CN105790265A - AC power flow constraint-based uncertainty unit commitment model and solving method - Google Patents

Abstract

The invention relates to an AC power flow constraint-based uncertainty unit commitment model and a solving method. With the minimum fuel cost sum for a thermal power unit in 24 hours as an optimization target, a chance-constrained method is adopted to describe uncertainty for wind power output, AC power flow model-based network security constraints are built, and thus, the AC power flow security constraint-based uncertainty unit commitment model is provided. In view of the problem that the model is hard to solve, the invention provides a random constraint ordinal optimization method for successfully realizing quick model solution. In comparison with the traditional DC power flow constraint-based uncertainty unit commitment model, the provided model of the invention can effectively reduce the risk of threshold crossing of power grid voltage after large-scale wind power access, effectiveness of current power generation planning decisions is enhanced, the built model can finely calculate system network losses, and data reference is provided for a dispatching person. In addition, the solved algorithm is greatly enhanced compared with the traditional algorithm.

Description

A kind of uncertain Unit Combination model and method for solving considering that AC power flow retrains

Technical field

The invention belongs to power system and automation research field, especially relate to a kind of uncertain Unit Combination model and method for solving considering that AC power flow retrains.

Background technology

The Unit Combination (SecurityConstraintUnitCommitment, SCUC) considering security constraint is the theoretical basis of power system scheduling decision a few days ago, is also that it formulates the Main Basis of generation schedule a few days ago.Along with exerting oneself in recent years, there is probabilistic wind-powered electricity generation to access on a large scale, and the requirement that becomes more meticulous that dispatching of power netwoks runs improves constantly, it is considered to the Power System Unit Commitment of wind power integration is increasingly becoming the study hotspot of people.

At present, in the conventional electric power system being left out wind power integration, it is common to the SCUC model of employing is to utilize the Network Security Constraints condition based on DC flow model that decision scheme is carried out safety and stability check.It is left out System Reactive Power and factors of voltage due to DC flow model, only calculates circuit effective power flow, thus model solution difficulty can be reduced greatly.And the power system after wind-powered electricity generation is accessed on a large scale, uncertainty is described by conventional scene method, chance constraint, robust optimization etc..But, when the access level of cluster wind-powered electricity generation is significantly high, often reactive-load compensation equipment cannot be unified regulation and control, therefore generated output factor still has bigger gap with line drop compared with conventional power unit, in addition the regulating power of wind-powered electricity generation is generally poor, thus causing that the hypotheses condition simplifying Network Security Constraints model is absent from.If continuing to adopt DC flow model to describe Network Security Constraints condition in SCUC model, then inevitably resulting in decision scheme effectiveness and reduce, voltage limit risk increases.

Additionally, SCUC model itself is a Nonlinear Mixed Integer Programming Problem with Complex Constraints, due under Uncertain environments, AC Ioad flow model itself cannot the solving of fast resolving, typically require by means of numerical algorithm consuming time, therefore, uncertain SCUC model introduces AC power flow constraint and will necessarily face the problem solving difficulty.And existing solve the definitiveness optimization problem with Complex Constraints based on Benders the having important advantages in that of solution throughway decomposed, for considering that wind-powered electricity generation is uncertain and the SCUC model of AC power flow constraint is then it is difficult to ensure that solution efficiency simultaneously.Therefore a kind of uncertain Unit Combination model and method for solving considering that AC power flow retrains of research is needed badly.

Summary of the invention

The present invention proposes a kind of uncertain Unit Combination model and method for solving considering that AC power flow retrains, and this model adopts chance constraint method to describe the uncertainty of wind power output, and based on AC Ioad flow model, network security is checked.Problem for model solution difficulty, it is proposed that a kind of random constraints sequence optimization method, successfully realizes the rapid solving to model.Embodiment based on the IEEE-118 node power system of amendment, demonstrate correctness and the effectiveness of the carried modeling method of the present invention, compare the uncertain SCUC model retrained with tradition based on DC power flow, the carried model of the present invention effectively reduces the risk of power grid out-of-limit after wind-powered electricity generation accesses on a large scale, improve the effectiveness of generation schedule decision-making a few days ago, the computing system via net loss that institute of the present invention established model can also become more meticulous simultaneously, thus providing data refer for dispatcher.

The technical solution adopted in the present invention is:

A kind of uncertain Unit Combination modeling method considering AC power flow security constraint, it is considered to the uncertainty of wind power output, and with AC power flow constraint, it is carried out Security Checking, make total operating cost minimum, its object function is:

$\min F_{Gt}=\underset{t=1}{\overset{24}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]---\left(1\right)$

In formula: F_GtFor the total operating cost of system, the generating set number in system is M, P_GitAnd U_GitRepresent that No. i-th unit is exerted oneself and start and stop state at the meritorious of t period respectively, U_Git=0 represents that generating set is in stopped status, U_Git=1 represents that generating set is in open state.Y_it(P_Git) for the operating cost of generating set, S_it(τ_i) for the start-up and shut-down costs of electromotor；

Wherein, the operating cost of generating set, start-up and shut-down costs concrete mathematical expression form as follows:

$Y_{it}\left(P_{Git}\right)={\mathrm{\α}}_i+{\mathrm{\β}}_i{P}_{Git}+{\mathrm{\γ}}_i{P}_{Git}^2---\left(2\right)$

$S_{it}\left({\mathrm{\τ}}_i\right)=S_{0i}+S_{1i}(1-e^{{\mathrm{\τ}}_i/{\mathrm{\ω}}_i})---\left(3\right)$

In formula: α_i、β_i、γ_iFor unit operation cost parameter.S_0i、S_1i、τ_iFor the downtime of i unit, ω_iFor start-up and shut-down costs parameter；

Decision variable generally meets following conventional constraint condition: system active power balance retrains；Generating set bound constraint of exerting oneself meritorious, idle；Minimum start-stop time retrains；Unit ramp loss；

Except above-mentioned conventional constraint condition, the Network Security Constraints condition that the present invention proposes is as follows:

-P_l,max≤P_l,t≤P_l,max(4)

-Q_l,max≤Q_l,t≤Q_l,max(5)

V_b,min≤V_bt≤V_b,max(6)

In formula: P_l,t、Q_l,tRespectively circuit l is at the active power of t period and reactive power, V_btFor the node b voltage in the t period, P_l,max、Q_l,maxFor maximum meritorious, the reactive capability allowed of circuit l respectively；V_b,min, V_b,maxThe minimax voltage that respectively node b allows；

Variable P in formula (4-6)_l,max、Q_l,max、V_btNeed to obtain through AC power flow equation.Its concrete form is as follows:

$\left\{\begin{array}{c}\mathrm{\Δ}P=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{cos\θ}}_{bc}+B_{bc}{\mathrm{sin\θ}}_{bc})\\ \mathrm{\Δ}Q=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{sin\θ}}_{bc}-B_{bc}{\mathrm{cos\θ}}_{bc})\end{array}\right.---\left(7\right)$

In formula: it is meritorious and idle that △ P, △ Q represent that node injects；B=1,2, LK are nodes；C ∈ b represents the node c being connected with node b；V represents node voltage；G_bc、B_bcIt is real part and the imaginary part of network admittance between node b and node c respectively；

Adopt the wind power output that chance constraint method describes in Optimization of Unit Commitment By Improved uncertain, the positive and negative spinning reserve risk indicator of constructing system respectively.

Q_d≤λ(13)

Q_u≤λ(14)

In formula: Q_d、Q_uRespectively positive and negative spinning reserve risk index；λ is spinning reserve risk threshold value, and the available minimum method of annual total cost of system call department converts after obtaining reliability standard and obtains, and generally takes between 0～10%.

A kind of uncertain Unit Combination model solution method considering AC power flow security constraint, from the feature of the uncertain SCUC model incorporating AC power flow constraint itself, it is proposed that a kind of random constraints sequence optimization method suitable in SCUC model solution.Discrete decision variable U for model_GitWith continuous decision variable P_Git, structure sequence optimizes rough model and accurate model respectively, is carrying out realizing while sequence compares the decoupling of MIXED INTEGER decision variable.

Step 1: Unit Commitment state solution space is screened by structure rough model in advance, constitutes sign set omega according to being uniformly distributed the N number of feasible solution of extraction, and the number of N is closely related with the size of solution space, in solution space less than 10⁸Time, the number of N generally selects 1000.Its concrete rough model is:

(1) power-balance constraint

Unit Commitment state vector need to ensure that the minimax generating capacity of start unit should meet load and the stand-by requirement of system, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gimax}+P_{Wt}\≥P_{Dt}+R_{Dp}---\left(15\right)$

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gi\min }+P_{Wt}\≤P_{Dt}-R_{Dn}---\left(16\right)$

In formula: R_Dp、R_DnAfter respectively considering wind power integration, the standby and negative spinning reserve of the positive rotation needed for system.

(2) unit ramp loss

Among rough model, unit ramping rate constraints is presented as in adjacent time interval unit maximum gradeability and maximum landslide ability sum is more than load variations absolute value, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{up}+P_{Gi\min }(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(17\right)$

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{down}+P_{Gi\min }(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(18\right)$

(3) Network Security Constraints

$\underset{i=1}{\overset{k-1}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gimax}+\underset{i=k+1}{\overset{N}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gi\min }+a_{l,i_k}\≤B_{l,t}---\left(19\right)$

In formula: a_l,t、B_l,tRespectively DC power flow coefficient matrices A, B_tIn l row element；K is integer, meets following constraints:

$\underset{i=1}{\overset{k-1}{\Σ}}(P_{Gi\max }-P_{Gi\min })\≤P_{Dt}-\underset{i=1}{\overset{M}{\Σ}}{P}_{Gi\min }\≤\underset{i=1}{\overset{k}{\Σ}}(P_{Gi\max }-P_{Gi\min })---\left(20\right)$

Step 2: utilize specific picking rule to pick out s solution from sign set further and need to ensure that the probability packet with at least α % is containing the individual enough good solution of k as the selected S that gathers, set S.

The present invention adopts and blind selects method to determine selected set S, and its mathematical model is:

$\begin{array}{c}P\left(\right|G\∩S|\≥k)\\ =\underset{j=k}{\overset{\min (g,s)}{\mathrm{\Σ}}}\underset{i=0}{\overset{s-j}{\mathrm{\Σ}}}\frac{C_g^j{C}_{N-g}^{s-i-j}}{C_N^{s-i}}{C}_s^i{q}^{s-i}{(1-q)}^i\≥\mathrm{\η}\end{array}---\left(21\right)$

In formula: P () is alignment probability, g is the number solved in enough good disaggregation G, s is the number solved in selected set S, k represents have at least k true enough good solution in selected set, η represents the probability comprising k enough good solution in selected set S, usual η take 0.95, q be in solution space real observation to the probability of feasible solution.

In above formula, enough good disaggregation G, selected set S, characterize the relation of set omega as shown in Figure 2.

Step 3: minimum for object function with unit operation totle drilling cost, it is considered to the constraints relevant to unit output, builds for continuous variable P_GitAccurate model, for selected set S in each Unit Commitment state, solve corresponding unit output and operating cost, and selected set sorted further, ask for optimal solution.Constructed sequence optimizes accurate model:

$\left\{\begin{array}{c}\min F_{Gt}=\underset{t=1}{\overset{24}{\Σ}}\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]\\ s.t.P_{Gi\min }\≤P_{Git}\≤P_{Gimax}\\ Q_{Gi\min }\≤Q_{Git}\≤Q_{Gimax}\\ -P_{l\max }\≤P_{lt}\≤P_{lmax}\\ -Q_{l\max }\≤Q_{lt}\≤Q_{l\max }\\ V_{bmix}\≤V_{bt}\≤V_{b\max }\end{array}\right.---\left(22\right)$

The solution throughway of accurate model is, by each Unit Combination state matrix U in selected set_GSubstitute in formula (21) as known parameters, utilize interior point method to solve corresponding generated power and exert oneself matrix P_G, and utilize F_GtSolution is ranked up, asks for optimal solution.

Compared with the conventional method, what the present invention proposed considers uncertain Unit Combination model and the method for solving of AC power flow constraint, has the following advantages and beneficial effect:

(1), the uncertain SCUC model incorporating AC power flow security constraint that the present invention proposes, adopt more precisely AC Ioad flow model that scheduling decision scheme is carried out safety and stability check, effectively prevent the appearance of the out-of-limit situation of system voltage, improve accuracy and the effectiveness of scheduling decision a few days ago.

(2), the computing system via net loss that can become more meticulous of institute of the present invention established model, thus providing data support for dispatcher.

(3), the present invention propose based on random constraints sequence optimize derivation algorithm, what effectively solve institute of the present invention established model solves a difficult problem, compared with traditional Benders decomposition method, has significant advantage in computational efficiency.

Accompanying drawing explanation

Fig. 1 is inventive algorithm general thought block diagram.

Fig. 2 is that the present invention retrains sequence concept of optimization schematic diagram.

Fig. 3 is the output of wind electric field curve chart in the embodiment of the present invention.

Fig. 4 is each moment active power loss in the embodiment of the present invention.

Detailed description of the invention

A kind of uncertain Unit Combination modeling method considering AC power flow security constraint, it is considered to the uncertainty of wind power output, and with AC power flow constraint, it is carried out Security Checking, make total operating cost minimum, its object function is:

$\min F_{Gt}=\underset{t=i}{\overset{24}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]---\left(1\right)$

In formula: F_GtFor the total operating cost of system, the generating set number in system is M, P_GitAnd U_GitRepresent that No. i-th unit is exerted oneself and start and stop state at the meritorious of t period respectively, U_Git=0 represents that generating set is in stopped status, U_Git=1 represents that generating set is in open state.Y_it(P_Git) for the operating cost of generating set, S_it(τ_i) for the start-up and shut-down costs of electromotor；

Wherein, the operating cost of generating set, start-up and shut-down costs concrete mathematical expression form as follows:

$Y_{it}\left(P_{Git}\right)={\mathrm{\α}}_i+{\mathrm{\β}}_i{P}_{Git}+{\mathrm{\γ}}_i{P}_{Git}^2---\left(2\right)$

$S_{it}\left({\mathrm{\τ}}_i\right)=S_{0i}+S_{1i}(1-e^{{\mathrm{\τ}}_i/{\mathrm{\ω}}_i})---\left(3\right)$

In formula: α_i、β_i、γ_iFor unit operation cost parameter.S_0i、S_1i、τ_iFor the downtime of i unit, ω_iFor start-up and shut-down costs parameter；

Decision variable generally meets following conventional constraint condition: system active power balance retrains；Generating set bound constraint of exerting oneself meritorious, idle；Minimum start-stop time retrains；Unit ramp loss；

Except above-mentioned conventional constraint condition, the Network Security Constraints condition that the present invention proposes is as follows:

-P_l,max≤P_l,t≤P_l,max(4)

-Q_l,max≤Q_l,t≤Q_l,max(5)

V_b,min≤V_bt≤V_b,max(6)

In formula: P_l,t、Q_l,tRespectively circuit l is at the active power of t period and reactive power, V_btFor the node b voltage in the t period, P_l,max、Q_l,maxFor maximum meritorious, the reactive capability allowed of circuit l respectively；V_b,min, V_b,maxThe minimax voltage that respectively node b allows；

Variable P in formula (4-6)_l,max、Q_l,max、V_btNeed to obtain through AC power flow equation.Its concrete form is as follows:

$\left\{\begin{array}{c}\mathrm{\Δ}P=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{cos\θ}}_{bc}+B_{bc}{\mathrm{sin\θ}}_{bc})\\ \mathrm{\Δ}Q=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{sin\θ}}_{bc}-B_{bc}{\mathrm{cos\θ}}_{bc})\end{array}\right.---\left(7\right)$

In formula: it is meritorious and idle that △ P, △ Q represent that node injects；B=1,2, LK are nodes；C ∈ b represents the node c being connected with node b；V represents node voltage；G_bc、B_bcIt is real part and the imaginary part of network admittance between node b and node c respectively；

Adopt the wind power output that chance constraint method describes in Optimization of Unit Commitment By Improved uncertain, the positive and negative spinning reserve risk indicator of constructing system respectively.

Q_d≤λ(13)

Q_u≤λ(14)

In formula: Q_d、Q_uRespectively positive and negative spinning reserve risk index；λ is spinning reserve risk threshold value, and the available minimum method of annual total cost of system call department converts after obtaining reliability standard and obtains, and generally takes between 0～10%.

A kind of uncertain Unit Combination model solution method considering AC power flow security constraint, from the feature of the uncertain SCUC model incorporating AC power flow constraint itself, it is proposed that a kind of random constraints sequence optimization method suitable in SCUC model solution.Discrete decision variable U for model_GitWith continuous decision variable P_Git, structure sequence optimizes rough model and accurate model respectively, is carrying out realizing while sequence compares the decoupling of MIXED INTEGER decision variable.

Step 1: Unit Commitment state solution space is screened by structure rough model in advance, constitutes sign set omega according to being uniformly distributed the N number of feasible solution of extraction, and the number of N is closely related with the size of solution space, in solution space less than 10⁸Time, the number of N generally selects 1000.Its concrete rough model is:

(1) power-balance constraint

Unit Commitment state vector need to ensure that the minimax generating capacity of start unit should meet load and the stand-by requirement of system, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gi\max }+P_{Wt}\≥P_{Dt}+R_{Dp}---\left(15\right)$

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gi\min }+P_{Wt}\≤P_{Dt}-R_{Dn}---\left(16\right)$

In formula: R_Dp、R_DnAfter respectively considering wind power integration, the standby and negative spinning reserve of the positive rotation needed for system.

(2) unit ramp loss

Among rough model, unit ramping rate constraints is presented as in adjacent time interval unit maximum gradeability and maximum landslide ability sum is more than load variations absolute value, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{up}+P_{Gi\min }(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(17\right)$

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{down}+P_{Gimin}(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(18\right)$

(3) Network Security Constraints

$\underset{i=1}{\overset{k-1}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gimax}+\underset{i=k+1}{\overset{N}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gi\min }+a_{l,i_k}\≤B_{l,t}---\left(19\right)$

In formula: a_l,t、B_l,tRespectively DC power flow coefficient matrices A, B_tIn l row element；K is integer, meets following constraints:

$\underset{i=1}{\overset{k-1}{\Σ}}(P_{Gi\max }-P_{Gi\min })\≤P_{Dt}-\underset{i=1}{\overset{M}{\Σ}}{P}_{Gi\min }\≤\underset{i=1}{\overset{k}{\Σ}}{P}_{Gi\max }-P_{Gi\min })---\left(20\right)$

Step 2: utilize specific picking rule to pick out s solution from sign set further and need to ensure that the probability packet with at least α % is containing the individual enough good solution of k as the selected S that gathers, set S.

The present invention adopts and blind selects method to determine selected set S, and its mathematical model is:

$\begin{array}{c}P\left(\right|G\∩S|\≥k)\\ =\underset{j=k}{\overset{\min (g,s)}{\Σ}}\underset{i=0}{\overset{s-j}{\Σ}}\frac{C_g^j{C}_{N-g}^{s-i-j}}{C_N^{s-i}}{C}_s^j{q}^{s-i}{(1-q)}^i\≥\mathrm{\η}\end{array}---\left(21\right)$

In formula: P () is alignment probability, g is the number solved in enough good disaggregation G, s is the number solved in selected set S, k represents have at least k true enough good solution in selected set, η represents the probability comprising k enough good solution in selected set S, usual η take 0.95, q be in solution space real observation to the probability of feasible solution.

In above formula, enough good disaggregation G, selected set S, characterize the relation of set omega as shown in Figure 2.

Step 3: minimum for object function with unit operation totle drilling cost, it is considered to the constraints relevant to unit output, builds for continuous variable P_GitAccurate model, for selected set S in each Unit Commitment state, solve corresponding unit output and operating cost, and selected set sorted further, ask for optimal solution.Constructed sequence optimizes accurate model:

$\left\{\begin{array}{c}\min F_{Gt}=\underset{t=1}{\overset{24}{\Σ}}\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]\\ s.t.P_{Gi\min }\≤P_{Git}\≤P_{Gimax}\\ Q_{Gi\min }\≤Q_{Git}\≤Q_{Gimax}\\ -P_{l\max }\≤P_{lt}\≤P_{lmax}\\ -Q_{l\max }\≤Q_{lt}\≤Q_{l\max }\\ V_{bmix}\≤V_{bt}\≤V_{b\max }\end{array}\right.---\left(22\right)$

The solution throughway of accurate model is, by each Unit Combination state matrix U in selected set_GSubstitute in formula (21) as known parameters, utilize interior point method to solve corresponding generated power and exert oneself matrix P_G, and utilize F_GtSolution is ranked up, asks for optimal solution.

Embodiment:

The present invention is for the IEEE-118 node power system of amendment, and this system comprises 54 fired power generating unit, 3 wind power plants, 91 load point, wherein wind energy turbine set lays respectively at node 14,54, on 95, its rated power respectively 100MW, 200MW, 250MW, gain merit and exert oneself as shown in Figure 3.In system, the stand-by requirement of conventional power unit positive rotation is the 8% of system peak load, and negative rotation turns 2% that stand-by requirement is system minimum load, and spinning reserve risk indicator is 0.01.Correlation computations all completes on Intel's Duo i3-3240 processor 3.40GHz, 4G memory computer, adopts Matlab8.0 and Cplex12.5 to be programmed solving to example.

1), model solution:

Spinning reserve needed for system that constrained procedure of improving the occasion is asked for and each wind energy turbine set is as shown in table 1.

The spinning reserve (MW) of each wind energy turbine set of table 1

Utilize the SCUC model incorporating AC power flow constraint and corresponding method for solving that the present invention builds, solve 24 hours interior plan for start-up and shut-downs of 54 generating sets in IEEE-118 node simulation example as shown in table 2.

The lower start and stop scheme of table 2 AC power flow constraint

The Transmission Loss of all right computing system in detail of institute of the present invention established model, thus providing the data support more become more meticulous for system coordinator.Utilize the Transmission Loss of the present invention built AC-SCUC 24 periods of model computing system as shown in Figure 4.

As shown in Figure 4, the variation tendency of system Transmission Loss and load variations trend are basically identical, maximum at 20, for 176.11MW, minimum at 4, for 29.71MW.

2), relative analysis:

(1), DC power flow security constraint and AC power flow security constraint relative analysis:

For verifying the advance of the put forward model of the present invention, utilize the SCUC model based on DC power flow Network Security Constraints that example of the present invention is solved, then utilize AC power flow constraint that simulation result is checked, its operating cost of relative analysis, the out-of-limit situation of power transmission network is as shown in table 3, represents with DC-SCUC and AC-SCUC respectively based on the SCUC model of DC power flow Network Security Constraints and institute of the present invention established model in table.

The constraint of table 3 direct current retrains comparative result with exchanging

As shown in Table 3, for system operation cost, the scheme that AC-SCUC model is formulated adds 57950 $ than DC-SCUC model, and in the scheduling slot of 24 hours, the circuit that two schemes cause out-of-limit number of gaining merit is 0, but the scheduling scheme formulated of DC-SCUC model occurs in that the situation of 105 minor node voltage out-of-limits.

It is shown that no matter be DC-SCUC model or the AC-SCUC model of present invention proposition, transmission line of electricity can be prevented effectively from and occur that effective power flow is out-of-limit.Because it is contemplated that more accurate voltage, the constraints such as idle, compared to DC-SCUC model, although the scheduling scheme cost that institute of the present invention established model is formulated increases to some extent, but it effectively prevent node voltage, and out-of-limit situation occurs, thus improving the effectiveness of scheduling decision scheme a few days ago, it is ensured that security of system reliability service.

Based on, under the scheduling scheme that DC-SCUC model is formulated, occurring that the node serial number of voltage out-of-limit and maximum more voltage limit are as shown in table 4.

Out-of-limit voltage inscribed by table 4 time each

As shown in Table 4, owing to DC-SCUC model fails node voltage constraint is checked, all there is low voltage condition in each moment in system.Wherein all there is voltage out-of-limit 24 periods in 29,31,53 3 branch roads, are the transmission cross-sections of whole system most fragile.

(2), wind energy turbine set is to Operation of Electric Systems impact analysis:

For analyzing the wind energy turbine set impact on Operation of Electric Systems safety, adopting DC-SCUC model to be optimized calculating to containing wind-powered electricity generation with the power system without wind-powered electricity generation respectively, and adopt AC Ioad flow model that simulation result is carried out Security Checking, its result is as shown in table 5.

Operation of Electric Systems is affected by table 5 wind energy turbine set

As shown in Table 5, after wind power integration, although can reduce the operating cost of system, it can also cause the out-of-limit nodes of system voltage increases, its reason is wind energy turbine set meeting absorbing reactive power to a certain extent in running.As can be seen here, under the situation that wind-powered electricity generation accesses on a large scale, tradition DC-SCUC model would become hard to ensure the effectiveness of scheduling decision a few days ago.

(3), the efficiency analysis that random constraints sequence optimizes:

For verifying the effectiveness of the carried derivation algorithm of the present invention, AC-SCUC model is solved by the COO algorithm being utilized respectively Benders decomposition method the most frequently used at present and present invention proposition, and it calculates time and optimizing result is as shown in table 6.

Table 6 derivation algorithm compares

As shown in Table 6, the COO algorithm that the present invention proposes with Benders decomposition method relatively, but has significant advantage in optimizing ability in computational efficiency.

Claims (3)

1. the uncertain Unit Combination modeling method considering AC power flow security constraint, it is characterised in that considering the uncertainty of wind power output, and with AC power flow constraint, it is carried out Security Checking, its object function is:

${\mathrm{minF}}_{Gt}=\underset{t=1}{\overset{24}{\Σ}}\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]---\left(1\right)$

In formula: F_GtFor the total operating cost of system, the generating set number in system is M, P_GitAnd U_GitRepresent that No. i-th unit is exerted oneself and start and stop state at the meritorious of t period respectively, U_Git=0 represents that generating set is in stopped status, U_Git=1 represents that generating set is in open state.Y_it(P_Git) for the operating cost of generating set, S_it(τ_i) for the start-up and shut-down costs of electromotor；

Wherein, the operating cost of generating set, start-up and shut-down costs concrete mathematical expression form as follows:

$Y_{it}\left(P_{Git}\right)={\mathrm{\α}}_i+{\mathrm{\β}}_i{P}_{Git}+{\mathrm{\γ}}_i{P}_{Git}^2---\left(2\right)$

$S_{it}\left({\mathrm{\τ}}_i\right)=S_{0i}+S_{1i}(1-e^{{\mathrm{\τ}}_i/{\mathrm{\ω}}_i})---\left(3\right)$

In formula: α_i、β_i、γ_iFor unit operation cost parameter, S_0i、S_1i、τ_iFor the downtime of i unit, ω_iFor start-up and shut-down costs parameter；

Decision variable generally meets following conventional constraint condition: system active power balance retrains；Generating set bound constraint of exerting oneself meritorious, idle；Minimum start-stop time retrains；Unit ramp loss；

Except above-mentioned conventional constraint condition, the Network Security Constraints condition that the method proposes is as follows:

-P_l,max≤P_l,t≤P_l,max(4)

-Q_l,max≤Q_l,t≤Q_l,max(5)

V_b,min≤V_bt≤V_b,max(6)

In formula: P_l,t、Q_l,tRespectively circuit l is at the active power of t period and reactive power, V_btFor the node b voltage in the t period, P_l,max、Q_l,maxFor maximum meritorious, the reactive capability allowed of circuit l respectively；V_b,min, V_b,maxThe minimax voltage that respectively node b allows；

Variable P in formula (4-6)_l,max、Q_l,max、V_btNeeding to obtain through AC power flow equation, its concrete form is as follows:

$\left\{\begin{array}{c}\mathrm{\Δ}P=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{cos\θ}}_{bc}+B_{bc}{\mathrm{sin\θ}}_{bc})\\ \mathrm{\Δ}Q=V_b\underset{c\∈b}{\Σ}{V}_c(G_{bc}{\mathrm{sin\θ}}_{bc}-B_{bc}{\mathrm{cos\θ}}_{bc})\end{array}\right.---\left(7\right)$

In formula: it is meritorious and idle that △ P, △ Q represent that node injects；B=1,2, LK are nodes；C ∈ b represents the node c being connected with node b；V represents node voltage；G_bc、B_bcIt is real part and the imaginary part of network admittance between node b and node c respectively；

Adopt the wind power output that chance constraint method describes in Optimization of Unit Commitment By Improved uncertain, the positive and negative spinning reserve risk indicator of constructing system respectively:

Q_d≤λ(13)

Q_u≤λ(14)

In formula: Q_d、Q_uRespectively positive and negative spinning reserve risk index；λ is spinning reserve risk threshold value, and the available minimum method of annual total cost of system call department converts after obtaining reliability standard and obtains, and generally takes between 0～10%.

2. the uncertain Unit Combination model solution method considering AC power flow security constraint, it is characterized in that, feature from the uncertain SCUC model itself incorporating AC power flow constraint, it is proposed that a kind of random constraints sequence optimization method suitable in SCUC model solution；Discrete decision variable U for model_GitWith continuous decision variable P_Git, structure sequence optimizes rough model and accurate model respectively, is carrying out realizing while sequence compares the decoupling of MIXED INTEGER decision variable.

3. a kind of uncertain Unit Combination model solution method considering AC power flow security constraint according to claim 2, it is characterised in that comprise the following steps:

Step 1: Unit Commitment state solution space is screened by structure rough model in advance, constitutes sign set omega according to being uniformly distributed the N number of feasible solution of extraction, and the number of N is closely related with the size of solution space, in solution space less than 10⁸Time, the number of N generally selects 1000, and its concrete rough model is:

(1), power-balance constraint:

Unit Commitment state vector need to ensure that the minimax generating capacity of start unit should meet load and the stand-by requirement of system, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gimax}+P_{Wt}\≥P_{Dt}+R_{Dp}---\left(15\right)$

$\underset{i=1}{\overset{M}{\Σ}}{U}_{Git}{P}_{Gi\min }+P_{Wt}\≤P_{Dt}-R_{Dn}---\left(16\right)$

In formula: R_Dp、R_DnAfter respectively considering wind power integration, the standby and negative spinning reserve of the positive rotation needed for system；

(2), unit ramp loss:

Among rough model, unit ramping rate constraints is presented as in adjacent time interval unit maximum gradeability and maximum landslide ability sum is more than load variations absolute value, it may be assumed that

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{up}+P_{Gi\min }(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(17\right)$

$\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{\mathrm{\ΔP}}_{Gi}^{down}+P_{Gi\min }(U_{Git}-U_{Git-1})\]\≥|P_{Dt}-P_{Dt-1}|---\left(18\right)$

(3) Network Security Constraints

$\underset{i=1}{\overset{k-1}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gi\max }+\underset{i=k+1}{\overset{N}{\Σ}}(a_{l,i_n}-a_{l,i_k})P_{Gi\min }+a_{l,i_k}\≤B_{l,t}---\left(19\right)$

In formula: a_l,t、B_l,tRespectively DC power flow coefficient matrices A, B_tIn l row element；K is integer, meets following constraints:

$\underset{i=1}{\overset{k-1}{\Σ}}(P_{Gi\max }-P_{Gi\min })\≤P_{Dt}-\underset{i=1}{\overset{M}{\Σ}}{P}_{Gi\min }\≤\underset{i=1}{\overset{k}{\Σ}}(P_{Gi\max }-P_{Gi\min })---\left(20\right)$

Step 2: utilize specific picking rule to pick out s solution from sign set further and need to ensure that the probability packet with at least α % is containing the individual enough good solution of k as the selected S that gathers, set S；

The present invention adopts and blind selects method to determine selected set S, and its mathematical model is:

$\begin{array}{c}P\left(\right|G\∩S|\≥k)\\ =\underset{j=k}{\overset{\min (g,s)}{\Σ}}\underset{i=0}{\overset{s-j}{\Σ}}\frac{C_g^j{C}_{N-g}^{s-i-j}}{C_N^{s-i}}{C}_s^i{q}^{s-i}{(1-q)}^i\≥\mathrm{\η}\end{array}---\left(21\right)$

In formula: P () is alignment probability, g is the number solved in enough good disaggregation G, s is the number solved in selected set S, k represents have at least k true enough good solution in selected set, η represents the probability comprising k enough good solution in selected set S, usual η take 0.95, q be in solution space real observation to the probability of feasible solution；

Step 3: minimum for object function with unit operation totle drilling cost, it is considered to the constraints relevant to unit output, builds for continuous variable P_GitAccurate model, for selected set S in each Unit Commitment state, solve corresponding unit output and operating cost, and selected set sorted further, ask for optimal solution.Constructed sequence optimizes accurate model:

$\left\{\begin{array}{c}{\mathrm{minF}}_{Gt}=\underset{t=1}{\overset{24}{\Σ}}\underset{i=1}{\overset{M}{\Σ}}\[U_{Git}{Y}_{it}\left(P_{Git}\right)+U_{Git}(1-U_{Git-1})S_{it}\left({\mathrm{\τ}}_i\right)\]\\ s.t.P_{Gi\min }\≤P_{Git}\≤P_{Gi\max }\\ Q_{Gi\min }\≤Q_{Git}\≤Q_{Gi\max }\\ -P_{l\max }\≤P_{lt}\≤P_{l\max }\\ -Q_{l\max }\≤Q_{lt}\≤Q_{l\max }\\ V_{bmix}\≤V_{bt}\≤V_{b\max }\end{array}\right.---\left(22\right)$

The solution throughway of accurate model is, by each Unit Combination state matrix U in selected set_GSubstitute in formula (21) as known parameters, utilize interior point method to solve corresponding generated power and exert oneself matrix P_G, and utilize F_GtSolution is ranked up, asks for optimal solution.