CN106712035A - Electric system economical dispatching method - Google Patents

Abstract

The present invention discloses an electric system economical dispatching method, and relates to the electric system optimizing operation technology field. The reliability of the economical dispatching of the electric system is improved. The method comprises: decoupling the electric system into a coordination center and a plurality of areas; according to the coordination center and a plurality of areas, establishing an electric system economical dispatching model, wherein the target of the electric system economical dispatching model is the sum of the total cost of power production of a conventional power unit in the electric system and the discarding cost of the new energy power generation load shedding penalty; decomposing the electric system economical dispatching model into an inter-area coordination model and a plurality of area dispatching models, wherein the area dispatching models are in one-to-one correspondence with the areas, the inter-area coordination model corresponds to the coordination center, and the inter-area coordination model performs distributed optimization of the boundary nodes of each area; and calculating the economical dispatching result of the electric system according to the inter-area coordination model and the area dispatching models.

Description

A kind of Economic Dispatch method

Technical field

The present invention relates to electric power system optimization running technology field, more particularly to a kind of Economic Dispatch method.

Background technology

At present, the proposition of energy environment issues, has promoted developing rapidly for new energy electric field, for example, wind energy electric field, the sun Energy electric field, tide electric field etc..However, because new energy electric field has larger randomness and fluctuation, thus resulting in and being incorporated with The power system of new energy electric field faces more severe form, and especially the economic load dispatching of power system faces larger choosing War.

The economic load dispatching of power system refers to ensure power system security, reliability service and meeting the quality of power supply, electricity consumption On the premise of needs, the unit output to each unit in power system is scheduled, and makes energy consumption, the running cost of whole power system With minimum, to obtain the economic benefit of maximum.The economic load dispatching of existing power system is generally made by control centre's unification, i.e., The economic load dispatching result of power system is obtained by the way of centralized optimization, for example, can be by conventional power unit in power system The total generating expense of (such as fired power generating unit, Hydropower Unit etc.) within dispatching cycle (is for example abandoned wind, abandons the sun with new energy is abandoned Can, abandon tide) generating cutting load rejection penalty sum as Economic Dispatch target, that is, solve power system in often Total generating expense of rule unit (such as fired power generating unit, Hydropower Unit etc.) within dispatching cycle (is for example abandoned wind, is abandoned with new energy is abandoned Solar energy, abandon tide) the minimum economic load dispatching result of generating cutting load rejection penalty sum, to realize that power system is entered to pass through Ji scheduling, now, control centre generally needs to obtain the whole network data of power system.However, with new energy electric field progressively It is incorporated to, the scale of power system constantly expands, when carrying out economic load dispatching to power system by control centre's unification, institute of control centre The whole network data of acquisition is huger and numerous and diverse, easily causes communication blockage and shortage of data, in turn results in the economy of power system The reliability of scheduling is poor.

The content of the invention

It is an object of the invention to provide a kind of Economic Dispatch method, for solving existing power system The poor problem of the reliability of economic load dispatching.

To achieve these goals, the present invention provides following technical scheme：

A kind of Economic Dispatch method, it is characterised in that including：

Step S100, it is Consultation Center and multiple region by power system decoupling；

Step S200, according to the Consultation Center and multiple regions, set up Economic Dispatch model, its In, the target of the Economic Dispatch model is total generating of the conventional power unit in scheduling duration in the power system Expense and abandon generation of electricity by new energy cutting load rejection penalty sum；

Step S300, it is interregional Coordination Model and multiple subdispatch by the Economic Dispatch model decomposition Model, wherein, multiple subdispatch models and multiple regions correspond, the interregional Coordination Model with it is described Consultation Center's correspondence, the interregional Coordination Model carries out distributed optimization to the boundary node in each region；

Step S400, according to the interregional Coordination Model and multiple subdispatch models, calculate power system Economic load dispatching result.

It is Consultation Center and multiple areas by power system decoupling in the Economic Dispatch method that the present invention is provided Domain, then sets up Economic Dispatch model, then by Economic Dispatch according to Consultation Center and multiple regions Model decomposition is interregional Coordination Model and multiple subdispatch models, is then adjusted according to interregional Coordination Model and multiple regions Degree model, calculates the economic load dispatching result of power system.Therefore, in the present invention, the economic load dispatching result of power system is calculated When, multiple subdispatch models are calculated respectively, each subdispatch model is distributed using interregional Coordination Model Formula optimizes, i.e., the economic load dispatching result in each region is calculated by the subdispatch model corresponding to the region, using region Between Coordination Model distributed optimization is carried out to the boundary node in the region, thus, Consultation Center utilize interregional Coordination Model When boundary node to each region carries out distributed optimization, Consultation Center need to only obtain the variable of the boundary node in each region, and Other variables in each region, i.e. Consultation Center need not be obtained need not obtain the whole network data of power system, therefore will not be because needing The whole network data to be obtained and cause communication blockage and shortage of data, such that it is able to improve the reliability of the economic load dispatching of power system Property.

Brief description of the drawings

Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes a part of the invention, this hair Bright schematic description and description does not constitute inappropriate limitation of the present invention for explaining the present invention.In the accompanying drawings：

Fig. 1 is the flow chart one of Economic Dispatch method in the embodiment of the present invention；

Fig. 2 is the flowchart 2 of Economic Dispatch method in the embodiment of the present invention；

Fig. 3 is power system decoupling schematic diagram in the embodiment of the present invention.

Specific embodiment

It is attached with reference to specification in order to further illustrate Economic Dispatch method provided in an embodiment of the present invention Figure is described in detail.

Fig. 1 is referred to, Economic Dispatch method provided in an embodiment of the present invention includes：

Step S100, it is Consultation Center and multiple region by power system decoupling.

Step S200, according to Consultation Center and multiple region, set up Economic Dispatch model, wherein, power train The target of system economic load dispatching model is for conventional power unit is in the total generating expense dispatched in duration in power system and abandons new energy hair TURP load rejection penalty sum.

Step S300, it is interregional Coordination Model and multiple subdispatch mould by Economic Dispatch model decomposition Type, wherein, multiple subdispatch models are corresponded with multiple regions, and interregional Coordination Model is corresponding with Consultation Center, region Between Coordination Model distributed optimization is carried out to the boundary node in each region.

Step S400, according to interregional Coordination Model and multiple subdispatch model, calculate the economic load dispatching of power system As a result.

Specifically, during by power system decoupling for Consultation Center and multiple region, can be using construction virtual region and multiple The method of boundary node variable processed, for example, refer to Fig. 3, is Consultation Center and two regions by power system decoupling As a example by illustrate, first with construction virtual region method power system is configured to two regions, respectively region a and area Domain b, two regions are by an interregional interconnector connection, one end of the interregional interconnector and the boundary node m of region a Connection, the other end of the interregional interconnector is connected with the boundary node n of region b, wherein, boundary node can be understood as certain The node that one region is connected with other regions, then using the method for replicating boundary node variable, the phase angle of boundary node m is become The phase angle variable of amount and boundary node n is replicated once, respectivelyWithWherein,WithBelong to region a, WithBelong to region b.When forming Consultation Center, the method using boundary node variable is replicated, by the phase angle variable of boundary node m Phase angle variable with boundary node n is replicated once again, respectivelyWithIn this way, to same boundary node, There is one group of corresponding variable to represent the phase angle of the boundary node for each relevant range and Consultation Center.

By power system decoupling for Consultation Center and multiple regions after, then the Consultation Center that can be formed according to decoupling and many Individual region, sets up Economic Dispatch model, wherein, the target of Economic Dispatch model can be set as electric power Conventional power unit is in the total generating expense dispatched in duration in system and abandons generation of electricity by new energy cutting load rejection penalty sum, that is, solve The economic load dispatching of the power system for obtaining result it is required that in power system total generating expense of the conventional power unit in scheduling duration and Abandon generation of electricity by new energy cutting load rejection penalty sum minimum, the Economic Dispatch model is the economy of whole power system All of parameter and interregional parameter, the i.e. whole network data including power system in scheduling model, including regional.

After completing Economic Dispatch model, cascade analytic approach using target and decomposed containing partially polymerized many cuttings Algorithm, is the interregional Coordination Model and one-to-one corresponding corresponding to Consultation Center by Economic Dispatch model decomposition In the multiple subdispatch models to individual region, Consultation Center receives variable (such as phase of the boundary node uploaded by each region Angle variable), and according to the variable in interregional Coordination Model coordinating calculating center corresponding to each boundary node, in then coordinating The variable corresponding to each boundary node is issued to corresponding region in the heart, is carried out with the boundary node to each region distributed excellent Change.

After being interregional Coordination Model and multiple subdispatch models by Economic Dispatch model decomposition, according to area Coordination Model and multiple subdispatch models between domain, by repeatedly dividing between interregional Coordination Model and each subdispatch model Cloth optimizes, and is calculated the economic load dispatching result of power system, the economic load dispatching result of power system by each region region Scheduling result is constituted.

From the foregoing, in Economic Dispatch method provided in an embodiment of the present invention, being by power system decoupling Consultation Center and multiple regions, then set up Economic Dispatch model according to Consultation Center and multiple regions, then will Economic Dispatch model decomposition is interregional Coordination Model and multiple subdispatch models, then according to interregional coordination Model and multiple subdispatch models, calculate the economic load dispatching result of power system.Therefore, in embodiments of the present invention, calculate During the economic load dispatching result of power system, multiple subdispatch models are calculated respectively, using interregional Coordination Model pair Each subdispatch model carries out the economic load dispatching result in distributed optimization, i.e. each region by the subdispatch corresponding to the region Model is calculated, and distributed optimization is carried out to the boundary node in the region using interregional Coordination Model, thus, Consultation Center When distributed optimization is carried out to the boundary node in each region using interregional Coordination Model, Consultation Center need to only obtain each region Boundary node variable, and other variables in each region need not be obtained, i.e., Consultation Center need not obtain the complete of power system Network data, thus will not because need obtain whole network data and caused by communication blockage and shortage of data, such that it is able to improve electric power The reliability of the economic load dispatching of system.

In addition, in embodiments of the present invention, the economic load dispatching result in each region is by the subdispatch corresponding to the region Model is calculated, and distributed optimization is carried out to the boundary node in the region using interregional Coordination Model, thus, Consultation Center When distributed optimization is carried out to the boundary node in each region using interregional Coordination Model, Consultation Center need to only obtain each region Boundary node variable, and other variables in each region need not be obtained, i.e., Consultation Center need not obtain the complete of power system Network data.It is thereby achieved that the independent scheduling in each region, and realize the protection of the data-privacy of some regions.

Furthermore, it is interregional Coordination Model and many by Economic Dispatch model decomposition in embodiments of the present invention Individual subdispatch model, will larger PROBLEM DECOMPOSITION be multiple small problems, then multiple small problems are entered respectively Row is calculated, thus can simplify the process of the economic load dispatching result for calculating power system, it is possible to is improved and is calculated power system The efficiency of economic load dispatching result, simultaneously as the negligible amounts of the parameter involved by each small problem, such that it is able to further Improve the reliability of the economic load dispatching of power system.

Fig. 1 and Fig. 2 is referred to, before step S100, Economic Dispatch method provided in an embodiment of the present invention Also include：

Step S10, determination carry out the scheduling duration of economic load dispatching to power system, and scheduling duration is averagely divided into n_T The individual period, wherein, n_T≥2。

For example, the scheduling duration for economic load dispatching being carried out to power system can be set as one day, i.e., 24 hour, will adjust Degree duration is averagely divided into n_TThe individual period, wherein, n_TIn the individual period, the duration of each period is identical, for example, can be by 24 hours 24 periods are divided into, are per hour a period, or, 96 periods can be divided into by 24 hours, it is within every 15 minutes one The individual period.

Please continue to refer to Fig. 2, after step sloo, before step S200, power system provided in an embodiment of the present invention Economic load dispatching method also includes：

Step S100 ', to each region setting prediction scene, and to new energy electric field region extraction multiple error fields Scape.

Specifically, using scene method to each region setting prediction scene, and the region with new energy electric field can be taken out Error scene is taken, for example, to the region with new energy electric field, 100 error scenes can be extracted, setting for prediction scene is completed After the fixed and extraction of error scene, when setting up Economic Dispatch model, Economic Dispatch model includes prediction The relevant parameter of scene and the relevant parameter of error scene, Economic Dispatch model consider the random of new energy electric field Property and waveform, and the random of new energy electric field in power system can be tackled with this Economic Dispatch model set up Property and fluctuation.

Please continue to refer to Fig. 2, after step S300, before step S400, power system provided in an embodiment of the present invention Economic load dispatching method also includes：

Step S300 ', by subdispatch model decomposition be regional prediction model of place and domain error model of place.

It is area by the subdispatch model decomposition in the region for the region with new energy electric field in step S300 ' Model of place and domain error model of place are predicted in domain, and in the economic load dispatching result of zoning, field is predicted in first zoning Scape model, is then carried out repeatedly random using domain error model of place to the result obtained after zoning prediction model of place Optimization.Therefore, in embodiments of the present invention, in the economic load dispatching result of zoning, also by one of the region big problem Two minor issues for corresponding respectively to predict scene and error scene are decomposed into, thus the economic load dispatching of zoning can be simplified The process of result, it is possible to improve the efficiency of the economic load dispatching result of zoning, simultaneously as involved by each small problem Parameter negligible amounts, such that it is able to improve the reliability of the economic load dispatching in region, and then further improve power system The reliability of economic load dispatching.

In above-described embodiment, Economic Dispatch model can be：

Object function：

Constraints：

The prediction context restrictions condition in region：

B_aP_a+D_aθ_a≤E_a；1≤a≤N (2)

The error scene constraints in region：

B_a,sP_a,s+D_a,sθ_a,s≤E_a,s+G_a,sP_a+H_a,sθ_a；1≤a≤N,1≤s≤S_a (3)

The constraints of Consultation Center：

Coupling constraint condition between Consultation Center and region：

In above-mentioned formula, f_aIt is the prediction scene total cost of region a；f_a,sFor the error scene of region a abandons generation of electricity by new energy Expense；N is the number in region；It is the number of conventional power unit in a of region；It is the number of region a new energy units；For In the number of the load bus of period t region a；S_aIt is the number of the error scene of region a；It is in period t region a, pre- The active power output of conventional power unit i under survey scene；WithThe generating cost coefficient of conventional power unit i in respectively region a；It is that, in period t, a new energy unit w in the case where scene is predicted in region abandon generation of electricity by new energy power；q_WFor abandoning for region a is new Energy generating rejection penalty coefficient；It is the cutting load power of region a load bus d in the case where scene is predicted in period t；q_D It is the cutting load rejection penalty coefficient of region a；p_sIt is the probability of the error scene s of region a, p_s=1/S_a；It is in the period T, region a the new energy unit w under error scene s abandon generation of electricity by new energy power；It is that region a is in error in period t The cutting load power of load bus d under scene s.

P_aFor region a predict scene under each conventional power unit day part matrix of exerting oneself, exert oneself matrix P_aUnit be region A conventional power unit i exerting oneself in period t in the case where scene is predicted, exert oneself matrix P_aForMatrix orMatrix； θ_aFor region a in the case where scene is predicted each node in the phase angle matrix of day part, node includes：Node (load section in a of region Point, non-load bus etc.), the boundary node that is connected with region a in the boundary node of region a and other regions, phase angle matrix θ_aUnit for region a predict scene under a certain node period t phase angle；B_a、D_aAnd E_aBe region a predict scene under Parameter matrix；P_a,sFor region a under error scene s each conventional power unit in the matrix of exerting oneself of day part, exert oneself matrix P_a,s's Unit is region a conventional power unit i exerting oneself in period t under error scene s, and exert oneself matrix P_a,sForMatrix orMatrix；θ_a,sFor region a under error scene s each node day part phase angle matrix, phase angle matrix θ_a,sUnit For region a under error scene s phase angle of a certain node in period t；B_a,s、D_a,s、E_a,s、G_a,sAnd H_a,sRegion a is in error Parameter matrix under scene s；TL^ab,aIt is the boundary node intersection being connected with region b in a of region；TL^ab,bIt is region b Zhong Yu areas The boundary node intersection that domain a is connected, and m and n is corresponding two boundary nodes of connecting line of join domain a and region b； For Consultation Center correspond to region a in boundary node m day part phase angle matrix, phase angle matrixUnit be Consultation Center pair Should in a of region phase angles of the boundary node m in period t；Boundary node n is in day part in corresponding to region a for Consultation Center Phase angle matrix, phase angle matrixUnit for Consultation Center correspond to region a in boundary node n period t phase angle；It is association Tune center correspond to region b in boundary node m day part phase angle matrix, phase angle matrixUnit correspond to for Consultation Center Phase angles of the boundary node m in period t in the b of region；For Consultation Center correspond to region b in boundary node n day part phase Angular moment battle array, phase angle matrixUnit for Consultation Center correspond to region b in boundary node n period t phase angle；It is region a Middle boundary node m day part phase angle matrix, phase angle matrixUnit be phase angles of the boundary node m in period t in a of region；For in a of region boundary node n day part phase angle matrix, phase angle matrixUnit in a of region boundary node n in period t Phase angle.

Above-mentioned Economic Dispatch model is compact, practically, in above-mentioned Economic Dispatch model,

The prediction context restrictions condition in region includes：

It is the matrix of exerting oneself of region a each conventional power units in the case where scene is predicted, matrix of exerting oneself in period tIt is row matrix Or column matrix, matrix of exerting oneselfUnit be that in period t, a conventional power unit i in the case where scene is predicted in region exert oneself；Be when The matrix of exerting oneself of section t, region a each new energy unit in the case where scene is predicted, matrix of exerting oneselfIt is row matrix or column matrix, exerts oneself MatrixUnit be that in period t, a new energy unit w in the case where scene is predicted in region exert oneself；It is that region a is pre- in period t The matrix of loadings of each load bus, matrix of loadings under survey sceneIt is row matrix or column matrix, matrix of loadingsUnit be when The load of section t, region a the load bus d in the case where scene is predicted；It is region a each new energy in the case where scene is predicted in period t Unit abandons generation of electricity by new energy power matrix, abandons generation of electricity by new energy power matrixIt is row matrix or column matrix, abandons new energy Generated output matrixUnit in period t, a new energy unit w in the case where predicting scene in region abandon generation of electricity by new energy power；It is the cutting load power matrix of region a each load buses in the case where scene is predicted, cutting load power matrix in period t It is row matrix or column matrix, cutting load power matrixUnit be region a load bus d in the case where scene is predicted in period t Load；B^aIt is the bus admittance matrix ignored branch resistance and set up to ground leg of region a；It is in period t, region The phase angle matrix of a each nodes in the case where scene is predicted, phase angle matrixIt is row matrix or column matrix, phase angle matrixUnit be when The phase angle of section t, region a a certain node in the case where scene is predicted；It is the active power output lower limit of conventional power unit i in a of region；It is area The active power output upper limit of conventional power unit i in a of domain；Be in period t, region a predict scene under new energy unit w it is active Exert oneself；It is the maximum active power output of the new energy unit w in period t region a；For in a of region conventional power unit i it is active go out Power climbing limitation；Limited for the active power output of conventional power unit i in a of region comes down；It is that, in period t-1, region a is in prediction The active power output of conventional power unit i under scene；N_JIt is the number of circuit relevant with region a in power system, circuit includes region a Internal wiring and join domain a and other regions interregional interconnector；Be the circuit j relevant with region a most Big transmission power value；It is the reactance value of the circuit j relevant with region a；It is the section of the circuit j under period t, prediction scene The phase angle of point j1；It is the phase angle of the node j2 of circuit j under period t, prediction scene；S_BOn the basis of be worth, S_B=100MW； It is conventional power unit i adjustable increments of exerting oneself in 10 minutes in a of region；It is that, in period t, region a is under error scene s The active power output of conventional power unit i.

The error scene constraints in region includes：

It is the matrix of exerting oneself of region a each conventional power units under error scene s, matrix of exerting oneself in period tIt is row Matrix or column matrix, matrix of exerting oneselfUnit be that, in period t, a conventional power unit i under the error scene s in region exert oneself； In period t, the matrix of exerting oneself of region a each new energy units under error scene s, matrix of exerting oneselfIt is row matrix or row square Battle array, matrix of exerting oneselfUnit be that, in period t, a new energy unit w under the error scene s in region exert oneself；It is in the period The matrix of loadings of t, region a each load bus under error scene s, matrix of loadingsIt is row matrix or column matrix, matrix of loadingsUnit be the load of region a load bus d under error scene s in period t；It is that, in period t, region a is by mistake Each new energy unit abandons generation of electricity by new energy power matrix under difference scene s, abandons generation of electricity by new energy power matrixIt is row matrix Or column matrix, abandon generation of electricity by new energy power matrixUnit in period t, region a new energy unit w under error scene s Abandon generation of electricity by new energy power；It is the cutting load power matrix of region a each load buses under error scene s in period t, Cutting load power matrixIt is row matrix or column matrix, cutting load power matrixUnit be that, in period t, region a exists The load of load bus d under error scene s；It is the phase angle matrix of region a each nodes under error scene s, phase in period t Angular moment battle arrayIt is row matrix or column matrix, phase angle matrixUnit be region a a certain nodes under error scene s in period t Phase angle；It is the active power output of region a new energy unit w under error scene s in period t；It is in period t, area The maximum active power output of domain a new energy unit w under error scene s；It is that, in period t-1, region a is under error scene s The active power output of conventional power unit i；It is the phase angle of the node j1 of circuit j under period t, error scene s；It is in the period The phase angle of the node j2 of circuit j under t, error scene s；It is phases of the region a in prediction scene lower boundary node m in period t Angle；It is phase angles of the region a in error scene s lower boundary nodes m in period t；It is that, in period t, region a is in prediction scene The phase angle of lower boundary node n；It is phase angles of the region a in error scene s lower boundary nodes n in period t.

The constraints of Consultation Center is specially：

Coupling constraint condition between Consultation Center and region is specially：

Be the period t Consultation Center correspond to region a in boundary node m phase angle；It is in period t Consultation Center Corresponding to the phase angle of boundary node m in the b of region；Be the period t Consultation Center correspond to region a in boundary node n phase angle；Be the period t Consultation Center correspond to region b in boundary node n phase angle.

Regional prediction model of place is：

Object function：

Constraints：

B_aP_a+D_aθ_a≤E_a；1≤a≤N (21)

The phase angles of the boundary node m in day part of region a is issued to for kth time distributed optimization iterative coordination center Matrix；It is phase angle matrixes of the boundary node n in day part that kth time distributed optimization iterative coordination center is issued to region a；Kth time distributed optimization iteration is corresponding to the coupling constraint condition between Consultation Center and region in day part Lagrange multiplier,It is the coupling that kth time distributed optimization iteration corresponds between Consultation Center and region Quadratic penalty function multiplier of the constraints in day part；It is region a intermediate variables corresponding with error scene aggregation group x, altogether X^aIt is individual；E is column matrix, and the unit of column matrix is 1；F_aIt is optimal cutling coefficient matrix；M_aAnd N_aIt is optimal cutling coefficient square Battle array；P_a ^TFor region a predict scene under each conventional power unit the matrix of exerting oneself of day part transposed matrix；It is region a pre- Transposed matrix of each node in the phase angle matrix of day part under survey scene.

Domain error model of place is：

Object function：

Constraints：

B_a,sP_a,s+D_a,sθ_a,s≤E_a,s+G_a,sP_a,l+H_a,sθ_a,l；1≤a≤N,1≤s≤S_a (24)

P_a,lIt is the l times random optimization iteration, the region a being calculated according to regional prediction model of place is predicting scene Under each conventional power unit day part matrix of exerting oneself；θ_a,lIt is the l times random optimization iteration, according to regional prediction model of place meter The region a for the obtaining phase angle matrixes of each node in day part in the case where scene is predicted.

Interregional Coordination Model is：

Object function is：

Constraints is：

It is kth time distributed optimization iteration, in being calculated and upload to coordination according to regional prediction model of place Phase angle matrixes of the boundary node m of the region a of the heart in day part；It is kth time distributed optimization iteration, according to regional prediction Model of place be calculated and upload to Consultation Center region a boundary node n day part phase angle matrix.

Please continue to refer to Fig. 2, in embodiments of the present invention, step S400 can include：

Step S410, setting power system in parameter initial value, initial value include Consultation Center in respectively with each region Corresponding initial distribution formula optimum results.Specifically, distributed optimization iterations k=1, arrange parameter can be setThat is, the 1st time distributed optimization iteration correspond to Consultation Center with Coupling constraint condition between region is the 100, the 1st distributed optimization iteration and corresponds in the Lagrange multiplier of day part Coupling constraint condition between Consultation Center and region is also 100 in the quadratic penalty function multiplier of day part, in period t, the 1st Secondary distributed optimization iteration is predicting the phase angle of scene lower boundary node m in period t region a, and the 1st time distributed optimization iteration exists Period t region a is 0 in the phase angle of prediction scene lower boundary node n.

Step S420, the regional prediction model of place according to each region, calculate the initial economic load dispatching result in each region, and Distributed optimization is carried out to the boundary node in each region using interregional Coordination Model, makes the initial economic load dispatching result in each region The first convergence criterion is satisfied by, wherein, the first convergence criterion is：

ε is convergence precision, ε=10^-3；It is kth time distributed optimization iteration, in period t, Consultation Center corresponds to area The phase angle of boundary node m in a of domain；It is kth time distributed optimization iteration, in period t region a in prediction scene lower boundary section The phase angle of point m；It is kth time distributed optimization iteration, the phase of boundary node m during Consultation Center corresponds to region b in period t Angle；It is kth time distributed optimization iteration, the phase angle of boundary node m in the b of region under prediction scene is in period t region a.

Regional prediction model of place according to each region, calculates the initial economic load dispatching result in each region, and utilizes region Between Coordination Model distributed optimization is carried out to the boundary node in each region, a preferable glug can be provided for follow-up calculating Bright day multiplier initial value, is easy to follow-up calculating, and reduce the calculating time.

Step S430, the regional prediction model of place according to each region, calculate the prediction economic load dispatching result in each region.I.e. According to the Lagrange multiplier initial value, the regional prediction model of place that are obtained in step S420, the prediction economy for calculating each region is adjusted Degree result.

Step S440, the domain error model of place according to each region, calculate the random optimization result in each region.Utilize Prediction economic load dispatching result, the domain error model of place being calculated in step S430, calculate the random optimization knot in each region Really.

Whether step S450, the prediction economic load dispatching result for judging each region are satisfied by with the random optimization result in each region Second convergence criterion；When meeting, the parameter of boundary node in the prediction economic load dispatching result in each region is uploaded in coordination The heart, performs step S460；When being unsatisfactory for, optimal cutling model is set up,

And the optimal cutling value in each region of random optimization result calculating using optimal cutling model and each region, Jiang Gequ The optimal cutling value correspondence in domain is incorporated to the constraints of regional prediction model of place, performs step S430.

Wherein, the second convergence criterion is：

Wherein,

f_a,lIt is the l times random optimization iteration, the prediction scene total cost of region a；It is the l times random optimization iteration, Phase angle matrixes of the boundary node m in day part in a of region；It is the l times random optimization iteration, boundary node n is each in a of region The phase angle matrix of period.

Optimal cutling model is：

π_a,s,lIt is the l times random optimization iteration, the dual variable of the constraints of domain error model of place in day part Matrix；X^aIt is by the number S of the error scene of region a_aThe number of the error scene aggregation group formed after average polymerization, each mistake Difference scene aggregation group includes S_a/X^aIndividual error scene.

It is pre- to the region by the region using the domain error model of place in region that step S430 is actual to step S450 The prediction economic load dispatching result for surveying the region that model of place is calculated carries out random optimization, when the prediction economy in each region is adjusted When degree result is satisfied by the second convergence criterion with the random optimization result in each region, show that the random optimization in each region is restrained, Then complete random optimization；When in the prediction economic load dispatching result in each region with the random optimization result in each region, wherein at least has When one prediction economic load dispatching result in region is unsatisfactory for the second convergence criterion with the random optimization result in the region, then show this The random optimization convergence in region, now then needs the domain error model of place for continuing with the region to the region by the region The prediction economic load dispatching result in the region that prediction model of place is calculated carries out random optimization, that is, carry out next time random excellent Change.In this way, by multiple random optimization, obtaining the prediction economic load dispatching result in optimal each region.

In the above-described embodiments, optimal cutling model is used S_aThe X formed after individual error scene average polymerization^aIndividual error Scene polymerization set constructor, in actual applications, optimal cutling model can also directly use S_aIndividual error scene directly carries out structure Make, specifically, optimal cutling model can be：

Wherein,It is region a intermediate variables corresponding with error scene s, common S_aIt is individual.

Now, regional prediction model of place can be：

Object function：

Constraints：

B_aP_a+D_aθ_a≤E_a；1≤a≤N (21)

Step S460, according to interregional Coordination Model, calculating corresponds respectively to the distributed optimization result in each region.Complete The prediction in the region using the domain error model of place in region to being calculated by the regional prediction model of place in the region After economic load dispatching result carries out random optimization, then border section in the prediction economic load dispatching result that Consultation Center is uploaded according to each region The parameter of point, and according to interregional Coordination Model, it is calculated distributed optimization result.

Step S470, the prediction economic load dispatching result for judging each region and the distributed optimization knot for corresponding respectively to each region Whether fruit is satisfied by the first convergence criterion；When meeting, using the prediction economic load dispatching result in each region as power system warp Ji scheduling result, performs step S480；When being unsatisfactory for, parameter more new model is set up, using parameter more new model, calculate and update Parameter afterwards, performs step S430.Wherein, parameter more new model is：

It is the coupling constraint that -1 distributed optimization iteration of kth corresponds between Consultation Center and region Lagrange multiplier of the condition in day part；- 1 distributed optimization iteration of kth is corresponding to Consultation Center The quadratic penalty function multiplier of coupling constraint condition between region in day part；α is regulation step parameter, 1≤α≤3, example Such as, α=1.05.

Step S460 and step S470 are actually that Consultation Center is saved using interregional Coordination Model to the border in each region Point carries out distributed optimization, to be calculated optimal regional prediction economic load dispatching result；When the prediction economic load dispatching in each region When result is satisfied by the first convergence criterion with the distributed optimization result for corresponding respectively to each region, now, Consultation Center is to each The distributed optimization convergence of the boundary node in region, then the prediction economic load dispatching result in each region collectively forms the warp of power system Ji scheduling result；When in prediction economic load dispatching result and the distributed optimization result for corresponding respectively to each region in each region, its In at least one region prediction economic load dispatching result with correspond to the region distributed optimization result be unsatisfactory for the first convergence During criterion, then show that Consultation Center does not restrain to the distributed optimization of the boundary node in each region, then need to be divided next time Cloth optimizes, and when carrying out distributed optimization next time, because parameter is updated according to parameter more new model, then needs again The prediction in the region using the domain error model of place in region to being calculated by the regional prediction model of place in the region Economic load dispatching result carries out random optimization.

The economic load dispatching result of step S480, output power system.

The above, specific embodiment only of the invention, but protection scope of the present invention is not limited thereto, and it is any Those familiar with the art the invention discloses technical scope in, change or replacement can be readily occurred in, should all contain Cover within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.

Claims (10)

1. a kind of Economic Dispatch method, it is characterised in that including：

Step S100, it is Consultation Center and multiple region by power system decoupling；

Step S200, according to the Consultation Center and multiple regions, set up Economic Dispatch model, wherein, institute The target for stating Economic Dispatch model is total generating expense of the conventional power unit in scheduling duration in the power system With abandon generation of electricity by new energy cutting load rejection penalty sum；

Step S300, it is interregional Coordination Model and multiple subdispatch mould by the Economic Dispatch model decomposition Type, wherein, multiple subdispatch models and multiple regions correspond, the interregional Coordination Model and the association Tune center correspondence, the interregional Coordination Model carries out distributed optimization to the boundary node in each region；

Step S400, according to the interregional Coordination Model and multiple subdispatch models, calculate the economy of power system Scheduling result.

2. Economic Dispatch method according to claim 1, it is characterised in that before the step S100, The Economic Dispatch method also includes：

Step S10, determination carry out the scheduling duration of economic load dispatching to the power system, and the scheduling duration is averagely divided It is n_TThe individual period, wherein, n_T≥2。

3. Economic Dispatch method according to claim 1, it is characterised in that after the step S100, Before the step S200, the Economic Dispatch method also includes：

Step S100 ', to each region setting prediction scene, and it is multiple to the region extraction with new energy electric field by mistake Difference scene.

4. Economic Dispatch method according to claim 3, it is characterised in that after the step S300, Before the step S400, the Economic Dispatch method also includes：

Step S300 ', by the subdispatch model decomposition be regional prediction model of place and domain error model of place.

5. Economic Dispatch method according to claim 4, it is characterised in that the step S400 includes：

Step S410, the initial value for setting parameter in the power system, the initial value include distinguishing in the Consultation Center Initial distribution formula optimum results corresponding with each region；

Step S420, the regional prediction model of place according to each region, calculate the initial economic load dispatching knot in each region Really, and using the interregional Coordination Model distributed optimization is carried out to the boundary node in each region, makes each region Initial economic load dispatching result be satisfied by the first convergence criterion；

Step S430, the regional prediction model of place according to each region, calculate the prediction economic load dispatching knot in each region Really；

Step S440, the domain error model of place according to each region, calculate the random optimization result in each region；

Whether step S450, the prediction economic load dispatching result for judging each region are satisfied by with the random optimization result in each region Second convergence criterion；

When meeting, the parameter of boundary node in the prediction economic load dispatching result in each region is uploaded in the coordination The heart, performs step S460；

When being unsatisfactory for, optimal cutling model is set up, and using the optimal cutling model and the random optimization in each region Result calculates the optimal cutling value in each region, and the optimal cutling value correspondence in each region is incorporated into the regional prediction The constraints of scape model, performs the step S430；

Step S460, according to the interregional Coordination Model, calculating corresponds respectively to the iteration distributed optimization in each region As a result；

Step S470, the prediction economic load dispatching result for judging each region are distributed with the iteration for corresponding respectively to each region Whether formula optimum results are satisfied by first convergence criterion；

When meeting, the prediction economic load dispatching result in each region as the economic load dispatching result of the power system is held Row step S480；

When being unsatisfactory for, parameter more new model is set up, using the parameter more new model, calculate the parameter after updating, perform institute State step S430；

Step S480, the economic load dispatching result for exporting the power system.

6. Economic Dispatch method according to claim 5, it is characterised in that the Economic Dispatch Model is：

Object function：

$\begin{array}{c}\min \underset{a=1}{\overset{N}{\Σ}}(f_a+\underset{s=1}{\overset{S_a}{\Σ}}{f}_{a,s})=\underset{a=1}{\overset{N}{\Σ}}\underset{t=1}{\overset{N_T}{\Σ}}\left\{\underset{i=1}{\overset{N_g^a}{\Σ}}\[{\mathrm{\α}}_i^a{\left(P_{i,t}^a\right)}^2+{\mathrm{\β}}_i^a{P}_{i,t}^a+{\mathrm{\γ}}_i^a\]+\underset{w=1}{\overset{N_w^a}{\Σ}}{q}_W{\mathrm{\ΔW}}_{w,t}^a+\underset{d=1}{\overset{N_d^a}{\Σ}}{q}_D{\mathrm{\ΔD}}_{d,t}^a\right\}\\ +\underset{a=1}{\overset{N}{\Σ}}\underset{t=1}{\overset{N_T}{\Σ}}\left\{\underset{s=1}{\overset{S_a}{\Σ}}{p}_s\[q_W\underset{w=1}{\overset{N_w^a}{\Σ}}{\mathrm{\ΔW}}_{w,t,s}^a+q_D\underset{d=1}{\overset{N_d^a}{\Σ}}{\mathrm{\ΔD}}_{d,t,s}^a\]\right\}\end{array};$

Constraints：

The prediction context restrictions condition in the region：

B_aP_a+D_aθ_a≤E_a；1≤a≤N；

The error scene constraints in the region：

B_a,sP_a,s+D_a,sθ_a,s≤E_a,s+G_a,sP_a+H_a,sθ_a；1≤a≤N,1≤s≤S_a；

The constraints of the Consultation Center：

Coupling constraint condition between the Consultation Center and the region：

Wherein, the prediction context restrictions condition in the region includes：

$P_{G,t}^a+(P_{W,t}^a-{\mathrm{\ΔW}}_{W,t}^a)-(P_{D,t}^a-{\mathrm{\ΔD}}_{D,t}^a)=B^a{\mathrm{\θ}}_t^a;1\≤a\≤N,1\≤t\≤N_T;$

${\underset{\‾}{P}}_i^a\≤P_{i,t}^a\≤{\stackrel{\‾}{P}}_i^a;1\≤a\≤N,1\≤t\≤N_T,1\≤i\≤N_g^a;$

$0\≤P_{w,t}^a\≤{\tilde{P}}_{w,t}^a;1\≤a\≤N,1\≤t\≤N_T,1\≤w\≤N_d^a;$

$r_{di}^a\≤P_{i,t}^a-P_{i,t-1}^a\≤r_{ui}^a;1\≤a\≤N,2\≤t\≤N_T,1\≤i\≤N_g^a;$

$-{\stackrel{\‾}{P}}_j^a\≤\frac{{\mathrm{\θ}}_{j1,t}^a-{\mathrm{\θ}}_{j2,t}^a}{x_j^a}{S}_B\≤{\stackrel{\‾}{P}}_j^a;1\≤a\≤N,1\≤t\≤N_T,1\≤j\≤N_J;$

$-{\mathrm{\Δ}}_i^a\≤P_{i,t}^a-P_{i,t,s}^a\≤{\mathrm{\Δ}}_i^a;1\≤i\≤N_g^a,1\≤t\≤N_T,1\≤s\≤S_a;$

The error scene constraints in the region includes：

$P_{G,t,s}^a+(P_{W,t,s}^a-{\mathrm{\ΔW}}_{W,t,s}^a)-(P_{D,t,s}^a-{\mathrm{\ΔD}}_{D,t,s}^a)=B^a{\mathrm{\θ}}_{t,s}^a;1\≤a\≤N,1\≤t\≤N_T,1\≤s\≤S_a;$

${\underset{\‾}{P}}_i^a\≤P_{i,t,s}^a\≤{\stackrel{\‾}{P}}_i^a;1\≤a\≤N,1\≤t\≤N_T,1\≤i\≤N_g^a;$

$0\≤P_{w,t,s}^a\≤{\tilde{P}}_{w,t,s}^a;1\≤a\≤N,1\≤t\≤N_T,1\≤w\≤N_d^a,1\≤s\≤S_a;$

$r_{di}^a\≤P_{i,t,s}^a-P_{i,t-1,s}^a\≤r_{ui}^a;1\≤a\≤N,2\≤t\≤N_T,1\≤i\≤N_g^a,1\≤s\≤S_a;$

$-{\stackrel{\‾}{P}}_j^a\≤\frac{{\mathrm{\θ}}_{j1,t,s}^a-{\mathrm{\θ}}_{j2,t,s}^a}{x_j^a}{S}_B\≤{\stackrel{\‾}{P}}_j^a;1\≤a\≤N,1\≤t\≤N_T,1\≤s\≤S_a,1\≤j\≤N_J;$

$\left\{\begin{array}{c}{\mathrm{\θ}}_{m,t}^a-{\mathrm{\θ}}_{m,t,s}^a=0\\ {\mathrm{\θ}}_{n,t}^a-{\mathrm{\θ}}_{n,t,s}^a=0\end{array}\right.;m\∈{\mathrm{TL}}^{ab,a},n\∈{\mathrm{TL}}^{ab,b},1\≤a\≤N,1\≤b\≤N,a\≠b,1\≤t\≤N_T,1\≤s\≤S_a;$

The constraints of the Consultation Center is specially：

Coupling constraint condition between the Consultation Center and the region is specially：

f_aIt is the prediction scene total cost of region a；f_a,sFor the error scene of region a abandons generation of electricity by new energy expense；N is the region Number；It is the number of conventional power unit in a of region；It is the number of region a new energy units；It is in period t region a Load bus number；S_aIt is the number of the error scene of region a；It is the routine in the case where scene is predicted in period t region a The active power output of unit i；WithThe generating cost coefficient of conventional power unit i in respectively region a；It is in the period T, region a the new energy unit w in the case where scene is predicted abandon generation of electricity by new energy power；q_WGeneration of electricity by new energy punishment is abandoned for region a Cost coefficient；It is the cutting load power of region a load bus d in the case where scene is predicted in period t；q_DFor cutting for region a is negative Lotus rejection penalty coefficient；p_sIt is the probability of the error scene s of region a, p_s=1/S_a；It is that region a is in error in period t New energy unit w's abandons generation of electricity by new energy power under scene s；It is region a load sections under error scene s in period t The cutting load power of point d；

P_aFor region a predict scene under each conventional power unit day part matrix of exerting oneself；θ_aFor region a is each in the case where scene is predicted Phase angle matrix of the node in day part；B_a、D_aAnd E_aIt is parameter matrixs of the region a in the case where scene is predicted；P_a,sIt is that region a is being missed Exert oneself matrix of each conventional power unit in day part under difference scene s；θ_a,sFor region a under error scene s each node in day part Phase angle matrix；B_a,s、D_a,s、E_a,s、G_a,sAnd H_a,sIt is parameter matrixs of the region a under error scene s；TL^ab,aFor in a of region The boundary node intersection being connected with region b；TL^ab,bIt is the boundary node intersection being connected with region a in the b of region, and m and n It is corresponding two boundary nodes of the connecting line of join domain a and region b；Border is saved in corresponding to region a for Consultation Center Phase angle matrixes of the point m in day part；For Consultation Center correspond to region a in boundary node n day part phase angle matrix； For Consultation Center correspond to region b in boundary node m day part phase angle matrix；Be Consultation Center correspond to region b in Phase angle matrixes of the boundary node n in day part；It is phase angle matrixes of the boundary node m in day part in a of region；It is region a Phase angle matrixes of the middle boundary node n in day part；

It is the matrix of exerting oneself of region a each conventional power units in the case where scene is predicted in period t；It is that region a is pre- in period t The matrix of exerting oneself of each new energy unit under survey scene；It is the load of region a each load buses in the case where scene is predicted in period t Matrix；It is that, in period t, a each new energy units in the case where scene is predicted in region abandon generation of electricity by new energy power matrix； It is the cutting load power matrix of region a each load buses in the case where scene is predicted in period t；B^aIgnore branch resistance for region a With the bus admittance matrix set up to ground leg；It is the phase angular moment of region a each nodes in the case where scene is predicted in period t Battle array；It is the active power output lower limit of conventional power unit i in a of region；It is the active power output upper limit of conventional power unit i in a of region； It is the active power output of region a new energy unit w in the case where scene is predicted in period t；It is the new energy unit in period t region a The maximum active power output of w；Climb for the active power output of conventional power unit i in a of region and limit；It is conventional power unit i in a of region The limitation of active power output landslide；It is the active power output of region a conventional power unit i in the case where scene is predicted in period t-1；N_JFor described The number of the circuit relevant with region a in power system, the internal wiring and join domain a of the circuit including region a and The interregional interconnector in other regions；It is the maximum transmission power value of the circuit j relevant with region a；It is have with region a The reactance value of the circuit j of pass；It is the phase angle of the node j1 of circuit j under period t, prediction scene；It is in period t, in advance The phase angle of the node j2 of circuit j under survey scene；S_BOn the basis of be worth, S_B=100MW；For in a of region conventional power unit i at 10 minutes Interior adjustable increment of exerting oneself；It is the active power output of region a conventional power unit i under error scene s in period t；

It is the matrix of exerting oneself of region a each conventional power units under error scene s in period t；In period t, region a is by mistake The matrix of exerting oneself of each new energy unit under difference scene s；Be in period t, region a under error scene s each load bus it is negative Lotus matrix；It is that, in period t, a each new energy units under error scene s in region abandon generation of electricity by new energy power matrix；It is the cutting load power matrix of region a each load buses under error scene s in period t；It is the region a in period t The phase angle matrix of each node under error scene s；It is that, in period t, a new energy unit w under the error scene s in region have Work(is exerted oneself；It is the maximum active power output of region a new energy unit w under error scene s in period t；It is in the period The active power output of t-1, region a the conventional power unit i under error scene s；It is the node of the circuit j under period t, error scene s The phase angle of j1；It is the phase angle of the node j2 of circuit j under period t, error scene s；It is that, in period t, region a is in prediction The phase angle of scene lower boundary node m；It is phase angles of the region a in error scene s lower boundary nodes m in period t；Be The phase angle of period t, region a in prediction scene lower boundary node n；It is that, in period t, region a is in error scene s lower boundary sections The phase angle of point n；

Be the period t Consultation Center correspond to region a in boundary node m phase angle；It is in period t Consultation Center's correspondence The phase angle of boundary node m in the b of region；Be the period t Consultation Center correspond to region a in boundary node n phase angle； Be the period t Consultation Center correspond to region b in boundary node n phase angle.

7. Economic Dispatch method according to claim 6, it is characterised in that

The regional prediction model of place is：

Object function：

Constraints：

B_aP_a+D_aθ_a≤E_a；1≤a≤N；

The phase angle matrixes of the boundary node m in day part of region a is issued to for kth time distributed optimization iterative coordination center；It is phase angle matrixes of the boundary node n in day part that kth time distributed optimization iterative coordination center is issued to region a；It is the coupling constraint bar that kth time distributed optimization iteration corresponds between the Consultation Center and the region Part day part Lagrange multiplier,Be kth time distributed optimization iteration correspond to the Consultation Center with The quadratic penalty function multiplier of coupling constraint condition between the region in day part；It is region a and error scene aggregation group x Corresponding intermediate variable, common X^aIt is individual；E is column matrix, and the unit of column matrix is 1；F_aIt is optimal cutling coefficient matrix；M_aAnd N_a It is optimal cutling coefficient matrix；P_a ^TFor region a predict scene under each conventional power unit the matrix of exerting oneself of day part transposition Matrix；For region a predict scene under each node the phase angle matrix of day part transposed matrix；

The domain error model of place is：

Object function：

$\min f_{a,s}=min\underset{t=1}{\overset{N_T}{\Σ}}\{\underset{s=1}{\overset{S_a}{\Σ}}{p}_s\[q_W\underset{w=1}{\overset{N_w^a}{\Σ}}{\mathrm{\ΔW}}_{w,t,s}^a+q_D\underset{d=1}{\overset{N_d^a}{\Σ}}{\mathrm{\ΔD}}_{d,t,s}^a\]\};$

Constraints：

B_a,sP_a,s+D_a,sθ_a,s≤E_a,s+G_a,sP_a,l+H_a,sθ_a,l；1≤a≤N,1≤s≤S_a；

P_a,lIt is the l times random optimization iteration, the region a being calculated according to the regional prediction model of place is predicting scene Under each conventional power unit day part matrix of exerting oneself；θ_a,lIt is the l times random optimization iteration, according to the regional prediction scene mould The region a that type the is calculated phase angle matrixes of each node in day part in the case where scene is predicted；

The interregional Coordination Model is：

Object function is：

Constraints is：

It is kth time distributed optimization iteration, the coordination is calculated and uploaded to according to the regional prediction model of place Phase angle matrixes of the boundary node m of the region a at center in day part；It is kth time distributed optimization iteration, according to the region Prediction model of place be calculated and upload to the Consultation Center region a boundary node n day part phase angular moment Battle array.

8. Economic Dispatch method according to claim 7, it is characterised in that

First convergence criterion is：

ε is convergence precision, ε=10^-3；It is kth time distributed optimization iteration, during in period t, Consultation Center corresponds to region a The phase angle of boundary node m；It is kth time distributed optimization iteration, in period t region a prediction scene lower boundary node m's Phase angle；It is kth time distributed optimization iteration, the phase angle of boundary node m during Consultation Center corresponds to region b in period t；It is kth time distributed optimization iteration, the phase angle of boundary node m in the b of region under prediction scene is in period t region a；

Second convergence criterion is：

$\frac{U_a-L_a}{U_a}<\mathrm{\&epsiv;};1\&le;a\&le;N;$

Wherein,

f_a,lIt is the l times random optimization iteration, the prediction scene total cost of region a；It is the l times random optimization iteration, region a Phase angle matrixes of the middle boundary node m in day part；It is the l times random optimization iteration, boundary node n is in day part in a of region Phase angle matrix.

9. Economic Dispatch method according to claim 7, it is characterised in that

The optimal cutling model is：

π_a,s,lIt is the l times random optimization iteration, the dual variable of the constraints of the domain error model of place in day part Matrix；X^aIt is by the number S of the error scene of region a_aThe number of the error scene aggregation group formed after average polymerization, each institute Stating error scene aggregation group includes S_a/X^aIndividual error scene.

10. Economic Dispatch method according to claim 7, it is characterised in that

The parameter more new model is：

It is the coupling that -1 distributed optimization iteration of kth corresponds between the Consultation Center and the region Lagrange multiplier of the constraints in day part；- 1 distributed optimization iteration of kth is corresponding to described The quadratic penalty function multiplier of coupling constraint condition between Consultation Center and the region in day part；α is regulation step parameter, 1≤α≤3。