CN109861305A - A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL - Google Patents
A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL Download PDFInfo
- Publication number
- CN109861305A CN109861305A CN201910096258.9A CN201910096258A CN109861305A CN 109861305 A CN109861305 A CN 109861305A CN 201910096258 A CN201910096258 A CN 201910096258A CN 109861305 A CN109861305 A CN 109861305A
- Authority
- CN
- China
- Prior art keywords
- power
- transmission
- distribution
- output
- active
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000009826 distribution Methods 0.000 title claims abstract description 78
- 230000005540 biological transmission Effects 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000005457 optimization Methods 0.000 claims abstract description 31
- 238000004146 energy storage Methods 0.000 claims abstract description 30
- 238000005096 rolling process Methods 0.000 claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 230000008569 process Effects 0.000 claims abstract description 3
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000007599 discharging Methods 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 3
- 238000013178 mathematical model Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000010248 power generation Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 125000004432 carbon atom Chemical group C* 0.000 claims description 2
- 238000013135 deep learning Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses the collaboration economic load dispatching methods of a kind of power transmission network of binding model PREDICTIVE CONTROL and power distribution network, this method is on the basis of the transmission & distribution optimal load flow of the long time scale based on master slave splitting method, short-term time scale is based on Model Predictive Control, optimized using multistep dynamic rolling, solve active and idle power output increment, the power output of the energy storage and distributed generation resource that access in major network and distribution is adjusted, the process for controlling power output is more smooth；Short-term forecast link is handled sequence data with LSTM neural network, is predicted using deep learning method, to improve prediction and control precision；The dispatching method can be realized the collaboration optimization of power transmission network and power distribution network, and the economical and efficient of the efficient consumption and power grid of realizing new energy is run.
Description
Technical Field
The invention relates to an economic dispatching method of a power system, in particular to a transmission and distribution cooperative economic dispatching method combining model prediction control.
Background
With the increasing penetration of distributed power sources in power systems, particularly in power distribution networks, in recent years, the operational control of power systems has become increasingly complex. In order to cope with the access of high permeability distributed power sources and energy storage devices, researchers have proposed the concept of an active power distribution network. The active power distribution network can carry out combined control on the power generator, the load, the energy storage and other resources in the distribution network, and flexible power flow control is carried out by utilizing a network topological structure. Because the output of the distributed power supply has strong uncertainty, how to reasonably arrange the power supply output and utilize a demand response technology is a problem to be solved urgently at present to realize the economic and safe operation of the active power distribution network. In addition, the transmission network is inevitably affected by renewable energy, and how to maintain the real-time balance of the power grid becomes a great challenge for the transmission network. In view of the above problems, it is necessary to propose a new control framework and a control method.
The coordinated economic dispatch of transmission and distribution networks can solve many of the problems described above. When sudden load increases cause voltage drops and frequency fluctuations in the transmission network, the distributed power and stored energy in the distribution network can support the transmission network by increasing the reactive output. This problem is helped by the transmission and distribution coordination when the transformer and corresponding line connecting the distribution and transmission network are congested.
Many of the methods currently used by electric power companies are implemented independently in transmission and distribution networks, lack global considerations, and are therefore difficult to achieve economic optimization. There is no way to reasonably adjust power output and implement demand response when there are multiple places with overload scenarios. Meanwhile, the current scheduling still depends on the experience of scheduling personnel in many situations, and the scheduling personnel are required to deeply understand and master the operation characteristics of the system and the equipment, so that the working pressure of the scheduling personnel is increased.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention aims to provide an economic dispatching method of a transmission and distribution system based on model prediction control by considering optimal power flow calculation and distributed power output control of the power system so as to realize global cooperative optimization and economic dispatching of the transmission network and the distribution network.
The technical scheme is as follows: a transmission and distribution cooperative economic dispatching method combined with model prediction control comprises the following contents:
(1) on a long time scale, obtaining an optimal power flow under transmission and distribution cooperation based on a master-slave splitting method; and the transmission network and the distribution network calculate and disassemble the global optimal power flow into optimal power flow calculation operator problems of a main network and a plurality of distribution networks by exchanging boundary information. According to actual conditions, the main network and the distribution network can respectively select proper methods for calculation.
(2) On a short time scale, ultra-short-term prediction is carried out on the output of the distributed power supply based on an LSTM neural network, a model prediction control method is adopted for carrying out short-term rolling optimization, and active and reactive output increments are solved;
(3) and taking the optimal power flow calculation result of the long-time scale as a reference value, taking the predicted value of the LSTM as a reference, and adjusting and correcting the output of the energy storage and distributed power supply accessed in the main network and the distribution network according to the short-term rolling optimization result.
In the step (1), a master-slave splitting method is utilized to decompose the power transmission and distribution system into a master system, a slave system and a boundary system, and a power transmission and distribution optimal power flow model taking the master system as a center is as follows:
mincM(uM,uB,xM,xB)+cs(uS,xB,xS)
constraint conditions are as follows:
fM(uM,xM,xB)＝0
gM(uM,xM,xB)＝0
fB(uB,xM,xB,xS)＝0
gB(uB,xB)≥0
fS(uS,xB,xS)＝0
gS(uS,xB,xS)≥0
wherein, cMAnd cSOptimized objective function for representing transmission and distribution networks respectivelyCounting; x is a state variable; u represents a control variable; superscripts M, B and S denote the master system, boundary system, and slave system, respectively;
the boundary power flow equation is equivalent to:
wherein,representing the active power flow from the main system into the border system,representing the active power flow into the slave system by the boundary system;representing reactive power flow into the boundary from the main system, fQ BSRepresenting the reactive power flow from the boundary system into the slave system;
by introducing auxiliary functionsEquivalently decomposing the boundary power flow equation into:
fMB(uB,xM,xB)＝yBS
yBS＝fBS(xB,xS)
constructing a Lagrange function of the transmission and distribution cooperative optimal power flow, namely an objective function:
wherein λ represents an equality constraint multiplier, and ω represents a non-negative multiplier that is not equality constraint; superscript T represents the transpose of the matrix; and solving the Lagrange equation through a heterogeneous decomposition algorithm to obtain the optimal power flow under the transmission and distribution coordination.
In the step (1), an economic dispatching model with the minimum active power optimization dispatching cost of the active power distribution network is adopted, namely:
in the formula: t is1Optimizing a scheduling period for a long time scale; m, NS and NG are respectively flexible load, energy storage device and controllable distributed power supply quantity; c. Cgrid(t) is the time-of-use electricity price of the power grid; pgrid(t) is the active power of the main network connecting line; cloadi(t) flexible load scheduling cost; cstarj(t) battery scheduling cost; cDG(t) is the controllable distributed power scheduling cost; the specific mathematical model for each scheduling cost is as follows:
(1.1) Flexible load scheduling cost
The relationship between the flexible load scheduling cost and the power variation is as follows:
in the formula, αi、βiRespectively scheduling cost coefficients for the flexible load; pload0An active initial value before flexible load scheduling is obtained; delta PloadiIs the flexible load power variation;
(1.2) scheduling cost of battery energy storage
The energy storage battery is respectively charged and discharged during operation, and the charging or discharging can affect the service life of the energy storage battery, and the scheduling cost of energy storage is assumed to be
In the formula, λessA scheduling cost coefficient for the battery; pstorj(t) is the charge and discharge power of the energy storage battery;
(1.3) controllable distributed power scheduling cost:
in the formula, ak、bk、ckScheduling a cost coefficient for the controllable distributed power supply; pDGkAnd (t) is the active power output of the controllable distributed power supply.
In the step (2), the ultra-short term prediction uses three types of data S1, W2 and S1W2, and respectively corresponds to the situation that only the new energy output data of the current day, only the weather data of the next day, the output data of the current day and the weather data of the next day are input together; setting the new energy processing data of the next day as S2 as a calibration value output by the model;
defining a data set to contain N samples, wherein each sample is composed of S1, W2 and S2 data, the sequence length of S1 and S2 is T, and the sequence length of W2 is 13T; w2ijRepresenting j dimension weather characteristic data of the ith time point; s1iRepresenting the new energy output data of the day at the ith time point, S2iThe same process is carried out; with S1W2 as input, the input sequence length of the model is T, and the corresponding input at each time i is S1i,W2i1,W2i2,…W2inFor n +1 dimensional data.
In the LSTM neural network structure in the step (2), input data are sequentially input into a model network according to a time sequence, a time sequence relation between the input data is dynamically captured through two layers of LSTM networks, and then the input data are sent into a full connection layer (FC) to dynamically output a new energy power generation power prediction result obtained through prediction.
Preferably, a Dropout layer is added to the first layer of the fully-connected layer to prevent overfitting.
Further preferably, a ReLU function is selected as an activation function of the model network, and the nonlinear expression capability of the network is improved.
In the step (2), if the active power output issued by the long-time scale is used as a reference value and the short-time scale optimization target is that the correction deviation of the active power output is minimum, establishing a secondary optimization performance index of the short-time scale active power optimization scheduling based on model prediction control as follows:
in the above formula, the first and second carbon atoms are,the active output reference value of each part of the power distribution network at the k + i moment is represented as follows:
p (k + i | k) represents the output of each controllable distributed power supply at the k moment and the predicted future k + i moment:
predicting the state of charge of the energy storage battery at the future k + i moment by using the soc (k + i | k) as the k moment; socminAnd socmaxRespectively an upper limit and a lower limit of the state of charge;predicting the future k + i moment net load for k moment; ploss(k + i | k) predicting the active loss of the system at the future k + i moment for the k moment;
ΔuT(k + i | k) predicting the active output variation column vector of each controllable distributed power supply at the future k + i moment for the k moment, and optimally solving the active variation at the future N moments by using a sequence quadratic programming method:
{ΔuT(k+1|k),ΔuT(k+2|k),…ΔuT(k+n|k)}；
issuing a first control variable column vector in the control variable sequence, and solving the active power output of the controllable distributed power supply, the energy storage and the flexible load of the active power distribution network at the moment of k + 1:
P(k+1|k)＝Po(k)+ΔuT(k+1|k)。
in the step (3), a closed-loop control is formed by taking the current actual active output value of the system as an initial value of a new round of rolling optimization scheduling:
P(k+1|k)＝Po(k)+ΔuT(k+1|k)
in the formula, Preal(k +1) after the active output value at the k moment is issued, acquiring the active output value at the k +1 moment through an actual measurement system; po(k +1) represents the initial value of the active output at the moment of k +1
Has the advantages that: compared with the prior art, the invention has the following remarkable progress: the new energy ultra-short term output prediction and model prediction control method based on the LSTM neural network realizes optimal closed-loop control performance in an optimal time period in the future, effectively corrects prediction errors and optimized scheduling result deviation generated by random factors through feedback correction, improves the precision of optimization control, and realizes the aims of multi-level coordination and progressive refinement and the maximum consumption of a distributed power supply. In addition, by using an optimal power flow algorithm based on master-slave splitting, cooperative scheduling between the power transmission network and the power distribution network is realized, support to the power transmission network under specific conditions is realized by using flexible loads, distributed power supplies and energy storage in the power distribution network, and the economical efficiency and safety of power grid operation are integrally improved.
Drawings
FIG. 1 is a schematic diagram of an LSTM neural network architecture for contribution prediction;
FIG. 2 is a schematic diagram of a power transmission and distribution system with a master-slave split architecture;
fig. 3 shows the optimization results.
Detailed Description
The key technology and specific implementation method of the present invention are described in detail below with reference to the accompanying drawings and specific embodiments. LSTM-based ultra-short term output prediction of new energy
(1) Data structure
The ultra-short term prediction needs to use three types of data, namely S1, W2 and S1W2, and respectively corresponds to the situation that only new energy output data of the current day, only weather data of the next day, output data of the current day and weather data of the next day are input together. The new energy processing data of the next day is set as S2 as a calibration value of the model output.
The data set is defined to contain N samples, each sample consisting of S1, W2, S2 data. Wherein the sequence lengths of S1 and S2 are T respectively, and the sequence length of W2 is 13T. W2ijAnd j dimension weather characteristic data representing the ith time point. S1iRepresenting the new energy output data of the day at the ith time point, S2iThe same is true. With S1W2 as input, the input sequence length of the model is T, and the corresponding input at each time i is S1i,W2i1,W2i2,…W2inFor n +1 dimensional data.
(2) Network architecture
The hidden layer structure is shown in fig. 1. t is t1…tnIs the input data of the model, t1'…tn' is the output data of the model. Input data are sequentially input in time sequenceAnd entering a model network, dynamically capturing the time sequence relation between input data through two layers of LSTM networks, sending the time sequence relation to a full connection layer (FC), and dynamically outputting a predicted new energy power generation power prediction result. As can be seen from the structure, t1Data at time' mainly covers t1Information of time of day, t2Data at time' mainly covers t1、t2And t1Information on' time of day, and so on.
According to the related experience, a Dropout layer is added to the FC first layer to prevent overfitting. And a ReLU function is selected as an activation function of the model, so that the nonlinear expression capability of the network is improved.
Secondly, calculating the optimal power flow of the transmission and distribution system
The power transmission and distribution system is decomposed into a main system, a slave system and a boundary system by using a master-slave splitting method, as shown in fig. 2. The boundary system can be considered as a high-voltage bus or a low-voltage bus on the distribution network side. A complete power transmission system is composed of a main system and a boundary system, and a corresponding power distribution system is composed of a slave system and a boundary system. The transmission and distribution optimal power flow model with the main system as the center can be described by the following formula:
mincM(uM,uB,xM,xB)+cs(uS,xB,xS) (1)
constraint conditions are as follows:
where x is a state variable such as the magnitude and phase angle of the voltage on each branch. u denotes control variables, such as active and reactive output power of renewable energy sources. c. CMAnd cSRespectively representing the optimization objective functions of the transmission network and the distribution network. M, B and the S superscript denote the master system, the border system and the slave system, respectively.
All the flows through the border system can be divided into three groups, one from the master system, one from the slave system, one from the generators, loads, etc. connected to the border system. The boundary power flow equation can be equivalent to:
wherein,representing the active power flow from the main system into the border system,representing the active power flow from the boundary system into the slave system.Representing the reactive power flow from the main system into the boundary,representing the reactive power flow from the boundary system into the slave system. By introducing auxiliary functionsThe formula (3) can be equivalently decomposed into
On the basis, a Lagrange function, namely an objective function, of the transmission and distribution cooperative optimal power flow can be constructed:
where λ represents an equality constraint multiplier and ω represents a non-negative multiplier that is not equality constraint. The superscript T represents the transpose of the matrix. And solving the Lagrange equation through a heterogeneous decomposition algorithm to obtain the optimal power flow under the transmission and distribution coordination.
Third, join in marriage net economic dispatch model
The economic dispatching model with the minimum active power optimization dispatching cost of the active power distribution network is adopted, namely:
in the formula: t is1Optimizing a scheduling period for a long time scale; m, NS and NG are respectively flexible load, energy storage device and controllable distributed power supply quantity; c. Cgrid(t) is the time-of-use electricity price of the power grid; pgrid(t) is the active power of the main network connecting line; cloadi(t) flexible load scheduling cost; cstarj(t) battery scheduling cost; cDGAnd (t) scheduling cost for the controllable distributed power supply. The specific mathematical model for each scheduling cost is as follows:
(1) flexible load scheduling cost
The relationship between the flexible load scheduling cost and the power variation is as follows:
in the formula αi、βiRespectively scheduling cost coefficients for the flexible load; pload0An active initial value before flexible load scheduling is obtained; delta PloadiIs the flexible load power variation.
(2) Scheduling cost of battery energy storage
The energy storage battery is respectively charged and discharged during operation, and the charging or discharging can affect the service life of the energy storage battery, and the scheduling cost of energy storage is assumed to be
In the formula: lambda [ alpha ]essA scheduling cost coefficient for the battery; pstorjAnd (t) is the charge and discharge power of the energy storage battery.
(3) Controllable distributed power source scheduling cost:
in the formula: a isk、bk、ckScheduling a cost coefficient for the controllable distributed power supply; pDGkAnd (t) is the active power output of the controllable distributed power supply.
Fourth, short-term scheduling model based on model predictive control
And (3) on the basis of a model prediction control principle in a short time scale, taking the transmission and distribution cooperative optimal power flow result as a reference value and taking the predicted value of the LSTM as a reference value, rolling to obtain an active power output increment of 15min in the future, and only executing a control instruction in the 1 st time period each time to correct the output of the distributed power supply.
(1) Short timescale optimization objective function
And taking the active output issued by the long-time scale as a reference value, and taking the short-time scale optimization target as the minimum correction deviation of the active output, and establishing a short-time scale active optimization scheduling secondary optimization performance index based on model prediction control as follows:
in the formula,the active output reference values of each part of the power distribution network at the k + i moment are represented, and can be calculated as follows:
p (k + i | k) represents the output of each controllable distributed power supply at the k moment and the predicted future k + i moment, and specifically comprises the following steps:
wherein: predicting the state of charge of the energy storage battery at the future k + i moment for the k moment by soc (k + i | k); socminAnd socmaxRespectively an upper limit and a lower limit of the state of charge;predicting the future k + i moment net load for k moment; plossAnd (k + i | k) predicting the active loss of the system at the future k + i moment for the k moment.
ΔuT(k + i | k) predicting the active output variation column vector of each controllable distributed power supply at the future k + i moment for the k moment, and optimally solving the active variation at the future N moments by using a sequence quadratic programming method:
{ΔuT(k+1|k),ΔuT(k+2|k),…ΔuT(k+n|k)} (14)
issuing a first control variable column vector in the control variable sequence, and solving the active power output of the controllable distributed power supply, the energy storage and the flexible load of the active power distribution network at the moment of k + 1:
P(k+1|k)＝Po(k)+ΔuT(k+1|k) (15)
(2) feedback correction (adjustment/correction)
Due to uncertainty of new energy output prediction, the advanced MPC control cannot ensure that wind power and photovoltaic output are the same as predicted values, so that deviation exists between the output of the controllable distributed power source issued in advance and the actual active output, and feedback correction is needed. The current actual active output value of the system is used as the initial value of a new round of rolling optimization scheduling, closed-loop control is formed, uncertainty of the system, wind power and photovoltaic is overcome, the new round of active output predicted value is more practical to fit, and the precision is higher, namely:
P(k+1|k)＝Po(k)+ΔuT(k+1|k) (16)
in the formula: preal(k +1) after the active output value at the k moment is issued, acquiring the active output value at the k +1 moment through an actual measurement system; poAnd (k +1) represents an initial value of the active power output at the moment k + 1.
Examples
Taking a transmission and distribution system consisting of an IEEE14 node transmission network and a modified IEEE5 node distribution network as an example, establishing a transmission and distribution cooperative global model according to the existing transmission and distribution network model, and performing simulation verification by using corresponding data.
The transmission and distribution cooperative scheduling model has the following specific implementation mode:
a. establishing a load, wind power and photovoltaic prediction model;
b. carrying out global optimal power flow calculation on the transmission and distribution network to generate a long-period scheduling plan;
c. solving a control variable by taking an active output actual value of the current distributed power supply of the distribution network as an initial value;
d. issuing a control variable, and adjusting the output of the distributed power supply, the stored energy and the flexible load;
e. and c, ending the optimization period and returning to the step a.
And (3) taking the main line power flow as an optimization result to show, wherein the active power flow optimization result is as follows (unit: MW):
table 1 main network active power flow optimization result of transmission and distribution system
By analyzing the calculation results, the power flow distribution of the power transmission network and the power distribution network can be effectively improved by the power transmission and distribution cooperative economic dispatching method based on model predictive control, meanwhile, the active marginal cost of each node of the main network is reduced, the operation cost of the system is reduced, and the overall operation economy and the new energy consumption capacity of the system are improved.
Claims (9)
1. A transmission and distribution cooperative economic dispatching method combined with model prediction control is characterized by comprising the following steps:
(1) on a long time scale, obtaining an optimal power flow under transmission and distribution cooperation based on a master-slave splitting method;
(2) on a short time scale, ultra-short-term prediction is carried out on the output of the distributed power supply based on an LSTM neural network, a model prediction control method is adopted for carrying out short-term rolling optimization, and active and reactive output increments are solved;
(3) and taking the optimal power flow calculation result of the long-time scale as a reference value, taking the predicted value of the LSTM as a reference, and adjusting and correcting the output of the energy storage and distributed power supply accessed in the main network and the distribution network according to the short-term rolling optimization result.
2. The transmission and distribution cooperative economic dispatching method according to claim 1, wherein: in the step (1), a master-slave splitting method is utilized to decompose the power transmission and distribution system into a master system, a slave system and a boundary system, and a power transmission and distribution optimal power flow model taking the master system as a center is as follows:
mincM(uM,uB,xM,xB)+cs(uS,xB,xS)
constraint conditions are as follows:
fM(uM,xM,xB)＝0
gM(uM,xM,xB)＝0
fB(uB,xM,xB,xS)＝0
gB(uB,xB)≥0
fS(uS,xB,xS)＝0
gS(uS,xB,xS)≥0
wherein, cMAnd cSRespectively representing the optimization objective functions of the transmission network and the distribution network; x is a state variable; u represents a control variable; superscripts M, B and S denote the master system, boundary system, and slave system, respectively;
the boundary power flow equation is equivalent to:
wherein,representing the active power flow from the main system into the border system,representing the active power flow into the slave system by the boundary system;representing reactive power flow into the boundary from the main system, fQ BSRepresenting the reactive power flow from the boundary system into the slave system;
by introducing auxiliary functionsEquivalently decomposing the boundary power flow equation into:
fMB(uB,xM,xB)＝yBS
yBS＝fBS(xB,xS)
constructing a Lagrange function of the transmission and distribution cooperative optimal power flow, namely an objective function:
wherein λ represents an equality constraint multiplier, and ω represents a non-negative multiplier that is not equality constraint; superscript T represents the transpose of the matrix; and solving the Lagrange equation through a heterogeneous decomposition algorithm to obtain the optimal power flow under the transmission and distribution coordination.
3. The transmission and distribution cooperative economic dispatching method according to claim 1, wherein: in the step (1), an economic dispatching model with the minimum active power optimization dispatching cost of the active power distribution network is adopted, namely:
in the formula: t is1For a long timeOptimizing a scheduling period; m, NS and NG are respectively flexible load, energy storage device and controllable distributed power supply quantity; c. Cgrid(t) is the time-of-use electricity price of the power grid; pgrid(t) is the active power of the main network connecting line; cloadi(t) flexible load scheduling cost; cstarj(t) battery scheduling cost; cDG(t) is the controllable distributed power scheduling cost; the specific mathematical model for each scheduling cost is as follows:
(1.1) Flexible load scheduling cost
The relationship between the flexible load scheduling cost and the power variation is as follows:
in the formula, αi、βiRespectively scheduling cost coefficients for the flexible load; pload0An active initial value before flexible load scheduling is obtained; delta PloadiIs the flexible load power variation;
(1.2) scheduling cost of battery energy storage
The energy storage battery is respectively charged and discharged during operation, and the charging or discharging can affect the service life of the energy storage battery, and the scheduling cost of energy storage is assumed to be
In the formula, λessA scheduling cost coefficient for the battery; pstorj(t) is the charge and discharge power of the energy storage battery;
(1.3) controllable distributed power scheduling cost:
in the formula, ak、bk、ckScheduling a cost coefficient for the controllable distributed power supply; pDGkAnd (t) is the active power output of the controllable distributed power supply.
4. The transmission and distribution cooperative economic dispatching method according to claim 1, wherein: in the step (2), the ultra-short term prediction uses three types of data S1, W2 and S1W2, and respectively corresponds to the situation that only the new energy output data of the current day, only the weather data of the next day, the output data of the current day and the weather data of the next day are input together; setting the new energy processing data of the next day as S2 as a calibration value output by the model;
defining a data set to contain N samples, wherein each sample is composed of S1, W2 and S2 data, the sequence length of S1 and S2 is T, and the sequence length of W2 is 13T; w2ijRepresenting j dimension weather characteristic data of the ith time point; s1iRepresenting the new energy output data of the day at the ith time point, S2iThe same process is carried out; with S1W2 as input, the input sequence length of the model is T, and the corresponding input at each time i is S1i,W2i1,W2i2,…W2inFor n +1 dimensional data.
5. The transmission and distribution cooperative economic dispatching method according to claim 1, wherein: in the LSTM neural network structure in the step (2), input data are sequentially input into a model network according to a time sequence, a time sequence relation between the input data is dynamically captured through two layers of LSTM networks, then the input data are sent into a full-connection layer, and a new energy power generation power prediction result obtained through prediction is dynamically output.
6. The transmission and distribution cooperative economic dispatching method according to claim 5, wherein: and adding a Dropout layer on the first layer of the full connection layer.
7. The transmission and distribution cooperative economic dispatching method according to claim 5, wherein: and selecting a ReLU function as an activation function of the model network.
8. The transmission and distribution cooperative economic dispatching method according to claim 1, wherein: in the step (2), if the active power output issued by the long-time scale is used as a reference value and the short-time scale optimization target is that the correction deviation of the active power output is minimum, establishing a secondary optimization performance index of the short-time scale active power optimization scheduling based on model prediction control as follows:
in the above formula, the first and second carbon atoms are,the active output reference value of each part of the power distribution network at the k + i moment is represented as follows:
p (k + i | k) represents the output of each controllable distributed power supply at the k moment and the predicted future k + i moment:
predicting the state of charge of the energy storage battery at the future k + i moment by using the soc (k + i | k) as the k moment; socminAnd socmaxRespectively an upper limit and a lower limit of the state of charge;predicting the future k + i moment net load for k moment; ploss(k + i | k) predicting the active loss of the system at the future k + i moment for the k moment;
ΔuT(k + i | k) predicting the active output variation column vector of each controllable distributed power supply at the future k + i moment for the k moment, and optimally solving the active variation at the future N moments by using a sequence quadratic programming method:
{ΔuT(k+1|k),ΔuT(k+2|k),…ΔuT(k+n|k)}；
issuing a first control variable column vector in the control variable sequence, and solving the active power output of the controllable distributed power supply, the energy storage and the flexible load of the active power distribution network at the moment of k + 1:
P(k+1|k)＝Po(k)+ΔuT(k+1|k)。
9. the transmission and distribution cooperative economic dispatching method as recited in claim 1, wherein: in the step (3), a closed-loop control is formed by taking the current actual active output value of the system as an initial value of a new round of rolling optimization scheduling:
P(k+1|k)＝Po(k)+ΔuT(k+1|k)
in the formula, Preal(k +1) after the active output value at the k moment is issued, acquiring the active output value at the k +1 moment through an actual measurement system; poAnd (k +1) represents an initial value of the active power output at the moment k + 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910096258.9A CN109861305A (en) | 2019-01-31 | 2019-01-31 | A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910096258.9A CN109861305A (en) | 2019-01-31 | 2019-01-31 | A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109861305A true CN109861305A (en) | 2019-06-07 |
Family
ID=66897209
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910096258.9A Pending CN109861305A (en) | 2019-01-31 | 2019-01-31 | A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109861305A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110334879A (en) * | 2019-07-11 | 2019-10-15 | 华北电力大学 | Power grid bus reactive load forecasting method and device |
| CN110390432A (en) * | 2019-07-22 | 2019-10-29 | 新奥数能科技有限公司 | Energy dynamically optimized scheduling method, apparatus and electronic equipment for comprehensive energy |
| CN111146785A (en) * | 2019-12-17 | 2020-05-12 | 华北电力大学 | Improved active set-based fast solving method for multi-source optimized scheduling calculation |
| CN111463794A (en) * | 2020-04-29 | 2020-07-28 | 山东大学 | Transmission and distribution network cooperative power flow control method and system with additional earth impedance |
| CN113141008A (en) * | 2021-04-23 | 2021-07-20 | 国网陕西省电力公司电力科学研究院 | Data-driven power distribution network distributed new energy consumption capacity assessment method |
| CN113688348A (en) * | 2021-07-16 | 2021-11-23 | 贵州万峰电力股份有限公司 | Controllable load distributed coordination control method, device and system based on dynamic network switching topology |
| CN118199138A (en) * | 2024-01-31 | 2024-06-14 | 江苏东港能源投资有限公司 | High-proportion new energy consumption method based on incremental power distribution network |
-
2019
- 2019-01-31 CN CN201910096258.9A patent/CN109861305A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 董雷等: "《基于模型预测控制的主动配电网多时间尺度动态优化调度》", 《中国电机工程学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110334879A (en) * | 2019-07-11 | 2019-10-15 | 华北电力大学 | Power grid bus reactive load forecasting method and device |
| CN110390432A (en) * | 2019-07-22 | 2019-10-29 | 新奥数能科技有限公司 | Energy dynamically optimized scheduling method, apparatus and electronic equipment for comprehensive energy |
| CN111146785A (en) * | 2019-12-17 | 2020-05-12 | 华北电力大学 | Improved active set-based fast solving method for multi-source optimized scheduling calculation |
| CN111463794A (en) * | 2020-04-29 | 2020-07-28 | 山东大学 | Transmission and distribution network cooperative power flow control method and system with additional earth impedance |
| CN113141008A (en) * | 2021-04-23 | 2021-07-20 | 国网陕西省电力公司电力科学研究院 | Data-driven power distribution network distributed new energy consumption capacity assessment method |
| CN113141008B (en) * | 2021-04-23 | 2023-06-13 | 国网陕西省电力公司电力科学研究院 | Data-driven power distribution network distributed new energy consumption capability assessment method |
| CN113688348A (en) * | 2021-07-16 | 2021-11-23 | 贵州万峰电力股份有限公司 | Controllable load distributed coordination control method, device and system based on dynamic network switching topology |
| CN113688348B (en) * | 2021-07-16 | 2024-02-27 | 贵州万峰电力股份有限公司 | Controllable load distributed coordination control method, device and system based on dynamic network switching topology |
| CN118199138A (en) * | 2024-01-31 | 2024-06-14 | 江苏东港能源投资有限公司 | High-proportion new energy consumption method based on incremental power distribution network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109861305A (en) | A kind of transmission & distribution collaboration economic load dispatching method of binding model PREDICTIVE CONTROL | |
| Ishraque et al. | Optimization of load dispatch strategies for an islanded microgrid connected with renewable energy sources | |
| CN108683179B (en) | Active power distribution network optimal scheduling method and system based on mixed integer linear programming | |
| Nguyen et al. | Exact optimal power dispatch in unbalanced distribution systems with high PV penetration | |
| Ranamuka et al. | Flexible AC power flow control in distribution systems by coordinated control of distributed solar-PV and battery energy storage units | |
| Velasquez et al. | Distributed model predictive control for economic dispatch of power systems with high penetration of renewable energy resources | |
| Meng et al. | Hierarchical SCOPF considering wind energy integration through multiterminal VSC-HVDC grids | |
| Maknouninejad et al. | Cooperative control for self-organizing microgrids and game strategies for optimal dispatch of distributed renewable generations | |
| CN112803505B (en) | Distributed voltage source converter cooperative control method and alternating current-direct current parallel-serial micro-grid | |
| CN113516278B (en) | Active power distribution network multi-time scale active and reactive power coordination optimization scheduling method and system | |
| CN113794244B (en) | Pricing and optimal energy scheduling method and system containing multi-microgrid active power distribution system | |
| Babaiahgari et al. | Coordinated control and dynamic optimization in DC microgrid systems | |
| CN112886624B (en) | Three-station-in-one substation energy storage device planning and designing system and method | |
| CN113783193B (en) | Rural energy supply and utilization system optimization regulation and control method and system based on edge-end cooperation | |
| CN117077974A (en) | Virtual power plant resource optimal scheduling method, device, equipment and storage medium | |
| CN104253439B (en) | Battery energy storage power station reactive power is distributed and control method | |
| Raducu et al. | Design and implementation of a hybrid power plant controller | |
| CN111555369A (en) | Medium-voltage and low-voltage collaborative optimization method for power distribution network | |
| CN109980656B (en) | Distributed reactive power optimization and voltage regulation and control method for power distribution network under two-layer cooperative architecture | |
| Wang et al. | Bi-stage operation optimization of active distribution networks with soft open point considering violation risk | |
| Sivadanam et al. | Dynamic performance enhancement of interconnected hybrid thermal power system in the presence of electric vehicles | |
| Zamani et al. | Integration of PV generation and storage on power distribution systems using MPC | |
| Wang et al. | Hierarchically coordinated optimization of power distribution systems with soft open points and electric vehicles | |
| Kim et al. | Optimal generation rescheduling for meshed AC/HIS grids with multi-terminal voltage source converter high voltage direct current and battery energy storage system | |
| CN108171384A (en) | One kind is based on composite particle swarm optimization algorithm microgrid energy management method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190607 |