CN104361401A - Step-by-step optimization method for real-time generation scheduling closed-loop control - Google Patents

Abstract

The invention discloses a step-by-step optimization method for real-time generation scheduling closed-loop control. The method includes the steps of acquiring various optimization data and determining an optimization space for real-time generation scheduling closed-loop control; respectively building an active power optimization model and a control mode optimization model for real-time generation scheduling closed-loop control; when unit adjusting capacity in a system is insufficient, enabling an active power optimization module to adjust unit output scheduling so as to achieve optimization and transfer of adjustable capacity; if the unit adjusting capacity still cannot satisfy requirements, enabling a control mode optimization module to adjust a control mode of a unit so as to supplement the adjustable capacity level of the system; after step-by-step optimization, obtaining a reasonable real-time generation scheduling closed-loop control result finally. The step-by-step optimization method has the advantages that more reasonable invoking of control resources by a power grid is benefited, and the power grid scheduling fining level is improved.

Description

A kind of step-by-step optimization method of real-time generation schedule closed-loop control

Technical field

The invention belongs to dispatching automation of electric power systems technical field, relate to a kind of step-by-step optimization method of real-time generation schedule closed-loop control.

Background technology

Dispatching of power netwoks is progressively gone forward one by one a multicycle and the process of refinement, comprises the link of multiple Different time scales such as monthly plan, a few days ago generation schedule, real-time generation schedule, Automatic Generation Control (AGC).Wherein, real-time generation schedule is according to the change of external condition, and rolling amendment is generation schedule a few days ago, for Automatic Generation Control provides basic point power more accurately, achieves plan a few days ago and being connected between Real-Time Scheduling, promotes the fine-grained management level of dispatching of power netwoks.Simultaneously, the randomness of large-scale grid connection wind-powered electricity generation and undulatory property, make to plan to there is relatively large deviation between Real-Time Scheduling, the scheduling mode that the time scale span that plan and AGC a few days ago should not be adopted directly to combine is larger a few days ago, thus highlighted the importance of real-time generation schedule in actual production.

The main task of Automatic Generation Control adjusts generated output when the load random variation of electric system, thus the frequency of maintenance electric system is in allowed band.Considering security and the economy of operation of power networks, require in Real-Time Scheduling, there is appropriate generating pondage, regulating the AGC that supplies for subsequent use to call for providing.The control model that genset in AGC link is preset according to it, is generally divided into 2 classes: the 1st class follows the tracks of generation schedule instruction operation, does not provide adjustment for subsequent use, i.e. scheduled mode; 2nd class is pressed electrical network real time control command and is participated in area control error (ACE) adjustment, provides adjustment for subsequent use, i.e. shaping modes.In Real-Time Scheduling, being the various uncertain factors in reply operation of power networks, needing setting unit unit operation in shaping modes, for eliminating ACE.

If can in real-time generation schedule link, according to ultra-short term comparatively accurately, arranged rational generation schedule (comprising unit allocation pattern), organize the Optimum Regulation of all kinds of unit allocation pattern and pondage, realize the closed-loop control of real-time generation schedule, both ensure that system-wide economical operation, and again in the requirement just reaching AGC control in advance, avoid AGC frequently to regulate unit.

Current real-time generation schedule establishment is relative with the coordination between AGC simple, lacks therebetween and combines closely.In scheduling is produced, real-time generation schedule establishment is main based on the meritorious plan of exerting oneself of unit, gives real-time controlling unit, as control objectives or the basic point power of unit in AGC according to ultra-short term after carrying out generation schedule adjustment.In AGC, rule of thumb being carried out the setting of unit allocation pattern by dispatcher, selecting a part of unit for following the tracks of real-time generation schedule, another part unit participates in the adjustment of district control deviation.

Real-time generation schedule does not take into full account the requirement of AGC adjusting function, can reduce the execution efficiency of real-time generation schedule on the one hand, the effect of impact plan in real time; To increase AGC controlled pressure on the other hand, AGC, only according to ACE feedback information afterwards, is difficult to keep comprehensive optimal adjustment ability at future time period.Along with electrical network scale constantly expands, operation of power networks is day by day complicated, the change of electrical network real time execution strengthens, peak load regulation network frequency modulation pressure day by day increases, dissimilar, different control objectives unit operation patten transformation is frequent, by the effect that extreme influence AGC controls in real time, also bring huge workload to operations staff.

The closed-loop control of real-time generation schedule, owing to considering AGC regulatory demand in planning link, what problem comprised unit meritoriously exerts oneself, regulate for subsequent use, control model three class decision content, further, the patten transformation of genset and active distribute (exerting oneself with for subsequent use) have coupling, genset meritorious is exerted oneself and is for subsequent usely also had coupling with regulating, and makes this problem have higher complexity, and there is no effective optimization method at present.

Summary of the invention

Object of the invention process is a kind of step-by-step optimization method providing real-time generation schedule closed-loop control, realize the integration between real-time generation schedule and Automatic Generation Control, by the reach of Generation Control link, considering the requirement of Real-Time Scheduling to generator set control pattern and pondage in real-time plan aspect, is real-time online scheduling reservation more reasonably distribution for subsequent use and the best regulating power of the overall situation.

For solving the problems of the technologies described above, the present invention proposes a kind of step-by-step optimization method being applicable to the real-time generation schedule closed-loop control of electrical network Real-Time Scheduling plan, it is characterized in that, comprise the following steps:

1) cycle needing to carry out real-time generation schedule closed-loop control is determined, carry out data encasement, in the acquisition cycle day part up-to-date load, regulate standby requirement, the original plan of turnaround plan, interconnection plan, each unit in the acquisition cycle, obtain the working control pattern of genset, to determine the optimization space of real-time generation schedule closed-loop control;

2) based on the working control pattern of genset, consider system balancing constraint, unit operation constraint, power system security constraints, with unit output deviation cost and system fading margin lax cost sum for subsequent use minimum for objective function, set up the active optimization model of real-time generation schedule closed-loop control, realize unit output plan and regulate combined optimization for subsequent use;

The active optimization model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}C\left(\right|p_{\mathrm{it}}-p_{\mathrm{it}0}\left|\right)+\underset{t=1}{\overset{T}{\mathrm{\Σ}}}{\mathrm{Mq}}_t$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{p}_{\mathrm{it}}=L_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}\≥Q_t-q_t$

p _it+r _it≤x _i0*R _i,max+(1-x _i0)P _i,max

p _it-r _it≥x _i0*R _i,min+(1-x _i0)P _i,min

-△ _i≤p _it-p _i(t-1)≤△ _i

0≤r _it≤τV _i

$\underset{\‾}{p_{\mathrm{ij}}}\≤p_{\mathrm{ij}}\left(t\right)\≤\stackrel{\‾}{p_{\mathrm{ij}}}$

Wherein, I is the unit number participating in Real-Time Scheduling in system, T be real-time generation schedule periodic packets containing time hop count, p _itfor unit i exerting oneself at period t; p _it0for unit i exerts oneself in the original plan of period t; C (| p _it-p _it0|) be the cost function of unit output absolute deviation, it is piecewise linear convex curve; q _tsystem fading margin for period t slack variable for subsequent use; M is system fading margin lax punishment cost for subsequent use; L _tfor the total load of system t period; r _itthe adjustment provided at period t for unit i is for subsequent use; Q _tfor system is in the adjustment standby requirement of period t; x _i0for the working control pattern of unit i, 0 represents scheduled mode, and 1 represents shaping modes; R _{i, max}and R _{i, min}represent the bound of unit i regulating power respectively; P _{i, max}and P _{i, min}represent the bound of unit i generated output respectively; △ _ifor the maximal value of the per period creep speed of unit i; τ is the unit response time of given permission; V _ifor the regulations speed of unit i; with p _ij represent the trend bound of branch road ij respectively, p _ijt () is for branch road ij is in the trend of t period;

Non-linear factor linearization in model is expressed, adopts linear programming technique to calculate the meritorious p that exerts oneself of unit day part within dispatching cycle _it, regulate r for subsequent use _itand the system fading margin of day part slack q for subsequent use _t;

3) according to step 2) active optimization result, judge that can working control pattern meet adjustment standby requirement; If the system fading margin of day part slack for subsequent use is always value added is zero, then working control pattern can meet adjustment standby requirement, arranges the control model x of unit i at period t _itfor its actual control model, enter step 5), otherwise enter step 4);

4) based on the active optimization result of real-time generation schedule closed-loop control, consider system fading margin standby requirement, minimum for objective function with the control model conversion cost sum of units all in system, set up the control model Optimized model of real-time generation schedule closed-loop control, Dynamic Selection provides and regulates genset for subsequent use;

The control model Optimized model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}(y_{\mathrm{it}}{S}_i+z_{\mathrm{it}}{K}_i)$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≥Q_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≤O_t$

x _it-x _i(t-1)＝y _it-z _it

y _it+z _it≤1

Wherein, x _itfor unit i is in the control model state of period t, 0/1 variable, 0 represents scheduled mode, and 1 represents shaping modes; y _itfor whether unit i is converted to shaping modes by scheduled mode at period t, 0/1 variable; z _itfor whether unit i is converted to scheduled mode by shaping modes at period t, 0/1 variable; S _ifor unit i to be converted to the priority cost of shaping modes by scheduled mode; K _ifor unit i to be converted to the priority cost of scheduled mode by shaping modes; Q _tfor system is in the adjustment standby requirement of period t; O _tfor system is in the adjustment standby requirement upper limit of period t;

Mixed integer programming approach is adopted to calculate the control model x of unit day part within dispatching cycle _it;

5) end is calculated; Obtain unit the meritorious of day part within dispatching cycle to exert oneself, it can be used as the basic point power of Automatic Generation Control link; Obtain the control model of unit day part within dispatching cycle, it can be used as the basic control mode of Automatic Generation Control link; Optimize and terminate.

Method of the present invention comprises following beneficial effect:

1) the present invention optimizes the real-time generation schedule close-loop control scheme that establishment meets dispatching of power netwoks service requirement, change and pattern is set at present based on artificial experience, under the prerequisite ensureing real-time Mass Control, by the reach of Generation Control link, the Optimum Regulation of unit allocation pattern and pondage is realized in real-time generation schedule aspect, be conducive to electrical network more rationally to call control resource, make full use of the regulating power of genset, improve the advanced pre-control ability of electrical network and genset On-line Control quality, alleviate management and running pressure.

2) this method is the real-time generation schedule Research on Closed Loop Control and trial of carrying out under actual electric network data, finds out the generation schedule optimization method for real-tim scheduling and control crucial requirement.This method is by the decoupling zero of active optimization and control model, establish active optimization model and the control model Optimized model of real-time generation schedule closed-loop control, through step-by-step optimization, solve in real-time generation schedule closed-loop control the optimization difficult problem that decision variable is many, decision space is large, the more rational close-loop control scheme of final acquisition, substitute original experience scheduling type scheme, contribute to the level that becomes more meticulous improving dispatching of power netwoks.

3) this method is by the step-by-step optimization of real-time generation schedule closed-loop control, draw the meritorious in real time planning and control mode scheme of exerting oneself of unit, these results can directly enter Automatic Generation Control link, contribute to realizing real-time generation schedule to be connected with the closed loop of Automatic Generation Control, improve the execution efficiency of real-time generation schedule, instruct the safety and economic operation of electric system better.

Accompanying drawing explanation

Fig. 1 is the step-by-step optimization method flow schematic diagram of a kind of real-time generation schedule closed-loop control of the present invention.

Embodiment

Below in conjunction with accompanying drawing, the present invention is further described.

The invention discloses a kind of step-by-step optimization method of real-time generation schedule closed-loop control.Here is a preferred case study on implementation of the present invention, and contain the optimizing process of the real-time generation schedule closed-loop control adopting the inventive method, its feature, object and advantage can be found out from the explanation of embodiment.

In real-time generation schedule closed loop control process, need to optimize the unit generation planning and control mode scheme determined in following 1 hour, perform for AGC link.Optimizing process requires to take into full account the information such as generation schedule, ultra-short term, the prediction of ultra-short term new forms of energy, interconnection plan, power failure plan, network topology a few days ago, make real-time generation schedule closed-loop control result can meet system balancing constraint, power system security constraints and unit operation constraint limit value, and meet the requirement of real-time control of AGC.

The step-by-step optimization method of real-time generation schedule closed-loop control of the present invention, the Optimum Regulation of all kinds of unit allocation pattern and pondage is considered in real-time generation schedule link, by the decoupling zero of active optimization and control model, establish active optimization model and the control model Optimized model of real-time generation schedule closed-loop control respectively.When AGC unit pondage in system is not enough, active optimization module can adjust the plan of exerting oneself of unit, realizes the optimization transfer of variable capacity; If when AGC unit pondage still cannot meet the demands, control model optimizes the control model that module can adjust unit further, the variable capacity level of replenishment system.Through step-by-step optimization, the rational real-time generation schedule closed-loop control result of final acquisition.

As shown in Figure 1, this method comprises the following steps:

1) optimization cycle of real-time generation schedule closed-loop control is following 1 hour, optimizes the period using 5 minutes as one, comprises 12 altogether and optimizes the period.Obtain the load of 12 periods in following 1 hour, regulate standby requirement, obtain the original plan of turnaround plan, interconnection plan, each unit, obtain the working control pattern of genset, to determine the optimization space of real-time generation schedule closed-loop control.

2) based on the working control pattern of genset, consider system balancing constraint, unit operation constraint, power system security constraints, with unit output deviation cost and system fading margin lax cost sum for subsequent use minimum for objective function, set up the active optimization model of real-time generation schedule closed-loop control;

The active optimization model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}C\left(\right|p_{\mathrm{it}}-p_{\mathrm{it}0}\left|\right)+\underset{t=1}{\overset{T}{\mathrm{\Σ}}}{\mathrm{Mq}}_t$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{p}_{\mathrm{it}}=L_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}\≥Q_t-q_t$

p _it+r _it≤x _i0*R _i,max+(1-x _i0)P _i,max

p _it-r _it≥x _i0*R _i,min+(1-x _i0)P _i,min

-△ _i≤p _it-p _i(t-1)≤△ _i

0≤r _it≤τV _i

$\underset{\‾}{p_{\mathrm{ij}}}\≤p_{\mathrm{ij}}\left(t\right)\≤\stackrel{\‾}{p_{\mathrm{ij}}}$

Wherein, I is the unit number participating in Real-Time Scheduling in system, T be real-time generation schedule periodic packets containing time hop count, p _itfor unit i exerting oneself at period t; p _it0for unit i exerts oneself in the original plan of period t; C (| p _it-p _it0|) be the cost function of unit output absolute deviation, it is piecewise linear convex curve; q _tsystem fading margin for period t slack variable for subsequent use; M is system fading margin lax punishment cost for subsequent use; L _tfor the total load of system t period; r _itthe adjustment provided at period t for unit i is for subsequent use; Q _tfor system is in the adjustment standby requirement of period t; x _i0for the working control pattern of unit i, 0 represents scheduled mode, and 1 represents shaping modes; R _{i, max}and R _{i, min}represent the bound of unit i regulating power respectively; P _{i, max}and P _{i, min}represent the bound of unit i generated output respectively; △ _ifor the maximal value of the per period creep speed of unit i; τ is the unit response time of given permission; V _ifor the regulations speed of unit i; with p _ij represent the trend bound of branch road ij respectively, p _ijt () is for branch road ij is in the trend of t period;

Non-linear factor linearization in model is expressed, adopts linear programming technique to calculate the meritorious p that exerts oneself of unit day part within dispatching cycle _it, regulate r for subsequent use _itand the system fading margin of day part slack q for subsequent use _t.

3) according to step 2) active optimization result, judge that can working control pattern meet adjustment standby requirement; If the system fading margin of day part slack for subsequent use is always value added is zero, then working control pattern can meet adjustment standby requirement, arranges the control model x of unit i at period t _itfor its actual control model, enter step 5), otherwise enter step 4);

4) based on the active optimization result of real-time generation schedule closed-loop control, consider system fading margin standby requirement, minimum for objective function with the control model conversion cost sum of units all in system, set up the control model Optimized model of real-time generation schedule closed-loop control;

The control model Optimized model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}(y_{\mathrm{it}}{S}_i+z_{\mathrm{it}}{K}_i)$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≥Q_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≤O_t$

x _it-x _i(t-1)＝y _it-z _it

y _it+z _it≤1

Wherein, x _itfor unit i is in the control model state of period t, 0/1 variable, 0 represents scheduled mode, and 1 represents shaping modes; y _itfor whether unit i is converted to shaping modes by scheduled mode at period t, 0/1 variable; z _itfor whether unit i is converted to scheduled mode by shaping modes at period t, 0/1 variable; S _ifor unit i to be converted to the priority cost of shaping modes by scheduled mode; K _ifor unit i to be converted to the priority cost of scheduled mode by shaping modes; Q _tfor system is in the adjustment standby requirement of period t; O _tfor system is in the adjustment standby requirement upper limit of period t;

Mixed integer programming approach is adopted to calculate the control model x of unit day part within dispatching cycle _it;

5) end is calculated; Obtain unit the meritorious of day part within dispatching cycle to exert oneself, it can be used as the basic point power of Automatic Generation Control link; Obtain the control model of unit day part within dispatching cycle, it can be used as the basic control mode of Automatic Generation Control link; Optimize and terminate.

System of the present invention comprise monthly, a few days ago, in real time, the dispatching of power netwoks planning and control function in multiple cycle such as AGC, real-time generation schedule closed-loop control is the key link of multicycle generation schedule coordinated operation, it realizes generation schedule and the integration between controlling in real time a middle or short term, for control provides basic point power and the control model scheme of unit in real time.Real-time generation schedule is on prospective project Data Integration basis, and the basis that application the present invention optimizes close-loop control scheme completes.

Analyze result of calculation, the adjustment of system future time period is for subsequent use all meets standby requirement, and the execution efficiency of real-time generation schedule is significantly improved, and the optimum results of real-time generation schedule closed-loop control is consistent with actual operating state.

The research of the step-by-step optimization of the real-time generation schedule closed-loop control that this method is carried out under actual electric network data and trial.The method optimizes the execute in steps toggle of module through active optimization module and control model, final acquisition is real-time generation schedule closed-loop control result more reasonably, contribute to electrical network more rationally to call control resource, reduce system operation cost, improve stationarity and the order of unit allocation.The method does not need the participation of a large amount of manpower, computing velocity can meet the needs of practical application, efficiently solve traditional real-time plan and AGC cooperation control needs a large amount of manpower, rely on artificial experience, efficiency is low, be difficult to the disadvantage effectively realizing real-time optimization regulation and control, there is promotion prospect widely.

Below be only the preferred embodiment of the present invention; be noted that for those skilled in the art; under the premise without departing from the principles of the invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (1)

1. a step-by-step optimization method for real-time generation schedule closed-loop control, is characterized in that, comprise the following steps:

1) cycle needing to carry out real-time generation schedule closed-loop control is determined, carry out data encasement, in the acquisition cycle day part up-to-date load, regulate standby requirement, the original plan of turnaround plan, interconnection plan, each unit in the acquisition cycle, obtain the working control pattern of genset, to determine the optimization space of real-time generation schedule closed-loop control;

2) based on the working control pattern of genset, consider system balancing constraint, unit operation constraint, power system security constraints, with unit output deviation cost and system fading margin lax cost sum for subsequent use minimum for objective function, set up the active optimization model of real-time generation schedule closed-loop control, realize unit output plan and regulate combined optimization for subsequent use;

The active optimization model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}C\left(\right|p_{\mathrm{it}}-p_{\mathrm{it}0}\left|\right)+\underset{t=1}{\overset{T}{\mathrm{\Σ}}}M{q}_t$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{p}_{\mathrm{it}}=L_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}\≥Q_t-q_t$

p _it+r _it≤x _i0*R _i,max+(1-x _i0)P _i,max

p _it-r _it≥x _i0*R _i,min+(1-x _i0)P _i,min

-Δ _i≤p _it-p _i(t-1)≤Δ _i

0≤r _it≤τV _i

$\underset{\‾}{p_{\mathrm{ij}}}\≤p_{\mathrm{ij}}\left(t\right)\≤\stackrel{\‾}{p_{\mathrm{ij}}}$

Wherein, I is the unit number participating in Real-Time Scheduling in system, T be real-time generation schedule periodic packets containing time hop count, p _itfor unit i exerting oneself at period t; p _it0for unit i exerts oneself in the original plan of period t; C (| p _it-p _it0|) be the cost function of unit output absolute deviation, it is piecewise linear convex curve; q _tsystem fading margin for period t slack variable for subsequent use; M is system fading margin lax punishment cost for subsequent use; L _tfor the total load of system t period; r _itthe adjustment provided at period t for unit i is for subsequent use; Q _tfor system is in the adjustment standby requirement of period t; x _i0for the working control pattern of unit i, 0 represents scheduled mode, and 1 represents shaping modes; R _{i, max}and R _{i, min}represent the bound of unit i regulating power respectively; P _{i, max}and P _{i, min}represent the bound of unit i generated output respectively; Δ _ifor the maximal value of the per period creep speed of unit i; τ is the unit response time of given permission; V _ifor the regulations speed of unit i; with p _ij represent the trend bound of branch road ij respectively, p _ijt () is for branch road ij is in the trend of t period;

Non-linear factor linearization in model is expressed, adopts linear programming technique to calculate the meritorious p that exerts oneself of unit day part within dispatching cycle _it, regulate r for subsequent use _itand the system fading margin of day part slack q for subsequent use _t;

3) according to step 2) active optimization result, judge that can working control pattern meet adjustment standby requirement; If the system fading margin of day part slack for subsequent use is always value added is zero, then working control pattern can meet adjustment standby requirement, arranges the control model x of unit i at period t _itfor its actual control model, enter step 5), otherwise enter step 4);

4) based on the active optimization result of real-time generation schedule closed-loop control, consider system fading margin standby requirement, minimum for objective function with the control model conversion cost sum of units all in system, set up the control model Optimized model of real-time generation schedule closed-loop control, Dynamic Selection provides and regulates genset for subsequent use;

The control model Optimized model of real-time generation schedule closed-loop control is:

Objective function:

$\min f=\underset{t=1}{\overset{T}{\mathrm{\Σ}}}\underset{i=1}{\overset{I}{\mathrm{\Σ}}}(y_{\mathrm{it}}{S}_i+z_{\mathrm{it}}{K}_i)$

Constraint condition:

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≥Q_t$

$\underset{i}{\overset{I}{\mathrm{\Σ}}}{r}_{\mathrm{it}}{x}_{\mathrm{it}}\≤O_t$

x _it-x _i(t-1)＝y _it-z _it

y _it+z _it≤1

Wherein, x _itfor unit i is in the control model state of period t, 0/1 variable, 0 represents scheduled mode, and 1 represents shaping modes; y _itfor whether unit i is converted to shaping modes by scheduled mode at period t, 0/1 variable; z _itfor whether unit i is converted to scheduled mode by shaping modes at period t, 0/1 variable; S _ifor unit i to be converted to the priority cost of shaping modes by scheduled mode; K _ifor unit i to be converted to the priority cost of scheduled mode by shaping modes; Q _tfor system is in the adjustment standby requirement of period t; O _tfor system is in the adjustment standby requirement upper limit of period t;

Mixed integer programming approach is adopted to calculate the control model x of unit day part within dispatching cycle _it;

5) end is calculated; Obtain unit the meritorious of day part within dispatching cycle to exert oneself, it can be used as the basic point power of Automatic Generation Control link; Obtain the control model of unit day part within dispatching cycle, it can be used as the basic control mode of Automatic Generation Control link; Optimize and terminate.