CN115995850A

CN115995850A - Collaborative scheduling optimization method and device for virtual power plant group

Info

Publication number: CN115995850A
Application number: CN202310200428.XA
Authority: CN
Inventors: 鞠立伟; 汪鹏; 杨莘博; 鲁肖龙; 孙杰; 樊伟
Original assignee: North China Electric Power University
Current assignee: Beijing Xinyuan Smart Internet Technology Co ltd; North China Electric Power University; Economic and Technological Research Institute of State Grid Hebei Electric Power Co Ltd
Priority date: 2023-03-06
Filing date: 2023-03-06
Publication date: 2023-04-21
Anticipated expiration: 2043-03-06
Also published as: CN115995850B

Abstract

The invention relates to a collaborative scheduling optimization method and device for a virtual power plant group, belongs to the technical field of power grids, and solves the problems that the competition and standby balance of a generator set exist in a daily stage and a real-time stage when the existing VPP collaborative operation is performed. The method comprises the following steps: a plurality of virtual power plant VPP co-operating systems: when the respective virtual power plants meet the self-load demands, the residual power generation capacity declares bidding price and the delivered power in a bidding transaction center in a bidding game mode, and simultaneously transmits bidding price and bidding energy to the VPPs of the respective virtual power plants; multi-time scale coordinated scheduling optimization of VPP: when the power dispatching plan is formulated, the daily forecast results of the WPP and the PV are considered, the daily forecast deviation of the WPP and the PV is corrected by combining the daily forecast values, and the real-time output of the random power supply is combined, so that the power is purchased from the rest VPP or UPG and the IBDR is called, and the real-time supply and demand balance of the power is ensured. The reliable supply of the self power is realized and the redundant power is sent to the UPG in the competitive price transaction mode.

Description

Collaborative scheduling optimization method and device for virtual power plant group

Technical Field

The invention relates to the technical field of power grids, in particular to a collaborative scheduling optimization method and device for a virtual power plant group.

Background

With the large-scale distributed energy sources (distributed energy resources, DERs) and flexible loads in the power system, the safe and stable operation of the power grid faces a brand new challenge, the traditional power generation scheduling method is difficult to adapt to new requirements, and new DERs control technology needs to be researched. The virtual power plant (Virtual power plant, virtual power plant VPP) can effectively aggregate a plurality of distributed power supplies, controlled loads and energy storage units by using advanced technology, realize the collaborative coordination and scheduling operation of all elements in the virtual power plant, and maximize environmental and economic benefits. Indeed, with the continued maturation and development of virtual power plant VPP technology, multiple virtual power plant cluster operations will become a necessary trend for single virtual power plant VPP structural extension and energy expansion.

At present, the research of domestic and foreign scholars around a virtual power plant mainly focuses on two aspects of uncertainty analysis and operation optimization model, and forms a representative research result, for example, in terms of uncertainty analysis, the research result mainly originates from two paths of uncertainty variable probability distribution curve simulation and uncertainty measurement, the scholars put forward a Latin hypercube sampling method to simulate and analyze the probability density of wind power and photovoltaic power generation, and the scholars put forward a fuzzy CVaR method to explore the load uncertainty, and further construct a random optimization model. The existing operation optimization method mainly considers operation constraints of different structural elements, determines a proper optimization target, builds a virtual power plant VPP operation optimization model, and builds an electric interconnection virtual power plant multi-target model by taking operation economic benefits, peak clipping and valley filling effects and carbon dioxide emission as optimization targets; and combining a two-stage optimization theory, and constructing a double-layer scheduling optimization model containing the virtual power plant by considering day-ahead scheduling and time-ahead scheduling. The existing scheme mainly takes intermittent and fluctuation of wind power and photovoltaic power generation into consideration, and establishes a daily scheduling optimization strategy of a virtual power plant, but when a plurality of virtual power plants operate cooperatively, the problem of generator set competition and standby balance exists in a daily stage and a real-time stage.

For the research of optimizing operation of the virtual power plant group, more consideration is given to the fact that when different virtual power plants VPPs meet own energy demands, surplus generated output is sent to a public power grid, and at the moment, the competitive price trading problem of a plurality of virtual power plants is formed. When a plurality of virtual power plants VPPs exist in the system, each virtual power plant VPP has an independent operation target, and in order to achieve the aim of energy supply and demand stability, an upper power grid needs to select different virtual power plants VPPs to provide required energy through optimization, and the optimization process is essentially a multi-virtual power plant VPP bidding game problem. How to construct a multi-time scale collaborative optimization operation model aiming at the operation characteristics of virtual power plant groups with different time scales is a key point for improving the virtual power plant VPP optimization operation. The virtual power plant group multi-time scale collaborative operation mode often has a plurality of targets and a plurality of optimization stages, so that the model is nonlinear, and how to optimally solve the related model is the key for formulating the virtual power plant optimization operation strategy. Compared with the traditional algorithm, the heuristic algorithm can obtain a better solution set, but has the problem of easy sinking into local optimum.

Disclosure of Invention

In view of the above analysis, the present invention aims to provide a collaborative scheduling optimization method and apparatus for a virtual power plant group, which are used for solving the problem that the existing multiple virtual power plants have the competition and standby balance of a generator set in the daily stage and the real-time stage when operating in a collaborative manner.

In one aspect, an embodiment of the present invention provides a collaborative scheduling optimization method for a virtual power plant group, including: a plurality of virtual power plant VPP co-operating systems: when the respective virtual power plants meet the self load demands, the residual power generation capacity declares bidding price and the output power in a bidding transaction center in a bidding game mode, the bidding transaction center integrates virtual power plant VPP bidding information and settles bidding energy shares and market price obtained by different virtual power plants VPP, and simultaneously transmits the bidding price and the bidding energy to each virtual power plant VPP, so that the internal energy supply and demand balance is ensured, and the redundant energy is sold to a public power grid to maximize the operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, an energy storage system ESS and flexible loads; multi-time scale collaborative scheduling optimization of virtual power plant VPPs: when a power dispatching plan is formulated, the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV are considered, the day-ahead prediction values are combined, the wind power plant WPP and the photovoltaic power plant PV prediction deviation is corrected by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load dispatching plan, and the real-time output of a random power supply is combined, so that the power supply and demand balance in real time is ensured in a mode of purchasing power from other virtual power plants VPPs or upper power grids UPG and calling excitation type demand response IBDR.

The beneficial effects of the technical scheme are as follows: the multi-time scale collaborative operation optimization model can effectively connect scheduling stages in different time scales, and consider the prediction power of WPP and PV in each stage, so that the reliable supply of self power can be realized, and meanwhile, redundant power is sent to UPG in a competitive price transaction mode, so that excess economic benefit is obtained. In addition, when emergency power supply service is required to be provided, IBDR, VPP and UPG can be called to realize the aim of VPP real-time power supply and demand balance.

Based on a further improvement of the above method, the multi-time scale collaborative scheduling optimization of the virtual power plant VPP comprises: day-ahead schedule optimization, day-in schedule optimization, and real-time schedule optimization, wherein the day-ahead schedule optimization: based on the wind power plant WPP and the photovoltaic power plant PV day-ahead predicted output, and considering interaction and coordination among the gas turbine power generation CGT, the energy storage system ESS and the flexible load, a scheduling plan with optimal power supply cost is obtained; the intra-day schedule optimization: combining the wind power plant WPP and the photovoltaic power plant PV intra-day predicted output, correcting the power generation CGT output of the energy storage system ESS and the gas turbine so as to eliminate the influence caused by the deviation of the wind power plant WPP and the photovoltaic power plant PV output as much as possible, calculating the unit power supply marginal cost and the unit power supply quantity in the virtual power plant VPP, reporting the bidding price and bidding electric energy to a transaction center, and finally obtaining the electric energy bidding transaction quantity and price returned by the transaction center; and the real-time scheduling optimization: and considering the real-time output of the wind power plant WPP and the photovoltaic power plant PV, and ensuring the real-time balance of power supply and demand on the basis of a daily electric energy scheduling plan and a competitive price transaction scheme, wherein when the real-time output of the wind power plant WPP and the photovoltaic power plant PV is smaller than the planned output, the electric energy is purchased from other virtual power plants VPPs or superior power grids UPG or the IBDR emergency output is called, and otherwise, the electric energy is sold or the IBDR electricity is increased.

Based on a further improvement of the above method, the day-ahead scheduling optimization includes: and combining the wind power plant WPP and the photovoltaic power plant PV day-ahead predicted power, and arranging power generation scheduling plans of different types of units by taking the minimum virtual power plant VPP operation cost as a target, wherein the virtual power plant VPP operation cost comprises wind and light energy discarding cost, gas turbine power generation CGT operation cost and energy storage loss cost.

Based on a further improvement of the above method, the day-ahead scheduling optimization further comprises: the objective function is calculated by the following equation 1:

equation 1

wherein ,C _WPP,t andC _PV,t respectively representing the moment of the wind power plant WPP and the moment of the photovoltaic power plant PVtIs characterized by that the wind-discarding cost and the light-discarding cost are set,C _CGT,t indicating that the gas turbine power generation CGT is at timetIs added to the running cost of the (a) and (b),C _ESS,t indicating the time of day of the ESS of the energy storage systemtIs added to the running cost of the (a) and (b),C _PB,t indicating that price type demand response PBDR is at timetThe implementation cost of (2); the total cost of operation of the virtual power plant VPP is calculated according to the following equation 2, equation 5:

equation 2

Wherein RE represents the wind power plant WPP and the photovoltaic plant PV,

indicating RE at timetIs used for generating the predicted output force of the engine,g _RE,t indicating that the RE is at timetIs used to control the actual output of the device,c _RE,t indicating that the RE is at time tIs added to the energy-discarding opportunity cost of the energy-discarding machine,p _RE,t at the moment for the REtIs connected with the internet for generating electricity;

equation 3

wherein ,a _CGT 、b _CGT 、c _CGT generating a CGT power generation output gas cost factor for the gas turbine,

and />

Representing the hot start cost and the cold start cost of the gas turbine power generation CGT respectively,u _CGT,t indicating that the gas turbine power generation CGT is at timetIs->

Generating CGT minimum down time for said gas turbine, < > for>

Indicating that the gas turbine power generation CGT is at timetIs not limited by the downtime of (a)，/>

Representing a cold start time of the gas turbine power generation CGT;

equation 4

wherein ,g _ESS,t is the moment the energy storage system ESS istCharge and discharge power wheng _ESS,t And < 0, indicating that the state of the energy storage system ESS is charged, and vice versa,ρ _ESS in order to adjust the coefficient of the power supply,

for the initial investment cost of the accumulator,N _t the service life of the storage battery is related to the depth of discharge;p _grid,t indicating the virtual power plant VPP at the momenttWith the electricity purchase and sale price of the public power grid,p _t indicating the time of day of the ESS of the energy storage systemtCharge and discharge price%>

and />

Respectively at the timetIs the charging and discharging efficiency of the ESS of the energy storage system, < >>

and />

Representing the charge-discharge loss cost of the unit electric quantity of the ESS;

equation 5

wherein ,

AndP _t respectively representing the price of the electric quantity before and after the PBDR at the time t,/>

AndL _t respectively shown intLoad demand before and after PBDR at time of day, attTime of day, deltaP _t and △L _PB , _t The price variable and the load variable before and after the PBDR are respectively, and the PBDR is elastically as follows according to the price of the power demand:

equation 6

wherein ,E _st representing power demand-price elasticity, the PBDR generated load fluctuation is calculated by the following formula:

equation 7

Calculating a virtual power plant VPP representative of the moment in time by the following formulatIs not required for the payload requirements of (2)M _t ：

Equation 8

wherein ,g _CGT,t indicating that the gas turbine power generation CGT is at timetIs characterized in that the power is generated by the power generation,η _CGT andη _ESS,t representing the output loss rates of the gas turbine power generation CGT and the energy storage system ESS respectively,g _UG,t representing the amount of power purchased by the virtual power plant VPP from the UPG at time t,L _t representing the load demand of the virtual power plant VPP at time t,u _PB,t indicating that the PBDR is at timet1 indicates that the PBDR is implemented, and 0 indicates that it is not implemented; calculating the output of the wind plant WPP and the photovoltaic plant PV by the following equation 9Power:

equation 9

Equation 10

wherein ,g _R for the WPP rated power of the wind power plant,v _t for the moment of timetWPP wind speed of the wind power plant,

at time for the wind power plant WPP tAvailable force, & gt>

At time for a photovoltaic plant PVtIs used to provide a positive pressure to the fluid,η _PV andS _PV for solar radiation efficiency and radiation area,θ _t to be at the momenttSolar radiation intensity; calculating a load demand constraint for the virtual power plant VPP operation by:

equation 11

Wherein the prediction errors of the wind power plant WPP and the photovoltaic power plant PV are respectively as followse _WPP,t Ande _PV,t wind power outputg _WPP,t Andg _PV,t the interval of (2) is [ (1 ]e _WPP,t )·g _WPP,t , (1+e _WPP,t )·g _WPP,t] and [(1-e _PV,t )·g _PV,t , (1+e _PV,t )·g _PV,t ]Selectinge _RE,t Substitution ofe _WPP,t Ande _PV,t ，g _RE,t substitution ofg _WPP,t Andg _PV,t ，g _RE,t will be distributed in [ (1 ]e _RE,t )·g _RE,t ，(1+e _RE,t )·g _RE,t ]，φ _RE Output loss rate of wind power or photovoltaic; as can be seen from the formula 11, when the uncertainty is strong, the load supply and demand imbalance is aggravated, usingθ _RE , _t and Γ_RE Correcting the load demand constraint to ensure the load supply and demand balance constraint, wherein the load demand constraint is corrected as follows:

equation 12

According to said formula 12 Γ _RE The introduction of the method provides a flexible risk decision tool for a decision maker to formulate a virtual power plant VPP scheduling scheme considering uncertainty according to the risk attitude of the decision maker.

Based on further improvements of the above method, the gas turbine power generation CGT operating constraint, the energy storage system ESS operating constraint, and the PBDR operating constraint are calculated by the following formulas, respectively: the gas turbine power generation CGT operation constraint is as follows:

Equation 13

Equation 14

Equation 15

Equation 16

wherein ,u _CGT,t indicating that gas turbine power generation CGT is at momenttIs used for the control of the operating state of the vehicle,

and />

Represents the minimum and maximum output power of the gas turbine power generation CGT, respectively, < >>

and />

Respectively representing downhill climbing power and uphill climbing power of the gas turbine power generation CGT, and +.>

and />

Respectively represent the time of the gas turbine power generation CGTt-continuous running time and continuous down time of-1, -/->

and />

Respectively representing the minimum start time and the minimum stop time of the gas turbine power generation CGT; the ESS operation constraints of the energy storage system are:

equation 17

Equation 18

Equation 19

wherein ,S _ESS,t indicating the time of day of the ESS of the energy storage systemtIs used for storing energy of the energy,

and />

Respectively representing the minimum charge-discharge power and the maximum charge-discharge power of the ESS of the energy storage system, +.>

and />

Respectively representing the minimum energy storage capacity and the maximum energy storage capacity of the ESS of the energy storage system, and in order to fully utilize the charging and discharging performances of the ESS of the energy storage system, the initial energy storage capacity of the ESS of the energy storage system in a dispatching cycle is utilizedS _ESS,0 And scheduling period end energy storageS _ESS,T All set to zero; the PBDR operating constraints are:

equation 20

wherein ,

indicating that the PBDR provides the maximum load fluctuation amount,φand a step of limiting the maximum load fluctuation amount provided by the PBDR and the load reduction specific gravity to avoid a phenomenon that a user transient response causes a peak-valley hang-up.

Based on further improvement of the method, the intra-day scheduling optimization comprises the steps of calculating unit power generation cost and residual power generation capacity of the virtual power plant VPP in each period according to the intra-day scheduling plan based on the wind power plant WPP and the photovoltaic power plant PV, and further constructing a multi-virtual power plant VPP power generation bidding game model by calling the energy storage system ESS, the PBDR or adjusting the gas turbine power generation CGT power generation capacity to cope with the pre-day predicted power deviation of the wind power plant WPP and the photovoltaic power plant PV and after finishing the deviation correction of the wind power plant WPP and the photovoltaic power plant PV.

Based on the further improvement of the method, the daily scheduling optimization further comprises the step of correcting the daily scheduling plan by taking the minimum fluctuation of wind and light output as a target, wherein the target function is as follows:

equation 21->

Equation 22

wherein ,△L _PB,tˊ Scheduling phase time before day for the PBDRtPlanned output at timetAt the location of the,g _PV,tˊ andg _WPP,tˊ representing the day-ahead scheduled output of the photovoltaic power plant PV and the wind power plant WPP respectively,gˊ _PV,tˊ andgˊ _WPP,tˊ representing the actual available output of the photovoltaic plant PV and the wind plant WPP, respectively, at the momenttAt the location of the body resistanceLˊ _PB,t Andgˊ _ESS,t the corrected output force of the PBDR and the ESS are respectively represented, and the setting is set t″=tAnd (2) energy storage system ESS capacity constraints are as follows:

equation 23

The modified ESS running output meets the constraint conditions of the formula 18 and the formula 19, and after the day-ahead scheduling is completed, the day-ahead scheduling of different virtual power plants is determined

、/>

、/>

、/>

and />

The method comprises the steps of carrying out a first treatment on the surface of the The remaining power supply capacity of each virtual power plant VPP is calculated by the following formula:

equation 24

wherein ,g _t to establish a unit energy supply cost when the internal load supply and demand balance premise of the virtual power plant VPP is satisfied according to the formula 1 and the formula 23 for the bidding output of the virtual power plant VPP participating in the power market at the time tC _VPP,t Setting the unit energy supply cost of the virtual power plant VPP at the time t, wherein the purpose of the different virtual power plants VPP participating in the bidding of the electric power market is to acquire excess profits, and if the bidding expectation profit rate isβ _VPP,t The price of the virtual power plant VPP to participate in the electric market bidding transaction is calculated by the following formula:

equation 25

Determining the daily bidding price and bidding quantity of the virtual power plant VPP according to a formula 24 and a formula 25, forming a plurality of bidding schemes when a plurality of virtual power plant VPPs participate in the power bidding, and carrying out energy transaction by the system according to the price as a selection standard until the energy balance is met; introducing a function argmax g (), in the context of a plurality of virtual power plants VPP participating in an electric power market bid, expressing an optimal strategy by the following formula:

Equation 26

Equation 27

Equation 28>

wherein ,L _UPG,t for UPG at timetAnd M represents the virtual plant VPP number.

Based on further improvement of the method, the real-time scheduling optimization comprises that the daily predicted power of the wind power plant WPP and the daily predicted power of the photovoltaic power plant PV still have deviation, so that the daily bidding game strategy is difficult to reach the optimal value, wherein when the actual tradable power is higher than the bidding power, the daily bidding scheme is still executed, otherwise, emergency standby resources are required to be called to cope with the problem of unbalanced load supply and demand, and the standby sources comprise IBDR, virtual power plant VPPs and three channels of an upper power grid.

Based on a further improvement of the above method, the real-time scheduling optimization further comprises: the objective function is expressed by the following formula:

equation 29

wherein ,R _m,t is virtual power plant VPPmAt the moment of timetIs used for the standby cost of the (a),

is virtual power plant VPPmAt the moment of timetEnergy purchased to the upper grid, +.>

Is virtual power plant VPPmAt the moment of timetElectric quantity price purchased to upper-level power grid and virtual power plant VPPnBid price is recorded as->

，/>

To be at the momentt，Virtual power plant VPPnFor virtual power plant VPPmReserve energy provided, +.>

Is a virtual power plant VPPmThe backup cost calculation using IBDR is as follows:

Equation 30

wherein ,

and />

Respectively at the moment of timet，Upper and lower spare capacity provided by IBDR, < + >>

And

prices corresponding to the upper and lower spare capacities, respectively; setting the virtual power plant VPP by the following formulamOptimal electricity purchasing combination:

equation 31

Equation 32

Equation 33

wherein ,

and />

At the moment for IBDRtThe minimum and maximum output force is provided, and the generated output force also satisfies the ascending and descending slopes about in the formulas 13 to 16 for IBDRBeam and start-stop time constraints.

In another aspect, an embodiment of the present invention provides a collaborative scheduling optimization apparatus for a virtual power plant group, including: the virtual power plant VPP cooperative operation module is used for reporting bidding price and external power supply quantity in a bidding transaction center in a bidding game mode after each virtual power plant meets self-load requirements, wherein the bidding transaction center integrates bidding information of the virtual power plant VPPs and settles bidding energy shares and market clearing prices of batches obtained by different virtual power plants VPPs, and simultaneously transmits the bidding price and the bidding energy to each virtual power plant VPP, so that internal energy supply and demand balance is ensured, and redundant energy is sold to a public power grid to maximize operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, energy storage systems ESS and flexible loads; and the multi-time-scale collaborative scheduling optimization module is used for considering the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV when the power scheduling plan is formulated, combining the day-ahead prediction values, correcting the prediction deviation of the wind power plant WPP and the photovoltaic power plant PV by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load scheduling plan, combining the real-time output of a random power supply, so as to purchase electric energy from other virtual power plants VPPs or upper power grids UPG and calling an excitation type demand response IBDR (hybrid power distribution R), and guaranteeing the real-time supply and demand balance of the electric energy.

Compared with the prior art, the invention has at least one of the following beneficial effects:

1. the multi-time scale collaborative operation optimization model can effectively link scheduling stages in different time scales, and consider the prediction out force of WPP and PV in each stage, so that the reliable supply of self power can be realized, and meanwhile, redundant power can be sent to UPG in a competitive price transaction mode, so that excess economic benefit is obtained. In addition, when emergency power supply service is required to be provided, IBDR, VPP and UPG can be called to realize the aim of VPP real-time power supply and demand balance.

2. The virtual power plant group collaborative scheduling optimization model can give consideration to the interaction relation of different VPPs in different stages, can establish a plurality of VPP optimal operation schemes in stages, namely establish an energy scheduling plan according to the predicted power before the day in the early stage, establish an optimal bidding strategy according to the predicted power before the day in the early stage by correcting the predicted power before the day scheduling plan, and achieve the real-time power supply and demand balance target by calling emergency power supply service in the real-time stage. In the MCP settlement mechanism, each VPPs will struggle to acquire transaction shares in a bid game in order to effectively improve its own operational benefits.

3. The improved chaotic ant colony algorithm can simulate bidding game processes of different subjects in stages, each subject belongs to a cooperative game problem in a day-ahead scheduling stage, each subject belongs to a non-cooperative game problem in a day-ahead scheduling stage, and each subject belongs to a non-cooperative game problem in a real-time scheduling stage. By utilizing the improved ant colony algorithm, different main body bidding game optimization is simulated in stages, so that the wind-light output fluctuation can be minimized while the overall energy supply cost of the micro-grid is minimum, multi-objective requirements such as maximizing bidding income and minimizing standby cost are realized, and a globally optimal virtual power plant group operation scheme is rapidly and accurately obtained.

4. And establishing a multi-time-scale collaborative operation optimization model of the virtual power plant group taking into consideration the multi-time-scale collaborative operation optimization model of the virtual power plant group by considering the collaborative relation of the three stages of the plurality of virtual power plants in the day, the day and the day, and fusing a chaotic search algorithm with high traversal solving precision and an ant colony algorithm with strong global optimizing capability to provide a chaotic ant colony optimizing algorithm for improving optimizing speed and solving efficiency.

In the invention, the technical schemes can be mutually combined to realize more preferable combination schemes. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and drawings.

Drawings

The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, and like reference numerals represent like parts throughout the several views:

FIG. 1 is a flow chart of a collaborative scheduling optimization method for a virtual power plant farm in accordance with an embodiment of the present invention;

FIG. 2 is a diagram of a multi-virtual power plant collaborative operation architecture according to an embodiment of the present invention;

FIG. 3 is a block diagram of a multiple virtual power plant project test system according to an embodiment of the invention;

FIG. 4 illustrates various virtual power plant load demands and WPP, PV available outputs for a typical load day according to an embodiment of the present invention;

FIG. 5 illustrates unit energy costs and the amount of power that can be delivered by different VPPs during an intra-day scheduling phase according to an embodiment of the present invention;

fig. 6 shows the price and quantity of electricity paid out by different VPPs at the day stage according to an embodiment of the invention.

Detailed Description

Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form a part hereof, and together with the description serve to explain the principles of the invention, and are not intended to limit the scope of the invention.

Referring to fig. 1, in one embodiment of the present invention, a collaborative scheduling optimization method for a virtual power plant group is disclosed, including: in step S102, a plurality of virtual power plant VPP cooperative operation systems: when the respective virtual power plants meet the self load demands, the residual power generation capacity declares bidding price and the output power in a bidding transaction center in a bidding game mode, the bidding transaction center integrates virtual power plant VPP bidding information and settles bidding energy shares and market price obtained by different virtual power plants VPP, and simultaneously transmits the bidding price and the bidding energy to each virtual power plant VPP, so that the internal energy supply and demand balance is ensured, and the redundant energy is sold to a public power grid to maximize the operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, an energy storage system ESS and flexible loads; and in step S104, the multi-time scale collaborative scheduling optimization of the virtual power plant VPP: when a power dispatching plan is formulated, the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV are considered, the day-ahead prediction values are combined, the wind power plant WPP and the photovoltaic power plant PV prediction deviation is corrected by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load dispatching plan, and the real-time output of a random power supply is combined, so that the power supply and demand balance in real time is ensured in a mode of purchasing power from other virtual power plants VPPs or upper power grids UPG and calling excitation type demand response IBDR.

Compared with the prior art, in the collaborative scheduling optimization method for the virtual power plant group, the multi-time-scale collaborative operation optimization model can effectively link scheduling stages in different time scales, and consider the prediction output force of WPP and PV in each stage, so that the reliable supply of self power can be realized, and meanwhile, redundant power is sent to UPG in a competitive transaction mode, so that excessive economic benefit is obtained. In addition, when emergency power supply service is required to be provided, IBDR, VPP and UPG can be called to realize the aim of VPP real-time power supply and demand balance.

Hereinafter, each step of the collaborative scheduling optimization method of a virtual power plant group according to an embodiment of the present invention will be described in detail with reference to fig. 1.

In step S102, a plurality of virtual power plant VPP cooperative operation systems: when the respective virtual power plants meet the self load demands, the residual power generation capacity declares bidding price and the output power in a bidding transaction center in a bidding game mode, the bidding transaction center integrates bidding information of virtual power plants VPP and settles bidding energy shares and market clearing prices obtained by different virtual power plants VPP, and meanwhile, the bidding price and the bidding energy are transmitted to each virtual power plant VPP, so that internal energy supply and demand balance is ensured, and excessive energy is sold to a public power grid to maximize operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, an energy storage system ESS and flexible loads.

In step S104, the multi-time scale collaborative scheduling optimization of the virtual power plant VPP: when a power dispatching plan is formulated, the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV are considered, the day-ahead prediction values are combined, the wind power plant WPP and the photovoltaic power plant PV are corrected by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load dispatching plan, and the real-time output of a random power supply is combined, so that the electric energy is purchased from the rest of the virtual power plant VPP or the upper power grid UPG, and the excitation type demand response IBDR is called, and the real-time supply and demand balance of the electric energy is ensured.

The multi-time scale collaborative scheduling optimization of the virtual power plant VPP comprises the following steps: day-ahead scheduling optimization, intra-day scheduling optimization and real-time scheduling optimization, wherein the day-ahead scheduling optimization: based on the wind power plant WPP and the photovoltaic power plant PV day-ahead predicted output, and in consideration of interaction and coordination among the gas turbine power generation CGT, the energy storage system ESS and the flexible load, a dispatching plan with optimal power supply cost is obtained; day scheduling optimization: the method comprises the steps of combining the wind power plant WPP and the photovoltaic power plant PV, correcting the power generation CGT output of an energy storage system ESS and a gas turbine so as to eliminate the influence caused by deviation of the wind power plant WPP and the photovoltaic power plant PV as much as possible, calculating unit power supply marginal cost and power delivery quantity in the virtual power plant VPP, reporting respective bidding price and bidding electric energy to a transaction center, and finally obtaining electric energy bidding transaction quantity and price returned by the transaction center; real-time scheduling optimization: and (3) considering the real-time output of the wind power plant WPP and the photovoltaic power plant PV, and ensuring the real-time balance of power supply and demand on the basis of a daily electric energy scheduling plan and an competitive price trading scheme, wherein when the real-time output of the wind power plant WPP and the photovoltaic power plant PV is smaller than the planned output, the electric energy is purchased from other virtual power plants VPPs or superior power grids UPG or IBDR emergency output is called, and otherwise the electric energy is sold or IBDR power is increased.

Specifically, the day-ahead schedule optimization includes: and combining wind power plant WPP and photovoltaic power plant PV day-ahead predicted power, and arranging power generation scheduling plans of different types of units by taking the minimum virtual power plant VPP operation cost as a target, wherein the virtual power plant VPP operation cost comprises wind and light energy abandoning cost, gas turbine power generation CGT operation cost and energy storage loss cost.

The day-ahead schedule optimization further comprises: the objective function is calculated by the following equation 1:

equation 1

wherein ,C _WPP,t andC _PV,t respectively representing the moment of wind power plant WPP and photovoltaic power plant PVtIs characterized by that the wind-discarding cost and the light-discarding cost are set,C _CGT,t indicating that gas turbine power generation CGT is at momenttIs added to the running cost of the (a) and (b),C _ESS,t indicating the time of day of the ESS of the energy storage systemtIs added to the running cost of the (a) and (b),C _PB,t indicating that price type demand response PBDR is at timetIs not limited by the implementation cost of the device.

The total cost of operation of the virtual power plant VPP is calculated according to the following equation 2, equation 5:

equation 2

Wherein RE represents a wind power plant WPP and a photovoltaic power plant PV,

indicating RE at timetIs used for generating the predicted output force of the engine,g _RE,t indicating RE at timetIs used to control the actual output of the device,c _RE,t indicating RE at timetIs added to the energy-discarding opportunity cost of the energy-discarding machine,p _RE,t for RE at timetIs connected with the internet for generating electricity;

equation 3

wherein ,a _CGT 、b _CGT 、c _CGT the CGT power generation output gas cost coefficient is generated for the gas turbine,

and />

The hot start cost and the cold start cost of the gas turbine power generation CGT are respectively represented,u _CGT,t indicating that gas turbine power generation CGT is at momenttIs->

Generating CGT minimum down time for gas turbine, < >>

Indicating that the gas turbine power generation CGT is at timetIs>

Representing a cold start time of the gas turbine power generation CGT;

equation 4

wherein ,g _ESS,t is the time of the ESS of the energy storage systemtCharge and discharge power wheng _ESS,t And < 0, indicating that the state of the energy storage system ESS is charged, and is discharged,ρ _ESS in order to adjust the coefficient of the power supply,

for the initial investment cost of the accumulator,N _t the service life of the storage battery is related to the depth of discharge;p _grid,t indicating that the virtual power plant VPP is at timetWith the electricity purchase and sale price of the public power grid,p _t indicating the time of day of the ESS of the energy storage systemtCharge and discharge price%>

and />

Respectively at the timetCharging and discharging efficiency of an energy storage system ESS of (E/S)>

and />

equation 5

wherein ,

andP _t respectively representing the charge price before and after PBDR at time t, < >>

AndL _t respectively shown intLoad demand before and after PBDR at time of day, attTime of day, deltaP _t and △L _PB , _t The price variable and the load variable before and after the PBDR are respectively, and the electricity demand price elasticity of the PBDR is as follows:

Equation 6

wherein ,E _st representing power demand-price elasticity, PBDR generated load fluctuations are calculated by the following formula:

equation 7

Calculating the moment of the representative virtual power plant VPP by the following formulatIs not required for the payload requirements of (2)M _t ：

Equation 8

wherein ,g _CGT,t indicating that gas turbine power generation CGT is at momenttIs characterized in that the power is generated by the power generation,η _CGT andη _ESS,t respectively representing the output loss rates of the gas turbine power generation CGT and the energy storage system ESS,g _UG,t representing the amount of power the virtual power plant VPP purchased from the UPG at time t,L _t representing the load demand of a virtual power plant VPP at time t，u _PB,t Indicating that PBDR is at time of dayt1 indicates that PBDR is implemented, and 0 indicates that PBDR is not implemented; the output power of the wind power plant WPP and the photovoltaic power plant PV is calculated by the following equation 9:

equation 9

Equation 10

wherein ,g _R for the WPP rated power of a wind power plant,v _t for the moment of timetWPP wind speed of the wind power plant,

at the moment for wind power plant WPPtAvailable force, & gt>

equation 11

Wherein the prediction errors of the WPP of the wind power plant and the PV of the photovoltaic power plant are respectivelye _WPP,t Ande _PV,t wind power output g _WPP,t Andg _PV,t the interval of (2) is [ (1 ]e _WPP,t )·g _WPP,t , (1+e _WPP,t )·g _WPP,t] and [(1-e _PV,t )·g _PV,t , (1+e _PV,t )·g _PV,t ]Selectinge _RE,t Substitution ofe _WPP,t Ande _PV,t ，g _RE,t substitution ofg _WPP,t Andg _PV,t ，g _RE,t will be distributed in [ (1 ]e _RE,t )·g _RE,t ，(1+e _RE,t )·g _RE,t ]，φ _RE Output loss rate of wind power or photovoltaic; as can be seen from the formula 11, when the uncertainty is strong, the load supply and demand imbalance is aggravated, usingθ _RE , _t and Γ_RE Correcting the load demand constraint to ensure the load supply and demand balance constraint, wherein the load demand constraint is corrected as follows:

equation 12

Γ is according to equation 12 _RE The introduction of the method provides a flexible risk decision tool for a decision maker to formulate a virtual power plant VPP scheduling scheme considering uncertainty according to the risk attitude of the decision maker.

Calculating a gas turbine power generation CGT operation constraint, an energy storage system ESS operation constraint and a PBDR operation constraint according to the following formulas: the gas turbine power generation CGT operation constraint is:

equation 13

Equation 14

Equation 15

Equation 16

and />

and />

and />

and />

Respectively representing the minimum start time and the minimum stop time of the gas turbine power generation CGT; the ESS operating constraints of the energy storage system are:

Equation 17

Equation 18

Equation 19

and />

and />

equation 20

wherein ,

indicating that the PBDR provides the maximum load fluctuation amount,φthe maximum load reduction specific gravity is indicated, and the peak-valley reverse hanging phenomenon caused by the user transient response is avoided by limiting the maximum load fluctuation amount provided by the PBDR and the load reduction specific gravity.

The daily scheduling optimization comprises the steps of based on a daily scheduling plan, according to daily predicted power of a wind power plant WPP and a photovoltaic power plant PV, generating power by calling an energy storage system ESS and PBDR or adjusting a gas turbine power generation CGT, coping with daily predicted power deviation of the wind power plant WPP and the photovoltaic power plant PV, calculating unit power generation cost and residual power generation capacity of a virtual power plant VPP in each period after finishing deviation correction of the wind power plant WPP and the photovoltaic power plant PV, and further constructing a multi-virtual power plant VPP power generation bidding game model.

The daily scheduling optimization further comprises the step of correcting a daily scheduling plan by taking the minimum fluctuation of wind and light output as a target, wherein the target function is as follows:

equation 21

Equation 22

wherein ,△L _PB,tˊ For PBDR day-ahead scheduling stage timetPlanned output at timetAt the location of the,g _PV,tˊ andg _WPP,tˊ representing the day-ahead dispatch output of the photovoltaic plant PV and the wind power plant WPP respectively,gˊ _PV,tˊ andgˊ _WPP,tˊ representing the actual available output of the photovoltaic plant PV and the wind plant WPP, respectively, at the momenttAt the location of the body resistanceLˊ _PB,t Andgˊ _ESS,t correction output force respectively representing PBDR and energy storage system ESS, and settingt″=tAnd (2) energy storage system ESS capacity constraints are as follows:

equation 23

、/>

、/>

、/>

and />

equation 24

wherein ,g _t to establish a unit energy supply cost when the internal load supply and demand balance premise of the virtual power plant VPP is satisfied according to the formula 1 and the formula 23 for the bidding output of the virtual power plant VPP participating in the power market at the time tC _VPP,t The unit energy supply cost of the virtual power plant VPPs at the time t is set, and different virtual power plants VPPs participate in the bidding of the power market to acquire excessive profits, if the bidding expectation profit rate is β _VPP,t The price of the virtual power plant VPP to participate in the electric market bidding transaction is calculated by the following formula:

equation 25

equation 26

Equation 27

Equation 28

Real-time scheduling optimization includes that the daily predicted power of the wind power plant WPP and the daily predicted power of the photovoltaic power plant PV still have deviation, so that the daily bidding game strategy is difficult to reach the optimal value, wherein when the actual tradable electricity quantity is higher than the bidding electricity quantity, a daily bidding scheme is still executed, otherwise emergency standby resources are required to be called to deal with the problem of unbalanced load supply and demand, and the standby sources comprise IBDR, virtual power plant VPPs and three channels of an upper power grid.

The real-time scheduling optimization further comprises: the objective function is expressed by the following formula:

Equation 29

，/>

equation 30

wherein ,

and />

And

equation 31->

Equation 32

Equation 33

wherein ,

and />

At the moment for IBDRtThe minimum and maximum output forces provided, for IBDR, also satisfy the up and down hill climbing constraints and the start-stop time constraints in equations 13-16.

In another embodiment of the present invention, a collaborative scheduling optimization device for a virtual power plant group is disclosed, including: the virtual power plant VPP cooperative operation module is used for reporting bidding price and output electric quantity in a bidding transaction center in a bidding game mode after the respective virtual power plant meets self-load requirements, the bidding transaction center integrates bidding information of the virtual power plant VPP and settles bidding energy share and market clear price of batches obtained by different virtual power plants VPP, and simultaneously transmits the bidding price and bidding energy to each virtual power plant VPP, so that internal energy supply and demand balance is ensured, and redundant energy is sold to a public power grid, and operation income is maximized, wherein the virtual power plant VPP comprises wind power plant WPP, photovoltaic power plant PV, gas turbine power generation CGT, an energy storage system ESS and flexible load; and the multi-time-scale collaborative scheduling optimization module is used for considering the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV when making a power scheduling plan, combining the day-ahead prediction values, correcting the prediction deviation of the wind power plant WPP and the photovoltaic power plant PV by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load scheduling plan, combining the real-time output of a random power supply, so as to purchase electric energy from other virtual power plants VPP or an upper power grid UPG and calling an excitation type demand response IBDR, and guaranteeing the real-time supply and demand balance of the electric energy.

According to the technical scheme, the optimal scheduling and operation problems of the virtual power plant are solved through a rolling hierarchical optimization thought of multi-time scale and multi-VPP cooperation. The heuristic of the technical scheme is to fully consider the output characteristic of the power supply: in the day-ahead dispatching stage, the output characteristics of various power supplies in the virtual power plants, such as wind power, photovoltaic randomness and intermittence characteristics, energy storage, a gas turbine and flexible load controllable characteristics are considered, so that the optimal dispatching in each virtual power plant is realized. Scheduling phase within day: mainly considering the energy mutual-aid requirement among the multiple VPPs, and realizing the supply and demand balance of electric energy through the non-cooperative game of each VPP. And (3) a real-time scheduling stage: the fine adjustment and correction of wind power output deviation are mainly considered, the specific mode is that the actual wind and light output is approximately calculated, electric energy is purchased from other VPP, UPG, IBDR, and otherwise, the electric energy is sold. The technical scheme is researched to comprise a virtual power plant consisting of WPP, PV, ESS, CGT and flexible loads; the technical scheme is that the optimal scheduling of various power supplies and loads in the virtual power plant is realized in the day-ahead stage; the concept of scheduling optimization in the day is that the electric energy among a plurality of VPPs is mutually balanced; the technical scheme utilizes the standby power supply to realize the correction of wind-light output deviation.

Hereinafter, a collaborative scheduling optimization method for a virtual power plant group according to an embodiment of the present invention will be described in detail by way of specific examples with reference to fig. 2 to 6.

a. And defining a multi-virtual power plant collaborative operation system. Referring to fig. 2, the virtual power plant mainly includes Wind Power Plant (WPP), photovoltaic power generation (Photovoltaic power, PV), energy storage system (Energy storage system, ESS), gas turbine power generation (Convention gas turbine, CGT), and flexible load. The flexibility load mainly relates to Price-type demand response (PBDR-based demand response) and Incentive-type demand response (IBDR-based demand response). Setting the internal unit equipment of the VPPs to belong to the same main body, carrying out power scheduling by different VPPs according to self-load demands, accounting the marginal energy supply cost and the energy which can be sent out by a unit, and analyzing the problem of energy exchange among multiple virtual power plants.

Considering a plurality of virtual power plant collaborative operation systems, when each VPP meets the self-load demand, the residual power generation capacity can report bidding price and externally-supplied power in a bidding transaction center in a bidding game mode, the bidding transaction center integrates VPP bidding information and settles bidding energy share and market price of different VPPs obtained by batch, and simultaneously, two information flows of the bidding price and the bidding energy are transmitted to each VPP, so that the purposes of ensuring internal energy supply and demand balance and selling the redundant energy to a public power grid and maximizing operation income can be realized.

b. Multiple time scales coordinated scheduling optimization of VPP. The WPP and the PV in the VPP have stronger random characteristics due to the influence of weather factors. When a power dispatching plan is formulated, the day-ahead prediction results (24 h) of WPP and PV are required to be fully considered, and the prediction deviation of WPP and PV is corrected by adjusting dispatching plans such as CGT, ESS, flexible load and the like in combination with the day-ahead (4 h) prediction values; and finally, combining the real-time (1 h) output of the random power supply to purchase electric energy from other VPPs or the upper power grid and call IBDR (integrated circuit) to ensure the real-time supply and demand balance of the electric energy. Thus, a three-stage cooperative operation mode of the virtual power plant group is formed, namely:

(1) And a day-ahead scheduling stage (24 h) which is a multi-power-source main body collaborative playing and optimizing problem in the virtual power plant. Predicting output based on WPP and PV day-ahead, and considering interaction coordination among CGT, ESS and flexible load to obtain a dispatching plan with optimal power supply cost;

(2) An intra-day scheduling stage (4 h), which is a multi-VPP non-cooperative gaming and optimization problem. Firstly, the WPP and the PV are combined to predict the force within the day, and the influence caused by the deviation of the WPP and the PV force is eliminated as much as possible by correcting the ESS and the CGT force. Secondly, calculating unit power supply marginal cost and the power quantity capable of being sent outwards in the VPP, reporting respective bidding price and bidding electric energy to a transaction center, and finally obtaining electric energy bidding transaction quantity and price returned by the transaction center;

(3) And a real-time scheduling stage (1 h), wherein the stage is a cooperative game problem of multiple standby sources. Specifically, on the basis of the daily electric energy scheduling plan and the bidding transaction scheme, if the real-time output of WPP (wind power plant) and PV (photovoltaic power plant) is smaller than the planned output, the electric energy is purchased from other VPP and UPG (Upper power grid) or the IBDR emergency output is called, otherwise, the electric energy is sold or the IBDR power is increased, so that the real-time balance of the electric power supply and demand is ensured.

1. Day-ahead scheduling optimization model

And in the day-ahead scheduling stage, the WPP and the PV day-ahead predicted power are combined, and the power generation scheduling plans of different types of units are arranged with the aim of minimum VPP running cost. The VPP operation cost mainly comprises wind and light energy abandoning cost, CGT operation cost and energy storage loss cost, and the specific objective function is as follows:

equation 1

wherein ,C _WPP,t andC _PV,t respectively representWPP and PV at time of daytAnd the cost of the waste wind and the cost of the waste light.C _CGT,t Indicating that the CGT is at timetStart-up and shut-down costs and fuel costs, i.e., operating costs.C _ESS,t Indicating the ESS at timetIncluding energy costs, battery costs, and opportunity return costs.C _PB,t Indicating that PBDR is at time of daytIs not limited by the implementation cost of the device.

Equation 2

Wherein RE stands for WPP or PV.

Indicating RE at timetIs a predicted power generation output.g _RE,t Indicating RE at timetIs used to control the actual output of the motor.c _RE,t Indicating RE at timetThe energy abandoning opportunity cost is mainly determined by factors such as unit construction cost, residual value, maintenance cost, labor cost and the like.p _RE,t For RE at timetIs connected with the internet for generating electricity.

Equation 3->

wherein ,a _CGT 、b _CGT 、c _CGT the gas cost coefficient for the CGT power generation output,

and />

The CGT hot start cost and cold start cost are represented, respectively.u _CGT,t Indicating that the CGT is at timetIs set in the operating state of (a). />

Minimum downtime for CGT. />

Indicating that the CGT is at timetIs not limited by the time required for the shutdown. />

Indicating the cold start time of the CGT.

Equation 4

In the formula 4 of the present invention,g _ESS,t is the ESS at the momenttCharge and discharge power wheng _ESS,t When < 0, it indicates that the state of the ESS is charged, and conversely, discharged,ρ _ESS in order to adjust the coefficient of the power supply,

is the initial investment cost of the battery.N _t The service life of the storage battery is related to the depth of discharge;p _grid,t indicating that VPP is at timetWith the electricity purchase and sale price of the public power grid,p _t indicating the ESS at timetCharge and discharge price%>

and />

Respectively at the timetESS charge-discharge efficiency of>

and />

Representing the cost of charge and discharge loss per unit of power of the ESS.

Equation 5

wherein ,

andP _t respectively at the time tBefore and after PBDR. />

and />

Respectively shown intLoad demand before and after time PBDR. At the position oftAt this time, the price variable and the load variable before and after PBDR are respectively expressed as deltaP _t and △L _PB , _t . PBDR is characterized in terms of electricity demand price flexibility as follows.

Equation 6

wherein ,E _st representing electricity demand-price elasticity. Further, load fluctuations generated by PBDR were calculated as follows:

equation 7

The total cost of operation of the VPP can be calculated according to equations 2 to 5, and further, constraints on the operation of different VPPs need to be considered. In consideration of uncertainty of WPP and PV output power, risks caused by deviation of WPP and PV predicted power are considered when a day-ahead power generation scheduling plan is carried out. The robust random optimization theory of uncertainty of the uncertainty parameter interval description is adopted, the method has the characteristic of less requirement on random variable probability distribution information, and meanwhile, decision tools can be provided for decision makers with different risk sensitivity degrees. Setting upM _t Representing the moment of VPPtThe specific calculation mode is as follows:

equation 8

wherein ,g _CGT,t indicating that the CGT is at timetIs used for generating power。η _CGT Andη _ESS,t the output loss rates of the CGT and ESS are shown, respectively.g _UG,t Indicating the amount of power the VPP purchased from the UPG at time t. L _t Indicating that VPP is at timetIs not required for the load demand of the vehicle.u _PB,t Indicating that PBDR is at time of daytThe implementation state of 0-1 variable, 1 indicating that PBDR is implemented, and not implemented otherwise. Further, considering the output power of WPP and PV, after natural wind and solar radiation intensity are obtained, the output power of WPP and PV can be calculated by Weibull distribution function and Beta distribution function, specifically as follows:

equation 9

Equation 10

wherein ,g _R rated for WPP.v _t For the moment of timetWPP wind speed of (2).

At the moment for WPPtIs used to generate the available force.

For the moment of PVtIs used to generate the available force.η _PV AndS _PV solar radiation efficiency and radiation area.θ _t To be at the momenttSolar radiation intensity.

However, the output power of WPP and PV has a strong dependence, and the prediction errors of WPP and PV are set as respectivelye _WPP,t Ande _PV,t wind power output powerg _WPP,t Andg _PV,t the usable interval is described as [ (1 ]e _WPP,t )·g _WPP,t , (1+e _WPP,t )·g _WPP,t] and [(1-e _PV,t )·g _PV,t , (1+e _PV,t )·g _PV,t ]. For easy analysis, selecte _RE,t Substitution ofe _WPP,t Ande _PV,t ，g _RE,t substitution ofg _WPP,t Andg _PV,t . In a corresponding manner,g _RE,t will be distributed ing _RE,t Will be distributed in [ (1 ]e _RE,t )·g _RE,t ，(1+e _RE,t )·g _RE,t ]. Thus, the load demand constraints of VPP operation are as follows:

equation 11

wherein ,φ _RE the output loss rate of wind power or photovoltaic power.

As can be seen from equation 11, when the uncertainty is strong, the load supply and demand imbalance is also aggravated, so as to ensure that the load supply and demand balance constraint is satisfied, using θ _RE , _t and Γ_RE Correcting the constraint conditions, wherein the constraint conditions are as follows:

equation 12

Γ is according to equation 12 _RE The introduction of the method can provide a flexible risk decision tool for a decision maker to formulate a VPP scheduling scheme considering uncertainty according to the risk attitude of the decision maker. The specific constraint conditions are as follows:

1) CGT operation constraint

Equation 13

Equation 14

Equation 15

Equation 16

wherein ,u _CGT,t indicating that the CGT is at timetIs set in the operating state of (a).

and />

Representing CGT minimum and maximum output power, respectively. />

and />

The downslope power and the upslope power of the CGT are respectively indicated. />

and />

Respectively represent the CGT at the momentt-1 continuous run time and continuous down time. />

and />

Representing the minimum start-up time and minimum shut-down time of the CGT, respectively.

2) ESS operation constraints

Equation 17

Equation 18

Equation 19

wherein ,S _ESS,t indicating the ESS at timetIs used for storing energy.

and />

Respectively representing the maximum charge-discharge power of the ESS.

and />

Representing the minimum and maximum stored energy of the ESS, respectively. In order to fully utilize the charge and discharge performance of the ESS, the ESS scheduling Zhou Qichu and the end energy storage energy of the scheduling period are set to be zero, namelyS _ESS,0 =S _ESS,T =0。

3) PBDR operation constraints

Equation 20

wherein ,

indicating that PBDR may provide maximum load fluctuation.φIndicating maximum load reduces specific gravity. By limiting the maximum load fluctuation amount and load shedding specific gravity provided by the PBDR, the phenomenon that the user transient response causes the "peak Gu Daogua" can be avoided. Also, the generated output provided by PBDR also needs to meet the start-stop time of equations 15 through 16.

2. Day scheduling phase

Based on the day-ahead schedule, WPP and PV day-ahead predicted power deviations are handled by invoking ESS, PBDR or adjusting CGT power generation based on WPP and PV day-ahead predicted power. After the WPP and PV deviation correction is completed, unit power generation cost and residual power generation capacity of the VPP in each period are measured and calculated, and then a multi-VPP power generation bidding game model is constructed. Firstly, correcting a day-ahead scheduling plan by taking the minimum fluctuation of wind and light output as a target, wherein a specific objective function is as follows:

equation 21

Equation 22

wherein ,△L _PB,tˊ For PBDR day-ahead scheduling stage timetPlanned output. Time of daytAt' point, the day-ahead scheduled outputs of PV and WPP are represented asg _PV,tˊ Andg _WPP,tˊ the actual available output of PV and WPP are respectively expressed asgˊ _PV,tˊ Andgˊ _WPP,tˊ . At the moment of timetAt the }' point, the corrected output forces of PBDR and ESS are denoted as delta, respectivelyLˊ _PB,t Andgˊ _ESS,t . In addition, the setting is thatt″=tAnd ± 1, then ESS capacity constraint is as follows:

equation 23

Also, the modified ESS operating force also satisfies the constraints of

equations

18 and 19. After the day-ahead schedule is completed, the day-ahead schedule of different virtual power plants can be determined at this time, i.e

、/>

、/>

、/>

And

. At this time, the remaining power supply capacity of each VPP may be calculated as follows:

equation 24

wherein ,g _t at the moment of VPPtCan participate in the bidding output of the electric power market. Further, the unit energy supply cost when the VPP internal load supply and demand balance premise is satisfied can be established according to the

formulas

1 and 23, and the setting can be madeC _VPP,t The unit energy cost at time t for VPP. The purpose of the different VPPs participating in the bidding of the electric power market is to obtain excess revenue, if the bidding expectation rate of return isβ _VPP,t The price quoted by the VPP to participate in the electric market bidding transaction is calculated as follows:

equation 25

The daily bidding price and bidding quantity of the VPP can be determined by the

formulas

24 and 25, in fact, when a plurality of VPPs participate in the power bidding in the system, a plurality of bidding schemes are formed, the system performs energy transaction according to the price as a selection standard until the energy balance is met, which is an infinitely repeated game process, and an independent bidding process exists in each period. The bidding process of each period needs to report the bidding quantity and bidding price. In the bidding process, if the operators of each VPP are enough to be able to offer reasonable prices, each individual will acquire ideal revenue in dynamic balance. The application introduces a function argmax g (), and under the situation that a plurality of VPPs participate in the bidding of the electric power market, the optimal strategy is as follows:

Equation 26

wherein ,VPPmThe bidding strategy is expressed asB _m ，VPPmThe optimal bidding strategy is recorded as

And the energy supply scheme in the optimal bidding strategy is expressed as +>

。

Transaction settlement is required after competitive price transaction is completed, common settlement mechanisms comprise a market price clearing MCP mechanism and a PAB mechanism, and in view of different characteristics of the two settlement mechanisms, the research adopts MCP and sets the lowest average energy supply cost as an optimization target, as shown in a formula 27.

Equation 27

Equation 28

wherein ,L _UPG,t for UPG at timetM represents the VPP number.

3. Real-time scheduling phase

Although scheduled to run on the day, the WPP and PV day-ahead predictive bias has been corrected and multi-VPPs developed for bid gaming optimization based on WPP and PV day-ahead predictive models. However, due to the strong uncertainty of WPP and PV output power, there may still be a bias in the intra-day predicted power of WPP and PV during the real-time scheduling phase, resulting in a less than optimal intra-day bidding game strategy. Wherein the intra-day bidding scheme is still executable when the actual tradable amount of electricity is higher than the bid amount of electricity. Otherwise, emergency standby resources need to be called to solve the problem of unbalanced supply and demand of the load. The standby sources mainly comprise IBDR, VPPs and three channels of an upper power grid. The objective function is shown in equation 29.

Equation 29

wherein ,R _m,t is VPPmAt the moment of timetSpare cost of (a) is provided.

Is VPPmAt the moment of timetEnergy purchased to the higher grid. />

Is VPPmAt the moment of timetAnd (5) purchasing electricity to the upper power grid. VPP (virtual private plane)nBid price is recorded as->

. Time of dayt，VPPnFor VPPmThe standby energy provided is expressed as +.>

，/>

Is VPP (virtual private plane)mThe backup cost calculation using IBDR is as follows:

equation 30

Wherein at the moment of timet，IBDR provides up and down spare capacities of

and />

The prices of the corresponding upper and lower spare capacities are respectively marked +.>

and />

. Finally, setting VPPmAn optimal electricity purchasing combination is adopted,expressed as formulas 31 to 33.

Equation 31

Equation 32

Equation 33

wherein ,

and />

At the moment for IBDRtProviding minimum and maximum force. Also, for IBDR, the power generation output thereof also needs to satisfy the uphill and downhill constraints and the start-stop time constraints in equations 13 to 16.

4. Model solving step

The virtual power plant multi-time-scale collaborative optimization operation model mainly covers 3 time scales, and the detailed analysis steps are as follows:

(1) In the day-ahead scheduling stage, a day-ahead scheduling plan is formulated based on day-ahead predicted values of the random power supply and the flexible load, and the electric quantity and the price of each VPP competitive price transaction are obtained;

(2) And in the day scheduling stage, based on the day predicted value of the random power supply and the flexible load, adopting formulas 25-27 to correct the day scheduling plan, then using

formulas

28 and 29 to calculate the electric quantity and price of the VPP which can actually participate in competitive price transaction, and using formulas 30-32 to establish the optimal competitive price strategy of the multi-VPP. The method adopts an ant colony algorithm to carry out optimizing solution to obtain an optimal transaction scheme of the multi-VPPs at the stage;

(3) And in the real-time scheduling stage, the daily scheduling plan deviation is corrected based on the actual values of the random power supply and the flexible load to obtain the electric quantity actually participating in competitive price transaction, and if the electric quantity shortage situation occurs, the real-time balance of power supply and demand is maintained in a mode of calling IBDR (customer-side IBDR) and purchasing electric energy from other VPPs or UPGs according to the principle of lowest standby cost.

5. Detailed description of the preferred embodiments

According to the method, a multi-virtual power plant demonstration project in a certain place in China is taken as an example system, effectiveness and applicability of the VPPs energy collaborative management model and the solving algorithm are verified, the regional incremental power distribution network structure is set, and the incremental power distribution network structure comprises a transformer substation 1 and a transformer substation 2. The transformer substation 1 comprises an outlet A and an outlet B which are respectively connected with the VPP 1 and the VPP 2. Substation 2 includes a C outlet connected to VPP 3. And setting two transformer stations to perform energy interaction, and connecting different VPPs with each other through transformer stations to realize energy interaction. The charge and discharge power, charge and discharge loss and initial energy storage capacity of the ESS system are respectively set to 0.12 MW, 4% and 0. FIG. 3 is a block diagram of a multiple virtual power plant project test system.

Among them, VPP 1 includes 6×0.4MW WPP, 3×0.15MW PV, 1×2.5MW CGT, and 1×1.0MW CGT, and matches 2×0.5 MW.h ESS. VPP 2 includes 5×0.2MW WPP, 2×0.2MW PV, and 1×2MWCGT, and matches 1×0.5MW h ESS. VPP 3 includes 4×0.4MW, 6×0.15MW PV, 1×2.5MWCGT and 1×2MWCGT, and matches 3×0.5MW h ESS. The specific parameters of the specific unit parameters selected by the CGT are shown in the table 1.

TABLE 1 CGT operating parameters

In order to simulate the optimized operation of different VPPs, the fan parameters are as followsv _in =3m/s，v _rated =14m/s,v _out =25m/sThe shape parameters and the dimension parameters phi=2,

solar photovoltaic radiation intensity parameterαAndβ0.39 and 8.54 respectively. Further, the proposed scene generation and reduction strategy yields 10 sets of typical simulation scenesThe method comprises the steps of selecting a field with the largest peak-valley difference as day-ahead prediction data, selecting a scene with the largest fluctuation as day-in prediction data, and selecting a scene with the largest occurrence probability as real-time data. The regional load demand on the typical load day is selected as the input data. The initial robust coefficient takes 0.5 and the prediction errors for pv and WPP are set to 0.04 and 0.05, respectively. The load demand of different VPPs at a typical load day and the WPP, PV available output are shown in figure 4.

Finally, the load fluctuation generated by setting the PBDR cannot exceed + -0.05 MW. At the same time, the available force that IBDR can provide cannot exceed ±0.1 MW, setting the re-real time correction phase. When the IBDR fails to meet enough backup requirements, the VPP may purchase electricity to other VPPs or to the upper grid. Setting the maximum iteration number of the algorithm as follows N _max =200. Randomly generating ants with individual quantity ofn=35, get parameters

，Ψ= 1.2501. Setting the minimum volatility coefficientρ _min =0.1, constantQ=1, and the proportion of variant individuals was set to be selected as 70% of the population size. And solving the provided multi-time-scale collaborative operation optimization model by acquiring the related data.

(1) Algorithm effectiveness analysis

And 5-class optimization algorithm is adopted to solve the multi-time-scale bidding game model, and the optimization solving result is shown in table 2.

Table 2 optimization solutions for various algorithms

As can be seen from table 2, the average number of convergence of ICACO, IACO, and ACO 3-type algorithms is less than that of GA and PSO algorithms, indicating that the global convergence capacity of the ACO algorithm is better than that of GA and PSO algorithms. In addition, ICACO algorithmF _cost 、F _Bidding 、R _VPP And the convergence times are superior to those of the GA algorithm and the PSO algorithm, and the optimizing efficiency of the chaotic search enhanced ACO algorithm is verified. Average convergence of ICACO algorithmThe number is less than that of ACO algorithm, and ICACO algorithm and IACO algorithmF _cost 、F _Bidding AndR _VPP the value is also superior to ACO algorithm, which verifies that the control parameter is improvedα、βAnd a volatilization factorρVarious performances of the ACO algorithm can be improved, and better results are obtained. In summary, the ICACO algorithm proposed by the present application can be used to design optimal bidding and running strategies for multi-VPPs.

(2) Analysis of the results of the calculation

Aiming at minimizing the power supply cost, and combining the day-ahead predicted output of the fluctuation renewable energy source to establish a day-ahead scheduling plan of the virtual power plant group. In order to pursue the minimum unit power supply cost, the VPP can call WPP and PV preferentially, the residual load is provided by CGT, but the uncertainty of the VPP can also bring great influence to the system, and in order to cope with the uncertainty of the WPP and PV, the system can call IBDR and ESS to provide standby service for VPP power generation scheduling.

TABLE 3 day-ahead schedule results for virtual power plant groups

According to table 3, in the day-ahead scheduling stage, although the power generation costs of WPP and PV are low, there are waste wind and waste light due to the fluctuation of the output thereof, and in the case of VPP 1, the total waste wind and waste light are 4.349mw·h and 0.651mw·h, respectively. Also, as VPP 3 has more PV power generation output during peak hours, ESS is invoked more to cope with the fluctuation of PV power generation output. From the power supply cost point of view, VPP 1 has the lowest unit power supply cost because of more WPP and PV output, and VPP 3 has higher unit power supply cost because of higher standby service cost because of the shortage of load supply and demand relationship in peak period despite more PV output.

2) Analysis of intra-day scheduling optimization results

The present section responds to WPP and PV intra-day predicted power deviations by modifying ESS schedule and CGT output based on WPP and PV intra-day predicted results. The CGT running state is determined by a day-ahead scheduling plan, and the generation power of the CGT is only modified at the stage. Because the WPP and the PV day-ahead predicted power deviate, the ESS balance deviation is preferentially called, and the real-time power supply and demand balance is maintained by correcting the CGT output. After the correction of the generated output is completed, the electric quantity and the generation cost of different VPPs which can participate in the competitive price transaction of the upper power grid can be calculated according to the daily scheduling plan of the multiple virtual power plants.

According to fig. 5, the unit supply costs and the biddable power amounts of the different VPPs are analyzed. VPP 1 has more WPP output, more power may be available to participate in the bidding transaction during the valley period, and VPP 3 has more PV output, more power may be available to participate in the bidding transaction during the peak period. From the unit power supply cost, VPP 1 has the lowest unit power supply cost because of more WPP and PV available, and VPP 2 and VPP 3 have certain fluctuation in unit power supply cost at different stages because of the difference of the output structures. For example, at 22:00-24:00, the VPP 2 unit supply cost is higher than VPP 3, and during the remaining period VPP 3 unit supply cost is higher than VPP 2. Further, according to the aim of lowest average power supply cost, different VPPs are matched to participate in competitive price transaction.

According to fig. 6, the price and the electric quantity share of different VPPs during the competitive price transaction are analyzed. In the valley period, since VPP 1 has more WPP power generation output and the unit power supply cost is lower, more competitive transaction shares are obtained, in the peak period, since VPP 3 has more PV power generation output, more transaction shares are obtained in the peak period, and in addition, since the unit power supply cost is lower than VPP 2 in 23:00-24:00, more transaction shares are obtained. Because the unit power supply cost of the VPP 2 is relatively low, more competitive price transaction shares can be obtained after the VPP 1 obtains the shares.

TABLE 4 competitive bidding power and competitive bidding yields for virtual Power plant groups during the day period

According to table 4, different VPP bid transaction results are analyzed. The price competition share obtained by VPP 1 is the greatest, so the price competition benefits are far higher than those obtained by VPP 2 and VPP 3, which are 3220.569 and 5506.485 respectively. The bid benefit of VPP 1 and VPP 3 is lower than VPP 2, but the benefit of CGT in the benefit of VPP 2 is higher than PV and WPP. VPP 3 has a higher PV available output, so the bid transaction revenue from PV is 110.486 and 131.059 yuan higher than VPP 1 and VPP 2, respectively. In general, in the bidding transaction process of a plurality of VPPs, the WPP and the PV have low power generation cost, so that the bidding game advantage of the VPP can be improved, and more shares can be obtained. The method is beneficial to improving the operation income of the VPP, promoting the larger-scale utilization of the WPP and the PV and realizing the optimization and the upgrading of the power supply structure.

(3) Real-time scheduling optimization result analysis

Based on the day-ahead scheduling plan and the day-ahead bidding game result, the deviation of the actual output of the fluctuation renewable energy source and the day-ahead predicted output is considered, and the real-time balance of the supply and demand of the electric energy is ensured by calling the output to the rest VPP and UPG or adopting an IBDR mode. If the emergency power dispatching is needed, the power supply cost at each moment in different modes is comprehensively considered, and the output scheme is corrected in real time with the aim of minimizing the cost. In order to cope with the real-time output deviation of the WPP and the PV, the VPPs can preferentially call the IBDR to provide flexible electric energy, and then the emergency power supply main body is reasonably and orderly called by comprehensively considering the corresponding cost when other VPPs and UPG are adopted for power supply. When the real-time output of the fluctuating renewable energy source is lower than expected, IBDR and other VPPs can be invoked, and if the power demand is still not met or the power supply cost is high, the VPPs can purchase electric energy from the UPG. In summary, VPPs can take into account the emergency power costs of IBDR, other VPPs and UPGs at various moments to make real-time corrections to the power scheduling strategy. The real-time stage virtual power plant group standby scheduling results are shown in table 5.

TABLE 5 real-time stage virtual Power plant Cluster Standby Dispatch results

As can be seen from Table 5, the real-time contribution of WPP and PV deviate from the predicted value, so IBDR, other VPPs and UPG are invoked. From the emergency power supply constitution of different VPPs, the power consumption of the VPP 1 to the VPP 2 is 0.43MW & h, the power consumption of the VPP 1 to the UPG is 0.39MW & h, the power consumption of the VPP 3 to the VPP 1 is-0.1 MW & h, the power consumption of the VPP 2 to the VPP 2 is 0.19MW & h and 0.16MW & h, and each VPPs purchases emergency power supply service to the IBDR so as to realize real-time power supply and demand balance.

The effect is that: the method and the device can quickly and accurately acquire the globally optimal virtual power plant group operation scheme. The reason for the effect: the provided virtual power plant group collaborative scheduling optimization model can give consideration to the interaction relation of different VPPs in different stages, can establish a plurality of VPP optimal operation schemes in stages, namely establish an energy scheduling plan according to the predicted power before the day in the early stage, establish an optimal bidding strategy according to the predicted power before the day in the early stage by correcting the predicted power before the day, and achieve the real-time power supply and demand balance target by calling emergency power supply service in the real-time stage.

Those skilled in the art will appreciate that all or part of the flow of the methods of the embodiments described above may be accomplished by way of a computer program to instruct associated hardware, where the program may be stored on a computer readable storage medium. Wherein the computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.

The present invention is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present invention are intended to be included in the scope of the present invention.

Claims

1. The collaborative scheduling optimization method for the virtual power plant group is characterized by comprising the following steps of:

a plurality of virtual power plant VPP co-operating systems: when the respective virtual power plants meet the self load demands, the residual power generation capacity declares bidding price and the output power in a bidding transaction center in a bidding game mode, the bidding transaction center integrates virtual power plant VPP bidding information and settles bidding energy shares and market price obtained by different virtual power plants VPP, and simultaneously transmits the bidding price and the bidding energy to each virtual power plant VPP, so that the internal energy supply and demand balance is ensured, and the redundant energy is sold to a public power grid to maximize the operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, an energy storage system ESS and flexible loads; and

virtual power plant VPP multi-time scale collaborative scheduling optimization: when a power dispatching plan is formulated, the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV are considered, the day-ahead prediction values are combined, the wind power plant WPP and the photovoltaic power plant PV prediction deviation is corrected by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load dispatching plan, and the real-time output of a random power supply is combined, so that the power supply and demand balance in real time is ensured in a mode of purchasing power from other virtual power plants VPPs or upper power grids UPG and calling excitation type demand response IBDR.

2. The method of collaborative scheduling optimization for a virtual power plant farm according to claim 1, wherein the multi-time scale collaborative scheduling optimization for virtual power plant VPP comprises: day-ahead scheduling optimization, day-in scheduling optimization, and real-time scheduling optimization, wherein,

the day-ahead schedule optimization: based on the wind power plant WPP and the photovoltaic power plant PV day-ahead predicted output, and considering interaction and coordination among the gas turbine power generation CGT, the energy storage system ESS and the flexible load, a scheduling plan with optimal power supply cost is obtained;

the intra-day schedule optimization: combining the wind power plant WPP and the photovoltaic power plant PV intra-day predicted output, correcting the power generation CGT output of the energy storage system ESS and the gas turbine so as to eliminate the influence caused by the deviation of the wind power plant WPP and the photovoltaic power plant PV output as much as possible, calculating the unit power supply marginal cost and the unit power supply quantity in the virtual power plant VPP, reporting the bidding price and bidding electric energy to a transaction center, and finally obtaining the electric energy bidding transaction quantity and price returned by the transaction center; and

the real-time scheduling optimization: and considering the real-time output of the wind power plant WPP and the photovoltaic power plant PV, and ensuring the real-time balance of power supply and demand on the basis of a daily electric energy scheduling plan and a competitive price transaction scheme, wherein when the real-time output of the wind power plant WPP and the photovoltaic power plant PV is smaller than the planned output, the electric energy is purchased from other virtual power plants VPPs or superior power grids UPG or the IBDR emergency output is called, and otherwise, the electric energy is sold or the IBDR electricity is increased.

3. The collaborative scheduling optimization method for a virtual power plant farm according to claim 2, wherein the day-ahead scheduling optimization comprises: and combining the wind power plant WPP and the photovoltaic power plant PV day-ahead predicted power, and arranging power generation scheduling plans of different types of units by taking the minimum virtual power plant VPP operation cost as a target, wherein the virtual power plant VPP operation cost comprises wind and light energy discarding cost, gas turbine power generation CGT operation cost and energy storage loss cost.

4. A method of collaborative scheduling optimization for a virtual power plant farm according to claim 3, wherein the day-ahead scheduling optimization further comprises:

the objective function is calculated by the following equation 1:

equation 1

wherein ,

and />

Respectively representing the moment of the wind power plant WPP and the moment of the photovoltaic power plant PVtIs a wind-discarding cost and a light-discarding cost, +.>

Indicating that the gas turbine power generation CGT is at timetRunning costs of->

Indicating the time of day of the ESS of the energy storage systemtRunning costs of->

Indicating that price type demand response PBDR is at timetThe implementation cost of (2); the total cost of operation of the virtual power plant VPP is calculated according to the following equation 2, equation 5:

equation 2

Wherein RE represents the wind power plant WPP and the photovoltaic plant PV,

Indicating RE at timetIs used for generating the predicted output force of the engine,

indicating that the RE is at timetIs>

Indicating that the RE is at timetEnergy abandoning opportunity cost of->

At the moment for the REtIs connected with the internet for generating electricity;

equation 3

wherein ,

、/>

、/>

generating a CGT power generation output gas cost coefficient for the gas turbine, < >>

And

representing the hot start cost and the cold start cost of the gas turbine power generation CGT respectively>

Indicating that the gas turbine power generation CGT is at timetIs->

Generating CGT minimum down time for said gas turbine, < > for>

Indicating that the gas turbine power generation CGT is at timetIs>

Representing a cold start time of the gas turbine power generation CGT;

equation 4

wherein ,

is the moment the energy storage system ESS istCharge and discharge power at the time when->

When the state of the ESS is charging or discharging, the state of the ESS is represented by +.>

For regulating the coefficient->

For the initial investment cost of the accumulator,

the service life of the storage battery is related to the depth of discharge; />

Indicating the virtual power plant VPP at the momenttElectricity purchase price with public power grid, +.>

Indicating the time of day of the ESS of the energy storage systemtCharge and discharge price%>

and />

Respectively at the timetIs the charging and discharging efficiency of the ESS of the energy storage system, < > >

and />

equation 5

wherein ,

and />

Respectively representing the price of the electric quantity before and after the PBDR at the time t,/>

and />

Respectively shown intLoad demand before and after PBDR at time of day, attTime of day (I)>

and />

The price variable and the load variable before and after the PBDR are respectively, and the PBDR is elastically as follows according to the price of the power demand:

equation 6->

wherein ,

representing power demand-price elasticity, the PBDR generated load fluctuation is calculated by the following formula:

equation 7

Calculating a virtual power plant VPP representative of the moment in time by the following formulatIs not required for the payload requirements of (2)

：

Equation 8

wherein ,

indicating that the gas turbine power generation CGT is at timetIs to generate power, < >>

and />

Output loss rates of the gas turbine power generation CGT and the energy storage system ESS are respectively represented by +.>

Representing the amount of electricity purchased by the virtual power plant VPP at time t from the UPG, +.>

Representing the load demand of the virtual power plant VPP at time t,/>

Indicating that the PBDR is at timet1 indicates that the PBDR is implemented, and 0 indicates that it is not implemented; the output power of the wind power plant WPP and the photovoltaic power plant PV is calculated by the following equation 9:

equation 9

Equation 10

wherein ,

rated power for WPP of said wind power plant, +.>

For the moment of timetWPP wind speed of wind power plant, +.>

At time for the wind power plant WPPtAvailable force, & gt>

At time for a photovoltaic plant PVtAvailable force, & gt>

and />

For solar radiation efficiency and radiation area +.>

To be at the momenttSolar radiation intensity; calculating a load demand constraint for the virtual power plant VPP operation by:

equation 11

Wherein the prediction errors of the wind power plant WPP and the photovoltaic power plant PV are respectively as follows

and />

Wind power output +.>

and />

The intervals of (2) are>

And

selecting->

Substitute->

and />

，/>

Substitute->

and />

，

Will be distributed in->

，/>

Output loss rate of wind power or photovoltaic; as can be seen from the above equation 11, when the uncertainty is strong, the load supply and demand imbalance is aggravated, using +.>

and />

Correcting the load demand constraint to ensure the load supply and demand balance constraint, wherein the load demand constraint is corrected as follows: />

Equation 1

According to the above-mentioned formula 12,

is introduced into decision maker to make decision according to own risk attitudeVirtual power plant VPP scheduling schemes that take into account uncertainty provide flexible risk decision tools.

5. The collaborative scheduling optimization method for a virtual power plant farm according to claim 4, wherein the gas turbine power generation CGT operating constraint, the energy storage system ESS operating constraint, and the PBDR operating constraint are calculated by the following formulas, respectively:

The gas turbine power generation CGT operation constraint is as follows:

equation 13

Equation 14

Equation 15

Equation 16

wherein ,

indicating that gas turbine power generation CGT is at momenttIs->

and />

and />

and />

and />

equation 17

Equation 18

Equation 19

wherein ,

indicating the time of day of the ESS of the energy storage systemtIs (are) stored energy>

and />

Respectively represent storageMinimum charge-discharge power and maximum charge-discharge power of ESS of system, and->

and />

Respectively representing the minimum energy storage capacity and the maximum energy storage capacity of the ESS of the energy storage system, and in order to fully utilize the charge and discharge performance of the ESS of the energy storage system, the initial energy storage capacity of the ESS scheduling period of the energy storage system is +.>

And scheduling period end energy storage +.>

All set to zero; the PBDR operating constraints are:

equation 20

wherein ,

indicating that said PBDR provides maximum load fluctuation amount, < >>

And a step of limiting the maximum load fluctuation amount provided by the PBDR and the load reduction specific gravity to avoid a phenomenon that a user transient response causes a peak-valley hang-up.

6. The collaborative scheduling optimization method for a virtual power plant group according to claim 4, wherein the daily scheduling optimization comprises calculating unit power generation cost and residual power generation capacity of a virtual power plant VPP in each period of time after finishing correction of the deviation of the wind power plant WPP and the photovoltaic power plant PV according to the daily scheduling plan, by calling the ESS, the PBDR or adjusting the CGT power generation capacity of the gas turbine power generation, according to the wind power plant WPP and the photovoltaic power plant PV, so as to construct a multi-virtual power plant VPP power generation bidding game model.

7. The collaborative scheduling optimization method for a virtual power plant farm according to claim 6, wherein the intra-day scheduling optimization further comprises targeting a minimum of wind and solar power fluctuation, modifying the pre-day scheduling plan, and the objective function is as follows:

equation 21

Equation 22

wherein ,

scheduling phase time before day for the PBDRtPlanned output at time +.>

At (I) a part of>

And

day-ahead dispatch output, respectively, representing the photovoltaic plant PV and the wind plant WPP>

and />

Representing the actual available output of the photovoltaic plant PV and the wind plant WPP, respectively, at the moment +. >

At (I) a part of>

and />

The corrected output force of the PBDR and the ESS are respectively indicated, and +.>

The ESS capacity constraints of the energy storage system are as follows:

equation 23

、/>

、/>

、/>

and />

equation 24

wherein ,

for the virtual power plant VPP to participate in the bidding output of the electric power market at the time t, establishing the unit energy supply cost when the internal load supply and demand balance premise of the virtual power plant VPP is satisfied according to the formula 1 and the formula 23, and adding ∈>

Setting the unit energy supply cost of the virtual power plant VPPs at the time t, wherein the purpose of the different virtual power plant VPPs participating in the bidding of the electric power market is to acquire excess profits, and if the expected profit rate of bidding is +.>

The price of the virtual power plant VPP to participate in the electric market bidding transaction is calculated by the following formula:

equation 25

Determining the daily bidding price and bidding quantity of the virtual power plant VPP according to a formula 24 and a formula 25, forming a plurality of bidding schemes when a plurality of virtual power plant VPPs participate in the power bidding, and carrying out energy transaction by the system according to the price as a selection standard until the energy balance is met; introducing functions

In the scenario where multiple virtual power plants VPP participate in the power market bidding, the optimal strategy is expressed by the following formula:

equation 26

wherein ,

bidding strategy for virtual power plant VPP m, < ->

Optimal bidding strategy for virtual power plant VPP m and will

Providing a scheme for energy in the optimal bidding strategy;

after the competitive price transaction is completed, transaction settlement is required, a market price clearing MCP is adopted, and the lowest average energy supply cost is set as an optimization target through the following formula;

equation 27

Equation 28

wherein ,

for UPG at timetAnd M represents the virtual plant VPP number.

8. The method of claim 7, wherein the real-time scheduling optimization includes that the predicted daily power of the wind power plant WPP and the photovoltaic power plant PV still have deviation, so that the daily bidding game strategy is difficult to reach the optimum, wherein when the actual tradable electricity quantity is higher than the bidding electricity quantity, the daily bidding scheme is still executed, and otherwise emergency standby resources are required to be called to cope with the unbalanced load supply and demand problem, and standby sources include three channels of IBDR, virtual power plant VPPs and upper grid.

9. The collaborative scheduling optimization method for a virtual power plant farm according to claim 8, wherein the real-time scheduling optimization further comprises:

The objective function is expressed by the following formula:

equation 29

wherein ,

is virtual power plant VPPmAt the moment of timetSpare cost of->

，/>

equation 30

wherein ,

and />

Respectively at the time ofEngravingt，Upper and lower spare capacity provided by IBDR, < + >>

and />

equation 31

Equation 32

Equation 33

wherein ,

and />

10. A collaborative scheduling optimization device for a virtual power plant group, comprising:

the virtual power plant VPP cooperative operation module is used for reporting bidding price and external power supply quantity in a bidding transaction center in a bidding game mode after each virtual power plant meets self-load requirements, wherein the bidding transaction center integrates bidding information of the virtual power plant VPPs and settles bidding energy shares and market clearing prices of batches obtained by different virtual power plants VPPs, and simultaneously transmits the bidding price and the bidding energy to each virtual power plant VPP, so that internal energy supply and demand balance is ensured, and redundant energy is sold to a public power grid to maximize operation income, wherein the virtual power plants VPP comprise wind power plants WPP, photovoltaic power plants PV, gas turbine power generation CGT, energy storage systems ESS and flexible loads; and

The multi-time scale collaborative scheduling optimization module is used for considering the day-ahead prediction results of the wind power plant WPP and the photovoltaic power plant PV when making a power scheduling plan, combining the day-ahead prediction values, correcting the prediction deviation of the wind power plant WPP and the photovoltaic power plant PV by adjusting the gas turbine power generation CGT, the energy storage system ESS and the flexible load scheduling plan, combining the real-time output of a random power supply, so as to purchase electric energy from other virtual power plants VPP or an upper power grid UPG and calling an excitation type demand response IBDR, and guaranteeing the real-time supply and demand balance of the electric energy.