CN106026077A - Wind power unit acceptance cost index determining method for power grid day-ahead wind power acceptance capability evaluation method based on multi-target optimization - Google Patents

Abstract

The invention discloses a kind of power grid based on multiple-objection optimization, wind-powered electricity generation receives the wind-powered electricity generation unit of capability assessment method that the indicator of costs is received to determine method a few days ago, based on power grid, wind-powered electricity generation receives the wind-powered electricity generation unit of capability assessment model Pareto optimal solution set to receive indicator of costs CApu a few days ago, i is calculated as follows: According to the definition that Pareto is optimal, respectively solves corresponding cost of electricity-generating and meet following relationship: CG, 1 amp; lt; CG,2 lt; .. CG, n wind-powered electricity generation receive the Pareto optimal solution set of capability assessment model to be made of n different solutions, receive the size of electricity Aw to be ranked up Pareto optimal solution set by wind-powered electricity generation, after sequence, there are following relationships: AW, 1 amp; lt; AW,2 lt; …AW,n. The present invention proposes wind-powered electricity generation unit on the basis of wind-powered electricity generation a few days ago receives capability assessment model Pareto optimal solution set and receives the indicator of costs, and can measure power grid is the cost price for receiving wind-powered electricity generation to pay.

Description

Electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago receives the wind-powered electricity generation list of capability assessment method Position receives the indicator of costs to determine method

The application is that application number " 201510031432.3 ", the applying date: 2015-01-21, title are " based on multiple-objection optimization Electrical network wind-powered electricity generation a few days ago receive capability assessment method " divisional application.

Technical field

The present invention relates to renewable energy power interconnection technology, be specifically related to a kind of electrical network based on multiple-objection optimization a few days ago Wind-powered electricity generation receives capability assessment method and wind-powered electricity generation unit to receive the indicator of costs.

Background technology

Nearly ten years, along with petering out of Fossil fuel and increasingly sharpening of environmental pollution, the whole world to development with Wind-powered electricity generation is that the regenerative resource of representative all gives enough attention.Ended for the end of the year 2012, through the high_speed development of several years, China The wind-powered electricity generation permeability of the subregion electrical network that wind-resources enriches district has reached higher level, sends out as the installation of power grid of West Inner Mongolia wind-powered electricity generation accounts for The ratio of electricity total installed capacity is up to 22.2%.Large-scale wind power is grid-connected adds the difficulty of scheduling decision, be degrading partial electric grid The quality of power supply, more seriously, when dispatching of power netwoks resource cannot balance the random fluctuation of wind power, in fact it could happen that serious " abandon wind ".2013, China's electrical network " abandons wind ", and electricity was up to 16,200,000,000 kilowatts, accounts for the 10% of wind-power electricity generation total amount then.

Along with " abandoning wind " phenomenon is day by day serious, electrical network is abandoned wind reason and has been carried out profound analysis by academia, and from multiple The wind-powered electricity generation of electrical network is received ability to be assessed by time angle, thus provides reference for scheduling decision.Document one is " electrically-based The Liaoning electric power grid of balance receives wind-powered electricity generation capability analysis " (Automation of Electric Systems, 2010, volume 34, the 3rd phase, page 86 was to 90 Page) by thinking, present stage causes the main cause of " abandoning wind " to be the restriction of conveying capacity and peak modulation capacity, as wind-powered electricity generation The grid-connected system load flow that causes, voltage stabilization, the problem such as the quality of power supply can solve inside partial electric grid, are still unlikely to restriction Wind-powered electricity generation is dissolved by whole electrical network.Document two " relevant issues of large-scale wind power access electrical network and measure " (China's electrical engineering Journal, volume 30, the 25th phase, page 1 to 9 page in 2010) compared for the power supply architecture of Sino-German two countries, it is believed that and power supply architecture is not It it is rationally the one of the main reasons causing extensive " abandoning wind ".Document three " considers the real-time wind electricity digestion energy of Network Security Constraints Force estimation " (Proceedings of the CSEE, 2013, volume 33, the 16th phase, page 23 to 29) considering Network Security Constraints On the basis of, receive ability to be assessed power grid wind from real time execution angle, and the wind-powered electricity generation analyzing emphatically each node connects Receive ability.Document four " wind electricity digestion capability appraisal procedure based on wind power prediction a few days ago " (electric power network technique, 2012, the 36th Volume, the 8th phase, page 69 to 75) from time angle analysis a few days ago, power grid wind receives ability, it is proposed that and wind-powered electricity generation can be dissolved " bag Network band " concept, provide useful reference to dispatcher.

The wind-powered electricity generation that document three, four proposes receives capability assessment method only to provide single assessment result, lays particular emphasis on displaying electrical network Theoretical maximum wind-powered electricity generation receive ability.Additionally, existing wind-powered electricity generation receives capability assessment model to have ignored wind-powered electricity generation in assessment completely Receive cost, thus the most do not propose corresponding wind-powered electricity generation and receive the indicator of costs.

Summary of the invention

It is an object of the invention to provide a kind of assessment more fully, easy electrical network based on multiple-objection optimization wind a few days ago Electricity receives capability assessment method.

The technical solution of the present invention is:

A kind of electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago receives capability assessment method, it is characterized in that: electrical network wind a few days ago Electricity receives capability assessment model to have wind-powered electricity generation to receive maximum two optimization aim minimum with conventional system cost of electricity-generating of ability, following institute Show:

Optimization aim 1:

Optimization aim 2:

In above formula, A_wReceive electricity for scheduling wind-powered electricity generation in a few days, deduct for scheduling wind-powered electricity generation prediction electricity in a few days and abandon wind-powered electricity generation Amount expectation；P_i,tFor unit i at the output of period t；u_iRepresenting that unit i is dispatching running status in a few days, " 0 " expression stops Machine, " 1 " represents start；T is scheduling slot number；F_w,tWind-powered electricity generation theoretical maximum for period t is exerted oneself, i.e. wind power prediction value；C_w,t " abandoning wind " electricity for period t is expected, with P_i,tAnd u_iRelevant；C_GCost of electricity-generating in a few days is being dispatched for conventional system；N is conventional The number of unit；f_i(P_i,t) it is the unit i fuel cost function at period t, can be by quadratic function matching；

The constraints of electrical network wind-powered electricity generation a few days ago receiving capability assessment model is as follows:

System active balance retrains:

$P_{d,t}-\underset{i=1}{\overset{N}{\Σ}}{u}_i{P}_{i,t}-P_{w,t}=0$

In above formula, P_d,tFor the predicted load of moment t, P_w,tWind-powered electricity generation electricity volume for moment t；

Conventional power unit units limits:

P_min,i≤P_i,t≤P_max,i

In above formula, P_max,i、P_min,iBe respectively the maximum of unit i, minimum technology is exerted oneself；

Climing constant:

P_i,t-P_i,t-1≤ΔTR_up,i

P_i,t-1-P_i,t≤ΔTR_down,i

In above formula, R_up,i、R_down,iIt is respectively unit i to increase most, subtract speed of exerting oneself；

Security of system retrains:

V_LOLP,t≤R_LOLP

In above formula, V_LOLP,tFor the load-loss probability of scheduling slot t, R_LOLPThe operational reliability level reached for expectation；

Wind power constraint:

P_w,t≤F_w,t。

Electrical network wind-powered electricity generation a few days ago receives capability assessment model, uses genetic algorithm based on non-dominant classification to ask it Solve, obtain the Pareto optimal solution set of this assessment models, specifically comprise the following steps that

Step 1, randomly generating the initial chromosome population of genetic algorithm, population scale is 10N；Use the two of a length of N Scale coding represents the item chromosome in chromosome population；Every chromosome gives N platform conventional power unit to be dispatched in a few days Running status, " 0 " represents shutdown, and " 1 " represents start；

Step 2, chromosome is carried out reliability assessment, the chromosome meeting security constraint is carried out economic load dispatching meter Calculate, calculate cost of electricity-generating C of conventional system on this basis_G, and calculate these chromosomes meeting security constraint in scheduling Wind-powered electricity generation amount of abandoning expectation C in a few days_w,t；

Step 3, being layered the chromosome in population by noninferior solution grade, level index is the least, noninferior solution layer grade The highest, during layering, the noninferior solution layer grade of the chromosome being unsatisfactory for security constraint is minimum；Assuming that population can be divided into m layer, to individual For body i, if its residing noninferior solution level is j, then the fitness V of this individuality_fit,iFor:

V_fit,i=10N-j

I=0,1 ..., 10N j=0,1 ..., m

Step 4, calculate the local congestion distance of each individuality in same noninferior solution layer.When local congestion distance calculates, individual It is divided into two classes: be in the individuality at sequence edge and the individuality that sequence is middle；For being in the individuality at sequence edge, its local Crowding distance directly composes a bigger numerical value so that it is obtain selection advantage；For being in the individuality in the middle of sequence, its local Crowding distance is the length sum on two limits of the rectangle constituted for summit with two adjacent individualities；

Step 5, the Noninferior Solution Set in parent population is directly copied to progeny population, as a part for progeny population； Carry out selecting operation to produce other individuality in progeny population, i.e. from parent according to ideal adaptation degree and local congestion distance Randomly select two individualities, if fitness value is different, then choose the individuality that fitness is big, if fitness is identical, then select local The individuality that crowding distance is bigger；Above-mentioned selection operation repeats, until forming progeny population.By certain probability, to filial generation Population carries out intersecting, mutation operation；

Step 6, repeated execution of steps 2 to 5, until algorithm meets the condition of convergence set in advance.

Use parsing probabilistic algorithm that each scheduling slot is abandoned wind-powered electricity generation amount expectation C_w,tWith load-loss probability V_LOLP,tCalculate, Its step is as follows:

Step 1, employing common probability distribution function represent wind power probability nature of random fluctuation near predictive value, its Probability density function is shown below respectively with accumulated probability distribution function:

$f\left(x\right)=\frac{\mathrm{\α}\mathrm{\β}\exp \[-\mathrm{\α}(x-\mathrm{\γ})\]}{{\{1+\exp \[-\mathrm{\α}(x-\mathrm{\γ})\]\}}^{\mathrm{\β}+1}}$

F (x)={ 1+exp [-α (x-γ)] }^-β

Step 2, employing normal distribution N (P_d,t,σ_d,t) represent load probability nature of random fluctuation near predictive value, and Use 7 discrete probabilistic points to normal distribution N (P_d,t,σ_d,t) carry out close approximation, it may be assumed that

Step 3, employing do not consider that double state model that improves of element reparation represents the random fault characteristic of unit, unit i Fault rate f at period t_i,tFor:

f_i,t=1-exp [-λ_i(T_LD+t)]≈λ_i(T_LD+t)

In formula, T_LDCapability evaluation pre-set time is received for wind-powered electricity generation；λ_iFailure rate for unit i.

Step 4, assume that period t has m platform unit to be in open state, in the situation ignoring more than two unit simultaneous faults Under, this period conventional power unit can be as follows by the discrete probabilistic expression formula of generating capacity:

$P\{G_t=G_j\}=p_j,j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in available generating capacity during normal condition, p for unit₀For corresponding probability, can be by following formula Calculate:

$G_0=\underset{j=1}{\overset{m}{\Σ}}{P}_{max,j}$

$p_0=\underset{j=1}{\overset{j=m}{\Π}}(1-f_{j,t})$

G_j(j=1,2 ... available generating capacity when being m) single unit fault, p_jFor corresponding probability.Assuming that fault machine The index of group is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_max,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) available generating capacity when being certain two unit simultaneous faults, p_jFor event The probability occurred.Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{max,k_1}-P_{max,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

Step 5, assume that period t has m platform unit to be in open state, in the situation ignoring more than two unit simultaneous faults Under, the discrete probabilistic expression formula that this total minimum technology of period conventional power unit is exerted oneself is as follows:

$P\{G_{min,t}=G_{min,j}\}=p_{min,j},j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in the total minimum technology of conventional power unit during normal condition for unit to exert oneself, p₀For corresponding general Rate, can be calculated by following formula:

$G_0=\underset{j=1}{\overset{m}{\Σ}}{P}_{min,j}$

$p_0=\underset{j=1}{\overset{i=m}{\Π}}(1-f_{j,t})$

G_j(j=1,2 ... the total minimum technology of conventional power unit when being m) single unit fault is exerted oneself, p_jFor corresponding probability. Assuming that the index of fault unit is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_min,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) it is that the total minimum technology of conventional power unit during certain two unit simultaneous faults goes out Power, p_jThe probability occurred for event.Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{min,k_1}-P_{min,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

Step 6, calculate each scheduling slot abandon wind-powered electricity generation amount expectation C_w,tWith load-loss probability V_LOLP,t, it is shown below:

$V_{LOLP,t}=\underset{l=1}{\overset{7}{\Σ}}\underset{j=0}{\overset{m(m+1)/2}{\Σ}}{p}_j{p}_{d,l}F\left(\frac{P_{d,l}-G_j}{G_{wind}}\right)$

if P_d,l-G_j<0 P_d,l-G_j=0

if P_d,l-G_j>G_wind P_d,l-G_j=G_wind

$C_{w,t}=\underset{l=1}{\overset{7}{\&Sigma;}}\underset{j=0}{\overset{m(m+1)/2}{\&Sigma;}}{p}_j{p}_{d,l}{\&Integral;}_{x_0}^1(x-x_0)f\left(x\right)dx$

$x_0=\frac{P_{d,l}-G_{min,j}}{G_{wind}}$

if x₀<0 x₀=0

if x₀>1 x₀=1.

A kind of electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago receives the wind-powered electricity generation unit of capability assessment method to receive cost to refer to Mark, is characterized in that: receive the wind-powered electricity generation unit of capability assessment model Pareto optimal solution set to receive cost based on electrical network wind-powered electricity generation a few days ago Index C_Apu,i, it is calculated as follows:

$C_{Apu,i}=\frac{C_{G,i}-C_{G,1}}{A_{W,i}-A_{W,1}},i=2,3,...n$

According to the definition that Pareto is optimum, each solve corresponding cost of electricity-generating and meet following relation:

C_G,1<C_G,2<…C_G,n

Wind-powered electricity generation receives the Pareto optimal solution set of capability assessment model to be made up of n different solution, receives electricity by wind-powered electricity generation A_wSize Pareto optimal solution set is ranked up, after sequence, there is following relation:

A_W,1<A_W,2<…A_W,n。

Beneficial effect: compared with prior art, the advantage that the present invention highlights includes: first, receives at electrical network wind-powered electricity generation a few days ago Capability evaluation considers the operating cost of power system, constructs wind-powered electricity generation a few days ago based on multiple-objection optimization and receive capability evaluation Model, model is actual closer to dispatching of power netwoks, and assessment is more fully；Secondly, existing assessment models is only capable of providing single assessment As a result, i.e. the theoretical maximum wind-powered electricity generation of system receives ability, and appraisal procedure disclosed by the invention can provide Pareto optimal solution set, This disaggregation is made up of the cost of a series of assessment results and correspondence；Finally, capability assessment model Pareto is received at wind-powered electricity generation a few days ago Proposing wind-powered electricity generation unit on the basis of optimal solution set and receive the indicator of costs, can weigh electrical network is the cost generation receiving wind-powered electricity generation to pay Valency.

Detailed description of the invention

The present invention is described further with embodiment below in conjunction with the accompanying drawings.

Fig. 1 is the flow chart of the present invention.

Fig. 2 is to abandon wind-powered electricity generation amount and cost of electricity-generating graph of a relation.

Embodiment 1

For receiving ability to be estimated power grid wind in time angle a few days ago, and the Cost Problems receiving wind-powered electricity generation is carried out Analyze, the invention discloses a kind of electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago and receive capability assessment method, and in this assessment Calculating wind-powered electricity generation unit on the basis of the Pareto optimal solution set that method provides and receive cost, its overall procedure is as shown in Figure 1.

Electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago receives two optimization aim following institute respectively of capability assessment model Show:

Optimization aim 1:

Optimization aim 2:

The optimization aim that above-mentioned two formulas are given is respectively wind-powered electricity generation and receives electricity maximum minimum with conventional system cost of electricity-generating.Formula In, A_wElectricity is received for scheduling wind-powered electricity generation in a few days；P_i,tFor unit i at the output of period t；u_iRepresent that unit i is in scheduling day Interior running status, " 0 " represents shutdown, and " 1 " represents start；T is scheduling slot number；F_w,tWind-powered electricity generation theoretical maximum for period t goes out Power, is given by wind power prediction system；C_w,t" abandoning wind " electricity for period t is expected, with P_i,tAnd u_iRelevant；C_GFor conventional system In scheduling cost of electricity-generating in a few days；N is the number of conventional power unit；f_i(P_i,t) it is the unit i fuel cost function at period t, can By the quadratic function matching shown in following formula:

$f_i\left(P_{i,t}\right)=a_i{P}_{i,t}^2+b_i{P}_{i,t}+c_i$

In above formula, a_i、b_iWith c_iIt is respectively the fuel cost coefficient of unit i.

The constraints of assessment models is as follows:

(1) system active balance constraint:

$P_{d,t}-\underset{i=1}{\overset{N}{\&Sigma;}}{u}_i{P}_{i,t}-P_{w,t}=0$

In above formula, P_d,tFor the predicted load of moment t, P_w,tWind-powered electricity generation electricity volume for moment t.

(2) conventional power unit units limits:

P_min,i≤P_i,t≤P_max,i

In above formula, P_max,i、P_min,iBe respectively the maximum of unit i, minimum technology is exerted oneself.

(3) Climing constant:

P_i,t-P_i,t-1≤ΔTR_up,i

P_i,t-1-P_i,t≤ΔTR_down,i

In above formula, R_up,i、R_down,iIt is respectively unit i to increase most, subtract speed of exerting oneself.

(4) security of system constraint:

V_LOLP,t≤R_LOLP

In above formula, V_LOLP,tFor the load-loss probability of scheduling slot t, operation risk can be quantified；R_LOLPThe fortune reached for expectation Row reliability level.

(5) wind power constraint:

P_w,t≤F_w,t

Above-mentioned model is Model for Multi-Objective Optimization, and two optimization aim are conflicted mutually, and the present invention uses based on non-dominant It is solved by the genetic algorithm of classification, can obtain the Pareto optimal solution set of this assessment models, specifically comprise the following steps that

Step 1, randomly generating the initial chromosome population of genetic algorithm, population scale is 10N.Use the two of a length of N Scale coding represents the item chromosome in chromosome population.Every chromosome gives N platform conventional power unit to be dispatched in a few days Running status, " 0 " represents shutdown, and " 1 " represents start；

Step 2, chromosome is carried out reliability assessment, the chromosome meeting security constraint is carried out economic load dispatching meter Calculate, calculate cost of electricity-generating C of conventional system on this basis_G, and calculate these chromosomes meeting security constraint in scheduling Wind-powered electricity generation amount of abandoning expectation C in a few days_w,t；

Step 3, being layered the chromosome in population by noninferior solution grade, level index is the least, noninferior solution layer grade The highest, during layering, the noninferior solution layer grade of the chromosome being unsatisfactory for security constraint is minimum.Assuming that population can be divided into m layer, to individual For body i, if its residing noninferior solution level is j, then the fitness V of this individuality_fit,iFor:

V_fit,i=10N-j

I=0,1 ..., 10N j=0,1 ..., m

Step 4, calculate the local congestion distance of each individuality in same noninferior solution layer.When local congestion distance calculates, individual It is divided into two classes: be in the individuality at sequence edge and the individuality (individual A, B and C in below figure) that sequence is middle.To being in For the individuality at sequence edge, its local congestion distance directly composes a bigger numerical value so that it is obtain selection advantage；To being in For individuality in the middle of sequence (the individual B in figure below), its local congestion distance is (such as in Fig. 2 with two adjacent individualities Body A and C) it is the length sum on two limits of the rectangle that summit is constituted.

Step 5, the Noninferior Solution Set in parent population is directly copied to progeny population, as a part for progeny population. Carry out selecting operation to produce other individuality in progeny population, i.e. from parent according to ideal adaptation degree and local congestion distance Randomly select two individualities, if fitness value is different, then choose the individuality that fitness is big, if fitness is identical, then select local The individuality that crowding distance is bigger.Above-mentioned selection operation repeats, until forming progeny population.According to certain probability to filial generation Population carries out intersecting, mutation operation.

Step 6, repeated execution of steps 2 to 5, until algorithm meets the condition of convergence set in advance.

In above-mentioned electricity based on nondominated sorting genetic algorithms wind-powered electricity generation a few days ago receives capability assessment model solution procedure, this Invention uses parsing probabilistic algorithm to calculate and abandons wind-powered electricity generation amount expectation C_w,tWith load-loss probability V_LOLP,t, specifically comprise the following steps that

Step 1, employing common probability distribution function represent wind power probability nature of random fluctuation near predictive value, its Probability density function is shown below respectively with accumulated probability distribution function:

$f\left(x\right)=\frac{\mathrm{\&alpha;}\mathrm{\&beta;}\exp \&lsqb;-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})\&rsqb;}{{\{1+\exp \&lsqb;-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})\&rsqb;\}}^{\mathrm{\&beta;}+1}}$

F (x)={ 1+exp [-α (x-γ)] }^-β

Step 2, employing normal distribution N (P_d,t,σ_d,t) represent that the probability distribution of load random fluctuation near predictive value is special Property (σ_d,tFor the standard deviation of load random fluctuation, typically within the 5% of predicted load).For avoiding complex convolution Computing, uses 7 discrete probabilistic points to normal distribution N (P_d,t,σ_d,t) approach, it may be assumed that

Step 3, employing do not consider that double state model that improves of element reparation represents the random fault characteristic of unit, unit i Fault rate f at period t_i,tFor:

f_i,t=1-exp [-λ_i(T_LD+t)]≈λ_i(T_LD+t)

In formula, T_LDCapability evaluation pre-set time is received for wind-powered electricity generation；λ_iFailure rate for unit i.

Step 4, assume that period t has m platform unit to be in open state, in the situation ignoring more than two unit simultaneous faults Under, this period conventional power unit can be as follows by the discrete probabilistic expression formula of generating capacity:

$P\{G_t=G_j\}=p_j,j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in available generating capacity during normal condition, p for unit₀For corresponding probability, can be by following formula Calculate:

$G_0=\underset{j=1}{\overset{m}{\&Sigma;}}{P}_{max,j}$

$p_0=\underset{j=1}{\overset{j=m}{\&Pi;}}(1-f_{j,t})$

G_j(j=1,2 ... available generating capacity when being m) single unit fault, p_jFor corresponding probability.Assuming that fault machine The index of group is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_max,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) available generating capacity when being certain two unit simultaneous faults, p_jFor event The probability occurred.Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{max,k_1}-P_{max,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

Step 5, assume that period t has m platform unit to be in open state, in the situation ignoring more than two unit simultaneous faults Under, the discrete probabilistic expression formula that this total minimum technology of period conventional power unit is exerted oneself is as follows:

$P\{G_{min,t}=G_{min,j}\}=p_{min,j},j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in the total minimum technology of conventional power unit during normal condition for unit to exert oneself, p₀For corresponding general Rate, can be calculated by following formula:

$G_0=\underset{j=1}{\overset{m}{\&Sigma;}}{P}_{min,j}$

$p_0=\underset{j=1}{\overset{j=m}{\&Pi;}}(1-f_{j,t})$

G_j(j=1,2 ... the total minimum technology of conventional power unit when being m) single unit fault is exerted oneself, p_jFor corresponding probability. Assuming that the index of fault unit is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_min,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) it is that the total minimum technology of conventional power unit during certain two unit simultaneous faults goes out Power, p_jThe probability occurred for event.Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{min,k_1}-P_{min,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

The minimum technology of step 6, once conventional power unit is exerted oneself with wind power sum more than load, due to conventional power unit peak regulation The restriction of ability, it will cause " abandoning wind ".Based on this, " abandoning wind " electricity of period t expects C_w,tCan be calculated by following formula:

$C_{w,t}=\underset{l=1}{\overset{7}{\&Sigma;}}\underset{j=0}{\overset{m(m+1)/2}{\&Sigma;}}{p}_j{p}_{d,l}{\&Integral;}_{x_0}^1(x-x_0)f\left(x\right)dx$

$x_0=\frac{P_{d,l}-G_{min,j}}{G_{wind}}$

if x₀<0 x₀=0

if x₀>1 x₀=1

And once actual load is more than available generating capacity and wind power sum, due to available generation capacity deficiency, it will Cause sub-load to have a power failure, based on this, the load-loss probability V of period t_LOLP,tCan be calculated by following formula:

$V_{LOLP,t}=\underset{l=1}{\overset{7}{\&Sigma;}}\underset{j=0}{\overset{m(m+1)/2}{\&Sigma;}}{p}_j{p}_{d,l}F\left(\frac{P_{d,l}-G_j}{G_{wind}}\right)$

if P_d,l-G_j<0 P_d,l-G_j=0

if P_d,l-G_j>G_wind P_d,l-G_j=G_wind

Existing wind-powered electricity generation receives appraisal procedure only to provide single assessment result, the theoretical maximum mainly stressing with showing electrical network Wind-powered electricity generation receives ability, and have ignored the Cost Problems that wind-powered electricity generation is received.Therefore, the invention discloses wind-powered electricity generation unit receives cost to refer to Mark, is the cost price receiving wind-powered electricity generation to pay for weighing electrical network.This index can receive capability assessment model to provide at wind-powered electricity generation Calculate on the basis of Pareto optimal solution set, specific as follows.

Assuming that the Pareto optimal solution set that wind-powered electricity generation receives capability assessment model is made up of n different solution, for convenience of description, Electricity A is received by wind-powered electricity generation_wSize Pareto optimal solution set is ranked up, after sequence, there is following relation:

A_W,1<A_W,2<…A_W,n

According to the definition that Pareto is optimum, each solve corresponding cost of electricity-generating and meet following relation:

C_G,1<C_G,2<…C_G,n

In Pareto optimal solution set after sequence, the 1st solves corresponding cost of electricity-generating C_G,1Minimum, but the wind-powered electricity generation of correspondence is received Electricity A_W,1Minimum.This is it is to say, dispatcher the most only considers that cost of electricity-generating is minimum, and have ignored completely wind-powered electricity generation electricity Receiving.Now, wind-powered electricity generation receives electricity A_W,1The wind power can naturally received for system, receiving cost is zero.As for Pareto Other of excellent solution concentration respectively solves, and for receiving more wind-powered electricity generation, system cost of electricity-generating goes out the lifting all showed in various degree.Obviously, The extra conventional system cost of electricity-generating increased may be regarded as wind-powered electricity generation and receives cost, and unit wind-powered electricity generation disclosed by the invention receives cost C_Apu,i Measurement system is the cost price receiving more wind-powered electricity generations to pay, and can be calculated as follows:

$C_{Apu,i}=\frac{C_{G,i}-C_{G,1}}{A_{W,i}-A_{W,1}},i=2,3,...n.$

Claims (1)

1. electrical network based on a multiple-objection optimization wind-powered electricity generation a few days ago receives the wind-powered electricity generation unit of capability assessment method to receive the indicator of costs Determine method, it is characterized in that: receive based on electrical network wind-powered electricity generation a few days ago the wind-powered electricity generation unit of capability assessment model Pareto optimal solution set to connect Receive indicator of costs C_Apu,i, it is calculated as follows:

$C_{Apu,i}=\frac{C_{G,i}-C_{G,1}}{A_{W,i}-A_{W,1}},i=2,3,...n$

According to the definition that Pareto is optimum, each solve corresponding cost of electricity-generating and meet following relation:

C_G,1<C_G,2<…C_G,n

Wind-powered electricity generation receives the Pareto optimal solution set of capability assessment model to be made up of n different solution, receives electricity A by wind-powered electricity generation_wBig Little Pareto optimal solution set is ranked up, after sequence, there is following relation:

A_W,1<A_W,2<…A_W,n；

Described electrical network based on multiple-objection optimization wind-powered electricity generation a few days ago receiving capability assessment method:

Electrical network wind-powered electricity generation a few days ago receives capability assessment model to have wind-powered electricity generation to receive ability maximum and minimum two of conventional system cost of electricity-generating Optimization aim, as follows:

Optimization aim 1:

Optimization aim 2:

In above formula, A_wReceive electricity for scheduling wind-powered electricity generation in a few days, deduct for scheduling wind-powered electricity generation prediction electricity in a few days and abandon the wind-powered electricity generation amount phase Hope；P_i,tFor unit i at the output of period t；u_iRepresenting that unit i is dispatching running status in a few days, " 0 " represents shutdown, " 1 " represents start；T is scheduling slot number；F_w,tWind-powered electricity generation theoretical maximum for period t is exerted oneself, i.e. wind power prediction value；C_w,tFor time " abandoning wind " electricity expectation of section t, with P_i,tAnd u_iRelevant；C_GCost of electricity-generating in a few days is being dispatched for conventional system；N is conventional power unit Number；f_i(P_i,t) it is the unit i fuel cost function at period t, can be by quadratic function matching；

The constraints of electrical network wind-powered electricity generation a few days ago receiving capability assessment model is as follows:

System active balance retrains:

$P_{d,t}-\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{u}_i{P}_{i,t}-P_{w,t}=0$

In above formula, P_d,tFor the predicted load of moment t, P_w,tWind-powered electricity generation electricity volume for moment t；

Conventional power unit units limits:

P_min,i≤P_i,t≤P_max,i

In above formula, P_max,i、P_min,iBe respectively the maximum of unit i, minimum technology is exerted oneself；

Climing constant:

P_i,t-P_i,t-1≤ΔTR_up,i

P_i,t-1-P_i,t≤ΔTR_down,i

In above formula, R_up,i、R_down,iIt is respectively unit i to increase most, subtract speed of exerting oneself；

Security of system retrains:

V_LOLP,t≤R_LOLP

In above formula, V_LOLP,tFor the load-loss probability of scheduling slot t, R_LOLPThe operational reliability level reached for expectation；

Wind power constraint:

P_w,t≤F_w,t；

Electrical network wind-powered electricity generation a few days ago receives capability assessment model, uses genetic algorithm based on non-dominant classification to solve it, asks Go out the Pareto optimal solution set of this assessment models, specifically comprise the following steps that

Step 1, randomly generating the initial chromosome population of genetic algorithm, population scale is 10N；Use the binary system of a length of N Item chromosome in coded representation chromosome population；Every chromosome gives N platform conventional power unit is dispatching operation in a few days State, " 0 " represents shutdown, and " 1 " represents start；

Step 2, chromosome is carried out reliability assessment, the chromosome meeting security constraint is carried out economic load dispatching calculating, Cost of electricity-generating C of conventional system is calculated on the basis of this_G, and calculate these chromosomes meeting security constraint in scheduling in a few days Abandon wind-powered electricity generation amount expectation C_w,t；

Step 3, being layered the chromosome in population by noninferior solution grade, level index is the least, and noninferior solution layer the highest grade, During layering, the noninferior solution layer grade of the chromosome being unsatisfactory for security constraint is minimum；Assuming that population can be divided into m layer, individual i is come Say, if its residing noninferior solution level is j, then the fitness V of this individuality_fit,iFor:

V_fit,i=10N-j

I=0,1 ..., 10N j=0,1 ..., m

Step 4, calculate the local congestion distance of each individuality in same noninferior solution layer；When local congestion distance calculates, individuality is divided into Two classes: be in the individuality at sequence edge and the individuality that sequence is middle；For being in the individuality at sequence edge, its local congestion Distance directly composes a bigger numerical value so that it is obtain selection advantage；For being in the individuality in the middle of sequence, its local congestion Distance is the length sum on two limits of the rectangle constituted for summit with two adjacent individualities；

Step 5, the Noninferior Solution Set in parent population is directly copied to progeny population, as a part for progeny population；According to Ideal adaptation degree and local congestion distance carry out selecting operation to produce other individuality in progeny population, i.e. random from parent Choose two individualities, if fitness value is different, then choose the individuality that fitness is big, if fitness is identical, then select local congestion The individuality that distance is bigger；Above-mentioned selection operation repeats, until forming progeny population；By certain probability, to progeny population Carry out intersecting, mutation operation；

Step 6, repeated execution of steps 2 to 5, until algorithm meets the condition of convergence set in advance；

Use parsing probabilistic algorithm that each scheduling slot is abandoned wind-powered electricity generation amount expectation C_w,tWith load-loss probability V_LOLP,tCalculate, its step Rapid as follows:

Step 1, employing common probability distribution function represent wind power probability nature of random fluctuation near predictive value, its probability Density function is shown below respectively with accumulated probability distribution function:

$f\left(x\right)=\frac{\mathrm{\&alpha;}\mathrm{\&beta;}\exp \&lsqb;-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})\&rsqb;}{{\{1+\exp \&lsqb;-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})\&rsqb;\}}^{\mathrm{\&beta;}+1}}$

F (x)={ 1+exp [-α (x-γ)] }^-β

Step 2, employing normal distribution N (P_d,t,σ_d,t) represent load probability nature of random fluctuation near predictive value, and use 7 discrete probabilistic points are to normal distribution N (P_d,t,σ_d,t) carry out close approximation, it may be assumed that

Step 3, use the double state model that improves not considering element reparation to represent the random fault characteristic of unit, unit i time Fault rate f of section t_i,tFor:

f_i,t=1-exp [-λ_i(T_LD+t)]≈λ_i(T_LD+t)

In formula, T_LDCapability evaluation pre-set time is received for wind-powered electricity generation；λ_iFailure rate for unit i；

Step 4, assume that period t has m platform unit to be in open state, in the case of ignoring more than two unit simultaneous faults, This period conventional power unit can be as follows by the discrete probabilistic expression formula of generating capacity:

$P\{G_t=G_j\}=p_j,j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in available generating capacity during normal condition, p for unit₀For corresponding probability, can be calculated by following formula:

$G_0=\underset{j=1}{\overset{m}{\&Sigma;}}{P}_{max,j}$

$p_0=\underset{j=1}{\overset{j=m}{\&Pi;}}(1-f_{j,t})$

G_j(j=1,2 ... available generating capacity when being m) single unit fault, p_jFor corresponding probability；Assuming that fault unit Index is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_max,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) available generating capacity when being certain two unit simultaneous faults, p_jOccur for event Probability；Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{max,k_1}-P_{max,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

Step 5, assume that period t has m platform unit to be in open state, in the case of ignoring more than two unit simultaneous faults, The discrete probabilistic expression formula that this total minimum technology of period conventional power unit is exerted oneself is as follows:

$P\{G_{\min ,t}=G_{\min ,j}\}=p_{min,j},j=0,1,2...,\frac{m(m+1)}{2}$

In above formula, G₀It is in the total minimum technology of conventional power unit during normal condition for unit to exert oneself, p₀For corresponding probability, can Calculated by following formula:

$G_0=\underset{j=1}{\overset{m}{\&Sigma;}}{P}_{\min ,j}$

$p_0=\underset{j=1}{\overset{j=m}{\&Pi;}}(1-f_{j,t})$

G_j(j=1,2 ... the total minimum technology of conventional power unit when being m) single unit fault is exerted oneself, p_jFor corresponding probability；Assuming that The index of fault unit is k, G_j、p_jIt is respectively as follows:

G_j=G₀-P_min,k

$p_j=p_0\frac{f_{k,t}}{1-f_{k,t}}$

G_j(j=m+1, m+2 ... m (m+1)/2) it is that the total minimum technology of conventional power unit during certain two unit simultaneous faults is exerted oneself, p_j The probability occurred for event；Assuming that the index of fault unit is k₁、k₂, G_j、p_jIt is respectively as follows:

$G_j=G_0-P_{min,k_1}-P_{\min ,k_2}$

$p_j=p_0\frac{f_{k_1,t}{f}_{k_2,t}}{(1-f_{k_1,t})(1-f_{k_2,t})}$

Step 6, calculate each scheduling slot abandon wind-powered electricity generation amount expectation C_w,tWith load-loss probability V_LOLP,t, it is shown below:

$V_{LOLP,t}=\underset{l=1}{\overset{7}{\&Sigma;}}\underset{j=0}{\overset{m(m+1)/2}{\&Sigma;}}{p}_j{p}_{d,l}F\left(\frac{P_{d,l}-G_j}{G_{wind}}\right)$

if P_d,l-G_j<0 P_d,l-G_j=0

if P_d,l-G_j>G_wind P_d,l-G_j=G_wind

$C_{w,t}=\underset{l=1}{\overset{7}{\&Sigma;}}\underset{j=0}{\overset{m(m+1)/2}{\&Sigma;}}{p}_j{p}_{d,l}{\&Integral;}_{x_0}^1(x-x_0)f\left(x\right)dx$

$x_0=\frac{P_{d,l}-G_{min,j}}{G_{wind}}$

if x₀<0 x₀=0

if x₀>1 x₀=1.