CN104779611A - Economic dispatch method for micro grid based on centralized and distributed double-layer optimization strategy - Google Patents

Abstract

The invention discloses an economic dispatch method for a micro grid based on a centralized and distributed double-layer optimization strategy. The method is an optimized dispatching method using economical efficiency as an objective under the running state of a single micro grid. The method comprises the following steps: in the centralized-distributed double-layer optimization, a centralized layer corresponds to an advanced optimization process within a predicted period, which is performed based on predicted data of the output power and the load of each uncontrollable unit in the micro grid, and the process of the method is completed by a centralized control center of the micro grid; the distributed optimization corresponds to a real-time optimization process performed based on real-time data of the output power and the micro grid load of each uncontrollable unit of the micro grid at each dispatch time in a predicted period length, and the optimization process is performed in a distributed manner and is completed by a micro power supply controller which is built in each unit of the micro grid. The economic dispatch method disclosed by the invention puts forwards that the economical efficiency of energy storage is not considered, namely the energy storage does not participate in the optimization process of the economical efficiency.

Description

Based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy

Technical field

The present invention relates to a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, belong to micro-capacitance sensor technical field.

Background technology

Along with electrical network scale increases rapidly, because traditional electrical network energy resource consumption is large, environmental pollution is serious, and via net loss is high, is more and more difficult to the requirement meeting user's high reliability and diversified power supply strategy.Distributed power generation has that energy utilization rate is high, environmental pollution is little, reliability is high, the plurality of advantages such as convenient and flexible installation and low cost high repayment, effectively can solve a lot of potential problems of traditional electrical entoilage.Internal combustion engine, fuel cell, miniature gas turbine, regenerative resource that it is fuel that distributed power source mainly comprises with gas or liquid, as: the distributed electrical source position dispersion of wind energy or solar power generation is flexibly, the features such as low-carbon environment-friendly have greatly adapted to resource and the electricity needs of dispersion, delay to cause the huge investment required for electrical power trans mission/distribution system maintenance due to the increase of load, reduced the cost depletions brought due to long-distance transmissions electric energy.In addition, distributed power source and traditional electrical network complement one another, and drastically increase the reliability of power supply.

Microgrid can make distributed power source run flexibly, efficiently, fully excavates value and the benefit of distributed power generation.Microgrid scale, between distributed power generation and bulk power grid, can connect buffering distributed power generation and bulk power grid, also can independent operating.It is the advanced form of distributed power source development.Microgrid is looked at problem from systematic point of view, by combinations such as generator, load, energy storage device and control device, forms a small-sized controlled electrical power trans mission/distribution system.Microgrid has certain energy management capabilities, becomes desirable selection by microgrid access distributed power source.DG in microgrid can be divided into intermittent power supply and continuity power supply two class by characteristics of output power, intermittent power supply comprises wind power generation and photovoltaic generation, its power output is larger by the impact of the natural conditions such as weather, there is obvious fluctuation and uncertainty, continuity power supply comprises miniature gas turbine and fuel cell etc., and it has primary energy supply relatively reliably and continuous print process regulating power.

Current electric power system scale and complexity are in continuous increase, and the data explosion that Power System Interconnection brings increases, and the network coordination problem that introducing generation of electricity by new energy brings, and are all the huge challenges to conventional electric power network analysis.And the present invention can solve problem above well.

Summary of the invention

The object of the invention there are provided a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, the method is applied to distributed scheduling mode in microgrid energy optimum management, in conjunction with traditional centralized scheduling mode, define double-deck centralized and distributed energy optimum management strategy.This distributed method can be transferred all objects that is scheduled and participate in scheduling calculation task, realizes calculation task dispersion treatment, takes full advantage of the computational resource of unit in micro-capacitance sensor.

The present invention solves the technical method that its technical problem takes: a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, micro-capacitance sensor economic dispatch optimizing process is divided into the centralized scheduling layer based on prediction data and the distributed scheduling layer based on real time data by the method, specifically comprises the steps:

Step 1: what the uncontrollable micro battery of micro-capacitance sensor in following predetermined period was collected by the control centre of micro-capacitance sensor dopes force data, exerting oneself of such as wind-powered electricity generation and photovoltaic generation unit, and the load prediction data that micro-capacitance sensor is total.

Step 2: micro-capacitance sensor control centre, according to information of forecasting, take economic optimum as target, under the prerequisite considering each constraints, adopt particle group optimizing method to calculate through row optimization, the distributed power source obtaining whole predetermined period is exerted oneself.Before the Real-Time Scheduling moment arrives, each DG unit is all exerted oneself by prediction and is produced.Its target function is:

min f＝f _DG+f _L

$f_{\mathrm{DG}}=\underset{t=1}{\overset{N}{\mathrm{\Σ}}}\underset{i=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}[U_{\mathrm{DG},i}^t(F_i^t\left(P_{\mathrm{DG},i}^t\right)+K\·P_{\mathrm{DG},i}^t)+c_{\mathrm{DG},i}^{\mathrm{on}}\·U_{\mathrm{DG},i*}^t]$

$f_L=\underset{t=1}{\overset{N}{\mathrm{\Σ}}}\underset{i=1}{\overset{Q_L}{\mathrm{\Σ}}}{c}_L(1-U_{L,i}^t)P_{\mathrm{Li}}^t$

Wherein f represents cost, and U represents state, and synchronization can only get 0 or 1.P represents power.DG represents controlled micro battery, N represents the time hop count divided a whole predetermined period, L represents and can cut off load, Q represents element number, K represents the maintenance cost of micro battery, and c represents price, and on represents the start and stop of micro battery, * represent the change of micro battery state, F represents the cost of electricity-generating function of micro battery.Note optimum results is with the plan being respectively each controlled micro battery is exerted oneself and predicts cutting load state.

The constraints that centralized optimization process is followed is as follows:

(1) power-balance constraint

$\underset{i=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{U}_{\mathrm{DG},i}^t{P}_{\mathrm{DG},i}^t=\underset{i}{\overset{Q_L}{\mathrm{\Σ}}}{U}_{L,i}^t{P}_{L,i}^t-\mathrm{\Σ}{P}_{\mathrm{unctrl}}^t$

Unctrl represents uncontrollable micro battery.

(2) state constraint

$\underset{j=0}{\overset{m}{\mathrm{\Σ}}}{U}_{\mathrm{Li}*}^{t+j}\≤1$

This formula represents can not repeatedly cut off a certain load within a continuous print m scheduling slot.

(3) DG units limits

$\underset{\‾}{P_{\mathrm{DG},i}}\≤P_{\mathrm{DG},i}\≤\stackrel{\‾}{P_{\mathrm{DG},i}}$

$\left|\frac{{\mathrm{\ΔP}}_{\mathrm{DG},i}}{\mathrm{\Δt}}\right|\≤r_i^{\max }$

be respectively the exert oneself upper limit and the lower limit of each DG unit. represent the maximum climbing rate of each unit.

Step 3: when the new Real-Time Scheduling moment arrives, centralized-control center detects the real time data of exerting oneself in real time of uncontrollable power supply and load.According to network service topological diagram particularly, determine the start node of distributed optimization.The error of this scheduling instance is calculated according to real time data and prediction data.And dope at start node amount force value adding error.Error delta P ^tcomputing formula be:

${\mathrm{\ΔP}}^t=\underset{i}{\overset{Q_L}{\mathrm{\Σ}}}{U}_{L,i}^{t,\mathrm{forecast}}(P_{L,i}^{t,\mathrm{realtime}}-P_{L,i}^t)-\mathrm{\Σ}(P_{\mathrm{unctrl}}^{t,\mathrm{realtime}}-P_{\mathrm{unctrl}}^t)$

Real time data with target on realtime.

After selected optimization start node, error is imported into its prediction and exert oneself, computing formula is:

$P_{\mathrm{DG},\mathrm{leader}}^{t,\mathrm{forecast}}=P_{\mathrm{DG},\mathrm{leader}}^{t,\mathrm{forecast}}+{\mathrm{\ΔP}}^t$

Before now starting to perform an action to step 5 micro battery, will regard a global variable as, can be revised by the optimizing process of each node in optimizing process, but before the process concludes, each DG does not adjust and exerts oneself.

Step 4: from start node, travels through by a certain traversal order the communication connected graph of each controlled micro battery composition of micro-capacitance sensor.The micro-capacitance sensor communication topology non-directed graph adjacency matrix of note t is get between 1 expression i and j and have syntople, getting 0 does not have.And if i=j, then micro-capacitance sensor centralized-control center controls the node traverses process of the whole network.Travel through from start node, the node traversed carries out 1 suboptimization calculating, and namely the optimization carried out of each node calculates is not parallel, but the sequencing following traversal is through row.Its target function is:

$\min F_{\mathrm{DG},i}\left(P_{\mathrm{DG},i}^t\right)+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{d}_{\mathrm{ij}}^t{F}_{\mathrm{DG},j}\left(P_{\mathrm{DG},j}^t\right)$

The constraints met is:

(1) power-balance constraint:

$P_{\mathrm{DG},i}^t+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{P}_{\mathrm{DG},j}^t=P_{\mathrm{DG},i}^{t,\mathrm{forecast}}+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{P}_{\mathrm{DG},j}^{t,\mathrm{forecast}}$

(2) DG units limits

$\underset{\‾}{P_{\mathrm{DG},i}}\≤P_{\mathrm{DG},i}\≤\stackrel{\‾}{P_{\mathrm{DG},i}}$

$\left|\frac{\mathrm{\Δ}{P}_{\mathrm{DG},i}}{\mathrm{\Δt}}\right|\≤r_i^{\max }$

Optimum results once just substitutes original by every node calculate afterwards and be adjacent node traveling through completely after net node, carry out convergence judgement, if do not restrain, repeat step 4, convergence then goes to step 5.

Step 5: each controlled micro battery is exerted oneself according to distributed real-time optimization result adjustment.Judge whether that a predetermined period terminates, if do not terminate, then when waiting for that next scheduling instance arrives, and go to step 3.Otherwise terminate.

Said method of the present invention is exerted oneself based on uncontrollable micro battery and the centralized optimization layer of prediction data of load.

Said method of the present invention is exerted oneself based on uncontrollable micro battery and the distributed optimization layer of real time data of load.

Beneficial effect:

1, concentrated layer of the present invention is based on the various prediction data of micro-capacitance sensor, can to controlled micro battery through row pre-scheduling, and according to reliable prediction data, and meet system and respectively retrain and the scheduling result drawn, its result data has reliability.

2, distributed optimization of the present invention is based on real time data and when meeting each constraint, can adjust forecast dispatching result, real-time uncertain demand can be met, make micro-capacitance sensor safe and reliable operation, its distributed optimization computational process, makes calculation task be dispersed to each scheduling node place, reduces the burden of micro-capacitance sensor control centre, and be easy to implement, there is feasibility.

3, the present invention can be applied to the isolated power grid state of the micro-capacitance sensor of various scale, fast response time in force.

4, Distributed Calculation of the present invention can utilize the account form of the same task of computational resource parallel processing of multiple variety classes loose coupling, and compared to the centralized calculating using supercomputer to carry out, it is more cheap, and makes full use of computational resource.

Accompanying drawing explanation

Fig. 1 is method flow diagram of the present invention.

Fig. 2 is the micro-capacitance sensor structural representation of the embodiment of the present invention.

Fig. 3 is the Communication topology schematic diagram of the micro-capacitance sensor of the embodiment of the present invention.

Embodiment

Below in conjunction with Figure of description, the invention is described in further detail.

As shown in Figure 2, micro-capacitance sensor of the present invention is made up of a typhoon power generator (that is: WT), photovoltaic generation unit (that is: PV), two miniature gas turbines (that is: MT), two parts such as diesel engine generator (that is: DE), a group of fuel cell (that is: FC), one group of energy-storage units (that is: Bat) and household loads.

As shown in Figure 1, a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, micro-capacitance sensor economic dispatch optimizing process is divided into the centralized scheduling layer based on prediction data and the distributed scheduling layer based on real time data by the method, specifically comprises the steps:

Step 1: what the uncontrollable micro battery of micro-capacitance sensor in following predetermined period was collected by the control centre of micro-capacitance sensor dopes force data, and total load.Predetermined period that example adopts is 24 hours, is a scheduling instance every 1 hour.The predicted value of typical micro-capacitance sensor family load is as shown in table 1.Wind power generation (WT) and photovoltaic generation (PV) predicted value of exerting oneself is as shown in table 2.

Table 1

Table 2

Step 2: micro-capacitance sensor control centre, according to the information of forecasting of step 1 gained, take economic optimum as target, under the prerequisite considering each constraints, calculate through row optimization, the distributed power source obtaining whole predetermined period is exerted oneself.Before the Real-Time Scheduling moment arrives, each DG unit is all exerted oneself by prediction and is produced.

The constraint of each generator unit is as shown in table 3.

Table 3

Through the calculating of centralized optimization, show that predictions in 24 hours of five controllable distributed power generation unit (being expressed as MT1, MT2, FC, DE1 and DE2) are exerted oneself as shown in table 4.

Table 4

Step 3: as shown in Figure 3, and to determine MT1 be distributed optimization start node to the micro-capacitance sensor Communication topology in embodiment.The new Real-Time Scheduling moment, when arriving, centralized-control center gathered the real time data of exerting oneself in real time of uncontrollable power supply and load.

For the 0th time, prediction data, real time data and error are as shown in table 5

Table 5

On the prediction margin of error being added MT1 is exerted oneself, therefore 0 time MT1 prediction exert oneself and become 29.9645kW, and replace prediction to exert oneself item corresponding in table with it.But before step 5 performs, do not carry out the actual adjustment of exerting oneself.The process of other scheduling instance by that analogy.

Step 4: can obtain adjacency matrix according to the micro-capacitance sensor communication topology figure of Fig. 3 is:

$A^t=\left[\begin{array}{ccccc}0& 1& 1& 0& 0\\ 1& 0& 0& 0& 1\\ 1& 0& 0& 1& 1\\ 0& 0& 1& 0& 0\\ 0& 1& 1& 0& 0\end{array}\right]$

Suppose that the traversal order according to 1-2-3-4-5 from node 1 (MT1) carries out distributed optimization.The target function of optimization when 0 on 1 node is:

$\begin{array}{c}{\min F}_{\mathrm{DG},1}\left(P_{\mathrm{DG},1}^0\right)+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{d}_{1j}^0{F}_{\mathrm{DG},j}\left(P_{\mathrm{DG},j}^0\right)\\ =\min F_{\mathrm{DG},1}\left(P_{\mathrm{DG},1}^0\right)+F_{\mathrm{DG},2}\left(P_{\mathrm{DG},2}^0\right)+F_{\mathrm{DG},3}\left(P_{\mathrm{DG},3}^0\right)\end{array}$

And meet power-balance constraint:

$P_{\mathrm{DG},1}^0+P_{\mathrm{DG},2}^0+P_{\mathrm{DG},3}^0=P_{\mathrm{DG},1}^{0,\mathrm{forecast}}+P_{\mathrm{DG},2}^{0,\mathrm{forecast}}+P_{\mathrm{DG},3}^{0,\mathrm{forecast}}$

Here can be regarded as one by the information in the table that global variable is formed with target amount on forecast, its can be revised by the optimum results of distributed optimization process each time, such as, after 1 node 0 time completes optimization, can use respectively in substitution list process on other nodes by that analogy.

After all nodes of the whole network all travel through and complete optimization calculating, carry out criteria for convergence, namely judge whether the absolute value of the knots modification be worth before and after 1 suboptimization process of exerting oneself of all nodes is less than a certain set-point.Namely be considered as restraining and going to step 5 if meet, otherwise repeat step 4, until result convergence.

Step 5: each controlled micro battery is exerted oneself according to distributed real-time optimization result adjustment.For the 0th time, after distributed optimization, show that the result of exerting oneself of each controlled micro battery is as table 6.

Table 6

DG node	MT1	DE1	FC	MT2	DE2
						Exert oneself (kW)	24.5330	0	22.9334	24.5330	0

Judge whether that a predetermined period terminates, if do not terminate, then when waiting for that next scheduling instance arrives, and go to step 3.Otherwise terminate.

Finally, the distributed optimization result of 24 hours is given in table 7.

Table 7

Claims (8)

1., based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that, described method comprises the steps:

Step 1: the micro-capacitance sensor prediction data in following predetermined period is collected by the control centre of micro-capacitance sensor, and what comprise uncontrollable micro battery dopes force data, that is: the exerting oneself of wind-powered electricity generation and photovoltaic generation unit, and the load data that micro-capacitance sensor is total;

Step 2: micro-capacitance sensor control centre is according to information of forecasting, take economic optimum as target, under the prerequisite considering each constraints, particle group optimizing method is adopted to calculate through row optimization, the distributed power source obtaining whole predetermined period is exerted oneself, and before the Real-Time Scheduling moment arrives, each DG unit is all exerted oneself by prediction and produced;

Step 3: when the new Real-Time Scheduling moment arrives, centralized-control center detects the real time data of exerting oneself in real time of uncontrollable power supply and load, according to network service topological diagram particularly, determine the start node of distributed optimization, calculate the error of this scheduling instance according to real time data and prediction data, and dope at start node amount force value adding error;

Step 4: from start node, the communication connected graph of each controlled micro battery composition of micro-capacitance sensor is traveled through by a certain traversal order, often traverse a node, this node just carries out primary particle group and optimizes calculating under constraints, and optimization aim self is adjacent exerting oneself of node, and uses optimum results to replace prediction to exert oneself, traveling through completely after net node, carry out convergence judgement, if do not restrain, repeat step 4, convergence then goes to step 5;

Step 5: each controlled micro battery is exerted oneself according to distributed real-time optimization result adjustment, judges whether that a predetermined period terminates, if do not terminate, then when waiting for that next scheduling instance arrives, and goes to step 3, otherwise terminate.

2. according to claim 1 a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that: described method is exerted oneself based on uncontrollable micro battery and the centralized optimization layer of prediction data of load.

3. according to claim 1 a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that: described method is exerted oneself based on uncontrollable micro battery and the distributed optimization layer of real time data of load.

4. according to claim 1 a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that, in the step 2 of described method, the target function of centralized optimization is:

minf＝f _DG+f _L

$f_{\mathrm{DG}}=\underset{t=1}{\overset{N}{\mathrm{\Σ}}}\underset{i=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}[U_{\mathrm{DG},i}^t(F_i^t\left(P_{\mathrm{DG},i}^t\right)+K\·P_{\mathrm{DG},i}^t)+c_{\mathrm{DG},i}^{\mathrm{on}}\·U_{\mathrm{DG},i*}^t]$

$f_L=\underset{t=1}{\overset{N}{\mathrm{\Σ}}}\underset{i=1}{\overset{Q_L}{\mathrm{\Σ}}}{c}_L(1-U_{L,i}^t)P_{\mathrm{Li}}^t$

Wherein f represents cost, and U represents state, and synchronization can only get 0 or 1, P represents power, and DG represents controlled micro battery, and N represents the time hop count divided a whole predetermined period, L represents and can cut off load, Q represents element number, and K represents the maintenance cost of micro battery, and c represents price, on represents the start and stop of micro battery, * represent the change of micro battery state, F represents the cost of electricity-generating function of micro battery, and note optimum results is with the plan being respectively each controlled micro battery is exerted oneself and predicts cutting load state.

5. according to claim 4 a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that, the constraints that the centralized optimization process of described method is followed comprises:

(1) power-balance constraint

$\underset{i=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{U}_{\mathrm{DG},i}^t{P}_{\mathrm{DG},i}^t=\underset{i}{\overset{Q_L}{\mathrm{\Σ}}}{U}_{L,i}^t{P}_{L,i}^t-\mathrm{\Σ}{P}_{\mathrm{unctrl}}^t$

Unctrl represents uncontrollable micro battery;

(2) state constraint

$\underset{j=0}{\overset{m}{\mathrm{\Σ}}}{U}_{\mathrm{Li}*}^{t+j}\≤1$

This formula represents can not repeatedly cut off a certain load within a continuous print m scheduling slot;

(3) DG units limits

$P_{\underset{\‾}{\mathrm{DG},i}}\≤P_{\mathrm{DG},i}\≤\stackrel{\‾}{P_{\mathrm{DG},i}}$

$\left|\frac{\mathrm{\Δ}{P}_{\mathrm{DG},i}}{\mathrm{\Δt}}\right|\≤r_i^{\max }$

p _{dG, i} , be respectively the exert oneself upper limit and the lower limit of each DG unit, represent the maximum climbing rate of each unit.

6. according to claim 1 a kind of based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that, the error delta P in the step 3 of described method ^tcomputing formula be:

$\mathrm{\Δ}{P}^t=\underset{i}{\overset{Q_L}{\mathrm{\Σ}}}{U}_{L,i}^{t,\mathrm{forecast}}(P_{L,i}^{t,\mathrm{realtime}}-P_{L,i}^t)-\mathrm{\Σ}(P_{\mathrm{unctrl}}^{t,\mathrm{realtime}}-P_{\mathrm{unctrl}}^t)$

Real time data with target on realtime;

After selected optimization start node leader, error is imported into its prediction and exert oneself, computing formula is:

$P_{\mathrm{DG},\mathrm{leader}}^{t,\mathrm{forecast}}=P_{\mathrm{DG},\mathrm{leader}}^{t,\mathrm{forecast}}+\mathrm{\Δ}{P}^t$

Before now starting to perform an action to step 5 micro battery, will regard a global variable as, can be revised by the optimizing process of each node in optimizing process, but before the process concludes, each DG does not adjust and exerts oneself.

7. according to claim 1 based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that, in the step 4 of described method, remember that the micro-capacitance sensor communication topology non-directed graph adjacency matrix of t is get between 1 expression i and j and have syntople, getting 0 does not have, and if i=j, then micro-capacitance sensor centralized-control center controls the node traverses process of the whole network; Travel through from leader node, the node traversed carries out 1 suboptimization calculating, and namely the optimization carried out of each node calculates is not parallel, but the sequencing following traversal is through row, and its target function is:

$\min F_{\mathrm{DG},i}\left(P_{\mathrm{DG},i}^t\right)+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{d}_{\mathrm{ij}}^t{F}_{\mathrm{DG},j}\left(P_{\mathrm{DG},j}^t\right)$

The constraints met is:

(1) power-balance constraint:

$P_{\mathrm{DG},i}^t+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{P}_{\mathrm{DG},j}^t=P_{\mathrm{DG},i}^{t,\mathrm{forecast}}+\underset{j=1}{\overset{Q_{\mathrm{DG}}}{\mathrm{\Σ}}}{P}_{\mathrm{DG},j}^{t,\mathrm{forecast}}$

(2) DG units limits

$P_{\underset{\‾}{\mathrm{DG},i}}\≤P_{\mathrm{DG},i}\≤\stackrel{\‾}{P_{\mathrm{DG},i}}$

$\left|\frac{\mathrm{\Δ}{P}_{\mathrm{DG},i}}{\mathrm{\Δt}}\right|\≤r_i^{\max }$

Optimum results once just substitutes original by every node calculate afterwards and be adjacent node

8. according to claim 1 based on centralized and micro-capacitance sensor economic dispatch method that is distributed type double optimisation strategy, it is characterized in that: described method is that micro-capacitance sensor economic dispatch optimizing process is divided into the centralized scheduling layer based on prediction data and the distributed scheduling layer based on real time data.