CN106712031A - Sequence-robustness optimizing dispatching system and dispatching method of active power distribution network considering uncertainty - Google Patents

Abstract

The invention discloses a sequence-robustness optimizing dispatching system and a dispatching method of an active power distribution network considering uncertainty. A dispatching framework comprises an upper level sequence optimizing model and a lower level robustness self-adaptive model. The dispatching method based on the dispatching framework comprises the following steps of for the traditional voltage regulation equipment in the active power distribution network, performing hourly optimizing dispatching, introducing a sequence optimizing theory, building an upper level optimizing model, and decreasing the influence of uncertainty to decisions; according to an upper level dispatching command, performing lower level optimizing within one hour, adopting a robustness self-adaptive optimizing method, and building a robustness self-adaptive active-reactive coordinated optimizing model, so as to realize the active control and real-time optimizing on the power distribution network. The dispatching method has the advantages that for the active power distribution network accessed with large-scale regenerated energy source, all controllable resources are called to provide the system support, the adverse effect by the output uncertainty of the regenerated energy source is fully decreased, and the reliable, safe, economic and efficient operation of the power distribution network is realized.

Description

Meter and probabilistic active distribution network it is sequential-robust Optimal Scheduling and scheduling Method

Technical field

It is more particularly to a kind of to count and probabilistic the present invention relates to the Scheduling Framework and dispatching method of active distribution network The sequential robust Optimal Scheduling of active distribution network and dispatching method.

Background technology

Positive Renewable Energy Development generation technology, be China readjust the energy structure, Economic Development Mode Conversion and realization The strategic choice of sustainable development, Distributed Power Generation (DG) has turned into Jiao of China's energy sustainable development key technology Point.The large-scale grid connection of distributed power source, the operation to power distribution network brings wide influence and huge challenge.

Existing active distribution network regulating strategy mainly includes：When single time section optimization is dispatched, dispatches and be many a few days ago Between scale coordination coordinate layering scheduling.Single time section optimization scheduling does not take into full account traditional adjusting device and distributed power source Difference, it is not enough to distribution actual motion directive significance；Scheduling a few days ago depends on the predicted value of distributed power source, and distributed electrical Often there is certain deviation in source actual value, and predicted time is more long with predicted value, and deviation is bigger, therefore operation plan is past a few days ago It is not high toward precision；Multiple Time Scales cooperation layering scheduling be still power distribution network security constraint it is out-of-limit after passive regulation, no It is real active control and real-time optimization, and it easily has that calculation times are more, the problem of high cost.

Therefore, prior art exist scheduling time scale it is too low, to distributed power source Uncertainty Management scarce capacity etc. Problem, it is impossible to ensure reliability, safety, economy, the Effec-tive Function of power distribution network.

The content of the invention

Goal of the invention：To overcome the shortcomings of that prior art is present, there is provided one kind uses double-deck stereo Scheduling Framework, hour The meter that is combined with Optimized Operation in hour of level Optimized Operation and probabilistic active distribution network it is sequential-robust Optimal Scheduling And dispatching method.

Technical scheme：One kind meter and probabilistic active distribution network it is sequential-robust Optimal Scheduling, including upper strata is excellent Change model and lower floor's Optimized model；

Upper strata Optimized model is optimized to traditional adjusting device, and lower suboptimization model carries out excellent in real time to distributed power source Change, lower floor's Optimized model receives the dispatch command of upper strata Optimized model, carry out the lower floor's optimization in hour, lower floor's Optimized model will Feedback of the information gives upper strata Optimized model.

Wherein, the upper strata Optimized model be sequential Optimized model, lower floor's Optimized model be robust adaptive it is active- Idle Coordination and Optimization Model；

The sequential Optimized model is：

Wherein, u represents control variables, and x represents state variable, and L (u, x)=0 represents equality constraint, and G (u, x)≤0 represents Inequality constraints, P_lossRepresent system losses；

The robust adaptive is active-and idle Coordination and Optimization Model is：

Wherein,It is the object function under undisturbed state,To there is state of disturbance Under object function, g (P, Q)=0 represent equality constraint, h (P, Q)≤0 represent inequality constraints；

The affine expression formula of maximum active power output of distributed power source is：

Wherein, p^maxIt is distributed power source maximum active power output,It is the predicted value of distributed power source maximum active power output,It is maximum perturbation amount, ε exerts oneself Discontinuous Factors for distributed power source, and Ω is uncertain collection；

The auto-adaptive function of the optimal active power output of distributed power source, is shown below：

P=p₀+p(ε) (4)

P (ε)=p_α1ε+p_α2ε^Tε+… (5)

Wherein, p is the optimal active power output of distributed power source, p₀For in the case of undisturbed distributed power source it is optimal active go out Power, p (ε) is auto-adaptive function of the distributed power source active power output with shock wave, p_α1It is the one of distributed power source active power output Rank disturbance quantity, p_α2It is the second order disturbance quantity of distributed power source active power output, ε^TIt is the transposition of ε；Wherein formula (5) is multinomial letter Number, the specific number of its item number is related to the expression of Optimized model；

The auto-adaptive function that distributed power source OPTIMAL REACTIVE POWER is exerted oneself, is shown below：

Q=q₀+q(ε) (6)

Q (ε)=q_α1ε+q_α2ε^Tε+… (7)

Wherein, Q exerts oneself for distributed power source OPTIMAL REACTIVE POWER, q₀For the OPTIMAL REACTIVE POWER of distributed power source in the case of undisturbed goes out Power, q (ε) is the idle auto-adaptive function exerted oneself with shock wave of distributed power source, q_α1One exerted oneself for distributed power source is idle Rank disturbance quantity, q_α2The second order disturbance quantity exerted oneself for distributed power source is idle；Wherein formula (7) is polynomial function, its item number tool Body number is related to the expression of Optimized model.

A kind of dispatching method based on above-mentioned Scheduling Framework, comprises the following steps：

(1) upper strata Optimized model is built

Traditional voltage adjusting device in for active distribution network, carries out hour level Optimized Operation, introduces sequential optimum theory, enters The sequential rolling optimization of row, builds upper strata Optimized model；

(2) lower floor's Optimized model is built

Based on upper strata dispatch command, the lower floor's optimization in hour is carried out, give full play to the system support energy of distributed power source Power, uncertain variables are processed into the form in interval, using robust adaptive optimization method, build robust adaptive it is active-nothing Work(Coordination and Optimization Model, i.e. lower floor's Optimized model.

Wherein, the step (1) further includes：

Formulate operation plan in (a) upper strata

When operation plan is formulated on upper strata, while consider traditional adjusting device and distributed power source, but main actions tradition from Adjusting device is dissipated, the regulation and control instruction of distributed power source is mainly derived from lower floor's optimization；

B () obtains operation plan

It is a dispatching cycle that upper strata optimizes with T, proceeds by Optimal Decision-making each dispatching cycle, obtains operation plan；

C () updates operation plan

If current time apart from the conventional discrete device action time more than 1 hour, every T₁Time is once rolled Optimization, updates operation plan；

If current time is less than 1 hour apart from the conventional discrete device action time, every T₂Time is once rolled Optimization, updates operation plan；

T before conventional discrete device action₃Time, then a suboptimization is carried out, update operation plan.

Upper strata optimization Mathematical Modeling be：

Wherein, u represents control variables, and x represents state variable, and L (u, x)=0 represents equality constraint, and G (u, x)≤0 represents Inequality constraints, P_lossRepresent system losses.

Wherein, upper strata optimization dispatching cycle T span be 1h~24h, T₁Span be 20min~1h, T₂ Span be 10min~20min, T₃Span be 3min~10min.

In addition, lower floor optimization only regulation and control distributed power source active power output and it is idle exert oneself, reality is carried out to distributed power source Shi Youhua；Dispatch command based on upper strata, lower floor optimization using robust adaptive it is active-idle coordination optimizing method, will not be true Determine variable and be processed as range format, build auto-adaptive function, build robust adaptive Optimized model, be used to ask for distributed power source Self Adaptive Control rule；The time granularity of optimization is T₄, i.e., every T₄Time carries out an Optimized Operation, obtains the T₄Time Interior operational order, T₄Span is 30s~5min；

In active distribution network, the maximum active power output of distributed power source is uncertain variables, its interval following institute of expression formula Show：

Wherein, P^maxDistributed power source maximum active power output is represented,The interval lower limit exerted oneself is represented,Represent interval The upper limit exerted oneself；

The affine expression formula of maximum active power output of distributed power source is：

Wherein, p^maxIt is distributed power source maximum active power output,It is the predicted value of distributed power source maximum active power output,It is maximum perturbation amount, ε exerts oneself Discontinuous Factors for distributed power source, and Ω is uncertain collection；

The auto-adaptive function of the optimal active power output of distributed power source, is shown below：

P=p₀+p(ε) (11)

P (ε)=p_α1ε+p_α2ε^Tε+… (12)

Wherein, p is the optimal active power output of distributed power source, p₀For in the case of undisturbed distributed power source it is optimal active go out Power, p (ε) is auto-adaptive function of the distributed power source active power output with shock wave, p_α1It is the one of distributed power source active power output Rank disturbance quantity, p_α2It is the second order disturbance quantity of distributed power source active power output, ε^TIt is the transposition of ε；Wherein formula (12) is multinomial Function, the specific number of its item number is related to the expression of Optimized model；

The auto-adaptive function that distributed power source OPTIMAL REACTIVE POWER is exerted oneself, is shown below：

Q=q₀+q(ε) (13)

Q (ε)=q_α1ε+q_α2ε^Tε+… (14)

Wherein, Q exerts oneself for distributed power source OPTIMAL REACTIVE POWER, q₀For the OPTIMAL REACTIVE POWER of distributed power source in the case of undisturbed goes out Power, q (ε) is the idle auto-adaptive function exerted oneself with shock wave of distributed power source, q_α1One exerted oneself for distributed power source is idle Rank disturbance quantity, q_α2The second order disturbance quantity exerted oneself for distributed power source is idle；Wherein formula (14) is polynomial function, its item number tool Body number is related to the expression of Optimized model；

Active distribution network robust adaptive is active-idle Coordination and Optimization Model, it is as follows：

Wherein,It is the object function under undisturbed state,To there is state of disturbance Under object function, g (P, Q)=0 represent equality constraint, h (P, Q)≤0 represent inequality constraints.

Beneficial effect：Compared with prior art, the present invention has advantages below：Fully have invoked the biography in active distribution network System adjusting device and distributed power source, and double-deck stereo Scheduling Framework is devised according to its feature；In the Optimized Operation of upper strata, draw Enter sequential optimum theory, it is ensured that under condition of uncertainty, the accuracy and validity of Optimal Decision-making, and compared to many times Scale coordination optimizes, and reduces operation times, reduces cost；In lower floor's Optimized Operation, ADAPTIVE ROBUST optimization is employed Method, realizes Optimal Decision-making and changes and adaptive change with disturbance quantity, it is ensured that the economy and security of distribution network operation； The present invention is suitable for the grid-connected active distribution network Optimized Operation of a large amount of distributed power sources.

Brief description of the drawings

Fig. 1 is Scheduling Framework schematic diagram of the present invention.

Fig. 2 is dispatching method isoboles of the present invention.

Specific embodiment

Further detailed description is done to the present invention below in conjunction with the accompanying drawings.

Regulating and controlling voltage resource is numerous in active distribution network, is typically divided into traditional voltage adjusting device and grid-connected renewable energy Source.Traditional voltage adjusting device is mostly discrete device, and dispatching cycle is long, and response speed is slow, and often restricted to action frequency, with Shnt capacitor group, load adjustable transformer are representative；Regenerative resource is mostly continuous device, can quick response, to action Number of times is not limited, but with randomness and it is uncertain the characteristics of, power supply is representative in a distributed manner.

The a large amount of grid-connected of regenerative resource brings substantial amounts of uncertainty, to be filled in the optimization operation of active distribution network Divide and consider probabilistic influence, while regenerative resource can provide active and reactive power support to power distribution network simultaneously, should fully adjust All maneuverable resources are moved, the effect of regenerative resource is given full play to, active control and real-time optimization to power distribution network is realized. Probabilistic influence and its supporting role to system is given full play to eliminate exerting oneself for regenerative resource, propose sequential-Shandong The dual-layer optimization adjustment and control system of rod self adaptation.

Sequential optimum theory is the Study on Decision-making Method for Optimization for uncertain dynamic system, and its task is in a scheduling week In phase, based on Optimized Operation plan before this and device action state, constrained according to the regular hour, operation plan is continued to repair Just, it is therefore an objective to realize the optimal control to uncertain system.

As shown in figure 1, a kind of meter and probabilistic active distribution network it is sequential-robust Optimal Scheduling, including upper strata Sequential optimization and lower floor robust adaptive optimization；Wherein, the sequential optimization in upper strata is mainly for the tradition in active distribution network Voltage adjusting device, carries out hour level Optimized Operation, and lower floor's robust adaptive optimization is carried out in hour mainly for distributed power source Real-time optimization；In addition, sequential optimization the in upper strata sends upper strata instruction, lower floor's robust adaptive optimization to lower floor's robust adaptive optimization Information of lower layer is fed back into the sequential optimization in upper strata；The sequential optimization in upper strata is divided into decision phase 1, decision phase 2 and decision phase 3, Wherein, time interval of the time interval of decision phase 1 to decision phase 2 more than decision phase 2 to the decision phase 3；Lower floor Shandong Rod adaptive optimization is divided into 3 periods, and a robust adaptive optimization is carried out every 5min.

A kind of dispatching method based on above-mentioned scheduling system, including step in detail below：

(1) for active distribution network in traditional voltage adjusting device, carry out hour level Optimized Operation, introduce it is sequential optimization reason By, sequential rolling optimization is carried out, build upper strata Optimized model；

(2) based on upper strata dispatch command, the lower floor's optimization in hour is carried out, gives full play to the system support of distributed power source Ability, uncertain variables are processed into the form in interval, using robust adaptive optimization method, build robust adaptive it is active- Idle Coordination and Optimization Model.

When operation plan is formulated on upper strata, while consider traditional adjusting device and distributed power source, but main actions tradition from Adjusting device is dissipated, the regulation and control instruction of distributed power source is mainly derived from lower floor's optimization.

As shown in Fig. 2 being the dispatching method isoboles from 0 point to 4 points, in step (1), the low of scheduling decision is considered Conservative degree, reduces cost and reduction operation times, upper strata optimization, for a dispatching cycle, were opened each dispatching cycle with 4 hours Beginning optimizes decision-making, obtains operation plan；If current time apart from the conventional discrete device action time more than 1 hour, A rolling optimization is carried out within 30 minutes, operation plan is updated；If current time is small less than 1 apart from the conventional discrete device action time When, then a rolling optimization is carried out within 15 minutes, update operation plan；5 minutes before conventional discrete device action, then carry out once Optimization, updates operation plan.

Upper strata optimizes with the minimum optimization aim of system losses, the minimum node injecting power for being each node of system losses Minimum, its expression formula is as follows：

Wherein, i is node serial number, and N is node number, p_iIt is the active injecting power of the node of node i, P_lossIt is system network Damage.

The trend Constraints of Equilibrium of system, is shown below：

S=diag [V] [Y]^*·[V]^* (17)

Wherein, S is node injecting power vector, and V is node voltage vector, and Y is bus admittance matrix.

Voltage magnitude is constrained, and is shown below：

Wherein, V_iIt is arbitrary node i voltage magnitudes,It is the lower limit of arbitrary node i voltage magnitudes,It is arbitrary node The upper limit of i voltage magnitudes.

Shnt capacitor group (CB) constraint is shown below：

Wherein, Q_i,CBIdle for CB is exerted oneself, Q_i,CB,stepIt is the reactive power that every group capacitor is provided, N_i,CBIt is the group of CB Count and it be nonnegative integer,It is the upper limit of capacitor bank number.

ULTC (OLTC) constrains as follows：

Wherein,It is the gear of OLTC,It is the upper limit of OLTC gears,It is OLTC gears Lower limit.

Distributed power source constraint is as follows：

Wherein,WithRespectively the active power output of distributed power source and idle exert oneself；WithRespectively It is the lower and upper limit of distributed power source active power output；WithThe respectively idle lower limit exerted oneself of distributed power source And the upper limit；It is distributed power source capacity.

Upper strata optimized mathematical model reduced form is as follows：

Wherein, u is control variables, including load transformer (OLTC) gear, shnt capacitor (CB) gear and distribution The active, idle of formula power supply is exerted oneself；X is state variable, including node voltage amplitude, capacity of trunk etc.；L (u, x)=0 is equation Constraint, G (u, x)≤0 is inequality constraints, P_lossIt is system losses.

In step (2), consider the accuracy of prediction and reduce calculation times, it is the time that lower floor optimized with 5 minutes Granularity, i.e., carry out an Optimized Operation in every 5 minutes, obtains the operational order in this 5 minutes.

Lower floor optimization only regulation and control distributed power source active power output and it is idle exert oneself, distributed power source is carried out in real time it is excellent Change.Dispatch command based on upper strata, lower floor optimization using robust adaptive it is active-idle coordination optimizing method, by uncertain change Amount is processed as range format, builds auto-adaptive function, builds robust adaptive Optimized model, be used to ask for distributed power source from Suitable solution rule, the time granularity of optimization is 5 minutes.

In active distribution network, the maximum active power output of distributed power source is uncertain variables, its interval following institute of expression formula Show：

Wherein,The respectively interval lower and upper limit exerted oneself.

The range format is expressed as the affine expression formula of maximum active power output of affine form, i.e. distributed power source, it is as follows It is shown：

Wherein, p^maxIt is distributed power source maximum active power output,It is the predicted value of distributed power source maximum active power output,It is maximum perturbation amount, ε exerts oneself Discontinuous Factors for distributed power source, and Ω is uncertain collection.

In the case, simple linear adaption rule can not well be fitted the optimized operation point of active distribution network, More accurate auto-adaptive function need to be used so that optimization instruction is more nearly the global optimum's operating point under every kind of possible operating mode.

The auto-adaptive function of the optimal active power output of distributed power source is a polynomial function, and it is expressed as follows shown in formula：

P=p₀+p(ε) (25)

P (ε)=p_α1ε+p_α2ε^Tε+… (26)

Wherein, p is the optimal active power output of distributed power source, p₀For in the case of undisturbed distributed power source it is optimal active go out Power, p (ε) is auto-adaptive function of the distributed power source active power output with shock wave, p_α1It is the one of distributed power source active power output Rank disturbance quantity, p_α2It is the second order disturbance quantity of distributed power source active power output, ε^TIt is the transposition of ε；

The auto-adaptive function that distributed power source OPTIMAL REACTIVE POWER is exerted oneself is similarly a nonlinear polynomial function, its expression Formula is as follows：

Q=q₀+q(ε) (27)

Q (ε)=q_α1ε+q_α2ε^Tε+… (28)

Wherein, Q exerts oneself for distributed power source OPTIMAL REACTIVE POWER, q₀For the OPTIMAL REACTIVE POWER of distributed power source in the case of undisturbed goes out Power, q (ε) is the idle auto-adaptive function exerted oneself with shock wave of distributed power source, q_α1One exerted oneself for distributed power source is idle Rank disturbance quantity, q_α2The second order disturbance quantity exerted oneself for distributed power source is idle；

The thought of ADAPTIVE ROBUST optimization is asking for the decision-making of adaptive change with disturbance variable change, i.e., it is solved It is a decision rule, rather than a decision value.Based on nonlinear distributed power source active power output and it is idle exert oneself it is adaptive Answer function, build active distribution network ADAPTIVE ROBUST it is active-idle Coordination and Optimization Model.The following institute of Mathematical Modeling of lower floor's control Show：

Wherein,It is the object function under undisturbed state,To there is state of disturbance Under object function, g (P, Q)=0 represent equality constraint, h (P, Q)≤0 represent inequality constraints.

Claims (6)

1. a kind of meter and probabilistic active distribution network it is sequential-robust Optimal Scheduling, it is characterised in that it is excellent including upper strata Change model and lower floor's Optimized model；

Upper strata Optimized model is optimized to traditional adjusting device, and lower floor's Optimized model carries out real-time optimization to distributed power source, Lower floor's Optimized model receives the dispatch command of upper strata Optimized model, carries out the lower floor's optimization in hour, and lower floor's Optimized model will be believed Breath feeds back to upper strata Optimized model.

2. a kind of meter according to claim 1 and probabilistic active distribution network it is sequential-robust Optimal Scheduling, its Be characterised by, the upper strata Optimized model be sequential Optimized model, lower floor's Optimized model be robust adaptive it is active-it is idle Coordination and Optimization Model；

The sequential Optimized model is：

Minf (u, x)=minP_loss

L (u, x)=0

G(u,x)≤0

Wherein, u represents control variables, and x represents state variable, and L (u, x)=0 represents equality constraint, and G (u, x)≤0 represents Formula is constrained, P_lossRepresent system losses；

The robust adaptive is active-and idle Coordination and Optimization Model is：

$\underset{p_0,q_0}{min}f(p_0,q_0)+\underset{p\left(\mathrm{\ϵ}\right),q\left(\mathrm{\ϵ}\right)}{min}\underset{\mathrm{\ϵ}}{max}f\left(p\right(\mathrm{\ϵ}),q(\mathrm{\ϵ}\left)\right)$

G (P, Q)=0

h(P,Q)≤0

P=p₀+ p (ε) p (ε)=p_α1ε+p_α2ε^Tε+…

Q=q₀+ q (ε) q (ε)=q_α1ε+q_α2ε^Tε+…

$p^{\max }=p_0^{\max }+p_{\mathrm{\α}}^{\max }\mathrm{\ϵ}$

ε∈Ω

Wherein,It is the object function under undisturbed state,For under having a state of disturbance Object function, g (P, Q)=0 represents equality constraint, and h (P, Q)≤0 represents inequality constraints；

The affine expression formula of maximum active power output of distributed power source is：

$p^{max}=p_0^{\max }+p_{\mathrm{\α}}^{\max }\mathrm{\ϵ},\mathrm{\ϵ}\∈\mathrm{\Ω}$

Wherein, p^maxIt is distributed power source maximum active power output,It is the predicted value of distributed power source maximum active power output, It is maximum perturbation amount, ε exerts oneself Discontinuous Factors for distributed power source, and Ω is uncertain collection；

The auto-adaptive function of the optimal active power output of distributed power source, is shown below：

P=p₀+p(ε)

P (ε)=p_α1ε+p_α2ε^Tε+…

Wherein, p is the optimal active power output of distributed power source, p₀It is the optimal active power output of distributed power source in the case of undisturbed, p (ε) is auto-adaptive function of the distributed power source active power output with shock wave, p_α1For the single order of distributed power source active power output is disturbed Momentum, p_α2It is the second order disturbance quantity of distributed power source active power output, ε^TIt is the transposition of ε；

The auto-adaptive function that distributed power source OPTIMAL REACTIVE POWER is exerted oneself, is shown below：

Q=q₀+q(ε)

Q (ε)=q_α1ε+q_α2ε^Tε+…

Wherein, Q exerts oneself for distributed power source OPTIMAL REACTIVE POWER, q₀For the OPTIMAL REACTIVE POWER of distributed power source in the case of undisturbed is exerted oneself, q (ε) is the idle auto-adaptive function exerted oneself with shock wave of distributed power source, q_α1The single order exerted oneself for distributed power source is idle is disturbed Momentum, q_α2The second order disturbance quantity exerted oneself for distributed power source is idle.

3. a kind of dispatching method of the scheduling system based on described in claim 1, it is characterised in that comprise the following steps：

(1) upper strata Optimized model is built

Traditional voltage adjusting device in for active distribution network, carries out hour level Optimized Operation, introduces sequential optimum theory, carries out sequence Rolling optimization is passed through, upper strata Optimized model is built；

(2) lower floor's Optimized model is built

Based on upper strata dispatch command, the lower floor's optimization in hour is carried out, give full play to the system enabling capabilities of distributed power source, will Uncertain variables are processed into the form in interval, using robust adaptive optimization method, build robust adaptive it is active-idle coordination Optimized model, i.e. lower floor's Optimized model.

4. a kind of dispatching method according to claim 3, it is characterised in that the step (1) specifically includes：

(11) operation plan is formulated on upper strata

When operation plan is formulated on upper strata, while consider traditional adjusting device and distributed power source, but main actions conventional discrete is adjusted Control equipment, the regulation and control instruction of distributed power source is mainly derived from lower floor's optimization；

(12) operation plan is obtained

It is a dispatching cycle that upper strata optimizes with T, proceeds by Optimal Decision-making each dispatching cycle, obtains operation plan；

(13) operation plan is updated

If current time apart from the conventional discrete device action time more than 1 hour, every T₁Time carries out a rolling optimization, Update operation plan；

If current time is less than 1 hour apart from the conventional discrete device action time, every T₂Time carries out a rolling optimization, Update operation plan；

T before conventional discrete device action₃Time, then a suboptimization is carried out, update operation plan.

Upper strata optimization Mathematical Modeling be：

Minf (u, x)=minP_loss

L (u, x)=0

G(u,x)≤0

Wherein, u represents control variables, and x represents state variable, and L (u, x)=0 represents equality constraint, and G (u, x)≤0 represents Formula is constrained, P_lossRepresent system losses.

5. a kind of dispatching method according to claim 4, it is characterised in that upper strata optimization dispatching cycle T value model It is 1h~24h to enclose, T₁Span be 20min~1h, T₂Span be 10min~20min, T₃Span be 3min~10min.

6. a kind of dispatching method according to claim 4, it is characterised in that lower floor's optimization only regulation and control distributed power source has Work(is exerted oneself and is exerted oneself with idle, and real-time optimization is carried out to distributed power source；Dispatch command based on upper strata, lower floor's optimization uses robust Self adaptation is active-idle coordination optimizing method, uncertain variables are processed as range format, and auto-adaptive function is built, build Shandong Rod adaptive optimization model, is used to ask for the Self Adaptive Control rule of distributed power source；

The time granularity of optimization is T₄, i.e., every T₄Time carries out an Optimized Operation, obtains the T₄Operational order in time, T₄ Span is 30s~5min；

In active distribution network, the maximum active power output of distributed power source is uncertain variables, and its interval expression formula is as follows：

$P^{max}=\[\underset{\‾}{P^{max}},\stackrel{\‾}{P^{max}}\]$

Wherein, P^maxDistributed power source maximum active power output is represented,P^max The interval lower limit exerted oneself is represented,Represent that interval is exerted oneself The upper limit；

The affine expression formula of maximum active power output of distributed power source is：

$p^{\max }=p_0^{\max }+p_{\mathrm{\α}}^{\max }\mathrm{\ϵ},\mathrm{\ϵ}\∈\mathrm{\Ω}$

Wherein, p^maxIt is distributed power source maximum active power output,It is the predicted value of distributed power source maximum active power output, It is maximum perturbation amount, ε exerts oneself Discontinuous Factors for distributed power source, and Ω is uncertain collection；

The auto-adaptive function of the optimal active power output of distributed power source, is shown below：

P=p₀+p(ε)

P (ε)=p_α1ε+p_α2ε^Tε+…

Wherein, p is the optimal active power output of distributed power source, p₀It is the optimal active power output of distributed power source in the case of undisturbed, p (ε) is auto-adaptive function of the distributed power source active power output with shock wave, p_α1For the single order of distributed power source active power output is disturbed Momentum, p_α2It is the second order disturbance quantity of distributed power source active power output, ε^TIt is the transposition of ε；

The auto-adaptive function that distributed power source OPTIMAL REACTIVE POWER is exerted oneself, is shown below：

Q=q₀+q(ε)

Q (ε)=q_α1ε+q_α2ε^Tε+…

Wherein, Q exerts oneself for distributed power source OPTIMAL REACTIVE POWER, q₀For the OPTIMAL REACTIVE POWER of distributed power source in the case of undisturbed is exerted oneself, q (ε) is the idle auto-adaptive function exerted oneself with shock wave of distributed power source, q_α1The single order exerted oneself for distributed power source is idle is disturbed Momentum, q_α2The second order disturbance quantity exerted oneself for distributed power source is idle；

Active distribution network robust adaptive is active-idle Coordination and Optimization Model, it is as follows：

$\underset{p_0,q_0}{min}f(p_0,q_0)+\underset{p\left(\mathrm{\ϵ}\right),q\left(\mathrm{\ϵ}\right)}{min}\underset{\mathrm{\ϵ}}{max}f\left(p\right(\mathrm{\ϵ}),q(\mathrm{\ϵ}\left)\right)$

G (P, Q)=0

h(P,Q)≤0

P=p₀+ p (ε) p (ε)=p_α1ε+p_α2ε^Tε+…

Q=q₀+ q (ε) q (ε)=q_α1ε+q_α2ε^Tε+…

$p^{max}=p_0^{\max }+p_{\mathrm{\α}}^{\max }\mathrm{\ϵ}$

ε∈Ω

Wherein,It is the object function under undisturbed state,For under having a state of disturbance Object function, g (P, Q)=0 represents equality constraint, and h (P, Q)≤0 represents inequality constraints.