CN1731645A - Voltage control method based on soft partitions in power system - Google Patents

Abstract

The invention relates to a voltage control method for electric system basing on soft division. The method comprises: real time data gathering in electrical network; partition online; establishing CSVC mathematical model for control area; calculating the control quantity of each CSVC mathematical model; outputting the control quantity of each control area; returning to circulating for CSVC controlling when the CSVC control cycle time out.

Description

In the electric power system based on the voltage control method of soft sectoring

Technical field

The invention belongs to electric power system automatism voltage control technical field, particularly in the electric power system based on the voltage control method of soft sectoring.

Background technology

Automatism voltage control (AVC, automatic voltage control) system is based on the overall voltage automated closed-loop control system of the modern power systems of computer, its basic principle is that the control generator reactive is exerted oneself by coordinating, load tap changer and reactive-load compensation equipment, guarantees safety, high-quality and the economy of power system operation.On control model, existing automatic voltage control system can be divided into substantially based on the control model of " optimal load flow " with based on control model two classes of layering and zoning.

Basic skills based on the control model of layering and zoning is that voltage control is divided into three levels: one-level voltage control (PVC, Primary Voltage Control), secondary voltage control (SVC, Secondary VoltageControl) and tertiary voltage control (TVC, Tertiary Voltage Control).Wherein, secondary voltage control is based on " maincenter bus " and " control area ", and its target is exactly by the control device in the one's respective area (exert oneself as, control generator reactive etc.), with the voltage control of maincenter bus on set point.Traditional secondary voltage control method is based on the hypothesis of electric weak coupling between each control area, but, along with power system development, coupling between the system is tight day by day, weak coupling hypothesis between original control subregion is difficult to guarantee, makes the effect of traditional secondary voltage control method be subjected to influence.

In order to address this problem, coordination secondary voltage control (CSVC, CoordinatedSecondary Voltage Control) has been proposed in the world.In coordinating secondary voltage control, because controlled generator number is greater than maincenter bus number, therefore except guaranteeing maincenter busbar voltage and set point deviation minimum, certain control degrees of freedom can also be arranged, utilize this degree of freedom to realize other coordination controlled target, the starting point of Here it is CSVC.

From " a kind of improved big line voltage control " (H.VU, P.PRUVOT, C.LAUNAY, et al.AnImproved Voltage Control on Large-scale Power System.IEEE Transactions onPower Systems, 1996,11 (3): 1295-1303) etc. a series ofly launch research and carry out in the document of engineering practice around coordinating secondary voltage control, typical secondary voltage control (CSVC) method of coordinating as can be seen, its concrete implementation step is:

1, adopt partition method that electrical network is divided into N control area based on electrical distance, its partition method is, calculate the sensitivity between each node voltage in the grid nodes model, thereby define the electrical distance between the different nodes, node in distance range is merged, obtain N node set at last, be N corresponding control area.Coordinate secondary voltage control (CSVC) in each control area separately, its implementation is: this N selects a control area the different Regional Control Center in its corresponding geographical position that the CSVC software and hardware system of this control area (hereinafter to be referred as zone C SVC system) is installed respectively: this control subregion and control structure are difficult to change.Therefore, the present invention is called " firmly " subregion with this subregion, this " firmly " control area structure as shown in Figure 1, represent the electrical network boundary line with heavy line among the figure, the boundary line of representing each control area with fine line, whole electrical network is divided into 6 control areas, has selected existing Regional Control Center installation region CSVC system on the corresponding geographical position in each control area:

2, in each control area, select maincenter bus and controlled generator: for example, can choose have big short circuit current in this zone bus as the maincenter bus, choose the generating set of getting in touch closely with the maincenter bus and having the generating set of larger capacity or automatic voltage regulator (AVR, Automatic Voltage Regulator) and remote signalling remote control equipment be installed as controlled generator;

3, each zone C SVC system gathers the real-time power network data of its control area, place respectively: from existing EMS (the Energy Management System of its region control centre, be abbreviated as EMS) obtain required network topology and device parameter, obtain the real time data that collects by the digital sensor of installing in power plant, the transformer station in this control area by the communication network of having built well simultaneously;

4, each zone C SVC system sets up the CSVC Mathematical Modeling (being generally a quadratic programming model) of each control area respectively: constructed one in the CSVC Mathematical Modeling except that the maincenter of realization busbar voltage deviation minimum, utilize unnecessary control degrees of freedom to realize the target function of the adjustment of reactive power flow is for example adopted the target function shown in the following formula simultaneously:

$\underset{\mathrm{\ΔU}}{\min }[{\mathrm{\λ}}_v{\left|\right|\mathrm{\α}(V_C-V_P)-C_V\mathrm{\ΔU}\left|\right|}^2+{\mathrm{\λ}}_q{\left|\right|\mathrm{\α}(Q_{\mathrm{ref}}-Q)-C_q\mathrm{\ΔU}\left|\right|}^2+{\mathrm{\λ}}_u{\left|\right|\mathrm{\α}(U_{\mathrm{ref}}-U)-\mathrm{\ΔU}\left|\right|}^2]$

In the formula, α is a ride gain, and U is a generator terminal voltage, and Δ U is the change value of generator terminal voltage, i.e. controlled quentity controlled variable, V _pBe the voltage of maincenter bus, V _cBe the set point of maincenter busbar voltage, Q is controlled generator reactive value of exerting oneself, Q _RefBe the set point that controlled generator reactive is exerted oneself, U _RefBe the set point of controlled generator terminal voltage, λ _v, λ _qAnd λ _uBe the weight of each component of target function, C _v, C _qBe respectively the sensitivity matrix of maincenter busbar voltage and the controlled generator reactive variable quantity of exerting oneself about the set end voltage variable quantity.

Wherein, V _p, U, Q be real-time collection capacity; V _c, Q _Ref, U _RefBe known quantity, optimize link by other and provide; α, λ _v, λ _q, λ _uBe known quantity, given by control experience; C _v, C _qBe known quantity, obtain by sensitivity calculations; Controlled quentity controlled variable Δ U obtains by finding the solution this CSVC Mathematical Modeling;

Complete CSVC Mathematical Modeling also comprises one group of constraints (omission), satisfies the requirement of working control to guarantee the controlled quentity controlled variable of trying to achieve;

5, each zone C SVC system finds the solution the CSVC Mathematical Modeling respectively, controlled amount Δ U: for example, the Mathematical Modeling in the step 4 can adopt the set algorithm that works (active set method) to find the solution controlled amount Δ U;

6, each zone C SVC system exports controlled quentity controlled variable Δ U respectively, be used to coordinate the one-level voltage controller: each zone C SVC system passes to the interface module (Interface module) that is installed on the controlled generator with controlled quentity controlled variable separately via communication network, interface module to controlled generator, is finished the control to controlled generator terminal voltage according to controlled quentity controlled variable output one-level voltage control duty setting signal;

7, set the CSVC control cycle in advance, step 3 forms a control closed loop to step 6, is undertaken by the control cycle circulation.

In above step, step 1 and 2 normally under off-line (offline) environment by manually finishing, step 3 is to be finished automatically by computer under (real time) environment in real time to 7.

The difference of existing various CSVC Mathematical Modelings is how to coordinate unnecessary control degrees of freedom and goes to realize adjustment to reactive power flow, divide from coordinating target, have two kinds substantially: 1. in coordinating target, require generator reactive exert oneself and set point between deviation minimum (the CSVC Mathematical Modeling of mentioning in the above-mentioned steps 4 promptly is that this is a kind of); 2. in coordinating target, it is poor with former vector behind one of the vector shift that generator reactive is exerted oneself, and requires deviation as far as possible little.

The limitation that above-mentioned two kinds of CSVC Mathematical Modelings exist is that the trade-off effect of first kind of CSVC Mathematical Modeling depends on how to provide the idle set point Q that exerts oneself _g ^Ref, this needs extra increase to calculate or specifies link by hand, and the idle set point Q that exerts oneself of controlled generator _g ^RefWith maincenter busbar voltage set point V _p ^RefIssuable contradiction also is the problem that can not be ignored; Second kind of CSVC Mathematical Modeling starting point is to guarantee as far as possible the idle torque equilibrium between the controlled generator, but do not consider how not keep more idle nargin.

In order to solve the deficiency of above-mentioned two kinds of CSVC Mathematical Modelings, the present application people Guo Qing comes, Sun Hongbin, Zhang Baiming, Li Qin etc. are in " research and the realization of Jiangsu electrical network AVC main station system " (Automation of Electric Systems, 2004,28 (22): proposed a kind of new CSVC Mathematical Modeling 83-87).This Mathematical Modeling realizes increasing generator reactive nargin by the new coordination target of definition, and makes the balanced more purpose of exerting oneself.The target function of this Mathematical Modeling is as follows:

$\underset{{\mathrm{\ΔQ}}_g}{\min }\{W_p{\left|\right|\mathrm{\α}\·(V_p-V_p^{\mathrm{ref}})+C_g\mathrm{\Δ}{Q}_g\left|\right|}^2+W_q{\left|\right|\frac{Q_g+{\mathrm{\ΔQ}}_g-Q_g^{\min }}{Q_g^{\max }-Q_g^{\min }}\left|\right|}^2\}---\left(1\right)$

In the formula, Δ Q _gRepresent the regulated quantity (controlled quentity controlled variable) that controlled generator reactive is exerted oneself; Q _g, Q _g ^MinAnd Q _g ^MaxRepresent controlled generator reactive currency, lower limit and the upper limit respectively; V _pAnd V _p ^RefCurrent voltage of expression maincenter bus and setting voltage; W _pAnd W _qBe weight coefficient, α is a gain coefficient; C _gBe of the sensitivity of maincenter busbar voltage to controlled generator reactive.Wherein, Q _g, V _pBe real-time collection capacity; V _p ^RefBe known quantity, optimize link by other and provide; α, W _p, W _qBe known quantity, given by control experience; Q _g ^Min, Q _g ^MaxBe known quantity, can directly obtain; C _gBe known quantity, obtain by sensitivity calculations; Controlled quentity controlled variable Δ Q _gObtain by finding the solution this CSVC Mathematical Modeling.

Deviation minimum after first the expression control of target function type (1) between maincenter busbar voltage and the set point, second expression of target function type (1) be the idle ratio of exerting oneself of control back generator, for certain generator, this ratio is more little, the idle nargin that this generator is described is big more, and appear in the target function with the form of quadratic sum, will impel each generator to develop to the idle more balanced direction of exerting oneself, help improving the voltage stability of electrical network.Complete CSVC Mathematical Modeling requires satisfying the minimization problem of finding the solution formula (1) under the situation of security constraints, and these constraints comprise:

$|C_{\mathrm{vg}}\·\mathrm{\Δ}{Q}_g|\≤\mathrm{\Δ}{V}_H^{\max }---\left(2\right)$

$V_H^{\min }\≤V_H+C_{\mathrm{vg}}\·{\mathrm{\ΔQ}}_g\≤V_H^{\max }---\left(3\right)$

$V_p^{\min }{\≤V}_p+C_g\·\mathrm{\Δ}{Q}_g{\≤V}_p^{\max }---\left(4\right)$

$Q_g^{\min }\≤Q_g+{\mathrm{\ΔQ}}_g{\≤Q}_g^{\max }---\left(5\right)$

In the formula, V _p, V _p ^MinAnd V _p ^MaxRepresent maincenter busbar voltage currency, lower limit and the upper limit respectively; Q _g, Q _g ^MinAnd Q _g ^MaxRepresent controlled generator reactive currency, lower limit and the upper limit respectively; V _H, V _H ^Min, V _H ^MaxWith Δ V _H ^MaxThe maximum adjustment amount of single step of representing currency, lower limit, the upper limit and the permission of controlled generator high-voltage side bus voltage respectively; C _VgBe of the sensitivity of controlled generator high-voltage side bus voltage to generator reactive.

In order to prevent that control operation from causing excessive fluctuation to electrical network, therefore in each step control, all the control step-length there is strict restriction, this is realized that by constraint equation (2) its physical meaning is V after the control just _HAdjustment amount be less than the maximum adjustment amount Δ of the single step V of permission _H ^MaxConstraint equation (3) and (4) have guaranteed can not cause V after the control _pAnd V _HProduce out-of-limitly, also can similarly add in the constraints for some other important busbar voltage.Constraint equation (5) has guaranteed that idle the exerting oneself of control back generator can be not out-of-limit.

No matter how the CSVC Mathematical Modeling is improved, and the basis of its subregion control all is weak coupling between the reactive voltage of each control area.Therefore, when power network development changed, the control area of original decoupling zero can be coupled, and perhaps can come in the control area of original coupling in decoupling zero, and this all will influence the control effect of existing various CSVC methods greatly.

As previously described, in the division of the structure of CSVC Mathematical Modeling and control area, all need meter sensitivity.Conventional sensitivity analysis only depends on the linearisation of electric power networks equation, do not consider the quasi-stationary physical responses of power equipment (as: generator, load etc.) to various control operations and disturbance, this is feasible in the electric network analysis of general static state, just can't be practical but will serve control decision.Sun Hongbin, Zhang Baiming, Xiang Niande is at " quasi-stationary sensitivity analysis method " (Proceedings of the CSEE, in April, 1999 V19N4, pp.9-13) proposed the new sensitivity method in, different with the sensitivity analysis method of the static state of routine, the new sensitivity method has been considered the quasi-stationary physical responses of electric power system, take into account the total variation between system's new and old stable state in control front and back, effectively improved the precision of sensitivity analysis.This method when generator is equipped with automatic voltage regulator (AVR), can think that this generator node is the PV node based on the PQ Decoupled Model of electric power system; And when generator is equipped with automatic reactive power and regulates (AQR) or automatic power factor and regulate (APFR), can think the identical PQ of the being node of this generator node with common load bus.In addition, the load voltage static characteristic is considered to node voltage once or conic section.The tide model M that is set up so just takes in these quasi-stationary physical responses naturally, thereby is quasi-stationary sensitivity based on the sensitivity that tide model M calculates.Under tide model M, establish the PQ node and PV node number is respectively N _PQAnd N _PV, quantity of state x is the voltage magnitude of PQ node $V_{\mathrm{PQ}}\∈R^{N_{\mathrm{PQ}}},$ Control variables u=[Q _PQV _PVT _K] ^T, wherein $Q_{\mathrm{PQ}}\∈R^{N_{\mathrm{PQ}}}$ Be the idle injection of PQ node, $V_{\mathrm{PV}}\∈R^{N_{\mathrm{PV}}}$ Be the voltage magnitude of PV node, $T_k\∈R^{N_T}$ Be transformer voltage ratio, the important variable of complying with $h={\left[\begin{array}{cc}{Q}_b& Q_{\mathrm{PV}}\end{array}\right]}^T,$ Q wherein _b∈ R ^bBe the branch road reactive power flow, $Q_{\mathrm{PV}}\∈R^{N_{\mathrm{PV}}}$ It is the idle injection of PV node.At this moment, idle tide model is arranged:

Q _PQ(V _PQ，V _PV，T _k)＝0 (6)

Q _b＝Q _b(V _PQ，V _PV，T _k) (7)

Q _PV＝Q _PV(V _PQ，V _PV，T _k) (8)

Can get the idle class Sensitivity calculation of quasi-stable state formula and see Table 1.

The quasi-stationary idle class sensitivity S of table 1 _{(x, h) u}Computing formula

Why the sensitivity that last table provides is different from conventional sensitivity, and be referred to as new sensitivity, be because node type (PV/PQ) is to determine according to the quasi-stable state physical responses of electric power system reality, taken into account the load voltage static characteristic among the tide model M simultaneously, this voltage static characteristic will obtain embodying in various Jacobian battle arrays about load bus voltage.Obviously, at this moment not had illusory reference node, to the PQ node, is idle injection Q _PQParticipate in control directly, and, be by voltage V to the PV node _PVServe as controlled quentity controlled variable.

Guo Qing comes, Sun Hongbin, Zhang Baiming, Wu Wenchuan is in " based on the reactive voltage subregion of reactive source control space cluster analysis " (Automation of Electric Systems, in May, 2005, V29, N10, P36-40) in research and proposed a kind of online adaptive " soft " partition method based on reactive source control space cluster analysis.This method is carried out subregion to electric power system and comprised structure reactive source space and carry out subregion two parts in reactive source control space: detailed process is as follows:

1, structure reactive source space:

(1). establishing has g generator node in the electrical network, constitute set G, and l the load bus for the treatment of subregion arranged, and constitutes set L.If the generator node j among the set G is the node of waiting to find the solution idle control sensitivity, its node type is set to the PQ node; Whether other generator nodes among the set G can keep set end voltage according to it in reactive power, correspondingly node type is made as PV node or PQ node.The B of PV node that formed augmentation " matrix, and strengthen number at the corresponding diagonal element of PV node place.

(2). at the B of step 1 structure " find the solution the sensitivity of each load bus voltage in the idle pair set L of injection of generator node j on the matrix basis.For example, for load bus i (i ∈ L), this sensitivity table is shown S _Ij

(3). for set each generator node among G completing steps (1) all, (2).

(4). by the coordinate space (being reactive source control space) of a g dimension of set G structure.Note load bus i coordinate in this g dimension space is (x _I1, x _I2..., x _Ig), j dimension coordinate x wherein _Ij=-log ₁₀(| S _Ij|).

(5). two treated subregion node m and n corresponding respectively coordinate vector (x _M1, x _M2..., x _Mg) and (x _N1, x _N2..., x _Ng), the distance between 2 adopts Euclidean distance to be defined as follows shown in the formula:

$D_{\mathrm{mn}}=\sqrt{{|x_{m1}-x_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="A20051009852700152.png"\; state="\{\{state\}\}">n</figure-callout>}1}|}^2+......+{|x_{\mathrm{mg}}-x_{\mathrm{ng}}|}^2}.$

2, use cluster algorithm to carry out subregion: based on the reactive source control space of structure in the above-mentioned steps 1, just can obtain coordinate and each internodal distance of each node, the algorithm of the cluster analysis inside reference mode identification this moment comes subregion is carried out in electric power system.For example, adopt the hierarchical clustering algorithm of cohesion to finish subregion and calculate, at first all nodes are all alone as a bunch of existence, and two bunches of per then step combined distance minimum meet the demands until final cluster number.

The sensitivity of calculating in the above-mentioned partition method is new sensitivity, has taken into account the quasi-stable state external characteristic of power equipment reactive voltage, can reflect the response characteristic of real system more accurately, has improved the computational accuracy of sensitivity analysis.By this partition method, idle basic decoupling zero between each control area of being divided on the one hand, can determine suitable subregion number automatically on the other hand, avoided the artificial where the shoe pinches of determining the subregion number, can be implemented in line self adaptation " soft " subregion, the composition of subregion number and subregion can change along with the development and change of electrical network automatically.

Because the limitation of existing CSVC method on control technology and system configuration even all adopt above-mentioned new technology in each concrete link of existing CSVC method, also still is difficult to guarantee control performance.For example, even in the step 1 of above-mentioned CSVC method, use online adaptive " soft " partitioning technique,, still be difficult to follow the tracks of the variation of electric network composition and operational mode under offline environment because this step 1 is to finish; In addition, according to existing CSVC method, need still to be the geographical distinct area CSVC system that distributes that goes up of each control area exploitation that this makes the control area in case after setting up, just be not easy to change.Therefore, existing CSVC method electrical network (as: Chinese electrical network) rapid for development and that operational mode often changes is difficult to guarantee control performance.

Summary of the invention

The objective of the invention is for overcoming the weak point of existing voltage control method, voltage control method based on soft sectoring has been proposed in a kind of new electric power system, this method can be according to the variation of electric network composition, self adaptation provides the system partitioning scheme under the current state, for setting up and find the solution the CSVC Mathematical Modeling in each control area, CSVC control is carried out in each control area, thereby be not required to be each control area stand-alone development zone C SVC system, not only better guarantee the control performance of CSVC, and reduced system's construction investment, be particularly suitable for grid structure and change frequent electrical network.

Based on the voltage control method of soft sectoring, may further comprise the steps in the electric power system that the present invention proposes:

1) obtains required real-time power network data from grid control centre;

2) utilize the real-time power network data that collect to carry out online subregion, at first according to criterion judge whether needs again subregion (concrete method can be for judging electric network compositions by some known topology analyzing methods, whether operational mode changes, perhaps whether the time that implement the subzone on the judging distance reaches prior specified value), set up as the criterion condition, then carry out subregion, perhaps by manually starting subregion (concrete partition method can select for use some known partition methods whole electrical network dynamically to be resolved into the control area of N idle mutual decoupling zero), and be each control area selection maincenter bus and controlled generator;

3) utilizing grid control centre is that each control area is set up CSVC (coordination secondary voltage control) Mathematical Modeling (this CSVC Mathematical Modeling can be selected the CSVC Mathematical Modeling that proposes in the existing research for use) respectively;

4) by grid control centre the CSVC Mathematical Modeling of each control area is found the solution controlled amount (can adopt existing mathematical method (for example, set algorithm works) to find the solution CSVC Mathematical Modeling, controlled amount) respectively;

5) each the control area controlled quentity controlled variable that will be tried to achieve by grid control centre is handed down to one-level voltage control link or the execution of local voltage controller by communication network;

6) arrive when the voltage control cycle one that configures in advance, return step 1: step 1 forms a closed loop to step 5, carries out voltage control by the voltage control loop cycle.

Said method can be achieved by utilize routine techniques to set up the voltage control system that a software combines with hardware at grid control centre.

Characteristics of the present invention and effect

In the electric power system proposed by the invention based on the structure of the voltage control method of soft sectoring as shown in Figure 2, among the figure, heavy line is represented the electrical network boundary line, chain-dotted line is represented each boundary line, control area that online adaptive partition is divided, do not need installation region CSVC system in each control area, the CSVC software and hardware system (hereinafter to be referred as the CSVC system) that only needs a total overall situation under the overall leadership is installed in grid control centre is implemented voltage control method of the present invention.Promptly obtain electrical network network topology, device parameter and real-time telemetry remote signalling data by existing EMS system in the grid control centre, in closed-loop control process of the present invention, its partition method is analyzed network configuration, the operational mode of electrical network according to the real time data that collects, when judging that the criterion condition that needs again subregion is set up, automatically carry out the online adaptive subregion, each control area of being divided is virtual " soft " control area, satisfies the condition of idle basic decoupling zero between each control area.Result according to online " soft " subregion sets up the CSVC Mathematical Modeling respectively for each control area automatically, and finds the solution the CSVC Mathematical Modeling.At last, find the solution the one-level voltage control link execution that the controlled quentity controlled variable that obtains is handed down to each control area, finish the CSVC closed-loop control.

Existing based on the CSVC method of " firmly " subregion in (see figure 1); subregion is finished under offline environment; and by manually being solidified in the control area; be geographical each difference installation region, control area CSVC system that distributes that goes up then; implement distributed control; the problem of bringing like this is when electric network composition or operational mode change; be difficult to reflect in time the variation in power grid control zone on the one hand; eliminate the coupling between the control area; variation along with the control area had both needed also to be difficult to protect existing investment for newly-increased control area increases zone C SVC system on the other hand.

Be based on the voltage control method of soft sectoring and the remarkable difference of existing CSVC method in the electric power system proposed by the invention:

The one, is in voltage control method proposed by the invention, the power grid control dividing region has been included in the CSVC closed loop controlling structure, the online adaptive subregion is as a step in the voltage control method, utilize the online real time data of obtaining, variation from motion tracking electric network composition or operational mode, carry out online adaptive " soft " subregion, therefore can adapt to the variation of electric network composition or operational mode, thereby decoupling zero mutually between each the CSVC control area that guarantees to be divided better guarantees the control performance of CSVC;

The 2nd, is in voltage control method proposed by the invention, the voltage control of all control areas all concentrates in the grid control centre voltage control system unified implementation by total overall situation under the overall leadership, thereby be not required to be each control area and build the distinct area voltage control system respectively, reduced investment;

Description of drawings

Fig. 1 is the partitioned organization figure of existing CSVC method based on " firmly " subregion.

Fig. 2 is the control area structure chart of voltage control method based on soft sectoring proposed by the invention.

Fig. 3 is the step block diagram of voltage control method based on soft sectoring proposed by the invention.

Fig. 4 is in the specific embodiment, IEEE 39 node system figure.

Fig. 5 is in the specific embodiment, IEEE 39 node systems 6 subregion schematic diagrames.

Fig. 6 is a comparison diagram as a result of implementing adaptive partition at two kinds of operational modes of Jiangsu electrical network.

Fig. 7 is in the specific embodiment, and control area V uses voltage control method based on soft sectoring proposed by the invention and existing secondary voltage control (SVC) the method control front and back generator reactive comparison diagram of exerting oneself.

Embodiment

Voltage control method based on soft sectoring in the electric power system that the present invention proposes reaches embodiment in conjunction with the accompanying drawings, is described in detail as follows:

Based on the voltage control method of soft sectoring as shown in Figure 3, may further comprise the steps in the electric power system that the present invention proposes:

1) real-time power network data acquisition: obtain required real-time power network data from grid control centre;

2) online subregion: utilize the real-time power network data that collect, at first according to criterion judge whether needs again subregion (concrete method can be for judging electric network compositions by some known topology analyzing methods, whether operational mode changes, perhaps whether the time that implement the subzone on the judging distance reaches prior specified value), set up as the criterion condition, then carry out subregion, perhaps by manually starting subregion (concrete partition method can select for use some known partition methods whole electrical network dynamically to be resolved into the control area of N idle mutual decoupling zero), and be each control area selection maincenter bus and controlled generator;

3) each control area is set up the CSVC Mathematical Modeling respectively: utilizing grid control centre is that each control area is set up CSVC Mathematical Modeling (this CSVC Mathematical Modeling can be selected the CSVC Mathematical Modeling that proposes in the existing research for use) respectively;

4) each control area is found the solution the CSVC Mathematical Modeling respectively, controlled amount: the CSVC Mathematical Modeling of each control area is found the solution respectively by grid control centre, controlled amount (can adopt existing mathematical method (for example, set algorithm works) to find the solution CSVC Mathematical Modeling, controlled amount);

5) each control area is exported controlled quentity controlled variable respectively: with each control area controlled quentity controlled variable of trying to achieve in the step 4, be handed down to one-level voltage control link or local voltage controller execution by communication network by grid control centre;

6) arrive when the voltage control cycle one that configures in advance, return step 1: step 1 forms a closed loop to step 5, carries out voltage control by the voltage control loop cycle.

Above-mentioned steps of the present invention is carried out emulation experiment with the IEEE39 node system and is made embodiment, further specifies as follows:

IEEE39 node system structure as shown in Figure 4, heavy line is represented bus (being the node in the nodal analysis method) among the figure, the other numeral node serial number of bus, downward arrow is represented load, band ' G ' character representation generator in the symbol circle.Its voltage control method specifically may further comprise the steps:

1) real-time power network data acquisition: obtain required real-time power network data and setting voltage control cycle from grid control centre;

For present embodiment, can directly import the data of IEEE39 node system, need connect in the grid control centre existing EMS system to obtain network topology, device parameter and real-time remote measure and communication value for practical power systems.

2) online subregion: utilize the real-time power network data that collect, at first according to criterion judge whether needs again subregion (concrete method can be for judging electric network compositions by some known topology analyzing methods, whether operational mode changes, perhaps whether the time that implement the subzone on the judging distance reaches prior specified value, perhaps by manually starting subregion etc.), set up as the criterion condition, then carry out subregion (concrete partition method can select for use some known partition methods whole electrical network dynamically to be resolved into the control area of N idle mutual decoupling zero), and for selecting maincenter bus and controlled generator (concrete system of selection can be selected some known methods for use, for example system of selection of existing CSVC method in the background technology) in each control area.

Because the period ratio of closed-loop control is shorter, subregion does not need so frequent operation, therefore, can carry out (being all after dates of N closed-loop control) after N time in closed-loop control, carries out a subzone again.In the present embodiment, online subregion can be in the cycle of operation of appointment (for example 10 cycles have been carried out in the CSVC closed-loop control) or the situation by switch changed position in the detection of grid, according to the real time data that collects the topological structure and the operational mode of electrical network are analyzed, when electric network composition or the bigger variation of operational mode generation, utilize known cluster partitioning algorithm based on reactive source control space, whole electrical network is carried out subregion again: determine the only number of partitions of electrical network at first automatically, then electrical network is carried out online adaptive partition.

In the present embodiment, adopt described in the background technology based on reactive source control space clustering partitioning algorithm, at first obtain subregion combined distance curve, automatically determine that optimally partitioned number is 6, then IEEE 39 node systems are divided into 6 control areas shown in Figure 5, dot each boundary line, control area among the figure, indicate each control area numbering with Roman number.

Selected corresponding maincenter bus and controlled generator in each control area, in the present embodiment, the Centroid of selecting each control area electrical distance selects all generators as controlled generator as the maincenter bus.The corresponding information of each control area is as shown in table 2:

Table 2IEEE39 node system secondary voltage control partition information

The control area	The maincenter bus	Controlled generator
			4 regional 5 zones 6,3 zones, 2 zones, 1 zone, zone	BUS-1 BUS-28 BUS-6 BUS-3 BUS-19 BUS-23	BUS-39 BUS-38 BUS-31，BUS-32 BUS-30，BUS-37 BUS-33，BUS-34 BUS-35，BUS-36

In order the necessity and the validity of carrying out the online adaptive subregion to be described better, to provide again that the result who utilizes this partitioning algorithm that two kinds of operational modes of Jiangsu electrical network are carried out subregion contrasts as shown in Figure 6 here.Among the figure, what different filling modes marked is different control areas, can see, when the operational mode of electrical network has greatly changed, be coupled during original subregion control area four (shown in Fig. 6 a) closely, carry out being broken down into two control areas four and control area five (shown in Fig. 6 b) that coupling is loose behind the online subregion, visible online subregion can reflect the variation of electric network composition and operational mode in time.

3) each control area is set up the CSVC Mathematical Modeling respectively: set up the CSVC Mathematical Modeling by grid control centre respectively for each control area, present embodiment adopts the existing CSVC Mathematical Modeling of taking into account the idle nargin equilibrium of exerting oneself, its target function as shown in the formula, promptly shown in the formula (1):

$\underset{{\mathrm{\&Delta;Q}}_g}{\min }\{W_p{\left|\right|\mathrm{\&alpha;}\&CenterDot;(V_p-V_p^{\mathrm{ref}})+C_g\mathrm{\&Delta;}{Q}_g\left|\right|}^2+W_q{\left|\right|\frac{Q_g+{\mathrm{\&Delta;Q}}_g-Q_g^{\min }}{Q_g^{\max }-Q_g^{\min }}\left|\right|}^2\}$

First of this target function guarantees maincenter busbar voltage and set point deviation minimum, and second to impel generator reactive to be tending towards nargin bigger, more balanced.

The security constraints of CSVC Mathematical Modeling adopts formula (2), (3), (4), (5) described in the background technology:

$|C_{\mathrm{vg}}\&CenterDot;\mathrm{\&Delta;}{Q}_g|\&le;\mathrm{\&Delta;}{V}_H^{\max }$

$V_H^{\min }\&le;V_H+C_{\mathrm{vg}}\&CenterDot;\mathrm{\&Delta;}{Q}_g\&le;V_H^{\max }$

$V_p^{\min }{\&le;V}_p+C_g\&CenterDot;\mathrm{\&Delta;}{Q}_g\&le;V_p^{\max }$

$Q_g^{\min }{\&le;Q}_g+{\mathrm{\&Delta;Q}}_g{\&le;Q}_g^{\max }$

By top constraint equation, can prevent that control operation from causing excessive fluctuation to electrical network, and guarantee can not cause controlled generator high side voltage V after the control _H, maincenter busbar voltage V _pAnd the idle Q that exerts oneself of controlled generator _gProduce out-of-limit.

As can be seen from Figure 7, after the CSVC Mathematical Modeling of the idle nargin equilibrium of exerting oneself is taken into account in employing, impelled regional interior generator operation, exert oneself more balanced state bigger, improved the voltage stability of electric power system in idle nargin.

In addition, the sensitivity of adopting in the CSVC Mathematical Modeling (is C _g, C _Vg) be the new sensitivity described in the background technology, improved the precision of control decision.

4) each control area is found the solution the CSVC Mathematical Modeling respectively, controlled amount: the present invention can adopt existing mathematical method (for example, set algorithm works) to find the solution the CSVC Mathematical Modeling of each control area, controlled amount by grid control centre.

In the present embodiment, adopt the existing set algorithm that works to find the solution the CSVC Mathematical Modeling of structure in the above-mentioned steps 3, try to achieve controlled quentity controlled variable Δ Q _g

5) each control area is exported controlled quentity controlled variable respectively:, be handed down to one-level voltage control link (or local voltage controller) respectively by communication network and carry out each control area controlled quentity controlled variable of trying to achieve in the step 4 by grid control centre.

Present embodiment is with the controlled quentity controlled variable Δ Q that tries to achieve in the step 4 _g, by the new sensitivity Matrix C _VgBe converted into controlled generator high-voltage side bus voltage Δ V _HSet point (C _VgBe of the sensitivity of controlled generator high-voltage side bus voltage to controlled generator reactive), as the formula (9)

ΔV _H＝C _vgΔQ _g (9)

At last, with controlled quentity controlled variable Δ V _HBeing handed down to each regional one-level voltage control link (or local voltage controller) carries out.

6) arrive when the voltage control cycle one that configures in advance, return step 1: step 1 forms a closed loop to step 5, is undertaken by the voltage control loop cycle that configures in advance.

The voltage control cycle can not waited at several seconds to several minutes, determine jointly by the control situation (as the response speed of one-level voltage control link, the delay that measures collection, controlled target etc.) of practical power systems, the multiple factors such as voltage static characteristic of load, for example can set 1 minute is a voltage control cycle, and circulation is carried out step 1 to step 5.

Claims (4)

1, in the electric power system that proposes of the present invention based on the voltage control method of soft sectoring, it is characterized in that, may further comprise the steps:

1) obtains required real-time power network data from grid control centre;

2) utilize the real-time power network data that collect, judge whether needs subregion again, set up, then carry out online subregion as the criterion condition according to criterion, and for selecting maincenter bus and controlled generator in each control area;

3) utilizing grid control centre is that each control area is set up coordination secondary voltage control Mathematical Modeling respectively;

4) by grid control centre the coordination secondary voltage control Mathematical Modeling of each control area is found the solution controlled amount respectively;

5) each the control area controlled quentity controlled variable that will be tried to achieve by grid control centre is handed down to one-level voltage control link or the execution of local voltage controller by communication network;

6) arrive when the voltage control cycle one that configures in advance, return step 1: step 1 forms a closed loop to step 5, carries out voltage control by the voltage control loop cycle.

2, in the electric power system as claimed in claim 1 based on the voltage control method of soft sectoring, it is characterized in that, described step 2) according to criterion judge whether needs again the concrete method of subregion for to judge by known topology analyzing method whether electric network composition, operational mode change, if change then subregion again; Perhaps whether the time that implement the subzone on the judging distance reaches prior specified value, if reach specified value, and subregion again then; Perhaps by manually starting subregion.

3, in the electric power system as claimed in claim 1 based on the voltage control method of soft sectoring, it is characterized in that, described step 2) partition method dynamically resolves into whole electrical network the control area of N idle mutual decoupling zero for adopting the cluster partitioning algorithm based on reactive source control space.

4, in the electric power system as claimed in claim 1 based on the voltage control method of soft sectoring, it is characterized in that, described step 3) is set up the method for coordinating the secondary voltage control Mathematical Modeling and is adopted the coordination secondary voltage control Mathematical Modeling of taking into account the idle nargin equilibrium of exerting oneself, and its target function is shown below:

$\underset{{\mathrm{\&Delta;Q}}_g}{\min }\{W_p{\left|\right|a\&CenterDot;(V_p-V_p^{\mathrm{ref}})+C_g{\mathrm{\&Delta;Q}}_g\left|\right|}^2+W_q{\left|\right|\frac{Q_g+{\mathrm{\&Delta;Q}}_g-Q_g^{\min }}{Q_g^{\max }-Q_g^{\min }}\left|\right|}^2\}$

Its constraints is:

$|C_{\mathrm{vg}}\&CenterDot;{\mathrm{\&Delta;Q}}_g|\&le;\mathrm{\&Delta;}{V}_H^{\max }$

$V_H^{\min }\&le;V_H+C_{\mathrm{vg}}\&CenterDot;\mathrm{\&Delta;}{Q}_g\&le;V_H^{\max }$

$V_p^{\min }\&le;V_p+C_g\&CenterDot;\mathrm{\&Delta;}{Q}_g\&le;V_p^{\max }$

$Q_g^{\min }\&le;Q_g+{\mathrm{\&Delta;Q}}_g\&le;Q_g^{\max }$

Images