CN102982388B - Day-ahead power system economical dispatching method - Google Patents

Abstract

The invention provides a day-ahead power system economical dispatching method which includes the following steps: step S1, modeling a linear programming mathematical model of the day-ahead power system economical dispatching method; step S2, building constraint conditions which enables the power system to satisfy a load balancing constraint, a machine group operating constraint and a contract power constraint, and forming a constraint matrix A by ordering the constraint conditions according to an attribute of variable relating to each constraint condition; step S3, using a predictor-corrector interior point method to get a result which can satisfy the constraint matrix A of an objective function, wherein the objective function of the mathematical model is that T is session number in the period of system dispatching, M is set number of system generators, pi (t) is active power of generator set i in a t period, and Ci [pi (t)] is an operating cost of the generator set i in the t period. The day-ahead power system economical dispatching method uses the predictor-corrector interior point method to solve and the predictor-corrector interior point method is low in an initial point requirement, good in astringency and rapid in astringency speed.

Description

A kind of method of Economic Dispatch a few days ago

Technical field

The present invention relates to field of power, be specifically related to a kind of method of Economic Dispatch a few days ago.

Background technology

Along with the propelling that intelligent grid is built, the requirement that power network resources is distributed rationally constantly strengthens, and this has higher requirement to management and running and operation plan.Economic load dispatching method is on day Unit Commitment generation schedule basis that Unit Combination is determined a few days ago a few days ago, take system minimum as target, according to ultra-short term prediction load, adjust in real time unit output, meet balancing the load constraint, unit operation constraint and contract Constraint, to realize energy saving economy scheduling.Economic load dispatching had obtained successful Application in Real-Time Scheduling field abroad a few days ago in recent years, almost can complete all analytical works of electrical production and scheduling with Unit Combination related tool a few days ago, and had also become at home research and application focus.

The optimization solution fado of the linear model of economic load dispatching a few days ago generally adopting both at home and abroad at present calculates by business mathematics software, and business mathematics software is common software, and the singularity for economic load dispatching model does not improve counting yield.

Summary of the invention

The present invention relates to a kind of method of Economic Dispatch a few days ago, described method comprises: step S1, and foundation makes the linear programming model of Economic Dispatch a few days ago of the operating cost F minimum of described electric system, and the objective function of described mathematical model is

t is the time hop count during system call, and M is system genset number, p _i(t) be genset i in the active power of t period, C _i[p _i(t)] be the operating cost of genset i in the t period; Step S2, sets up constraint condition and makes described electric system meet balancing the load constraint, unit operation constraint and contract Constraint, and the attribute of the variable relating to according to constraint condition described in each sorts to described constraint condition, forms constraint matrix A; Step S3, is used prediction-correction interior point to solve the result of the described Economic Dispatch linear programming model that is met described constraint matrix A.

In the first preferred embodiment provided by the invention: the described constraint condition that makes described electric system meet balancing the load constraint, unit operation constraint and the foundation of contract Constraint in described step S2 comprises:

Unit cost constraint, system loading Constraints of Equilibrium, the constraint of unit output bound, unit ramp loss and contract Constraint.

In the second preferred embodiment provided by the invention: described C _i[p _i(t)] be quadratic function: C _i[p _i(t)]=a _ip _i ²(t)+b _ip _i(t)+c _i, wherein, a _i, b _i, c _ifor unknown quadratic function coefficient;

Operating cost C by described genset i in the t period _i[p _i(t) quadratic function] is divided into l section and is write as l inequality linear restriction, obtains described unit cost constraint: C _i[p _i(t)]>=k _i,rp _i(t)+m _i,r, k _i,r, m _i,rbe the coefficient of r section linear function, r=1 ..., l;

Described system loading Constraints of Equilibrium is:

p _d(t) be the total load of system t period;

Described unit output bound is constrained to:

be respectively the upper and lower limit of genset i output power;

Described unit ramp loss is :-Δ p _i≤ p _i(t)-p _i(t-1)≤Δ p _i, Δ p _ifor per period of genset i can be added and subtracted the maximal value of power;

Described contract Constraint is:

be respectively the upper and lower limit of mono-day contract electric weight of genset i.

In the 3rd preferred embodiment provided by the invention: the variable that described unit output bound constraint and described unit cost constraint unit relate to comprises the power and variable of single described genset;

The variable that described unit ramp loss relates to comprises the power and variable of the described genset of two periods;

The variable that described contract Constraint relates to comprises the power and variable of all periods described in single genset;

The variable that described system loading Constraints of Equilibrium relates to comprises the power and variable of all described genset in the period;

The method that the attribute of the variable relating to according to constraint condition described in each in described step S2 sorts to described constraint condition is:

Described unit output bound constraint and described unit cost constraint unit are combined to the independent constrain set Q that regards single described genset in the period as, according to the period, be arranged in order the constrain set Q of each genset, then by the period, be arranged in order the described unit ramp loss of each genset, finally arrange contract Constraint and system loading Constraints of Equilibrium;

Described each constraint condition described constraint matrix A of rear formation that sorted.

In the 4th preferred embodiment provided by the invention: the canonical form of the linear programming model of Economic Dispatch a few days ago of setting up in described step S1 is:

minF＝cx

$s.t.\left\{\begin{array}{c}\mathrm{Ax}=b\\ x\≥0\end{array}\right.$

Wherein, x is variable, and c, b are vector, and A is constraint matrix;

In described step S3, using prediction-correction interior point to solve to obtain the result of described Economic Dispatch linear programming model is to solve according to the dual problem of described Economic Dispatch linear programming model, and the dual problem of the canonical form of described Economic Dispatch linear programming model is:

max?w ^Tb

$s.t.\left\{\begin{array}{c}{w}^{T}A-w_s=c\\ w_s\≥0\end{array}\right.$

Wherein, w, w _svariable for dual problem.

In the 5th preferred embodiment provided by the invention: the concrete grammar that uses prediction-correction interior point to solve the result that obtains described Economic Dispatch linear programming model in described step S3 is:

Step S301, from a point set out, the Newton direction of correction is:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\begin{array}{}\end{array}\left[\begin{array}{c}\mathrm{\Δx}\\ \mathrm{\Δw}\\ {\mathrm{\Δw}}_s\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ \mathrm{\μe}-\mathrm{\ΔX\ΔZe}\end{array}\right]$

Wherein, Δ x, Δ w, Δ w _sfor iteration correction, I is unit matrix, and e is vector of unit length, and Z, X, Δ Z, Δ X are diagonal matrix, and diagonal element is by w _s, x, Δ w _s, Δ x forms, μ is barrier parameter;

Step S302, the compensate for clearance while calculating the k time iteration

and judge the condition of convergence, if compensate for clearance gap ^kbe less than convergence precision ε, export optimum solution, finish to calculate; Otherwise continue, calculate affine compensate for clearance gap _aff;

Calculate described affine compensate for clearance gap _affcomprise:

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\Δx}}_{\mathrm{aff}}\\ {\mathrm{\Δx}}_{\mathrm{aff}}\\ {\mathrm{\Δx}}_{\mathrm{aff}}\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]$

Calculate affine step delta x _aff, Δ w _aff, Δ w _saff, then according to equation:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}_{\mathrm{aff}}^j\right)}\right|{\mathrm{\&Delta;x}}_{\mathrm{aff}}^j<0\}\\ a_2=0.995\min \left\{\frac{{-w}_s^j}{{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j}\right|{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

Calculate affine iteration step length α ₁, α ₂, wherein, num is the element number of variable x;

According to formula: ${\mathrm{gap}}_{\mathrm{aff}}={(w_s^k+{\mathrm{\&alpha;}}_2\mathrm{\&Delta;}{w}_{\mathrm{saff}})}^T(x^k+{\mathrm{\&alpha;}}_1{\mathrm{\&Delta;x}}_{\mathrm{aff}})$ Calculate described affine compensate for clearance gap _aff;

Step S303, dyscalculia parameter μ:

$\mathrm{\&mu;}=\frac{{\mathrm{gap}}_{\mathrm{aff}}}{2\mathrm{num}}\min \{{\left(\frac{{\mathrm{gap}}_{\mathrm{aff}}}{{\mathrm{gap}}^k}\right)}^2,0.2\}$

Step S304, with described barrier parameter μ, the described affine step delta w according to trying to achieve _saffwith Δ x _affobtain the Newton direction of described correction, according to the Newton direction of described correction, calculate iteration step length

obtain k+1 point;

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right]=\left[\begin{array}{c}0\\ 0\\ \mathrm{\&mu;e}-{\mathrm{\&Delta;X}}_{\mathrm{aff}}{\mathrm{\&Delta;Z}}_{\mathrm{aff}}e\end{array}\right]$

Wherein, Δ Z _aff, Δ X _afffor diagonal matrix, diagonal element is by Δ w _saff, Δ x _affform.

The Newton direction that obtains described correction is:

$\left[\begin{array}{c}\mathrm{\&Delta;x}\\ \mathrm{\&Delta;w}\\ {\mathrm{\&Delta;w}}_s\end{array}\right]=\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{saff}}\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right].$

According to equation:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1^k=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}^j\right)}\right|{\mathrm{\&Delta;x}}^j<0\}\\ {\mathrm{\&alpha;}}_2^k=0.995\min \left\{\frac{{-w}_s^j}{\left({\mathrm{\&Delta;w}}_s^j\right)}\right|{\mathrm{\&Delta;w}}_s^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

Calculate iteration step length k+1 point is:

$\left[\begin{array}{c}{x}^{k+1}\\ w^{k+1}\\ {w_s}^{k+1}\end{array}\right]=\left[\begin{array}{c}{x}^k\\ w^k\\ {w_s}^k\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&alpha;}}_1^k\mathrm{\&Delta;x}\\ {\mathrm{\&alpha;}}_2^k\mathrm{\&Delta;w}\\ {\mathrm{\&alpha;}}_2^k{\mathrm{\&Delta;w}}_s\end{array}\right].$

In the 6th preferred embodiment provided by the invention: convergence precision ε described in described step S302 is 10 ^-6.

The beneficial effect of the method for a kind of Economic Dispatch a few days ago provided by the invention comprises:

1, the method for a kind of Economic Dispatch a few days ago provided by the invention, the time period that the power of the assembling unit variable relating to according to each constraint condition relates to, constraint condition is sorted according to the period, form constraint matrix, effectively reduce newly-increased first quantity that injects, improve the efficiency of interior point method.

2, this few days ago the method for Economic Dispatch use prediction-correction interior point to solve, prediction-correction interior point to initial point require low, convergence is good, fast convergence rate, to solve electric system economy a few days ago to adjust the efficient accurate method of reading, be particularly suitable for the linear programming problem of Solving Large Scale Sparse matrix, so be well suited for economic load dispatching a few days ago, calculate.

Accompanying drawing explanation

Be illustrated in figure 1 the process flow diagram of the method for a kind of Economic Dispatch a few days ago provided by the invention.

Embodiment

The invention provides a kind of method of Economic Dispatch a few days ago, as shown in Figure 1, as shown in Figure 1, the method comprises the process flow diagram of the method:

Step S1, foundation makes the linear programming model of Economic Dispatch a few days ago of the operating cost F minimum of this electric system, and the objective function of this mathematical model is

Wherein, T is the time hop count during this electric power system dispatching, and M is system genset number, p _i(t) be genset i in the active power of t period, C _i[p _i(t)] be the operating cost of genset i in the t period.

Step S2, sets up constraint condition and makes electric system meet balancing the load constraint, genset operation constraint and contract Constraint, and the attribute of the variable relating to according to each constraint condition sorts to constraint condition, forms constraint matrix A.

Step S3, is used prediction-correction interior point to solve the result of the objective function that is met constraint matrix A.

Embodiment mono-:

The embodiment mono-of the method for a kind of Economic Dispatch a few days ago provided by the invention is a kind of embodiment of method of Economic Dispatch a few days ago.

Concrete, the constraint condition that makes electric system meet balancing the load constraint, genset operation constraint and the foundation of contract Constraint in step S2 comprises unit cost constraint, system loading Constraints of Equilibrium, the constraint of unit output bound, unit ramp loss and contract Constraint.

Genset i is at the operating cost C of t period _i[p _i(t)] be generally quadratic function: C _i[p _i(t)]=a _ip _i ²(t)+b _ip _i(t)+c _i, wherein, a _i, b _i, c _ifor unknown quadratic function coefficient.

Operating cost C to genset i in the t period _i[p _i(t) quadratic function] carries out piece-wise linearization, when segments is abundant, just the quafric curve of former quadratic function can be approached with arbitrary accuracy, while being divided into l section, this quafric curve can be write as l inequality linear restriction, and this linear restriction is unit cost constraint: C _i[p _i(t)]>=k _i,rp _i(t)+m _i,r, wherein, k _i,r, m _i,rbe r(r=1 ..., the l) coefficient of section linear function.

System loading Constraints of Equilibrium is:

p _d(t) be the total load of system t period, current domestic p _d(t) the Generation Side load prediction for comprising network loss.

Unit output bound is constrained to: be respectively the upper and lower limit of genset i output power.

Unit ramp loss is :-Δ p _i≤ p _i(t)-p _i(t-1)≤Δ p _i, Δ p _ifor per period of genset i can be added and subtracted the maximal value of power.

Contract Constraint is: be respectively the upper and lower limit of mono-day contract electric weight of genset i.

The canonical form of the linear programming model of Economic Dispatch a few days ago of setting up in step S1 is:

minF＝cx

$s.t.\left\{\begin{array}{c}\mathrm{Ax}=b\\ x\&GreaterEqual;0\end{array}\right.$

Wherein, x is variable, and c, b are vector, and A is constraint matrix.

Iterative Matrix shape when the canonical form of this mathematical model is calculated is as ARA ^t, wherein R is diagonal matrix, can find out, different to the sequence of the constraint condition comprising in constraint matrix A, to ARA ^tthe injection unit quantity producing in the time of factorization has appreciable impact, and for reducing as far as possible newly-increased first quantity that injects, the attribute of the variable relating to according to each constraint condition sorts to described constraint condition, and method is as follows:

The variable that the constraint of unit output bound and unit cost constraint relate to comprises the power and variable of single genset, without room and time coupling, therefore can combine the independent constrain set Q that regards single unit in the period as.

The variable that unit ramp loss relates to comprises the power and variable of the genset of two periods, the variable that contract Constraint relates to comprises the power and variable of single all periods of genset, and the variable that system loading Constraints of Equilibrium relates to comprises the power and variable of all genset in the period.Therefore be first arranged in order the constrain set Q of each genset by the period, then by the period, arrange successively the unit ramp loss of each genset, finally arrange contract Constraint and system loading Constraints of Equilibrium.After each constraint condition has sorted, formed constraint matrix A.

In step S3, use prediction-correction interior point to solve the result that obtains Economic Dispatch linear programming model, concrete, be to solve according to the dual problem of Economic Dispatch linear programming model, the dual problem of the canonical form of this Economic Dispatch linear programming model is:

max?w ^Tb

$s.t.\left\{\begin{array}{c}{w}^{T}A-w_s=c\\ w_s\&GreaterEqual;0\end{array}\right.$

Wherein, w, w _svariable for dual problem.

Adopt prediction-proofread and correct former-antithesis path trace interior point method solves above problem, the first step is that forecast obtains affine direction, second step is to proofread and correct to obtain the Newton direction revised.

Step S301, from a point

set out, the Newton direction of correction is:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\begin{array}{}\end{array}\left[\begin{array}{c}\mathrm{\&Delta;x}\\ \mathrm{\&Delta;w}\\ {\mathrm{\&Delta;w}}_s\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ \mathrm{\&mu;e}-\mathrm{\&Delta;X\&Delta;Ze}\end{array}\right]$

In formula, Δ x, Δ w, Δ w _sfor iteration correction, I is unit matrix, and e is vector of unit length, and Z, X, Δ Z, Δ X are diagonal matrix, and diagonal element is by w _s, x, Δ w _s, Δ x forms, μ is barrier parameter.

In the first step, forecast that the process that obtains affine direction comprises: step S302, the compensate for clearance while calculating the k time iteration

and judge the condition of convergence, if compensate for clearance gap ^kbe less than convergence precision ε, export optimum solution, finish to calculate; Otherwise continue, calculate affine compensate for clearance gap _aff.

In step S302, convergence precision ε is 10 ^-6.Calculate affine compensate for clearance gap _affmethod be:

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;x}}_{\mathrm{aff}}\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]$

Calculate affine step delta x _aff, Δ w _aff, Δ w _saff, then calculate affine iteration step length α ₁, α ₂:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}_{\mathrm{aff}}^j\right)}\right|{\mathrm{\&Delta;x}}_{\mathrm{aff}}^j<0\}\\ a_2=0.995\min \left\{\frac{{-w}_s^j}{{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j}\right|{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

In formula, num is the element number of variable x.

Calculate affine compensate for clearance gap _aff:

${\mathrm{gap}}_{\mathrm{aff}}={(w_s^k+{\mathrm{\&alpha;}}_2\mathrm{\&Delta;}{w}_{\mathrm{saff}})}^T(x^k+{\mathrm{\&alpha;}}_1{\mathrm{\&Delta;x}}_{\mathrm{aff}})$

The method of calculating barrier parameter μ in step S303 is:

$\mathrm{\&mu;}=\frac{{\mathrm{gap}}_{\mathrm{aff}}}{2\mathrm{num}}\min \{{\left(\frac{{\mathrm{gap}}_{\mathrm{aff}}}{{\mathrm{gap}}^k}\right)}^2,0.2\}$

In formula, gap ^kcompensate for clearance for last point.

Second step is that the method for proofreading and correct the Newton direction that obtains correction is:

Step S304, according to the described barrier parameter μ trying to achieve, described affine step delta w _saffwith Δ x _affobtain the Newton direction of described correction:

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right]=\left[\begin{array}{c}0\\ 0\\ \mathrm{\&mu;e}-{\mathrm{\&Delta;X}}_{\mathrm{aff}}{\mathrm{\&Delta;Z}}_{\mathrm{aff}}e\end{array}\right]$

Wherein, Δ Z _aff, Δ X _afffor diagonal matrix, diagonal element is by Δ w _saff, Δ x _affform.

The Newton direction that finally obtains revising is:

$\left[\begin{array}{c}\mathrm{\&Delta;x}\\ \mathrm{\&Delta;w}\\ {\mathrm{\&Delta;w}}_s\end{array}\right]=\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{saff}}\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right]$

According to equation:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1^k=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}^j\right)}\right|{\mathrm{\&Delta;x}}^j<0\}\\ {\mathrm{\&alpha;}}_2^k=0.995\min \left\{\frac{{-w}_s^j}{\left({\mathrm{\&Delta;w}}_s^j\right)}\right|{\mathrm{\&Delta;w}}_s^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

Calculate iteration step length

k+1 point is:

$\left[\begin{array}{c}{x}^{k+1}\\ w^{k+1}\\ {w_s}^{k+1}\end{array}\right]=\left[\begin{array}{c}{x}^k\\ w^k\\ {w_s}^k\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&alpha;}}_1^k\mathrm{\&Delta;x}\\ {\mathrm{\&alpha;}}_2^k\mathrm{\&Delta;w}\\ {\mathrm{\&alpha;}}_2^k{\mathrm{\&Delta;w}}_s\end{array}\right]$

Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit, although the present invention is had been described in detail with reference to above-described embodiment, those of ordinary skill in the field are to be understood that: still can modify or be equal to replacement the specific embodiment of the present invention, and do not depart from any modification of spirit and scope of the invention or be equal to replacement, it all should be encompassed in the middle of claim scope of the present invention.

Claims (7)

1. a method for Economic Dispatch a few days ago, is characterized in that, described method comprises:

Step S1, foundation makes the linear programming model of Economic Dispatch a few days ago of the operating cost F minimum of described electric system, and the objective function of described mathematical model is

T is the time hop count during system call, and M is system genset number, p _i(t) be genset i in the active power of t period, C _i[p _i(t)] be the operating cost of genset i in the t period;

Step S2, sets up constraint condition and makes described electric system meet balancing the load constraint, unit operation constraint and contract Constraint, and the attribute of the variable relating to according to constraint condition described in each sorts to described constraint condition, forms constraint matrix A;

Step S3, is used prediction-correction interior point to solve the result of the described Economic Dispatch linear programming model that is met described constraint matrix A.

2. the method for claim 1, is characterized in that, the described constraint condition that makes described electric system meet balancing the load constraint, unit operation constraint and the foundation of contract Constraint in described step S2 comprises:

Unit cost constraint, system loading Constraints of Equilibrium, the constraint of unit output bound, unit ramp loss and contract Constraint.

3. method as claimed in claim 2, is characterized in that,

Described C _i[p _i(t)] be quadratic function: C _i[p _i(t)]=a _ip _i ²(t)+b _ip _i(t)+c _i, wherein, a _i, b _i, c _ifor unknown quadratic function coefficient;

Operating cost C by described genset i in the t period _i[p _i(t) quadratic function] is divided into l section and is write as l inequality linear restriction, obtains described unit cost constraint: C _i[p _i(t)]>=k _i,rp _i(t)+m _i,r, k _i,r, m _i,rbe the coefficient of r section linear function, r=1 ..., l;

Described system loading Constraints of Equilibrium is:

p _d(t) be the total load of system t period;

Described unit output bound is constrained to:

be respectively the upper and lower limit of genset i output power;

Described unit ramp loss is :-Δ p _i≤ p _i(t)-p _i(t-1)≤Δ p _i, Δ p _ifor per period of genset i can be added and subtracted the maximal value of power;

Described contract Constraint is: be respectively the upper and lower limit of mono-day contract electric weight of genset i.

4. method as claimed in claim 3, is characterized in that,

The variable that described unit output bound constraint and described unit cost constraint unit relate to comprises the power and variable of single described genset;

The variable that described unit ramp loss relates to comprises the power and variable of the described genset of two periods;

The variable that described contract Constraint relates to comprises the power and variable of single all periods of genset;

The variable that described system loading Constraints of Equilibrium relates to comprises the power and variable of all described genset in the period;

The method that the attribute of the variable relating to according to constraint condition described in each in described step S2 sorts to described constraint condition is:

Described unit output bound constraint and described unit cost constraint unit are combined to the independent constrain set Q that regards single described genset in the period as, according to the period, be arranged in order the constrain set Q of each genset, then by the period, be arranged in order the described unit ramp loss of each genset, finally arrange contract Constraint and system loading Constraints of Equilibrium;

Each constraint condition described constraint matrix A of rear formation that sorted.

5. the method for claim 1, is characterized in that, the canonical form of the linear programming model of Economic Dispatch a few days ago of setting up in described step S1 is:

minF＝cx

$s.t.\left\{\begin{array}{c}\mathrm{Ax}=b\\ x\&GreaterEqual;0\end{array}\right.$

Wherein, x is variable, and c, b are vector, and A is constraint matrix;

In described step S3, using prediction-correction interior point to solve to obtain the result of described Economic Dispatch linear programming model is to solve according to the dual problem of described Economic Dispatch linear programming model, and the dual problem of the canonical form of described Economic Dispatch linear programming model is:

max?w ^Tb

$s.t.\left\{\begin{array}{c}{w}^{T}A-w_s=c\\ w_s\&GreaterEqual;0\end{array}\right.$

Wherein, w, w _svariable for dual problem.

6. method as claimed in claim 5, is characterized in that, the concrete grammar that uses prediction-correction interior point to solve the result that obtains described Economic Dispatch linear programming model in described step S3 is:

Step S301, from a point

set out, the Newton direction of correction is:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\begin{array}{}\end{array}\left[\begin{array}{c}\mathrm{\&Delta;x}\\ \mathrm{\&Delta;w}\\ {\mathrm{\&Delta;w}}_s\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ \mathrm{\&mu;e}-\mathrm{\&Delta;X\&Delta;Ze}\end{array}\right]$

Wherein, Δ x, Δ w, Δ w _sfor iteration correction, I is unit matrix, and e is vector of unit length, and Z, X, Δ Z, Δ X are diagonal matrix, and diagonal element is by w _s, x, Δ w _s, Δ x forms, μ is barrier parameter;

Step S302, the compensate for clearance while calculating the k time iteration

and judge the condition of convergence, if compensate for clearance gap ^kbe less than convergence precision ε, export optimum solution, finish to calculate; Otherwise continue, calculate affine compensate for clearance gap _aff;

Calculate described affine compensate for clearance gap _affcomprise:

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;x}}_{\mathrm{aff}}\end{array}\right]=\left[\begin{array}{c}b-{\mathrm{Ax}}^0\\ c-{\left(w^0\right)}^{T}A+w_s\\ -\mathrm{XZ}\end{array}\right]$

Calculate affine step delta x _aff, Δ w _aff, Δ w _saff, then according to equation:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}_{\mathrm{aff}}^j\right)}\right|{\mathrm{\&Delta;x}}_{\mathrm{aff}}^j<0\}\\ a_2=0.995\min \left\{\frac{{-w}_s^j}{{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j}\right|{\mathrm{\&Delta;w}}_{\mathrm{saff}}^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

Calculate affine iteration step length α ₁, α ₂, wherein, num is the element number of variable x;

According to formula: ${\mathrm{gap}}_{\mathrm{aff}}={(w_s^k+{\mathrm{\&alpha;}}_2\mathrm{\&Delta;}{w}_{\mathrm{saff}})}^T(x^k+{\mathrm{\&alpha;}}_1{\mathrm{\&Delta;x}}_{\mathrm{aff}})$ Calculate described affine compensate for clearance gap _aff;

Step S303, dyscalculia parameter μ:

$\mathrm{\&mu;}=\frac{{\mathrm{gap}}_{\mathrm{aff}}}{2\mathrm{num}}\min \{{\left(\frac{{\mathrm{gap}}_{\mathrm{aff}}}{{\mathrm{gap}}^k}\right)}^2,0.2\}$

Step S304, with described barrier parameter μ, the described affine step delta w according to trying to achieve _saffwith Δ x _affobtain the Newton direction of described correction, according to the Newton direction of described correction, calculate iteration step length

obtain k+1 point;

Solving equation:

$\left[\begin{array}{ccc}A& 0& 0\\ 0& A^T& -I\\ Z& 0& X\end{array}\right]\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right]=\left[\begin{array}{c}0\\ 0\\ \mathrm{\&mu;e}-{\mathrm{\&Delta;X}}_{\mathrm{aff}}{\mathrm{\&Delta;Z}}_{\mathrm{aff}}e\end{array}\right]$

Δ Z _aff, Δ X _afffor diagonal matrix, diagonal element is by Δ w _saff, Δ x _affform;

The Newton direction that obtains described correction is:

$\left[\begin{array}{c}\mathrm{\&Delta;x}\\ \mathrm{\&Delta;w}\\ {\mathrm{\&Delta;w}}_s\end{array}\right]=\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{aff}}\\ {\mathrm{\&Delta;w}}_{\mathrm{saff}}\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{co}}\\ {\mathrm{\&Delta;w}}_{\mathrm{sco}}\end{array}\right]$

According to equation:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_1^k=0.995\min \left\{\frac{{-x}^j}{\left({\mathrm{\&Delta;x}}^j\right)}\right|{\mathrm{\&Delta;x}}^j<0\}\\ {\mathrm{\&alpha;}}_2^k=0.995\min \left\{\frac{{-w}_s^j}{\left({\mathrm{\&Delta;w}}_s^j\right)}\right|{\mathrm{\&Delta;w}}_s^j<0\}\end{array}\right.(j=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,\mathrm{num})$

Calculate iteration step length k+1 point is:

$\left[\begin{array}{c}{x}^{k+1}\\ w^{k+1}\\ {w_s}^{k+1}\end{array}\right]=\left[\begin{array}{c}{x}^k\\ w^k\\ {w_s}^k\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&alpha;}}_1^k\mathrm{\&Delta;x}\\ {\mathrm{\&alpha;}}_2^k\mathrm{\&Delta;w}\\ {\mathrm{\&alpha;}}_2^k{\mathrm{\&Delta;w}}_s\end{array}\right].$

7. method as claimed in claim 6, is characterized in that, convergence precision ε described in described step S302 is 10 ^-6.