{i,\min }\&le;\underset{j\&element;l}{\&sigma;}{q}_j\&le;\underset{i\&element;g}{\&sigma;}{q}_{i,\max }\end{array}\right.---\left(20\right)$

In formula (20), g is to have recovered power supply node, and l is the set of load bus, p_jThe burden with power amount recovered for node, q_jFor The load or burden without work amount that node recovers, p_{I, max}For the active injecting power of maximum of node, p_{I, min}Minimum active injection work(for node Rate, q_{I, max}For the idle injecting power of maximum of node, q_{I, min}The idle injecting power of minimum for node.

16. methods as claimed in claim 14 are it is characterised in that described utilization area networking reconstruction and its constraints Sub-system carries out networking between system and recovers, comprising:

A. obtain the direct-to-ground capacitance c of the amount, loading and every circuit in n subsystem_ij ^*, open so that each subsystem is black Galvanic electricity Yuan Wei has recovered rack in area, and as the starting point of searching route, remainder is not recover rack in area, and initializes k =1；Initialize the ξ in each subsystem₁=1, ξ₂=0；

B. it is equivalent to a node o by having recovered rack in the area of subsystem k_kIf, ξ₂=0, then with o_kCentered on obtain radius be r_kInterior all belong to the power supply node not recovering rack in area and load bus, and described power supply node and load bus are put into Destination node collection i_kIf, ξ₂=1, then with o_kCentered on obtain radius be r_kInterior all still unrecovered power supply nodes and load section Point, and described power supply node and load bus are put into destination node collection i_k；

C. the weights of circuit in direct-to-ground capacitance as path, call dijkstra algorithm to obtain i_kIn each destination node shortest path Footpath, and described shortest path is put into shortest path collection s_k；

If d. ξ₁=1, then obtain the main footpath t of subsystem k network_kWith load restoration rate r_1kIf, t_k＞ 2/3t_max, then by ξ₁Set to 0, if r_1k＞ 50%, then by ξ₂Put 1；

E. determine s respectively_kIn each path area networking reconstruction value, and delete the path being unsatisfactory for constraints；

F. from s_kThe minimum path l of middle selection region networking reconstruction value, is incorporated in subsystem k；

G. sub-system k carries out trend verification, if trend convergence, goes to step h；Trend does not restrain, in s_kDelete with subsystem k Path l, goes to step f；

H. whether verification subsystem k and adjacent subsystems m meet in the difference on the frequency putting both sides arranged side by side, voltage amplitude value difference and phase angle difference Requiring side by side, if meeting, subsystem k and m being merged, n=n-1, k=1；If being unsatisfactory for, k=(k+1) %n；

If i. load restoration rate r of subsystem k_k＞ 70%, and subsystem number n=1, algorithm stops；Otherwise go to step b.