CN104537576A - Pre-control and dispatching model and method for surplus water probabilities in inter-basin hydropower station group - Google Patents

Abstract

The invention discloses a pre-control and dispatching model and method for surplus water probabilities in an inter-basin hydropower station group. The maximum surplus water probabilities of hydropower stations above season regulation performance in the inter-basin hydropower station group are minimized to serve as optimization objectives, an 'inter-basin hydropower stations-power station' two-layer optimization and coordination system structure is adopted, and load optimal distribution is directly performed on all the hydropower stations. Equalization of risks and benefits of multiple power generation bodies can be achieved. Accordingly, the new method is provided for joint optimization dispatching of the power generation body hydropower group.

Description

Water probability pre-control scheduling model and dispatching method is abandoned across the equilibrium of basin water station group

Technical field

The present invention relates to one and abandon water probability pre-control scheduling model and dispatching method across the equilibrium of basin water station group, belong to water conservancy and hydropower dispatching technique field.

Background technology

When two basin water station groups all access same electrical network, between two basin water station groups, there is electric power contact closely.For any instant, the gross generation sum in two basins is determined relatively, and electrical network gives the generated energy of first GROUP OF HYDROPOWER STATIONS large, then give the generated energy of second GROUP OF HYDROPOWER STATIONS then little, vice versa.When flood season two, basin all faced larger water, two GROUP OF HYDROPOWER STATIONS all wish to strengthen its generated energy, reduced abandon water probability and abandon the water yield by the mode increasing generating flow.When the two basin retaining of flood end, all need to expect to reduce generated energy, to raise reservoir level as early as possible, reduce generating water consumption rate.Therefore, often can there is scheduling intent conflict in flood season in both sides.Now, current Technical Solving is according to the deviation between the current level in each power station and flood season limit level, and facing the size of period reservoir inflow and Incoming water quantity, the roughly degree of risk of water is abandoned in each power station of fuzzy discrimination by rule of thumb, and formulates each hydropower station plan accordingly.Because the probability of water is abandoned in each power station of the non-quantitative analysis of this mode, therefore cannot realize becoming more meticulous across basin water station group load distribution in flood season, be difficult to effectively to avoid Hydropower Stations to occur unnecessary to abandon water.

Summary of the invention

One is the object of the present invention is to provide to abandon water probability pre-control scheduling model and dispatching method across the equilibrium of basin water station group, the rational distribution method of gross generation between two each power stations, basin across basin water station group is proposed, two each power stations, basin are consistent or state the most balanced at the water probability of abandoning in the scheduling end of term, solve basin in flood season two each hydropower station sequencing problem.

For achieving the above object, the technical solution used in the present invention is as follows:

Abandon water probability pre-control scheduling model across the equilibrium of basin water station group, the maximum water probability of abandoning across power station each in basin water station group minimizes as objective function by described model, that is:

J＝min{max[F _i(V _i,n)]} i＝1,2,…,S (1)

Wherein: if i-th power station is more than season adjusting function, F _iwater probability is abandoned for this power station, otherwise, F _ivalue is zero; V _i,nbe i-th power station scheduling end of term water level, n represents the scheduling end of term; S is the power station sum across basin water station group.

Aforesaid model meets following constraint condition:

Water balance equation:

V _i,j+1＝V _i,j+(I _i,j-Q _i,j)Δt _ji＝1,2,…,S；j＝1,2,…,T (2)

Electric quantity balancing equation:

$E_t=\underset{i=1}{\overset{S}{\mathrm{\Σ}}}{E}_i---\left(3\right)$

Flow coupling constraint:

I _i+1,j＝Q _i,j+R _i+1,ji＝1,2,…,N-1；j＝1,2,…,T (4)

I _i+1,j＝Q _i,j+R _i+1,ji＝N+1,N+2,…,N+M-1；j＝1,2,…,T (5)

Reservoir level retrains:

Z _i,min≤Z _i,j≤Z _i,maxi＝1,2,…,S；j＝1,2,…,T (6)

Storage outflow retrains:

Q _i,min≤Q _i,j≤Q _i,maxi＝1,2,…,S；j＝1,2,…,T (7)

Output of power station retrains:

N _i,min≤N _i,j≤N _i,maxi＝1,2,…,S；j＝1,2,…,T (8)

Wherein, V _i,jrepresent i-th power station jth period storage capacity; Δ t _jrepresent jth Period Length; T indicates T period; E _iit is the generated energy in i-th power station; E _tfor the gross generation across basin water station group; I _i,j, R _i,jrepresent i-th power station jth period total reservoir inflow, interval reservoir inflow respectively; N is the power station quantity of first Hydropower Stations; M is the power station quantity of second GROUP OF HYDROPOWER STATIONS; Z _i,j, Z _{i, min}, Z _{i, max}be respectively i-th power station jth period reservoir level, allow minimum pool level and the highest reservoir level; Q _i,j, Q _{i, min}, Q _{i, max}be respectively i-th power station jth period storage outflow, the restriction of minimum and maximum storage outflow; N _i,j, N _{i, min}, N _{i, max}be respectively i-th power station jth period meritorious to exert oneself, minimum and maximum meritorious restriction of exerting oneself.

Abandon water probability pre-control dispatching method across the equilibrium of basin water station group, comprise the following steps:

(1): obtain each power station real-time database water level and across basin water station group gross generation E _t;

(2): according to the final convergence precision requirement of power energy allocation, and take into account consideration Algorithm for Solving rate request, determine power energy allocation scheme iteration running orbit convergence criterion epsilon _ewith initial generated energy discrete steps Δ E;

(3): all power stations will be given across basin water station group gross generation according to installed capacity proportional allocations, and form original allocation scheme;

(4): in conjunction with original allocation scheme, calculate respectively regulate in each season above power station scheduling the end of term abandon water probability;

(5) that: from the above power station of the first season adjusting function, compares itself and the above power station of the second season adjusting function abandons water probability; If the above power station of the first season adjusting function above hydroelectric station surplus water the second season of likelihood ratio adjusting function to abandon water probability large, then proceed to step (6); If the first season adjusting function above hydroelectric station surplus water probability is less than the above hydroelectric station surplus water probability of the second season adjusting function, then proceed to step (11);

(6): the above hydropower station amount of the first season adjusting function increases Δ E, and adopt and determine water model with electricity and calculate generated energy, then proceed to step (7);

(7): if average generated output is greater than unit installed capacity after the above power station of the first season adjusting function increases generated energy Δ E, then calculate the generated energy in corresponding schedule periods according to unit installed capacity; If reservoir level is lower than permission minimum pool level after increasing generated energy Δ E, then according to permission minimum pool level inverse generated energy and increment thereof, proceed to step (8);

(8): if there is below one or more season adjusting function power station between the above power station of two seasons adjusting function, then utilize principle of water balance to calculate the gross generation in below these adjusting function power station successively in season, and calculate the difference of this gross generation and the gross generation in corresponding power station in original allocation scheme, be designated as Δ E ', proceed to step (9);

(9): the generated energy in the above power station of the second season modulability performance is reduced Δ E+ Δ E ', and employing determines with electricity the generating flow that water model calculates this power station, then proceeds to step (10);

(10): if in step (9), the former generated energy E in the above power station of the second season adjusting function ₂be less than Δ E+ Δ E ', then this hydropower station amount reduces E ₂, the generated energy in the above power station of the first season adjusting function and power station reduces Δ E+ Δ E '-E ₂; If reduce E ₂after this power station reservoir level exceed permission peak level, then according to permission, peak level is counter pushes away generated energy, then proceeds to step (12);

(11): the above hydropower station amount of the first season adjusting function reduces Δ E, calculate below two power stations middle season the adjusting function gross generation in power station and the difference Δ E ' of the gross generation in corresponding power station in original allocation scheme equally, and the generated energy in above for the second season adjusting function power station is increased Δ E+ Δ E ', then proceed to step (12);

(12): successively between two the above power station of adjusting function of more all seasons abandon water probability, and regulate the generated energy in the above power station of two seasons adjusting function and middle power station;

(13): after having traveled through the above power station of all season adjusting functions, the generating share { E in the above power station of each season adjusting function is recorded ₁, E ₂..., E _n;

(14): the maximum generating watt luffing Δ E judging the original allocation scheme of the scheme that (13) obtain and step (3) _maxif, Δ E _max> ε _e, then step (5) is repeated to step (13); If Δ E _max< ε _ebut Δ E > ε _e, then make Δ E=Δ E/2, repeat step (5) to step (13); If Δ E _max< ε _eand Δ E < ε _e, then termination of iterations process;

(15): according to the generation capacity allocation scheme in each power station last, and carry out the minimum water consumption generating calculating of single storehouse successively, obtain power generation process and the generated energy in each power station.

In aforesaid step (4), regulate in season above power station scheduling the end of term the computing method of abandoning water probability be:

Definition of probability power station being abandoned during [t, t+ Δ t] water is hydroelectric station surplus water probability;

If chopped-off head power station, then need first to calculate critical reservoir inflow I ' ₁, then in the natural reservoir inflow probability distribution curve in chopped-off head power station, pass through interpolation, obtain reservoir inflow and be greater than this critical reservoir inflow I ' ₁probability, be the probability that water is abandoned in this chopped-off head power station, wherein, critical reservoir inflow I ' ₁be defined as:

I′ ₁＝Q _1,gen+(V _1,l-V _1,0)/Δt

In formula, V _{1, l}for the corresponding storage capacity of Limited Water Level of Reservoir in Flood Season, Q _{1, gen}for unit completely sends out flow, V _1,0for the current time storage capacity in power station;

If be non-chopped-off head power station, then need first to calculate reservoir inflow R ' between critical section, then in interval reservoir inflow probability distribution curve, pass through interpolation, obtain interval reservoir inflow stochastic variable R _ibe greater than the probability of reservoir inflow R ' between this critical section, be the probability that water is abandoned in this power station, wherein, between critical section, reservoir inflow R ' is defined as:

$R^{\&prime;}=Q_{i,\mathrm{gen}}-{\stackrel{\&OverBar;}{Q}}_{i,\mathrm{prev}}+(V_{i,l}-V_{i,0})/\mathrm{\&Delta;t}$

In formula, Q _{i, gen}be that the unit in i-th grade of non-chopped-off head power station completely sends out flow, V _i,lbe that the corresponding storage capacity in position of restricting water supply is interrogated, V in i-th grade of non-chopped-off head power station _{i, 0}be the storage capacity of i-th grade of non-chopped-off head power station current time, for total storage outflow expectation value of the upper pond in i-th grade that has calculated non-chopped-off head power station.

By adopting technique scheme, the present invention has following beneficial effect:

Pass through the present invention, can according to across the Current Library water level of basin water station group and two given basin gross generations, the generated energy being calculated each power station by computer software Automatic Optimal optimizes distribution share, make each power station scheduling the end of term face abandon the target that water probability reaches equalization the most, there is provided theory support for two basin power generation in the flood seasons amount optimizations distribute, Hydropower Stations can be avoided to greatest extent to abandon the situation of water.In addition, the present invention is also applicable to multiple electric main body GROUP OF HYDROPOWER STATIONS, realizes the equalization of multiple generating main body risk and interests, for multiple electric main body GROUP OF HYDROPOWER STATIONS joint optimal operation provides a kind of new method.

Accompanying drawing explanation

Fig. 1 is across basin water station group schematic diagram in embodiments of the invention.

Embodiment

Now with embodiment, the present invention is described in further detail by reference to the accompanying drawings.

Equilibrium is abandoned the solution of water probability associating pre-control dispatching method and is mass-sended electric sequencing problem across basin water power station flood season, by the reasonable distribution of gross generation between each power station, makes each power station be in state the most balanced at the water probability of abandoning in the scheduling end of term.

The present invention adopts " across basin water station group-power station " bilevel optimization to coordinate architecture, directly between all power stations, carry out load optimal distribution, not only avoid the definition that basin entirety abandons water probability, and coordinate level owing to decreasing one, the solution efficiency of problem can be made significantly to promote.The present invention considers when mathematical modeling between internal water power station, basin exists hydraulic connection, and there is not the characteristic of hydraulic connection between power station, two different basins.

One, water probability pre-control scheduling model is abandoned in equilibrium

Equilibrium abandons the scheduling of water probability associating pre-control for solving the Optimal Control Problem of scheduling end of term water level, is based upon in the past on Optimized Scheduling of Hydroelectric Power theory, and expands its achievement system.When utilizing equilibrium to abandon water probability associating pre-control optimizing scheduling each power station end water level, still be optimized calculating according to maximum generating watt model/maximum generation Benefit Model in the past in the schedule periods of each power station, also can adopt the routine dispactching method in conjunction with artificial optimization's experience.

Maximum water probability of abandoning across power station each in basin water station group is minimized as objective function, that is:

J＝min{max[F _i(V _i,n)]} i＝1,2,…,S (1)

In formula: if i-th power station is more than season adjusting function, F _iwater probability is abandoned for this power station, otherwise, F _ivalue is zero; S is the power station quantity across basin water station group, V _i,nbe i-th power station scheduling end of term water level, be generally all be divided into n period in reservoir operation, initial is called V ₀, after be followed successively by V ₁, V ₂..., V _n, so V _i,nfor the end of term water level in i power station.

For the Hydropower Stations of 2 shown in Fig. 1 composition across basin water station group, N is the power station quantity of first Hydropower Stations, and M is the power station quantity of second Hydropower Stations, then S=N+M.

It is as follows that above-mentioned objective function need meet constraint condition:

Water balance equation:

V _i,j+1＝V _i,j+(I _i,j-Q _i,j)Δt _ji＝1,2,…,S；j＝1,2,…,T (2)

Electric quantity balancing equation:

$E_t=\underset{i=1}{\overset{S}{\mathrm{\&Sigma;}}}{E}_i---\left(3\right)$

Flow coupling constraint:

I _i+1,j＝Q _i,j+R _i+1,ji＝1,2,…,N-1；j＝1,2,…,T (4)

I _i+1,j＝Q _i,j+R _i+1,ji＝N+1,N+2,…,N+M-1；j＝1,2,…,T (5)

Hydropower station water level retrains:

Z _i,min≤Z _i,j≤Z _i,maxi＝1,2,…,S；j＝1,2,…,T (6)

Storage outflow retrains:

Q _i,min≤Q _i,j≤Q _i,maxi＝1,2,…,S；j＝1,2,…,T (7)

Output of power station retrains:

N _i,min≤N _i,j≤N _i,maxi＝1,2,…,S；j＝1,2,…,T (8)

Wherein, V _i,jrepresent i-th power station jth period storage capacity; Δ t _jrepresent jth Period Length; T indicates T period; E _iit is the generated energy in i-th power station; E _tfor the gross generation across basin water station group; I _i,j, R _i,jrepresent i-th power station jth period reservoir inflow, interval reservoir inflow respectively; Z _i,j, Z _{i, min}, Z _{i, max}be respectively i-th power station jth period reservoir level, allow minimum pool level and the highest reservoir level; Q _i,j, Q _{i, min}, Q _{i, max}be respectively i-th power station jth period storage outflow, the restriction of minimum and maximum storage outflow, wherein minimum and maximum storage outflow restriction is determined according to unit conveyance capacity and requirements of comprehensive utilization; N _i,j, N _{i, min}, N _{i, max}be respectively i-th power station jth period meritorious to exert oneself, minimum and maximum meritorious restriction of exerting oneself, wherein minimum and maximum gain merit to exert oneself to limit determine according to unit output scope and network load requirement.

Two, water probability pre-control scheduling model method for solving is abandoned in equilibrium

For abandon across the equilibrium of basin water station group water probability pre-control scheduling model there is the period after validity, the feature that dynamic programming method cannot be used to solve, the present invention proposes to improve discrete differential progressive optimal algorithm (Discrete DifferentialProgressive Optimal Algorithm, DDPOA), under the condition keeping solving precision, effectively model solution speed is improved.

Of the present inventionly abandon water probability pre-control dispatching method across the equilibrium of basin water station group, comprise the following steps:

Step1: obtain each power station real-time database water level and across basin water station group gross generation E _t.

Step2: according to the final convergence precision requirement of power energy allocation, and take into account consideration Algorithm for Solving rate request, determine power energy allocation scheme iteration convergence criterion epsilon _ewith initial generated energy discrete steps Δ E (Δ E > ε _e).

Step3: all power stations will be given across basin water station group gross generation according to installed capacity proportional allocations, and form original allocation scheme.

Step4: in conjunction with original allocation scheme, calculate respectively regulate in each season above power station scheduling the end of term abandon water probability;

The computing method of abandoning water probability are:

Definition of probability power station being abandoned during [t, t+ Δ t] water is hydroelectric station surplus water probability.

If chopped-off head power station, then need first to calculate critical reservoir inflow I ' ₁, then in the natural reservoir inflow probability distribution curve in chopped-off head power station, pass through interpolation, obtain reservoir inflow and be greater than this critical reservoir inflow I ' ₁probability, be the probability that water is abandoned in this chopped-off head power station, wherein, critical reservoir inflow I ' ₁be defined as:

I′ ₁＝Q _1,gen+(V _1,l-V _1,0)/Δt

In formula, the corresponding storage capacity V of Limited Water Level of Reservoir in Flood Season _{1, l}flow Q is completely sent out with unit _{1, gen}can obtain from the design data in power station, the current time storage capacity V in power station _1,0, can obtain according to measured water level and water level-capacity curve interpolation calculation.

If be non-chopped-off head power station, then need first to calculate reservoir inflow R ' between critical section, then in interval reservoir inflow probability distribution curve, pass through interpolation, obtain interval reservoir inflow stochastic variable R _ibe greater than the probability of reservoir inflow R ' between this critical section, be the probability that water is abandoned in this power station, wherein, between critical section, reservoir inflow R ' is defined as:

$R^{\&prime;}=Q_{i,\mathrm{gen}}-{\stackrel{\&OverBar;}{Q}}_{i,\mathrm{prev}}+(V_{i,l}-V_{i,0})/\mathrm{\&Delta;t}$

In formula, Q _{i, gen}be that the unit in i-th grade of non-chopped-off head power station completely sends out flow, V _i,lbe that the corresponding storage capacity in position of restricting water supply is interrogated, Q in i-th grade of non-chopped-off head power station _{i, gen}and V _i,lall can obtain from hydroelectric station design data, V _{i, 0}be the storage capacity of i-th grade of non-chopped-off head power station current time, can obtain according to measured water level and water level-capacity curve interpolation calculation, for total storage outflow expectation value of the upper pond in i-th grade that has calculated non-chopped-off head power station.

Step5: from the above power station of the first season adjusting function, that compares itself and the above power station of the second season adjusting function abandons water probability; If the above power station of the first season adjusting function above hydroelectric station surplus water the second season of likelihood ratio adjusting function to abandon water probability large, then proceed to step 6; If the first season adjusting function above hydroelectric station surplus water probability is less than the above hydroelectric station surplus water probability of the second season adjusting function, then proceed to step 11.

The above hydropower station amount of Step6: the first season adjusting function increases Δ E, and adopts and determine water model with electricity and calculate generating flow, then proceeds to step 7.

Step7: if average generated output is greater than unit installed capacity after the above power station of the first season adjusting function increases generated energy Δ E, then calculate the generated energy in corresponding schedule periods according to unit installed capacity; If reservoir level is lower than permission minimum pool level after increasing generated energy Δ E, then according to permission minimum pool level inverse generated energy and increment thereof, proceed to step 8.

Step8: if there is below one or more season adjusting function power station between the above power station of two seasons adjusting function, then utilize principle of water balance to calculate the gross generation in below these adjusting function power station successively in season, and calculate the difference of this gross generation and the gross generation in corresponding power station in original allocation scheme, be designated as Δ E ', proceed to step 9.

Step9: the generated energy in the above power station of the second season modulability performance is reduced Δ E+ Δ E ', and employing determines with electricity the generating flow that water model calculates this power station, then proceeds to step 10.

Step10: if in step 9, the former generated energy E in the above power station of the second season adjusting function ₂be less than Δ E+ Δ E ', then this hydropower station amount reduces E ₂, the generated energy in the above power station of the first season adjusting function and power station reduces Δ E+ Δ E '-E ₂; If reduce E ₂after this power station reservoir level exceed permission peak level, then according to permission, peak level is counter pushes away generated energy, then proceeds to step 12.

The above hydropower station amount of Step11: the first season adjusting function reduces Δ E, calculate below two power stations middle season the adjusting function gross generation in power station and the difference Δ E ' of the gross generation in corresponding power station in step 3 original allocation scheme equally, and the generated energy in above for the second season adjusting function power station is increased Δ E+ Δ E ', then proceed to step 12.

Step12: successively between two the above power station of adjusting function of more all seasons abandon water probability, and regulate the generated energy in the above power station of two seasons adjusting function and middle power station.

Step13: after having traveled through the above power station of all season adjusting functions, records the generated energy share { E in the above power station of each season adjusting function ₁, E ₂, ', E _n.

Step14: the maximum generating watt luffing Δ E of the scheme that determining step 13 obtains and step 3 original allocation scheme _maxif, Δ E _max> ε _e, then step 5 is repeated to step 13; If Δ E _max< ε _ebut Δ E > ε _e, then make Δ E=Δ E/2, repeat step 5 to step 13; If Δ E _max< ε _eand Δ E < ε _e, then termination of iterations process.

Step15: according to the generation capacity allocation scheme in each power station last, and carry out the minimum water consumption generating calculating of single storehouse successively, obtain power generation process and the generated energy in each power station.

That only considers the above power station of season adjusting function inside objective function of the present invention abandons water probability, but needs when water balanced calculation to consider all power stations.For power station below season adjusting function, dispatch according to the principle going out warehouse-in balance: when power station reservoir level is lower than flood season limit level, pay the utmost attention to retaining, if power station reservoir level stores still have unnecessary water to flood season limit level, then using unnecessary water as abandoning water treatment; When power station reservoir level maintains flood season limit level, if reservoir inflow is less than its unit completely send out flow, then generate electricity according to reservoir inflow; If reservoir inflow is greater than unit completely send out flow, then excess traffic is as abandoning discharge.

Claims (4)

1. abandon water probability pre-control scheduling model across the equilibrium of basin water station group, it is characterized in that, described model will minimize as objective function across the maximum water probability of abandoning in power station each in basin water station group, that is:

J＝min{max[F _i(V _i,n)]}i＝1,2,…,S (1)

Wherein: if i-th power station is more than season adjusting function, F _iwater probability is abandoned for this power station, otherwise, F _ivalue is zero; V _i,nbe the i-th season adjusting function above power station scheduling end of term water level, n represents the scheduling end of term; S is the power station sum across basin water station group.

2. according to claim 1ly abandon water probability pre-control scheduling model across the equilibrium of basin water station group, it is characterized in that, described model meets following constraint condition:

Water balance equation:

V _i,j+1＝V _i,j+(I _i,j-Q _i,j)Δt _ji＝1,2,…,S；j＝1,2,…,T (2)

Electric quantity balancing equation:

$E_t=\underset{i=1}{\overset{S}{\mathrm{\&Sigma;}}}{E}_i---\left(3\right)$

Flow coupling constraint:

I _i+1,j＝Q _i,j+R _i+1,ji＝1,2,…,N-1；j＝1,2,…,T (4)

I _i+1,j＝Q _i,j+R _i+1,ji＝N+1,N+2,…,N+M-1；j＝1,2,…,T (5)

Reservoir level retrains:

Z _i,min≤Z _i,j≤Z _i,maxi＝1,2,…,S；j＝1,2,…,T (6)

Storage outflow retrains:

Q _i,min≤Q _i,j≤Q _i,maxi＝1,2,…,S；j＝1,2,…,T (7)

Output of power station retrains:

N _i,min≤N _i,j≤N _i,maxi＝1,2,…,S；j＝1,2,…,T (8)

Wherein, V _i,jrepresent i-th power station jth period storage capacity; Δ t _jrepresent jth Period Length; T indicates T period; E _iit is the generated energy in i-th power station; E _tfor the gross generation across basin water station group; I _i,j, R _i,jrepresent i-th power station jth period total reservoir inflow, interval reservoir inflow respectively; N is the power station quantity of first Hydropower Stations, and M is the power station quantity of second Hydropower Stations, S=M+N; Z _i,j, Z _{i, min}, Z _{i, max}be respectively i-th power station jth period reservoir level, allow minimum pool level and the highest reservoir level; Q _i,j, Q _{i, min}, Q _{i, max}be respectively i-th power station jth period storage outflow, the restriction of minimum and maximum storage outflow; N _i,j, N _{i, min}, N _{i, max}be respectively i-th power station jth period meritorious to exert oneself, minimum and maximum meritorious restriction of exerting oneself.

3., based on the dispatching method abandoning water probability pre-control scheduling model across the equilibrium of basin water station group described in claim 1 or 2, it is characterized in that, comprise the following steps:

(1): obtain each power station real-time database water level and across basin water station group gross generation E _t;

(2): according to the final convergence precision requirement of power energy allocation, and take into account consideration Algorithm for Solving rate request, determine power energy allocation scheme iteration running orbit convergence criterion epsilon _ewith initial generated energy discrete steps Δ E;

(3): all power stations will be given across basin water station group gross generation according to installed capacity proportional allocations, and form original allocation scheme;

(4): in conjunction with original allocation scheme, calculate respectively regulate in each season above power station scheduling the end of term abandon water probability;

(5) that: from the above power station of the first season adjusting function, compares itself and the above power station of the second season adjusting function abandons water probability; If the above power station of the first season adjusting function above hydroelectric station surplus water the second season of likelihood ratio adjusting function to abandon water probability large, then proceed to step (6); If the first season adjusting function above hydroelectric station surplus water probability is less than the above hydroelectric station surplus water probability of the second season adjusting function, then proceed to step (11);

(6): the above hydropower station amount of the first season adjusting function increases Δ E, and adopt and determine water model with electricity and calculate generated energy, then proceed to step (7);

(7): if average generated output is greater than unit installed capacity after the above power station of the first season adjusting function increases generated energy Δ E, then calculate the generated energy in corresponding schedule periods according to unit installed capacity; If reservoir level is lower than permission minimum pool level after increasing generated energy Δ E, then according to permission minimum pool level inverse generated energy and increment thereof, proceed to step (8);

(8): if there is below one or more season adjusting function power station between the above power station of two seasons adjusting function, then utilize principle of water balance to calculate the gross generation in below these adjusting function power station successively in season, and calculate the difference of this gross generation and the gross generation in corresponding power station in original allocation scheme, be designated as Δ E ', proceed to step (9);

(9): the generated energy in the above power station of the second season modulability performance is reduced Δ E+ Δ E ', and employing determines with electricity the generating flow that water model calculates this power station, then proceeds to step (10);

(10): if in step (9), the former generated energy E in the above power station of the second season adjusting function ₂be less than Δ E+ Δ E ', then this hydropower station amount reduces E ₂, the generated energy in the above power station of the first season adjusting function and power station reduces Δ E+ Δ E '-E ₂; If reduce E ₂after this power station reservoir level exceed permission peak level, then according to permission, peak level is counter pushes away generated energy, then proceeds to step (12);

(11): the above hydropower station amount of the first season adjusting function reduces Δ E, calculate below two power stations middle season the adjusting function gross generation in power station and the difference Δ E ' of the gross generation in corresponding power station in original allocation scheme equally, and the generated energy in above for the second season adjusting function power station is increased Δ E+ Δ E ', then proceed to step (12);

(12): successively between two the above power station of adjusting function of more all seasons abandon water probability, and regulate the generated energy in the above power station of two seasons adjusting function and middle power station;

(13): after having traveled through the above power station of all season adjusting functions, the generating share { E in the above power station of each season adjusting function is recorded ₁, E ₂..., E _n;

(14): the maximum generating watt luffing Δ E judging the original allocation scheme of the scheme that (13) obtain and step (3) _maxif, Δ E _max> ε _e, then step (5) is repeated to step (13); If Δ E _max< ε _ebut Δ E > ε _e, then make Δ E=Δ E/2, repeat step (5) to step (13); If Δ E _max< ε _eand Δ E < ε _e, then termination of iterations process;

(15): according to the generation capacity allocation scheme in each power station last, and carry out the minimum water consumption generating calculating of single storehouse successively, obtain power generation process and the generated energy in each power station.

4. dispatching method according to claim 3, is characterized in that, in described step (4), regulate in season above power station scheduling the end of term the computing method of abandoning water probability be:

Definition of probability power station being abandoned during [t, t+ Δ t] water is hydroelectric station surplus water probability;

If chopped-off head power station, then need first to calculate critical reservoir inflow I ' ₁, then in the natural reservoir inflow probability distribution curve in chopped-off head power station, pass through interpolation, obtain reservoir inflow and be greater than this critical reservoir inflow I ' ₁probability, be the probability that water is abandoned in this chopped-off head power station, wherein, critical reservoir inflow I ' ₁be defined as:

I′ ₁＝Q _1,gen+(V _1,l-V _1,0)/Δt

In formula, V _{1, l}for the corresponding storage capacity of Limited Water Level of Reservoir in Flood Season, Q _{1, gen}for unit completely sends out flow, V _1,0for the current time storage capacity in power station;

If be non-chopped-off head power station, then need first to calculate reservoir inflow R ' between critical section, then in interval reservoir inflow probability distribution curve, pass through interpolation, obtain interval reservoir inflow stochastic variable R _ibe greater than the probability of reservoir inflow R ' between this critical section, be the probability that water is abandoned in this power station, wherein, between critical section, reservoir inflow R ' is defined as:

$R^{\&prime;}=Q_{i,\mathrm{gen}}-{\stackrel{\&OverBar;}{Q}}_{i,\mathrm{prev}}+(V_{i,l}-V_{i,0})/\mathrm{\&Delta;t}$

In formula, Q _{i, gen}be that the unit in i-th grade of non-chopped-off head power station completely sends out flow, V _i,lbe that the corresponding storage capacity in position of restricting water supply is interrogated, V in i-th grade of non-chopped-off head power station _{i, 0}be the storage capacity of i-th grade of non-chopped-off head power station current time, for total storage outflow expectation value of the upper pond in i-th grade that has calculated non-chopped-off head power station.