CN105024398A - Optimization scheduling method based on optimal wind power confidence - Google Patents

Abstract

The invention relates to the field of scheduling and control in a new energy power generation technology, and discloses an optimization scheduling method based on optimal wind power confidence. The optimization scheduling method takes a wind power prediction power output value and a probability density distribution of the wind power prediction power output value as a basis, an electric power system optimization scheduling model, which takes wind power confidence, machine set start-stop and a power output value as decision variables and takes a fact that an economic scheduling cost in a confidence interval and risk cost outside the confidence interval are minimum as an objective function, is established, and the electric power system optimization scheduling model is solved through an application of an improved particle swarm algorithm. The optimization scheduling method determines power output arrangement of the system machine set in the condition of the optimal wind power confidence, the optimal wind power acceptance interval and the optimal economic acceptance interval, and electric power optimization scheduling containing wind power can select wind power confidence and an wind power confidence interval which have comprehensive economy and risk optima, so economy and risk balance of optimization scheduling is achieved, a reference is provided for a scheduling department to arrange a current power output plane, and the optimal economical benefit is achieved.

Description

A kind of Optimization Scheduling based on optimum wind-powered electricity generation confidence level

Technical field

The invention belongs to the scheduling field in generation of electricity by new energy technology, be specifically related to a kind of Optimization Scheduling based on optimum wind-powered electricity generation confidence level.

Background technology

Along with the continuous increase of wind-electricity integration scale, Unit Commitment and Economic Dispatch Problem need the uncertainty considering wind-powered electricity generation, and the uncertainty of wind-powered electricity generation can adopt the interval meeting confidence degree to describe.Current research empirically determines confidence interval usually, lacks enough theoretical foundations.But, different confidence intervals is corresponding different system operation cost and risk cost.When confidence level is large, wind-powered electricity generation prediction is exerted oneself interval large, and now, in order to tackle wind-powered electricity generation in the uncertainty in interval of exerting oneself, needed for electrical network, cost is higher, but interval large owing to exerting oneself, and the possibility dropped on outside interval is lower, and namely now risk is less, and risk cost is lower; Otherwise confidence level hour, wind-powered electricity generation prediction is exerted oneself interval little, and now risk cost is higher, and the probabilistic system operation cost of reply wind-powered electricity generation is lower.By the research of the optimum confidence level of wind-powered electricity generation, can determine that wind-powered electricity generation Optimum Economic is received interval, realize the overall operation cost minimization of system, show that Optimum Economic receives the arrangement of exerting oneself of interval lower system unit, for traffic department arranges the plan of exerting oneself a few days ago to provide reference.

Wind-powered electricity generation itself has very strong uncertainty, and corresponding with it, the confidence level in each moment is not identical, and wind-powered electricity generation confidence wind-powered electricity generation forecast interval is also not in relation to wind-powered electricity generation, and to dope force curve symmetrical up and down.Therefore, if entire system operating cost is now minimum, on traditional scheduler model basis, using the wind-powered electricity generation confidence level in each moment and every platform unit output simultaneously as decision variable, determine that wind-powered electricity generation Optimum Economic is received interval thus, and then make the total operating cost of system minimum.

The present invention adopts the algorithm of the population of improvement, using every platform unit exert oneself and the confidence level in per moment is simultaneously brought particle cluster algorithm into as optimizing particle and is solved; Particle is revised simultaneously, particle is controlled on restrained boundary, by the form of penalty function, risk amount is retrained.When unit output cannot meet system power Constraints of Equilibrium, power shortage is reallocated.Modified particle swarm optiziation can make particle as much as possible at feasible zone or carry out optimizing close within the scope of feasible zone as far as possible, improves the precision of particle cluster algorithm, speed and optimizing ability.To be how the technical issues that need to address of the present invention with the Solve problems that this method solves optimum confidence level.

Summary of the invention

One is the object of the present invention is to provide to solve optimum wind-powered electricity generation confidence level, make the optimum confidence level derivation algorithm based on modified particle swarm optiziation that the entire system operating cost under this confidence level is minimum, optimum wind-powered electricity generation confidence level can be determined, have wind-powered electricity generation to receive arrangement of exerting oneself that is interval and the interval lower system unit of Optimum Economic receiving most, for traffic department arranges the plan of exerting oneself a few days ago to provide reference, realize the balance of Optimized Operation economy and risk with this, realize optimum economic benefit simultaneously.

In order to achieve the above object, technical scheme of the present invention is:

Based on a method for solving for the optimum wind-powered electricity generation confidence level of modified particle swarm optiziation, its step is:

(1) target function is set up;

This model is minimum for target function with risk cost outside economic dispatch cost in confidential interval and confidential interval, such as formula shown.Wherein, wind-powered electricity generation confidence level and Unit Commitment, exert oneself as decision variable.In formula, the 1st is the economic dispatch cost in interval, and the 2nd is the risk cost outside interval.Risk cost outside interval comprises pressure and abandons eolian and cutting load cost, all adopts risk expectation to calculate, such as formula with shown in formula.Further, stochastic variable w is supposed _tmeet normal distribution, such as formula shown.

$F=min\{(\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt1}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct1})\}---\left(1\right)$

$p_{wlt1}={\∫}_{p_{wt\max }}^{p_{w\max }}{w}_t\·\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t---\left(2\right)$

$p_{lct1}={\∫}_0^{p_{wt\min }}{w}_t\·\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t---\left(3\right)$

$f_{w_t}\left(w_t\right)=\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}---\left(4\right)$

In formula: a _i, b _i, c _irepresent the operation cost parameter of unit, p _i,trepresent i-th unit (i=1,2 ..., N) exert oneself, β is for abandoning wind loss coefficient, and γ is cutting load loss coefficient, p _wlt1represent and abandon eolian, p outside interval _lct1represent the cutting load cost outside interval, p _wtmax, p _wtminrepresent that the wind-powered electricity generation of t predicts interval upper and lower bound of exerting oneself, p respectively _wmaxrepresent wind farm grid-connected capacity.

(2) constraints is set up

Constraints comprises power-balance constraints, unit output constraint, wind power output constraint, Climing constant and Reserve Constraint.

Power-balance constraints:

$\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+p_{wt}=p_{lt},t=1,\mathrm{...}T---\left(5\right)$

Unit output retrains:

p _imin≤p _i,t≤p _imax(6)

Wind power output retrains:

p _wtmin≤p _wt≤p _wtmax(7)

Climing constant:

-Δ _i,dT ₆₀≤p _i,t-p _i,t-1≤Δ _i,uT ₆₀(8)

Reserve Constraint:

$\underset{i=1}{\overset{n}{\Σ}}min(p_{imax}-p_{i,t},{\mathrm{\Δ}}_{i,u}{T}_{15})\≥p_{wt}-p_{wt\min }---\left(9\right)$

$\underset{i=1}{\overset{n}{\Σ}}min(p_{i,t}-p_{i\min },{\mathrm{\Δ}}_{i,d}{T}_{15})\≥p_{wtmax}-p_{wt}---\left(10\right)$

Wherein, p _i,tfor the plan of i-th unit of t period is exerted oneself, p _wtfor the prediction of t period wind-powered electricity generation is exerted oneself, p _ltfor t period load prediction is exerted oneself; p _imin, p _imaxminimum and the maximum technology being respectively i-th unit is exerted oneself; p _wtmin, p _wtmaxbe respectively the minimum and maximum value of t period wind power output; Δ _i,d, Δ _i,ube respectively downwards and the upwards creep speed of i-th unit, getting the climbing time is here T ₆₀, namely 1 hour.

(3) particle cluster algorithm

By modified particle swarm optiziation, the constraint of unit bound and Climing constant, Reserve Constraint and system power Constraints of Equilibrium are adjusted in the present invention:

The constraint of unit bound and Climing constant adjustment:

After each iteration of particle upgrades, particle is controlled on restrained boundary, shown in (11), make particle can meet the constraint of unit bound and Climing constant;

$x_i^k=\left\{\begin{array}{c}\max (p_{i\min },p_{i,t-1}-{\mathrm{\Δ}}_{i,d}{T}_{60})\\ \begin{array}{cc}if& x_i^k\≤\max (p_{i\min },p_{i,t-1}-{\mathrm{\Δ}}_{i,d}{T}_{60})\end{array}\\ \min (p_{i\max },p_{i,t-1}+{\mathrm{\Δ}}_{i,u}{T}_{60})\\ \begin{array}{cc}if& x_i^k\≥\min (p_{i\max },p_{i,t-1}{+\mathrm{\Δ}}_{i,u}{T}_{60})\end{array}\end{array}\right.---\left(11\right)$

Wherein, k represents kth time iteration, and i represents machine group #, represent i-th unit exerting oneself in kth time iteration, p _{i, t-1}for exerting oneself of t-1 period i-th unit.

Reserve Constraint adjusts:

By the method for penalty function, the amount of abandoning wind and cutting load in interval is retrained, shown in (12);

$f=f_0+\mathrm{\λ}(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt0}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct0})---\left(12\right)$

Wherein, λ is penalty coefficient, gets a maximum, p _wlt0, p _lct0be respectively excessive due to wind power output interval and value of abandoning wind and cutting load that is that cause, β is for abandoning wind loss coefficient, and γ is cutting load loss coefficient, f ₀for target function, namely system synthesis originally, and such as formula shown, t is for running the period, and span gets 24 from 1.P _wlt0, p _lct0computing formula as follows:

$p_{wlt0}=max(0,p_{wtmax}-p_{wt}-\underset{i=1}{\overset{I}{\Σ}}min(p_{i,t}-p_{imin},T_{15}{\mathrm{\Δ}}_{i,d}))---\left(13\right)$

$p_{lct0}=max(0,p_{wt}-p_{wtmin}-\underset{i=1}{\overset{I}{\Σ}}min(p_{imax}-p_{i,t},T_{15}{\mathrm{\Δ}}_{i,u}))---\left(14\right)$

Wherein, p _imax, p _iminrepresent that maximum, the minimum technology of unit i are exerted oneself respectively, p _wtrepresent wind power output, Δ _i,u, Δ _i,drepresent the rising of unit i, the constraint of decline climbing rate respectively, T ₁₅represent 15 minutes.

System power Constraints of Equilibrium adjusts:

When unit output cannot meet system power Constraints of Equilibrium, power shortage is now formula (15), and power shortage is carried out unit output reallocation;

${\mathrm{\Δp}}_t=\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+p_{wt}-p_{lt}---\left(15\right)$

By the above-mentioned equation of PSO Algorithm, obtain the optimal Confidence Interval of optimum unit output scheme, optimum confidence level and correspondence thereof, obtain minimum entire system operating cost.

In described step (), the corresponding wind-powered electricity generation of confidence level in each moment is received interval, the risk that wind-powered electricity generation receives the direct influential system of interval difference to run, what produce when namely wind-powered electricity generation fluctuates outward probabilisticly abandons wind and the such risk cost of cutting load, wind-powered electricity generation confidence level is joined in the target function of system operation cost with the form of risk cost, such as formula (16);

$\begin{array}{c}F=\left(\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i\right)+F_1+F_2\\ =\left(\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i\right)+\mathrm{\λ}(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt0}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct0})+(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt1}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct1})\end{array}---\left(16\right)$

In formula, F ₁represent the risk cost outside interval, F ₂represent the risk cost in interval.

In described step (three), for making wind-powered electricity generation prediction curve receive fluctuation in interval at the wind-powered electricity generation of optimum as much as possible, the reserve capacity for system retrains, and wherein every platform unit can be raised the p that exerts oneself _imax-p _i,twith every 15 minutes upwards climbing amount Δs _i,ut ₁₅smaller value carry out cumulative summation, the constraint in structural formula (17), this makes to receive interval lower limit by wind-powered electricity generation curve can be made to cover wind-powered electricity generation to the adjustment of unit process.Every platform unit can be lowered the p that exerts oneself simultaneously _imax-p _i,twith every 15 minutes upwards climbing amount Δs _i,ut ₁₅smaller value carry out cumulative summation, the constraint in structural formula (18), this makes to receive the interval upper limit by wind-powered electricity generation curve can be made to cover wind-powered electricity generation to the adjustment of unit process.

$\underset{i=1}{\overset{n}{\Σ}}min(p_{i\max }-p_{i,t},{\mathrm{\Δ}}_{i,u}{T}_{15})\≥p_{wt}-p_{wtmin}---\left(17\right)$

$\underset{i=1}{\overset{n}{\Σ}}min(p_{i,t}-p_{imin},{\mathrm{\Δ}}_{i,d}{T}_{15})\≥p_{wtmax}-p_{wt}---\left(18\right)$

In step (three), concrete solution procedure is as follows:

(1) input wind-powered electricity generation and dope force value and load prediction goes out force value, scope of exerting oneself and the climbing restriction of unit is set;

(2) population dimension is set, Population Size, and iterations;

(3) using unit output and the wind-powered electricity generation confidence level in per moment as particle, the position range of each particle is set.

(4) position of initialization particle, according to successive dynasties optimal solution and the total optimization solution of the size initialization population of fitness function;

(5) by formula (19) and (20) the more position of new particle and speed,

$v_i^{k+1}={\mathrm{wv}}_i^k+c_1{r}_1({\mathrm{pbest}}_i-x_i^k)+c_2{r}_2(gbest-x_i^k)---\left(19\right)$

$x_i^{k+1}=x_i^k+v_i^{k+1}---\left(20\right)$

Wherein, for the speed of particle i when kth time iteration; c ₁, c ₂for accelerator coefficient; r ₁, r ₂it is equally distributed random number between 0 ~ 1; for the position of particle i when kth time iteration; Pbest _ifor the optimal solution that particle i itself finds; Gbest is the optimal solution of whole population; W is inertia coeffeicent, generally sets by the method for formula (21);

$w=w^{max}-Cur\mathrm{Cou}nt\·\frac{w^{max}-w^{min}}{MaxCount}---\left(21\right)$

Wherein, w ^maxgenerally get 0.9, w ^mingenerally getting 0.4, CurCount is current iteration number of times, and MaxCount is maximum iteration time.

(6) whether the particle upgraded according to formula formula (19) and (20) inspection meets the constraints requirement of formula (5) ~ (15), if do not met, regenerate particle rapidity, upgrade position, until meet constraints;

(7) according to fitness function, the successive dynasties optimal solution of particle and total optimization solution are upgraded;

(8) judge whether current iteration number of times reaches maximum, if do not have, then repeat step (5) ~ (7), otherwise stop particle optimizing, export result of calculation.

The invention has the beneficial effects as follows:

(1) with the scheduling cost of unit and risk cost minimum for optimization aim, with wind-powered electricity generation confidence level and Unit Commitment, exert oneself as decision variable, realize the balance of Optimized Operation economy and risk.

(2) solving by optimum confidence level, the optimum giving wind-powered electricity generation is received interval, and Optimum Economic receives the arrangement of exerting oneself of interval lower system unit, for traffic department arranges the plan of exerting oneself a few days ago to provide reference.

(3) the grain seed group algorithm of application enhancements solves, and makes particle as much as possible at feasible zone or carry out optimizing close within the scope of feasible zone as far as possible, improves precision and the speed of particle cluster algorithm, improve the optimizing ability of particle cluster algorithm.The result solved also demonstrates the practical operation situation that institute's established model meets system unit, gives when system operation conditions the unknown, and the defining method of the optimum confidence level of wind-powered electricity generation, has more practical significance.

Accompanying drawing explanation

Below in conjunction with the drawings and specific embodiments, the present invention is described in further details.

Fig. 1 is the flow chart of the inventive method;

Fig. 2 is the structured flowchart of particle cluster algorithm;

Embodiment

The technological means realized for making the present invention, creation characteristic, reaching object and effect is easy to understand, below in conjunction with embodiment, setting forth the present invention further.

With reference to figure 1, a kind of Optimization Scheduling based on optimum wind-powered electricity generation confidence level of the present invention, its method is as follows:

In the present invention, wind-powered electricity generation predicts that the research strategy of exerting oneself is: predict according to wind-powered electricity generation a few days ago the p that exerts oneself _{pre, t}the error ε that exerts oneself is predicted with wind-powered electricity generation _w,t, can output of wind electric field be obtained:

p _wt＝p _pre,t+ε _w,t(22)

Wherein wind-powered electricity generation predicts the error ε that exerts oneself _w,tmeet normal distribution.

The bound of wind-powered electricity generation prediction curve can be expressed as:

p _wtmin＝max(0,p _wt+f ^-1( α _t)) (23)

$p_{wtmax}=min(p_{wmax},p_{wt}+f^{-1}({\stackrel{\‾}{\mathrm{\α}}}_t))---\left(24\right)$

In formula, f ^-1for the inverse function of Normal probability distribution function, t=1,2 ..., 24, μ _tfor not wind-powered electricity generation confidence level in the same time.When the particle cluster algorithm of application enhancements solves, to the parameter of interval bound of exerting oneself α _tbe handled as follows:

${\underset{\‾}{\mathrm{\α}}}_t=\{\begin{array}{cc}{e}^{-2},& if{\underset{\‾}{\mathrm{\α}}}_t\≤0\\ 0.5-e^{-2},& if{\underset{\‾}{\mathrm{\α}}}_t>0.5\end{array}---\left(25\right)$

${\stackrel{\‾}{\mathrm{\α}}}_t=\left\{\begin{array}{cc}0.5+e^{-2},& if{\underset{\‾}{\mathrm{\α}}}_t\≤0.5\\ 1-e^{-2},& if{\underset{\‾}{\mathrm{\α}}}_t>1\end{array}\right.---\left(26\right)$

Target function

The basis of power system optimal dispatch model is that wind-powered electricity generation prediction is a few days ago exerted oneself p _{pre, t}the error ε that exerts oneself is predicted with wind-powered electricity generation _w,t, obtain wind-powered electricity generation prediction curve.The optimization aim of optimum wind-powered electricity generation confidence level scheduling is wind-powered electricity generation confidence level by determining each moment and Unit Commitment scheme, realizes the balance of Optimized Operation economy and risk with economic dispatch cost in minimum confidential interval and the outer risk cost of confidential interval.

The target function of optimum wind-powered electricity generation confidence level Optimization Scheduling comprises the economic dispatch cost in confidential interval and the risk cost outside confidential interval.Economic dispatch cost wherein in confidential interval determines primarily of the scheme of exerting oneself of unit, outside confidential interval abandon confidential interval that eolian and cutting load cost determine primarily of wind-powered electricity generation confidence level, the installed capacity of wind energy turbine set determines.Target function, such as formula shown in (27), abandons eolian and cutting load cost respectively such as formula shown in (28), formula (29) outside interval.Further, stochastic variable w is supposed _tmeet normal distribution, shown in (30).

$F=\min \{\left(\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i\right)+(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt1}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct1})\}---\left(27\right)$

$p_{wlt1}={\∫}_{p_{wt\max }}^{p_{w\max }}{w}_t\·\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t---\left(28\right)$

$p_{lct1}={\∫}_0^{p_{wt\min }}{w}_t\·\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t---\left(29\right)$

$f_{w_t}\left(w_t\right)=\frac{1}{\sqrt{2\mathrm{\π}}}\·e^{-\frac{{w_t}^2}{2}}---\left(30\right)$

In above formula: a _i, b _i, c _irepresent the operation cost parameter of unit, p _i,trepresent i-th unit (i=1,2 ..., N) exert oneself, β is for abandoning wind loss coefficient, and γ is cutting load loss coefficient, p _wlt1represent and abandon eolian, p outside interval _lct1represent the cutting load cost outside interval, p _wtmax, p _wtminrepresent that the wind-powered electricity generation of t predicts interval bound of exerting oneself, p respectively _wmaxrepresent wind farm grid-connected capacity.

Constraints

Constraints comprises power-balance constraints, unit output constraint, wind power output constraint, Climing constant and Reserve Constraint.

Power-balance constraints:

$\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+p_{wt}=p_{lt},t=1,\mathrm{...}T---\left(31\right)$

Unit output retrains:

p _imin≤p _i,t≤p _imax(32)

Wind power output retrains:

p _wtmin≤p _wt≤p _wtmax(33)

Climing constant:

-Δ _i,dT ₆₀≤p _i,t-p _i,t-1≤Δ _i,uT ₆₀(34)

Reserve Constraint:

$\underset{i=1}{\overset{n}{\Σ}}min(p_{imax}-p_{i,t},{\mathrm{\Δ}}_{i,u}{T}_{15})\≥p_{wt}-p_{wt\min }---\left(35\right)$

$\underset{i=1}{\overset{n}{\Σ}}min(p_{i,t}-p_{i\min },{\mathrm{\Δ}}_{i,d}{T}_{15})\≥p_{wt\max }-p_{wt}---\left(36\right)$

Wherein, p _i,tfor the plan of i-th unit of t period is exerted oneself, p _wtfor the prediction of t period wind-powered electricity generation is exerted oneself, p _ltfor t period load prediction is exerted oneself; p _imin, p _imaxminimum and the maximum technology being respectively i-th unit is exerted oneself; p _wtmin, p _wtmaxbe respectively the minimum and maximum value of t period wind power output; Δ _i,d, Δ _i,ube respectively downwards and the upwards creep speed of i-th unit, getting the climbing time is here T ₆₀, namely 1 hour.

Method for solving

The present invention adopts modified particle swarm optiziation to solve model, and Fig. 2 is the structured flowchart of PSO Algorithm.

In particle cluster algorithm, when process is containing the optimization problem of constraints, the form normally constraints being changed into penalty function joins in fitness function in population, has that computational efficiency is high, the advantage of strong robustness.For making particle as much as possible at feasible zone or carry out optimizing close within the scope of feasible zone as far as possible, obtain the optimal solution meeting constraints, in the present invention, the unit bound constraint in constraints and Climing constant, Reserve Constraint and system power Constraints of Equilibrium adjusted:

The constraint of unit bound and Climing constant adjustment:

After each iteration of particle upgrades, particle is controlled on restrained boundary, such as formula (37), make particle can meet the constraint of unit bound and Climing constant;

$x_i^k=\left\{\begin{array}{c}\max (p_{i\min },p_{i,t-1}-{\mathrm{\Δ}}_{i,d}{T}_{60})\\ \begin{array}{cc}if& x_i^k\≤\max (p_{i\min },p_{i,t-1}{-\mathrm{\Δ}}_{i,d}{T}_{60})\end{array}\\ \min (p_{i\max },p_{i,t-1}+{\mathrm{\Δ}}_{i,u}{T}_{60})\\ \begin{array}{cc}if& x_i^k\≥\min (p_{i\max },p_{i,t-1}{+\mathrm{\Δ}}_{i,u}{T}_{60})\end{array}\end{array}\right.---\left(37\right)$

Wherein, k represents kth time iteration, and i represents machine group #, represent i-th unit exerting oneself in kth time iteration, p _{i, t-1}for exerting oneself of t-1 period i-th unit.

Reserve Constraint adjusts:

By the method for penalty function, the amount of abandoning wind and cutting load in interval is retrained, shown in (38);

$f=f_0+\mathrm{\λ}(\underset{t=1}{\overset{T}{\Σ}}\mathrm{\β}\·p_{wlt0}+\underset{t=1}{\overset{T}{\Σ}}\mathrm{\γ}\·p_{lct0})---\left(38\right)$

Wherein, λ is penalty coefficient, gets a maximum, p _wlt0, p _lct0be respectively excessive due to wind power output interval and value of abandoning wind and cutting load that is that cause, β is for abandoning wind loss coefficient, and γ is cutting load loss coefficient, f ₀for target function, namely system synthesis originally, and shown in (27), t is for running the period, and span gets 24 from 1.P _wlt0, p _lct0computing formula as follows:

$p_{wlt0}=max(0,p_{wtmax}-p_{wt}-\underset{i=1}{\overset{I}{\Σ}}min(p_{i,t}-p_{imin},T_{15}{\mathrm{\Δ}}_{i,d}))---\left(39\right)$

$p_{lct0}=max(0,p_{wt}-p_{wtmin}-\underset{i=1}{\overset{I}{\Σ}}min(p_{imax}-p_{i,t},T_{15}{\mathrm{\Δ}}_{i,u}))---\left(40\right)$

Wherein, p _imax, p _iminrepresent that maximum, the minimum technology of unit i are exerted oneself respectively, p _wtrepresent wind power output, Δ _i,u, Δ _i,drepresent the rising of unit i, the constraint of decline climbing rate respectively, T ₁₅represent 15 minutes.

System power Constraints of Equilibrium adjusts:

When unit output cannot meet system power Constraints of Equilibrium, power shortage is now formula (41);

${\mathrm{\Δp}}_t=\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+p_{wt}-p_{lt}---\left(41\right)$

And power shortage is carried out unit output reallocation, concrete steps are as follows:

1) ask if | Δ p _t| < ε (ε is a very little positive number), then turn 7);

2) corresponding Unit Economic tiny increment is tried to achieve, if Δ p according to initially exerting oneself of each unit _t> 0, then arrange unit according to tiny increment order from big to small; If Δ p _t< 0, then arrange unit according to tiny increment order from small to large;

3) establish i=1, i is the machine group # after queuing up here;

4) p is established _{temp, t}=p _i,t, p _i,t=p _i,t-Δ p _t, then according to formula (41) adjustment p _i,tvalue, meet various constraints to make it.

5) Δ p is established _t=Δ p _t+ p _i,t-p _{temp, t}if, | Δ p _t| < ε, then turn 7), otherwise turn next step;

6) if i < is N, then establish i=i+1, then turn 4), otherwise turn next step;

7) terminate.

The concrete solution procedure of particle cluster algorithm is as follows:

(1) input wind-powered electricity generation and dope force value and load prediction goes out force value, scope of exerting oneself and the climbing restriction of unit is set;

(2) population dimension is set, Population Size, and iterations;

(3) using unit output and the wind-powered electricity generation confidence level in per moment as particle, the position range of each particle is set.

(4) position of initialization particle, according to successive dynasties optimal solution and the total optimization solution of the size initialization population of fitness function;

(5) by formula (42) and (43) the more position of new particle and speed,

$v_i^{k+1}={\mathrm{wv}}_i^k+c_1{r}_1({\mathrm{pbest}}_i-x_i^k)+c_2{r}_2(gbest-x_i^k)---\left(42\right)$

$x_i^{k+1}=x_i^k+v_i^{k+1}---\left(43\right)$

Wherein, for the speed of particle i when kth time iteration; c ₁, c ₂for accelerator coefficient; r ₁, r ₂it is equally distributed random number between 0 ~ 1; for the position of particle i when kth time iteration; Pbest _ifor the optimal solution that particle i itself finds; Gbest is the optimal solution of whole population; W is inertia coeffeicent, generally sets by the method for formula (44);

$w=w^{\max }-\mathrm{Cu}rCount\&CenterDot;\frac{w^{\max }-w^{\min }}{MaxCount}---\left(44\right)$

Wherein, w ^maxgenerally get 0.9, w ^mingenerally getting 0.4, CurCount is current iteration number of times, and MaxCount is maximum iteration time.

(6) the constraints requirement of formula (31) ~ formula (41) whether is met according to the particle that formula (42) ~ formula (43) inspection upgrades, if do not met, regenerate particle rapidity, upgrade position, until meet constraints;

(7) according to fitness function, the successive dynasties optimal solution of particle and total optimization solution are upgraded;

(8) judge whether current iteration number of times reaches maximum, if do not have, then repeat step (5) ~ (7), otherwise stop particle optimizing, export result of calculation.

The present invention to exert oneself and based on probability density distribution by wind power prediction, with wind-powered electricity generation confidence level and Unit Commitment, exert oneself as decision variable, consider to abandon the risk cost that wind and cutting load cause simultaneously, minimum for target function with risk cost outside economic dispatch cost in confidential interval and confidential interval, set up the Optimal Operation Model based on the optimum confidence level of electric power system, and the particle cluster algorithm of application enhancements solves.This research can be determined optimum wind-powered electricity generation confidence level, have wind-powered electricity generation to receive arrangement of exerting oneself that is interval and the interval lower system unit of Optimum Economic receiving most, for traffic department arranges the plan of exerting oneself a few days ago to provide reference, realizes optimum economic benefit simultaneously.

More than show and describe general principle of the present invention and principal character and advantage of the present invention.The technical staff of the industry should understand; the present invention is not restricted to the described embodiments; what describe in above-described embodiment and specification just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.Application claims protection range is defined by appending claims and equivalent thereof.

Claims (4)

1. based on an Optimization Scheduling for optimum wind-powered electricity generation confidence level, it is characterized in that, comprise following step:

(1) wind-powered electricity generation prediction data and unit performance parameter is gathered;

(2) with confidence level and unit output for decision variable sets up target function model;

This target function model is minimum for target function with risk cost outside economic dispatch cost in confidential interval and confidential interval, and described objective function Equation is as follows:

$F=min\{(\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+(\underset{t=1}{\overset{T}{\&Sigma;}}\mathrm{\&beta;}\&CenterDot;p_{wlt1}+\underset{t=1}{\overset{T}{\&Sigma;}}\mathrm{\&gamma;}\&CenterDot;p_{lct1})\}---\left(1\right)$

Wherein, wind-powered electricity generation confidence level and Unit Commitment, exert oneself as decision variable; In formula (1), the 1st for the economic dispatch cost in interval, the 2nd for the risk cost outside interval; Risk cost outside interval comprises pressure and abandons eolian p _wlt1with cutting load cost p _lct1;

(3) ask for pressure and abandon eolian p _wlt1with cutting load cost p _lct1, and abandon eolian p according to pressure _wlt1with cutting load cost p _lct1substitute into the target function that formula (1) asks for unit operation cost and risk cost minimization; Utilize risk expectation computing formula to calculate pressure and abandon eolian p _wlt1with cutting load cost p _lct1; This risk expectation computing formula is as follows: establish stochastic variable w _tmeet probability density function, i.e. normal distribution then calculate pressure and abandon eolian p _wlt1with cutting load cost p _lct1expectation;

$p_{wlt1}={\&Integral;}_{p_{wtmax}}^{p_{wmax}}{w}_t\&CenterDot;\frac{1}{\sqrt{2\mathrm{\&pi;}}}\&CenterDot;e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t\text{---}\left(2\right)$

$p_{lct1}={\&Integral;}_0^{p_{wtmin}}w\&CenterDot;\frac{1}{\sqrt{2\mathrm{\&pi;}}}\&CenterDot;e^{-\frac{{w_t}^2}{2}}{\mathrm{dw}}_t\text{---}\left(3\right)$

$f_{w_t}\left(w_t\right)=\frac{1}{\sqrt{2\mathrm{\&pi;}}}\&CenterDot;e^{-\frac{{w_t}^2}{2}}---\left(4\right)$

Wherein, a _i, b _i, c _irepresent the operation cost parameter of unit, p _i,trepresent i-th unit output, wherein, i=1,2 ..., N, β are for abandoning wind loss coefficient, and γ is cutting load loss coefficient, p _wlt1represent and abandon eolian, p outside interval _lct1represent the cutting load cost outside interval, p _wtmax, p _wtminrepresent that the wind-powered electricity generation of t predicts interval bound of exerting oneself, p respectively _wmaxrepresent wind farm grid-connected capacity;

(4) constraints is set up;

Constraints comprises power-balance constraints, unit output constraint, wind power output constraint, Climing constant and Reserve Constraint;

Power-balance constraints:

$\underset{i=1}{\overset{I}{\&Sigma;}}{p}_{i,t}+p_{wt}=p_{lt},t=1,...T---\left(5\right)$

Unit output retrains:

p _imin≤p _i,t≤p _imax(6)

Wind power output retrains:

p _wtmin≤p _wt≤p _wtmax(7)

Climing constant:

-Δ _i,dT ₆₀≤p _i,t-p _i,t-1≤Δ _i,uT ₆₀(8)

Reserve Constraint:

$\underset{i=1}{\overset{n}{\&Sigma;}}min(p_{imax}-p_{i,t},{\mathrm{\&Delta;}}_{i,u}{T}_{15})\&GreaterEqual;p_{wt}-p_{wt\min }---\left(9\right)$

$\underset{i=1}{\overset{n}{\&Sigma;}}min(p_{i,t}-p_{i\min ,}{\mathrm{\&Delta;}}_{i,d}{T}_{15})\&GreaterEqual;p_{wtmax}-p_{wt}---\left(10\right)$

Wherein, p _i,tfor the plan of i-th unit of t period is exerted oneself, p _wtfor the prediction of t period wind-powered electricity generation is exerted oneself, p _ltfor t period load prediction is exerted oneself; p _imin, p _imaxminimum and the maximum technology being respectively i-th unit is exerted oneself; p _wtmin, p _wtmaxbe respectively the minimum and maximum value of t period wind power output; Δ _i,d, Δ _i,ube respectively downwards and the upwards creep speed of i-th unit, get climbing time T ₆₀it is 1 hour;

(5) by modified particle swarm optiziation, adjustment has been entered to the unit bound constraint in step (4) and Climing constant, Reserve Constraint and system power Constraints of Equilibrium;

The constraint of unit bound and Climing constant adjustment: after each iteration of particle upgrades, controlled by particle on restrained boundary, make particle can meet the constraint of unit bound and Climing constant; Unit bound constraint after adjustment and Climing constant formula are:

$x_i^k=\left\{\begin{array}{c}\max (p_{i\min },p_{i,t-1}-{\mathrm{\&Delta;}}_{i,d}{T}_{60})\\ \begin{array}{cc}if& x_i^k\&le;(p_{i\min },p_{i,t-1}-{\mathrm{\&Delta;}}_{i,d}{T}_{60})\end{array}\\ \min (p_{i\max },p_{i,t-1}+{\mathrm{\&Delta;}}_{i,u}{T}_{60})\\ \begin{array}{cc}if& x_i^k\&GreaterEqual;\min (p_{i\max },p_{i,t-1}+{\mathrm{\&Delta;}}_{i,u}{T}_{60})\end{array}\end{array}\right.---\left(11\right)$

Wherein, k represents kth time iteration, and i represents machine group #, represent i-th unit exerting oneself in kth time iteration, p _{i, t-1}for exerting oneself of t-1 period i-th unit;

Reserve Constraint adjusts: by the method for penalty function, retrain the amount of abandoning wind and cutting load in interval, this constraint formulations is;

$f=f_0+\mathrm{\&lambda;}(\underset{t=1}{\overset{T}{\&Sigma;}}\mathrm{\&beta;}\&CenterDot;p_{wlt0}+\underset{t=1}{\overset{T}{\&Sigma;}}\mathrm{\&gamma;}\&CenterDot;p_{lct0})---\left(12\right)$

Wherein, λ is penalty coefficient, gets maximum; p _wlt0, p _lct0be respectively excessive due to wind power output interval and value of abandoning air quantity and cutting load that is that cause, β is for abandoning wind loss coefficient, and γ is cutting load loss coefficient, f ₀for target function, namely system synthesis originally, and such as formula shown, t is for running the period, and span gets 24 from 1; p _wlt0, p _lct0computing formula as follows:

$p_{wlt0}=max(0,p_{wtmax}-p_{wt}-\underset{i=1}{\overset{I}{\&Sigma;}}min(p_{i,t}-p_{i\min },T_{15}{\mathrm{\&Delta;}}_{i,d}))---\left(13\right)$

$p_{lct0}=max(0,p_{wt}-p_{wtmin}-\underset{i=1}{\overset{I}{\&Sigma;}}min(p_{imax}-p_{i,t},T_{15}{\mathrm{\&Delta;}}_{i,u}))---\left(14\right)$

Wherein, p _imax, p _iminrepresent that maximum, the minimum technology of unit i are exerted oneself respectively, p _wtrepresent wind power output, Δ _i,u, Δ _i,drepresent the rising of unit i, the constraint of decline climbing rate respectively, T ₁₅represent 15 minutes;

System power Constraints of Equilibrium adjusts: when unit output cannot meet system power Constraints of Equilibrium, power shortage is now Δ p _t, power shortage Δ p _tcomputing formula is:

${\mathrm{\&Delta;p}}_t=\underset{i=1}{\overset{I}{\&Sigma;}}{p}_{i,t}+p_{wt}-p_{lt}---\left(15\right)$

And power shortage is carried out unit output reallocation, concrete steps are as follows:

1) ask if | Δ p _t| < ε (ε is a very little positive number), then go to step 7);

2) corresponding Unit Economic tiny increment is tried to achieve, if Δ p according to initially exerting oneself of each unit _t> 0, then arrange unit according to tiny increment order from big to small; If Δ p _t< 0, then arrange unit according to tiny increment order from small to large;

3) establish i=1, i is the machine group # after queuing up here;

4) p is established _{temp, t}=p _i,t, p _i,t=p _i,t-Δ p _t, then according to power shortage Δ p _tcomputing formula adjustment p _i,tvalue, to make p _i,tmeet various constraints;

5) Δ p is established _t=Δ p _t+ p _i,t-p _{temp, t}if, | Δ p _t| < ε, then go to step 7), otherwise turn next step;

6) if i < is N, then establish i=i+1, then go to step 4), otherwise turn next step;

7) terminate;

(6) by the PSO Algorithm target function model of step (five), obtain the optimal Confidence Interval of optimum unit output scheme, optimum confidence level and correspondence thereof, obtain minimum entire system operating cost.

2. the Optimization Scheduling based on optimum wind-powered electricity generation confidence level according to claim 1, it is characterized in that, in described step (two), the corresponding wind-powered electricity generation of confidence level in each moment is received interval, the risk that wind-powered electricity generation receives the direct influential system of interval difference to run, namely the probabilistic risk cost of abandoning wind and cutting load produced when wind-powered electricity generation fluctuates outward, wind-powered electricity generation confidence level is joined in the target function of system operation cost with the form of risk cost, obtains object function as follows:

$\begin{array}{c}F=(\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{I}{\mathrm{\&Sigma;}}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+F_1+F_2\\ =(\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{I}{\mathrm{\&Sigma;}}}{a}_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+\mathrm{\&lambda;}(\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\mathrm{\&beta;}\&CenterDot;p_{wlt0}+\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\mathrm{\&gamma;}\&CenterDot;p_{lct0})+(\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\mathrm{\&beta;}\&CenterDot;p_{wlt1}+\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}\mathrm{\&gamma;}\&CenterDot;p_{lct1})\end{array}---\left(16\right)$

Wherein, F ₁represent the risk cost outside interval, F ₂represent the risk cost in interval.

3. the Optimization Scheduling based on optimum wind-powered electricity generation confidence level according to claim 1, it is characterized in that, in described step (five), fluctuation in interval is received at the wind-powered electricity generation of optimum as much as possible for making wind-powered electricity generation prediction curve, reserve capacity for system retrains, and wherein every platform unit can be raised the p that exerts oneself _imax-p _i,twith every 15 minutes upwards climbing amount Δs _i,ut ₁₅smaller value carry out cumulative summation, the constraint in constructive formula (17), this makes to receive interval lower limit by wind-powered electricity generation curve can be made to cover wind-powered electricity generation to the adjustment of unit process; Every platform unit can be lowered the p that exerts oneself simultaneously _imax-p _i,twith every 15 minutes upwards climbing amount Δs _i,ut ₁₅smaller value carry out cumulative summation, the constraint in constructive formula (18), cover wind-powered electricity generation by making wind-powered electricity generation curve to the adjustment of unit process and receive the interval upper limit;

$\underset{i=1}{\overset{n}{\&Sigma;}}min(p_{imax}-p_{i,t}{\mathrm{\&Delta;}}_{i,u}{T}_{15})\&GreaterEqual;p_{wt}-p_{wtmin}---\left(17\right)$

$\underset{i=1}{\overset{n}{\&Sigma;}}min(p_{i,t}-p_{imin,}{\mathrm{\&Delta;}}_{i,d}{T}_{15})\&GreaterEqual;p_{wtmax}-p_{wt}---\left(18\right).$

4. the Optimization Scheduling based on optimum wind-powered electricity generation confidence level according to claim 1, is characterized in that, in step (six), concrete solution procedure is as follows:

(1) input wind-powered electricity generation and dope force value and load prediction goes out force value, scope of exerting oneself and the climbing restriction of unit is set;

(2) population dimension is set, Population Size, and iterations;

(3) using unit output and the wind-powered electricity generation confidence level in per moment as particle, the position range of each particle is set;

(4) position of initialization particle, according to successive dynasties optimal solution and the total optimization solution of the size initialization population of fitness function;

(5) by formula (19) and formula (20) the more position of new particle and speed,

$v_i^{k+1}={\mathrm{wv}}_i^k+c_1{r}_1({\mathrm{pbest}}_i-x_i^k)+c_2{r}_2(gbest-x_i^k)---\left(19\right)$

$x_i^{k+1}=x_i^k+v_i^{k+1}---\left(20\right)$

Wherein, for the speed of particle i when kth time iteration; c ₁, c ₂for accelerator coefficient; r ₁, r ₂it is equally distributed random number between 0 ~ 1; for the position of particle i when kth time iteration; Pbest _ifor the optimal solution that particle i itself finds; Gbest is the optimal solution of whole population; W is inertia coeffeicent, sets by formula (21) method;

$w=w^{\max }-CurCount\&CenterDot;\frac{w^{\max }-w^{\min }}{MaxCount}---\left(21\right)$

(6) the constraints requirement of formula (5) ~ formula (15) whether is met according to the particle that formula (19) ~ formula (20) inspection upgrades, if do not met, regenerate particle rapidity, upgrade position, until meet constraints;

(7) according to fitness function, the successive dynasties optimal solution of particle and total optimization solution are upgraded;

(8) judge whether current iteration number of times reaches maximum, if do not have, then repeat step (5) ~ (7), otherwise stop particle optimizing, export result of calculation.