CN103824123A - Novel distribution network battery energy storage system optimal allocation algorithm - Google Patents

Abstract

The invention discloses a novel distribution network battery energy storage system optimal allocation algorithm. According to characteristics of a differential evolution algorithm which is simple in operation, convenient in discrete variable processing and characteristics of a primal-dual interior point method which is good in robustness and fast in calculation speed, a mixed solution algorithm which uses the differential evolution algorithm as a framework for optimizing the discrete variables and uses the primal-dual interior point method to optimize continuous variables is provided. In each iteration of differential evolution, the primal-dual interior point method is adopted for optimizing the continuous variable of each body in a population, and an objective function value serves as the fitness value of the body. The analysis result of an example shows that the mixed algorithm is applied to linear/nonlinear mixed optimal problems containing discrete-continuous variables, the searching ability is strong and the numerical stability is good.

Description

A kind of new power distribution network energy-storage system of accumulator distribute algorithm rationally

Technical field

The present invention relates to power distribution network batteries to store energy field, particularly a kind of new power distribution network energy-storage system of accumulator distribute algorithm rationally.

Background technology

Tradition electric energy cannot store, and in power scheduling, must meet in real time the energy equilibrium between generating and electricity consumption.Along with the development of energy storage technology, stored energy application is progressively feasible in electric system, and it provides a kind of effective method to electrical power storage.Along with the increase of power distribution network load peak-valley difference, power supply growing tension, utilizes the energy storage can peak load shifting, improves the utilization factor of controller switching equipment, delays electrical network upgrading, alleviates mains supply pressure, and therefore energy storage will have broad application prospects in following power distribution network.

Energy storage technology is mainly divided into four classes such as physics energy storage, Power Flow, batteries to store energy and phase-change accumulation energy, and wherein batteries to store energy has that the speed of discharging and recharging is fast, efficiency is high, long service life, can realize extensive utilization and the advantages such as geographical conditional request is low be received to more concern and research.Therefore energy-storage system of accumulator is more suitable for urban distribution network.For accumulator, capacity and power can separate configurations, and an extensive batteries to store energy equipment is made up of power model and the energy module of multiple unit.As required, can accumulators accumulator system power model and the energy module of configuration different units number.In power distribution network, energy-storage system of accumulator is generally arranged on middle voltage distribution networks side, uses as the peaking power source of load side, can give full play to so the low storage of energy storage advantage occurred frequently, also can reduce the electric energy loss regulating in load process simultaneously.

Differential evolution algorithm is a kind of heuristic search algorithm based on Swarm Evolution, and this algorithm is applied at aspects such as electric system Electric Power Network Planning, load economical distribution and idle work optimizations at present.Differential evolution algorithm adjusting parameter simple to operate is few, processes discrete variable convenient, have global optimizing ability, but computing velocity is slow.Former dual interior point computing velocity is fast, optimizing ability is strong but processing discrete variable difficulty.Hybrid algorithm based on differential evolution and interior point method can solve the linear/non-linear hybrid optimization problem that contains discrete-continuous variable, can and overcome deficiency separately in conjunction with the advantage of differential evolution and interior point method.

Summary of the invention

Goal of the invention: technical matters to be solved by this invention is for capacity and power configuration problem that batteries to store energy device is installed in power distribution network, proposes a kind of hybrid algorithm based on differential evolution and interior point method for solving the configuration of power distribution network accumulator and realizing the optimization that day part charge/discharge operates.

Technical scheme: the present invention for achieving the above object, adopts following technical scheme: a kind of new power distribution network energy-storage system of accumulator distribute algorithm rationally, it is characterized in that, described method realizes according to the following steps successively in calculating:

(1) obtain power distribution network and energy-storage system of accumulator parameter information.

(2) Population Size, maximum evolutionary generation, zoom factor, the intersection factor of setting differential evolution algorithm; Set slack variable, Lagrange multiplier and penalty factor initial value and the convergence precision of former dual interior point;

(3) using the power model number of energy-storage system of accumulator and energy module number as discrete variable to be optimized, continuous variable peak clipping rate together participates in the optimization of differential evolution algorithm, generates initial population;

(4) adopt former dual interior point to be optimized each individual continuous variable of population, i.e. the charge/discharge power of day part accumulator, and the target function value that calculates, as individual fitness, obtains global optimum's individuality of initial population with this after optimizing;

(5) carry out Population Variation operation according to following formula:

v _i,j(t+1)＝x _r1,j(t)+F(x _r2,j(t)-x _r3,j(t))

(6) carry out interlace operation according to following formula:

$u_{i,j}(t+1)=\left\{\begin{array}{c}{v}_{i,j}(t+1),\mathrm{ifrand}\left(\mathrm{0,1}\right)\≤C_R\mathrm{orj}=\mathrm{randi}(1,n)\\ x_{i,j}(t+1),\mathrm{else}\end{array};\right.$

(7) keep discrete variable and the continuous variable of each individuality that above-mentioned differential evolution algorithm obtains constant, adopt the continuous variable (be day part accumulator charge/discharge power) of former dual interior point to the each individuality of population to be optimized, and the target function value f that optimization is obtained is as the individual fitness of population, global optimum's individuality of Population Regeneration;

(8) select operation according to following formula:

$x_i(t+1)=\left\{\begin{array}{c}{u}_i(t+1),\mathrm{iff}\left(u_i(t+1)\right)\≤f\left(x_i\left(t\right)\right)\\ x_i\left(t\right),\mathrm{else}\end{array};\right.$

(9) judge whether to reach maximum evolutionary generation, if so, Output rusults, quits a program; If not, put evolutionary generation t and add 1, return to (5) and continue iteration.

As optimization, described parameter information comprises that daily load is meritorious, the year operation expense of the cost of the cost of the year load growth rate in tou power price, power distribution network transfer cost, electrical network upgrading construction cost, power distribution network area, inflation rate, rate of discount, battery cell's charge/discharge power, unit capacity, energy storage specific power, accumulator tenure of use, year days running etc.;

Beneficial effect: the present invention compared with prior art: the power distribution network energy-storage system of accumulator method of distributing rationally that the present invention proposes, life cycle management based on energy storage is set up the Economic Benefit Model of power distribution network batteries to store energy, has taken into account energy storage and has been realized arbitrage, reduces transfer cost, delayed the electrical network upgrading profit of three aspects: and fixed investment cost, the operation expense of energy storage by the charge/discharge operation of different time.For the linear/non-linear optimization problem that contains continuously-mixed discrete variable, algorithm effectively combines the advantage of differential evolution algorithm and interior point method herein, adopt differential evolution algorithm to optimize power model group number and the energy module group number of discrete variable accumulator, and the peak clipping rate of continuous variable energy storage, in the iteration each time of differential evolution algorithm, adopt former dual interior point to be optimized and to optimize an individual fitness assessment energy storage 24 period charge/discharge power, thereby the algorithm that is can effectively be processed hybrid optimization problem, and have concurrently and process that discrete-continuous variable is convenient and robustness is good, the feature that numerical stability is strong.

Accompanying drawing explanation

Fig. 1 is the inventive method process flow diagram;

Fig. 2 is the accumulator that the present invention is based on structural representation at power distribution network;

Fig. 3 is tou power price curve synoptic diagram of the present invention;

Fig. 4 is power distribution network transfer cost schematic diagram of the present invention;

Fig. 5 is the average daily load curve schematic diagram of power distribution network of the present invention;

Fig. 6 is that the present invention adopts NAS(sodium-sulphur battery) convergence curve schematic diagram;

Fig. 7 is the charge/discharge curve schematic diagram of NAS of the present invention.

Embodiment

Differential evolution (differentialevolution, DE) algorithm is proposed in nineteen ninety-five by StornR and PriceK, it is a kind of heuristic search algorithm based on Swarm Evolution, and this algorithm is applied at aspects such as electric system Electric Power Network Planning, load economical distribution and idle work optimizations at present.The same with intelligent algorithm, first DE algorithm will generate at random initialization population X (0)=[x in variable edge bound constrained ₁(0) ..., x _n(0)], x wherein _i(0)=[x _i1(0) ..., x _in(0)] ^t, N is population quantity .. is dimension.

Select three individual x current t is random in for population _r1(t), x _r2(t), x _r3(t), must make a variation according to following formula

v _i(t+1)＝x _r1(t)+F(x _r2(t)-x _r3(t))，r1≠r2≠r3≠i

In formula: .., x _r2and x (t) _r3(t) three different individualities for choosing at random in population, subscript represents current population at individual; v _i(t+1) be i the individuality of variation population; F is zoom factor, characterizes individual degree of variation, generally gets [0.4,1.0].

In order to increase the diversity of population at individual, initial population and variation population are carried out to interlace operation:

$u_{i,j}(t+1)=\left\{\begin{array}{c}{v}_{i,j}(t+1),\mathrm{if}:\mathrm{rand}\left(\mathrm{0,1}\right)\≤C_R\mathrm{orj}=\mathrm{randi}(1,n)\\ x_{i,j}(t+1),\mathrm{else}\end{array};\right.$

In formula: u _i,j(t+1) be i j the individual component of intersection population; Rand (0,1) is for obeying [0,1] equally distributed random number; Randi (1, N) be [1 ..., n] between the integer chosen at random; C _rfor intersecting the factor, in the middle of characterizing, individual component replaces the probability of current individual component, C _r∈ [0,1].

Finally, from current population and Experimental population, carry out preferably according to fitness value.For maximization problems, be greater than original individuality if test individual fitness value, test is individual replaces original individuality, otherwise retains original individuality, forms initial population of future generation.

$x_i(t+1)=\left\{\begin{array}{cc}{u}_i(t+1)& f\left(u_i(t+1)\right)>f\left(x_i\left(t\right)\right)\\ x_i(t+1)& f\left(u_i(t+1)\right)\≤f\left(x_i\left(t\right)\right)\end{array}\right.$

In formula: f (*) is ideal adaptation degree functional value.

Former dual interior point (primal-dualinteriorpointmethod, PDIPM) proposed in 1984 by Karmarkar, because its convergence rapidly, the advantages such as strong robustness have won the favor of optimizing field scholar, are widely used in the field such as Optimal Power Flow Problems, state estimation.Interior point method also expands to nonlinear optimization field by linear optimization field originally.

Optimization Problems In Power Systems generally comprises equality constraint and inequality constrain, and its mathematical model can be described with following formula:

$\left\{\begin{array}{cc}\min & f\left(x\right)\\ s.t.& h\left(x\right)=0\\ & \underset{\‾}{g}\≤g\left(x\right)\≤\stackrel{\‾}{g}\end{array}\right.$

Wherein x is variable to be optimized, and f (x) is objective function, and h (x), g (x) are respectively equality constraint and inequality constrain,

with gbe respectively the bound of g (x).

The basic ideas of interior point method are: inequality constrain is converted into equality constraint by slack variable, recycling Lagrange multiplier is incorporated into constraint in objective function, and slack variable is retrained with barrier function method.For OPF problem, structure Lagrangian function is as follows:

$L=f\left(x\right)-y^{T}h\left(x\right)-z^T[g\left(x\right)-l-\underset{\‾}{g}]-w^T[g\left(x\right)+u-\stackrel{\‾}{g}]-\mathrm{\μ}\underset{j=1}{\overset{r}{\mathrm{\Σ}}}\ln \left(l_j\right)-\mathrm{\μ}\underset{j=1}{\overset{r}{\mathrm{\Σ}}}\ln \left(u_j\right)$

In formula: y=[y ₁..., y _m] ^tfor the Lagrange multiplier of equality constraint, z=[z ₁..., z _r] ^t, w=[w ₁..., w _r] ^tfor the Lagrange multiplier of inequality constrain, l=[l ₁..., l _r] ^t, u=[u ₁..., u _r] ^tfor the slack variable of inequality constrain, μ is the penalty factor of barrier function, and μ is obstacle constant, the number that r is inequality constrain.

KKT (Karush-Kuhn-Tucker) condition of this problem is:

The saddle point of LagrangianL is the smallest point of objective function.

$\left\{\begin{array}{c}{L}_x={\▿}_{x}f\left(x\right)-{\▿}_{x}h\left(x\right)y-{\▿}_{x}g\left(x\right)(z+w)=0\\ L_y=h\left(x\right)=0\\ L_z=g\left(x\right)-l-\underset{\‾}{g}=0\\ L_w=g\left(x\right)+u-\stackrel{\‾}{g}=0\\ L_l=z-\mathrm{\μ}{U}^{-1}e=0\\ L_u=-w-{\mathrm{\μL}}^{-1}=0\end{array}\right.$

In formula: ▽ _xf (x) is 1 order derivative of f (x) to x, ▽ _xh (x), ▽ _xg (x) is respectively the Jacobian matrix of h (x), g (x).L＝diag(l ₁,…,l _r)，U＝diag(u ₁,…,u _r)，Z＝diag(z ₁,…,z _r)，W＝diag(w ₁,…,w _r)，L ^-1＝diag(1/l ₁,…,1/l _r)，U ^-1＝diag(1/u ₁,…,1/u _r)，e＝[1,…,1] ^T。

By in formula KKT condition latter two equation can be in the hope of obstacle constant

definition C _gap=l ^tz-u ^tw.

Practice is general to be adopted

$\mathrm{\μ}=\mathrm{\σ}\frac{C_{\mathrm{Gap}}}{2r}$

Wherein σ is called Center Parameter, generally gets 0.1.Nonlinear System of Equations in KKT condition can solve by the inferior method of newton-pressgang, by its linearization, can obtain:

$\left[\begin{array}{cc}{H}^{\′}& {\▿}_{x}h\left(x\right)\\ {\▿}_x^{T}h\left(x\right)& 0\end{array}\right]\left[\begin{array}{c}\mathrm{\Δx}\\ \mathrm{\Δy}\end{array}\right]=\left[\begin{array}{c}{L}_x^{\′}\\ -L_y\end{array}\right]$

$\left[\begin{array}{cc}I& L^{-1}Z\\ 0& I\end{array}\right]\left[\begin{array}{c}\mathrm{\Δz}\\ \mathrm{\Δl}\end{array}\right]=\left[\begin{array}{c}-L^{-1}{L}_l^{\mathrm{\μ}}\\ L_z+{\▿}_x^{T}g\left(x\right)\mathrm{\Δx}\end{array}\right]$

$\left[\begin{array}{cc}I& U^{-1}W\\ 0& I\end{array}\right]\left[\begin{array}{c}\mathrm{\Δw}\\ \mathrm{\Δu}\end{array}\right]=\left[\begin{array}{c}-U^{-1}{L}_u^{\mathrm{\μ}}\\ -L_w-{\▿}_x^{T}g\left(x\right)\mathrm{\Δx}\end{array}\right]$

In formula: Δ x, Δ y, Δ z, Δ l, Δ u, Δ w are the correction of x, y, z, l, u, w,

${L^{\′}}_x=L_x+{\▿}_{x}g\left(x\right)[L^{-1}(L_l^{\mathrm{\μ}}+{\mathrm{ZL}}_z)+U^{-1}(L_u^{\mathrm{\μ}}+{\mathrm{WL}}_w)],$

$H^{\′}=H-{\▿}_{x}g\left(x\right)[L^{-1}Z-U^{-1}W]{\▿}_x^{T}g\left(x\right),$

$H=-[{\▿}_x^{2}f\left(x\right)-{\▿}_x^{2}h\left(x\right)y-{\▿}_x^{2}g\left(x\right)(z+w)].$

The above-mentioned three prescription journeys of solving equation can obtain the correction of the k time iteration of interior point method.

Differential evolution algorithm adjusting parameter simple to operate is few, processes discrete variable convenient, have global optimizing ability, but computing velocity is slow.Former dual interior point computing velocity is fast, optimizing ability is strong but processing discrete variable difficulty.The relative merits of comprehensive this two classes algorithm, propose the hybrid algorithm of a kind of combination DE and PDIPM herein, and numerical results shows, this algorithm the convergence speed is fast, and optimizing ability is strong.

The core concept of hybrid algorithm is: take DE algorithm as framework, optimize rated power P, rated capacity E and the peak clipping rate λ of discrete variable accumulator, in the iterative process each time of DE, adopt interior point method individuality to be carried out to optimization and the fitness assessment of continuous variable (energy storage 24 period charge/discharge power).

Consider that the charge/discharge operation of energy storage by different time realizes arbitrage, reduces transfer cost, delays the income of electrical network upgrading three aspects:.Below in conjunction with accompanying drawing, the technical scheme of invention is elaborated:

Within one day, be divided into 24 periods, " low storage the is occurred frequently " arbitrage in energy storage one day is

$B_1=\underset{i=1}{\overset{24}{\mathrm{\Σ}}}(P_{\mathrm{dis}}\left(i\right)U_{\mathrm{dis}}\left(i\right)-P_{\mathrm{ch}}\left(i\right)U_{\mathrm{ch}}\left(i\right))p_e\left(i\right).$

In formula: i is period sequence number, P _dis(i), P _ch(i) be charge power and the discharge power of i period accumulator, U _dis, U _chfor the charging and discharging state variable of i period accumulator, p _e(i) be the electricity price of the i period shown in Fig. 3.

It is pointed out that charging and discharging state variable meets constraint:

$\left\{\begin{array}{c}{U}_{\mathrm{dis}}\left(i\right)+U_{\mathrm{ch}}\left(i\right)=1\\ U_{\mathrm{dis}}\left(i\right)\×U_{\mathrm{ch}}\left(i\right)=0\end{array}\right.$

In the electricity market separating in transmission & distribution, electric power for, just need distribution company to pay certain electric energy transfer cost.The transshipment charge difference of different time as shown in Figure 4, and general and load positive correlation.Utilize the difference of different time electric energy transfer cost, by the transmission time of energy storage transfer part electric energy, reduce transfer cost.Install additional after batteries to store energy, the transfer cost that can reduce in a month is

${\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="HDA0000462231580000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_2=\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}(P_{\mathrm{dis}}\left(i\right)-P_{\mathrm{ch}}\left(i\right))p_r\left(i\right)$

In formula: p _r(i) be delivery of electrical energy expense in i period power distribution network as shown in Figure 4.

In power distribution network, when a certain line load increase year after year exceed capacity, need to upgrade to electrical network.The classic method of distribution upgrading is to increase the transformer of building or change the outfit, or distribution line is transformed, to meet the demand of load growth.The cost of distribution upgrading is higher, and installs after energy storage by peak load shifting, realizes peak load regulation network, can improve grid equipment utilization factor, delay electrical network upgrading, thereby reduce electric grid investment construction cost.If the annual growth of load is τ, the peak clipping rate of energy storage is λ, installs the year number that can delay electrical network upgrading after energy storage additional to be

$\mathrm{\&Delta;n}=\frac{{\log }_{10}(1+\mathrm{\&lambda;})}{{\log }_{10}(1+\mathrm{\&tau;})}$

The income that delays electrical network upgrading is

$B_3=C_{\mathrm{inv}}(1-{\left(\frac{1+i_r}{1+d_r}\right)}^{\mathrm{\&Delta;n}})$

In formula: C _invfor electrical network upgrading construction cost, i _rfor inflation rate, d _rfor rate of discount.

The overall life cycle cost of energy storage mainly comprises fixed investment cost and operation expense.Fixed investment cost can be expressed as

C ₁＝C _pP+C _eE

In formula: C _pfor the cost of energy storage specific power, the specified charge/discharge power that P is energy storage device, C _efor the cost of energy storage unit capacity, the rated capacity that E is energy storage device.

The year operation expense of BESS is main relevant to the rated power of energy-storage battery:

C ₂＝C _mP

In formula: C _mfor the year operation expense of energy storage unit's charge/discharge power.

Total economic benefit in energy storage life cycle management is installed in power distribution network after accumulator is

$B=\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}{B}_{1}D{\left(\frac{1+i_r}{1+d_r}\right)}^t+\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}{B}_{2}M{\left(\frac{1+i_r}{1+d_r}\right)}^t+B_3-C_1-\underset{t=1}{\overset{T}{\mathrm{\&Sigma;}}}{C}_2{\left(\frac{1+i_r}{1+d_r}\right)}^t$

In formula: in the use time that t is BESS, in the life-span that T is BESS, D is the number of days that utilizes of BESS 1 year, and M is the moon umber comprising for a year.

For energy-storage battery can be recycled, accumulator needs to meet energy conservation in a charge/discharge cycle,

$\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}(P_{\mathrm{dis}}\left(i\right)-P_{\mathrm{ch}}\left(i\right)\mathrm{\&eta;})=0$

In formula: the energy conversion efficiency that η is batteries to store energy, for the energy storage battery of different medium, η is difference to some extent.

The accumulator system of having determined for rated capacity and specified charge/discharge power must meet the constraint of maximum charge/discharge power in the process of its operation, and the constraint of total charge volume.

P-P _dis(i)≥0，i＝1,...,24

P-P _ch(i)≥0，i＝1,...,24

$\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{P}_{\mathrm{dis}}\left(i\right)\&le;E$

In addition, by discharging and recharging of energy storage, system equivalent load should be not more than the load peak after peak clipping:

$P_{\mathrm{load}}\left(i\right)-P_{\mathrm{dis}}\left(i\right)+P_{\mathrm{ch}}\left(i\right)\&le;(1-\mathrm{\&lambda;})\stackrel{\&OverBar;}{P_{\mathrm{load}}}$

In formula: P _load(i) be the load of i period as shown in Figure 5,

it is the load peak of 24 periods.

Be objective function to the maximum with economic benefit total in energy storage life cycle management herein, the Economic Benefit Model of power distribution network BESS is expressed as

obj.max.B(x,y)

s.t.h(x)=0

$g\left(x\right)\&le;\stackrel{\&OverBar;}{g}\left(y\right)$

In formula: x=[P _dis(1), P _ch(1) ..., P _dis(24), P _ch(24)] ^t, represent that one day day part BESS's discharges and recharges power, y=[E, P, λ] ^t, h (x)=0 is equality constraint,

for inequality constrain,

be the function about y, represent the upper limit of inequality constrain.

Power distribution network batteries to store energy distribute rationally with day part charge/discharge operation optimization problem in, adopt DE algorithm to be optimized discrete variable P, E and continuous variable λ, then for each individual definite P, E and λ, optimize day part energy storage charge/discharge power by PDIPM, concrete steps are as follows:

(1) obtain power distribution network and energy-storage system of accumulator parameter information.Comprise: daily load is meritorious, the year operation expense of the cost of the cost of the year load growth rate in tou power price, power distribution network transfer cost, electrical network upgrading construction cost, power distribution network area, inflation rate, rate of discount, battery cell's charge/discharge power, unit capacity, energy storage specific power, accumulator tenure of use, year days running etc.;

(2) program initialization.Comprise: Population Size, maximum evolutionary generation, zoom factor, the intersection factor of setting differential evolution algorithm; Set slack variable, Lagrange multiplier and penalty factor initial value and the convergence precision of former dual interior point;

(3) generate at random initial population.Using the power model number of energy-storage system of accumulator and energy module number as discrete variable to be optimized, continuous variable peak clipping rate together participates in the optimization of differential evolution algorithm, generates initial population; (4) initial value of each individual optimal solution and initial population globally optimal solution in calculating population.Adopt former dual interior point to be optimized each individual continuous variable of population, i.e. the charge/discharge power of day part accumulator, and the target function value that calculates, as individual fitness, obtains global optimum's individuality of initial population with this after optimizing;

(5) carry out Population Variation operation according to following formula:

v _i,j(t+1)＝x _r1,j(t)+F(x _r2,j(t)-x _r3,j(t))

(6) carry out interlace operation according to following formula:

$u_{i,j}(t+1)=\left\{\begin{array}{c}{v}_{i,j}(t+1),\mathrm{ifrand}\left(\mathrm{0,1}\right)\&le;C_R\mathrm{orj}=\mathrm{randi}(1,n)\\ x_{i,j}(t+1),\mathrm{else}\end{array}\right.$

(7) optimize continuous variable.Keep discrete variable and the continuous variable of each individuality that above-mentioned differential evolution algorithm obtains constant, adopt the continuous variable (be day part accumulator charge/discharge power) of former dual interior point to the each individuality of population to be optimized, and the target function value f that optimization is obtained is as the individual fitness of population, global optimum's individuality of Population Regeneration;

(8) select operation according to following formula:

$x_i(t+1)=\left\{\begin{array}{c}{u}_i(t+1),\mathrm{iff}\left(u_i(t+1)\right)\&le;f\left(x_i\left(t\right)\right)\\ x_i\left(t\right),\mathrm{else}\end{array}\right.$

(9) judge whether to reach maximum evolutionary generation, if so, Output rusults, quits a program; If not, put evolutionary generation and add 1, return to (5) and continue iteration.

The present invention adopts the hybrid algorithm based on differential evolution algorithm and former antithesis adytum algorithm to solve distributing rationally with the charge/discharge of day part of power distribution network energy-storage system of accumulator and operates.By Simulation Example, verify the validity of algorithm of the present invention.

Introduce one embodiment of the present of invention below:

The present invention adopts the hybrid algorithm based on DE and PDIPM to carry out simulation calculation to the distribution network system take Fig. 3-Fig. 5 as feature, and correlation parameter is as shown in table 1.NAS (sodium-sulphur battery) the effect energy-storage battery of selecting current widespread use, NAS correlation parameter is as shown in table 2.

Table 1 distribution network system correlation parameter

Table 2NAS correlation parameter

Cell power module and energy module are taken as respectively 200kW and 360kWh.The specified charge/discharge power setting of accumulator is 1-15MW (power model 5-75 group), and rated capacity is set as 20-70MWh (energy module 55-195 group), and peak clipping rate is set as 0-10%.Set the population scale N=100 of DE algorithm, maximum evolutionary generation K _max=100, zoom factor is got C _r=0.9, the factor of intersecting is got F=0.8.

Optimum results is that power model configures 41 groups, energy module and configures 134 groups, and the rated power of NAS is 8.2MW, and rated capacity is 48.24MWh.In life cycle management, the total benefit of batteries to store energy is 720446 $.

This convergence of algorithm curve as shown in Figure 6.

Can find out, iterative process can reach balance before 40 generations, and this has also verified that invention algorithm has good convergence.Algorithm of the present invention combine DE algorithm regulate search capability strong, be applicable to hybrid variable optimization and PDIPM robustness is good, computing velocity is fast advantage, the hybrid algorithm better numerical value stability proposing, optimizing ability is strong.

With negative number representation NAS charging, positive number represents electric discharge, and the filling of 24 period accumulators/electric curve as shown in Figure 7.

Claims (2)

1. New power distribution network energy-storage system of accumulator distribute algorithm rationally, it is characterized in that, described method is to realize according to the following steps successively in computing machine:

   (1) obtain power distribution network and energy-storage system of accumulator parameter information;

   (2) program initialization, comprises and sets the Population Size of differential evolution algorithm, maximum evolutionary generation, zoom factor, the intersection factor; Set slack variable, Lagrange multiplier and penalty factor initial value and the convergence precision of former dual interior point;

   (3) using the power model number of energy-storage system of accumulator and energy module number as discrete variable to be optimized, continuous variable peak clipping rate together participates in the optimization of differential evolution algorithm, generates initial population;

   (4) adopt former dual interior point to be optimized each individual continuous variable of population, i.e. the charge/discharge power of day part accumulator, and the target function value that calculates, as individual fitness, obtains global optimum's individuality of initial population with this after optimizing;

   (5) carry out Population Variation operation according to following formula:

   v _i,j(t+1)＝x _r1,j(t)+F(x _r2,j(t)-x _r3,j(t))；

   (6) carry out interlace operation according to following formula:

   $u_{i,j}(t+1)=\left\{\begin{array}{c}{v}_{i,j}(t+1),\mathrm{ifrand}\left(\mathrm{0,1}\right)\&le;C_R\mathrm{orj}=\mathrm{randi}(1,n)\\ x_{i,j}(t+1),\mathrm{else}\end{array};\right.$

   (7) keep discrete variable and the continuous variable of each individuality that above-mentioned differential evolution algorithm obtains constant, adopt the continuous variable (be day part accumulator charge/discharge power) of former dual interior point to the each individuality of population to be optimized, and the target function value f that optimization is obtained is as the individual fitness of population, global optimum's individuality of Population Regeneration;

   (8) select operation according to following formula:

   $x_i(t+1)=\left\{\begin{array}{c}{u}_i(t+1),\mathrm{iff}\left(u_i(t+1)\right)\&le;f\left(x_i\left(t\right)\right)\\ x_i\left(t\right),\mathrm{else}\end{array};\right.$

   (9) judge whether to reach maximum evolutionary generation, if so, Output rusults, quits a program; If not, put evolutionary generation t and add 1, return to (5) and continue iteration.
2. New power distribution network energy-storage system of accumulator according to claim 1 distribute algorithm rationally, it is characterized in that, described parameter information comprises that daily load is meritorious, the year operation expense of the cost of the cost of the year load growth rate in tou power price, power distribution network transfer cost, electrical network upgrading construction cost, power distribution network area, inflation rate, rate of discount, battery cell's charge/discharge power, unit capacity, energy storage specific power, accumulator tenure of use, year days running etc.

Images