CN109638883B - Power grid voltage control method and device, computer equipment and storage medium - Google Patents
Power grid voltage control method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109638883B CN109638883B CN201811641746.5A CN201811641746A CN109638883B CN 109638883 B CN109638883 B CN 109638883B CN 201811641746 A CN201811641746 A CN 201811641746A CN 109638883 B CN109638883 B CN 109638883B
- Authority
- CN
- China
- Prior art keywords
- grid
- power supply
- voltage
- connected point
- power
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000002347 injection Methods 0.000 claims abstract description 156
- 239000007924 injection Substances 0.000 claims abstract description 156
- 238000005457 optimization Methods 0.000 claims abstract description 73
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000004891 communication Methods 0.000 claims abstract description 29
- 230000006870 function Effects 0.000 claims description 106
- 230000009467 reduction Effects 0.000 claims description 28
- 238000004590 computer program Methods 0.000 claims description 13
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 6
- 230000007547 defect Effects 0.000 description 4
- 238000011217 control strategy Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/381—Dispersed generators
-
- H02J3/383—
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
- Y02E10/56—Power conversion systems, e.g. maximum power point trackers
Abstract
The application relates to a power grid voltage control method, a power grid voltage control device, computer equipment and a storage medium. The method comprises the following steps: constructing an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply; training an initial convolutional neural network according to a voltage measured value and an injection power measured value of a distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point; obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period; and controlling the voltage of the power grid according to the adjusted first moment injection power, the adjusted second corresponding relation and the adjusted first corresponding relation. And controlling the voltage of the distributed power grid by controlling the output of the injected power.
Description
Technical Field
The present disclosure relates to the field of power grid voltage control technologies, and in particular, to a power grid voltage control method, an apparatus, a computer device, and a storage medium.
Background
With the rapid consumption of traditional fossil energy and the increasing prominence of environmental issues, renewable energy has been rapidly developed. Among them, power supplies such as photovoltaic power, wind power, and the like have wide distribution, are renewable, have no pollution, and have wide application prospects, for example, a large number of Distributed power supplies (DG) such as photovoltaic power, wind power, and the like are connected to a power grid.
However, the output of the distributed power supply highly depends on environmental changes, and the distributed power supply has great randomness, volatility and uncertainty, so when a large number of distributed power supplies such as photovoltaic power supplies, wind power supplies and the like are connected to a power grid, the fluctuation range of the power grid voltage is increased, the voltage distribution variance is increased, and great difficulty is brought to the power grid voltage control.
Disclosure of Invention
In view of the above, it is necessary to provide a grid voltage control method, apparatus, computer device and storage medium for solving the above technical problems.
A method of grid voltage control, the method comprising:
constructing an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply;
training an initial convolutional neural network according to the voltage measured value and the injection power measured value of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point;
obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period;
and controlling the voltage of the power grid according to the adjusted first moment injection power, the adjusted second corresponding relation and the adjusted first corresponding relation.
In one embodiment, the constructing an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply includes:
constructing an initial objective function according to the voltage of the grid-connected point of the distributed power supply and the injection power of the grid-connected point of the distributed power supply;
determining the communication constraint condition according to the voltage constraint condition of the grid-connected point of the distributed power supply and the inverter capacity of the grid-connected point of the distributed power supply;
and determining the optimized objective function according to the communication constraint condition and the initial objective function.
In one embodiment, the training of the initial convolutional neural network according to the measured values of the voltage and the injection power of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relationship between the voltage and the injection power of the distributed power supply grid-connected point includes:
inputting the measured injection power value in the first preset period into the initial convolutional neural network to obtain a training result;
obtaining a loss function according to the training result and the voltage measured value of the distributed power supply grid-connected point;
if the loss function does not meet the preset condition, returning to the step of inputting the measured injection power value in the first preset period into the initial convolutional neural network;
and if the loss function meets a preset condition, obtaining the first corresponding relation according to the initial convolutional neural network.
In one embodiment, the obtaining a second corresponding relationship between the first-time injection power and the second-time injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relationship, and the measured value of the injection power of the grid-connected point of the distributed power supply in a second preset period includes:
substituting the first corresponding relation into the optimization objective function to obtain a dimension reduction optimization objective function, wherein the dimension reduction optimization objective function is related to the injection power;
and obtaining the second corresponding relation according to the dimension reduction optimization objective function and the measured value of the grid-connected point injection power of the distributed power supply in the second preset period.
In one embodiment, the obtaining the second corresponding relationship according to the dimension reduction optimization objective function and the measured value of the injection power of the grid-connected point of the distributed power supply in the second preset period includes:
obtaining the gradient of the dimension reduction optimization objective function to the injection power according to the dimension reduction optimization objective function and the measured value of the injection power of the grid-connected point of the distributed power supply in the second preset period;
and obtaining the second corresponding relation according to the dimension reduction optimization objective function and the gradient.
In one embodiment, the controlling the grid voltage according to the adjusted first time injection power, the second corresponding relationship, and the first corresponding relationship includes:
determining the injection power at the second moment according to the adjusted injection power at the first moment and the second corresponding relation;
obtaining the voltage at the second moment according to the injection power at the second moment and the first corresponding relation;
and controlling the power grid voltage at the second moment by using the voltage at the second moment.
In one embodiment, the injected power includes reactive power, active power;
if the adjustable reactive power capacity exists in the grid-connected point of the distributed power supply, the reactive power capacity of the grid-connected point of the distributed power supply is adjusted;
controlling the voltage of the power grid according to the adjusted reactive power capacity of the grid-connected point of the distributed power supply;
if the grid-connected point of the distributed power supply has no adjustable reactive power capacity, adjusting the active power capacity of the grid-connected point of the distributed power supply;
and controlling the voltage of the power grid according to the adjusted active power capacity of the grid-connected point of the distributed power supply.
A grid voltage control apparatus, the apparatus comprising:
the optimization objective function building module is used for building an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply;
the first corresponding relation determining module is used for training the initial convolutional neural network according to the voltage measured value and the injection power measured value of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point;
the second corresponding relation determining module is used for obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period;
and the power grid voltage control module is used for controlling the power grid voltage according to the adjusted first moment injection power, the second corresponding relation and the first corresponding relation.
A computer device comprising a memory storing a computer program and a processor implementing the steps of the method of any one of claims 1 to 7 when the computer program is executed.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
According to the power grid voltage control method, the power grid voltage control device, the computer equipment and the storage medium, an optimization objective function is constructed according to the voltage of the grid-connected point of the distributed power supply, the injection power and the communication constraint condition. And determining a first corresponding relation between the voltage and the injection power of the grid-connected point of the distributed power supply and a second corresponding relation between the injection power of the grid-connected point of the distributed power supply at the first moment and the injection power of the distributed power supply at the second moment. And gradually realizing the control of the first moment injection power on the second moment injection power, the control of the second moment injection power on the second moment voltage and the control of the second moment voltage on the power grid voltage. And furthermore, the control of the power grid voltage by adjusting the power is realized. The power grid voltage is controlled by adjusting the injected power, and the defect that the distributed power supply voltage is difficult to control due to randomness, fluctuation and uncertainty is overcome.
Drawings
FIG. 1 is a schematic diagram of a power grid system including a distributed power supply in one embodiment;
FIG. 2 is a schematic flow chart of a grid voltage control method according to an embodiment;
FIG. 3 is a schematic flow chart illustrating the construction of an optimization objective function in one embodiment;
FIG. 4 is a schematic flow chart illustrating obtaining a first correspondence in one embodiment;
FIG. 5 is a diagram illustrating a convolutional neural network structure, according to an embodiment;
FIG. 6 is a flowchart illustrating a process of obtaining a second mapping relationship in one embodiment;
FIG. 7 is a flowchart illustrating a process of obtaining a second mapping relationship in one embodiment;
FIG. 8 is a schematic diagram of a process for controlling a grid voltage according to an embodiment;
FIG. 9 is a schematic diagram of convolutional neural network voltage control in one embodiment;
FIG. 10 is a schematic flow chart of the injection power control grid voltage in one embodiment;
FIG. 11 is a diagram of a computer device in one embodiment;
FIG. 12 is an example system diagram of an IEEE33 node incorporating distributed photovoltaic in one embodiment;
fig. 13 is a schematic diagram of the distribution of the grid-connected point voltage of the distributed power supply under different control strategies.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It will be understood that the terms "first," "second," and the like as used in this application may be used herein to describe various correspondences and time relationships, but are not limited by these terms. These terms are only used to distinguish one correspondence or time relationship from another.
The power grid voltage control method provided by the application can be applied to a power grid system shown in fig. 1, and fig. 1 is a schematic diagram of the power grid system including a distributed power supply, and the power grid system includes a control center, the distributed power supply, a communication line and a branch for connecting the distributed power supply to a power grid. The control center can be but not limited to various personal computers and notebook computers, the distributed power supply is divided into an observable distributed power supply (observable DG) and an unobservable distributed power supply (unobservable DG), a branch of the distributed power supply connected to the power grid is divided into a branch with a known access mode and a branch with an unknown access mode, the known branch can be understood as an observable branch, and the unknown branch can be understood as an unobservable branch. The nodes in the system can be divided into nodes N which are four remote to the access scheduling main station0And the rest of the nodes NXAll nodes N in the system can be represented as:
wherein the content of the first and second substances,
representing a collection of appreciable controllable nodes with distributed power access,
representing a collection of nodes with distributed power access but not observable and controllable,
representing a set of appreciable controllable nodes without distributed power access,
representing a collection of nodes that are not distributed power accessed but are not observable and uncontrollable.
All-node N and observable distributed power supply grid-connected point
The set of injected powers is (P, Q) and
namely, it is
Wherein, P represents the active power set injected by all nodes in the power grid system, Q represents the reactive power set injected by all nodes in the power grid system,
showing the active power set injected into the grid system by the observable and controllable distributed power grid-connected point,
and the reactive power set of the grid-connected point injection power grid system of the observable and controllable distributed power supply is represented.
Specifically, the control center can realize the control of the power grid voltage according to the adjustment of the injection power of the observable and controllable distributed power nodes. The observable and controllable distributed power supply comprises an observable and controllable node power supply, and the access mode of the power supply to the power grid is observable and controllable, namely the branch of the power supply to the power grid is known. And further constructing an optimization objective function for the distributed power source nodes meeting the controllable requirements, and solving the optimization objective function to obtain the corresponding relation between the power and the power of the distributed power source nodes at different moments and the corresponding relation between the power and the voltage. The control of the voltage of the power grid is realized by controlling the power of the distributed power nodes, and the defect that the voltage of the distributed power is difficult to control due to randomness, fluctuation and uncertainty is overcome.
In one embodiment, as shown in fig. 2, a method for controlling a grid voltage is provided, which is exemplified by applying the method to the observable and controllable node with distributed power access in fig. 1, and includes the following steps:
The voltage refers to the power supply voltage of the distributed power supply when the grid-connected point supplies power to the power grid system, and the injected power refers to the power input by the distributed power supply when the grid-connected point supplies power to the power grid system.
Wherein, the first preset period refers to a set specific time period. The first correspondence relationship is a correspondence relationship between the grid-connected point Voltage of the distributed power supply and the injected power under a Convolutional Neural Network, and the first correspondence relationship under the Convolutional Neural Network may be referred to as a Voltage Convolutional Neural Network (VCNN).
Wherein, the second preset period refers to another set specific time period. The second preset period has no time relation with the first preset period, so that the accuracy of the output result is further improved, the first preset period can be different from the second preset period, and the data processing is carried out by adopting different preset periods, so that the occurrence of iterative errors is avoided. The first time refers to a specific sampling time, and the second time refers to another sampling time which has a certain association relation with the first time. For example, the current time is 9:00, the first time may be 9:00, and the second time may be one hour after the first time, and then the second time is 10: 00. The specific association relationship between the first time and the second time is set according to the actual control requirement, and is not specifically limited herein. Specifically, according to step 202 and step 204, a second corresponding relationship between the grid-connected point injection powers of the distributed power supplies at different times can be further obtained according to the first corresponding relationship within the constraint condition of the optimization objective function.
And 208, controlling the power grid voltage according to the adjusted first time injection power, the adjusted second corresponding relation and the adjusted first corresponding relation.
Specifically, the first corresponding relationship and the second corresponding relationship are obtained in step 202, step 204, and step 206. The second corresponding relation is the corresponding relation between the first moment injection power and the second moment injection power of the grid-connected point of the distributed power supply, namely when the first moment injection power of the grid-connected point of the distributed power supply changes, the second moment injection power changes along with the first moment injection power. And according to the first corresponding relation, when the grid-connected point injection power of the distributed power supply is changed, the voltage is changed along with the change of the grid-connected point injection power of the distributed power supply. Therefore, the voltage of the grid-connected point of the distributed power supply can be changed by adjusting the injection power of the grid-connected point of the distributed power supply at one moment. Therefore, the adjustment of the injected power at the first moment can realize the control of the voltage of the power supply grid-connected point, and further realize the control of the power grid voltage.
In an actual power grid system, the medium and low voltage communication system is often not perfect due to the limitation of construction investment. Taking distributed photovoltaic as an example, about half of distributed photovoltaic with the voltage class of 10kV and above is accessed to the dispatching master station system with relevant information. Meanwhile, the modeling precision of the medium and low voltage grid system is usually far lower than that of the power transmission grid, the problem that the state estimation is often not accurate enough greatly limits the application of the voltage control technology in the existing medium and low voltage grid system.
According to the power grid voltage control method, an optimization objective function is constructed according to the voltage of a grid-connected point of the distributed power supply, the injection power and the communication constraint condition. Further, a first correspondence between distributed power supply grid-connected point voltage and injected power is determined. And determining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply by optimizing the objective function, the first corresponding relation and the measured value of the injection power of the grid-connected point of the distributed power supply in a second preset period. And finally, controlling the voltage of the power grid according to the adjusted first moment injection power, the second corresponding relation and the first corresponding relation, and overcoming the defect that the voltage of the power grid after the distributed power supply is connected to the power grid is difficult to control.
In one embodiment, as shown in fig. 3, the constructing an optimization objective function according to the voltage, the injection power, and the communication constraint conditions of the distributed power supply grid-connected point includes:
The initial optimization objective function is the superposition of the sum of squares of the voltage of the grid-connected point of the distributed power supply, the injection power fitting voltage and the voltage reference value, and is specifically represented as the relationship between the initial optimization objective function and the voltage of the grid-connected point of the distributed power supply, the injection power fitting voltage and the voltage reference value.
The initial optimization objective function is obtained in the step 2022, and is limited by the communication constraint conditions of the power grid system, and the specific communication constraint conditions are determined by the specific requirements of the distributed power grid connection point on the voltage and the injection power. Specifically, the voltage variation range of the grid-connected point of the distributed power supply is determined according to conditions such as the requirement of the grid-connected point of the distributed power supply on the voltage, the communication requirement of the distributed power supply generation equipment and the power grid system at the grid-connected point of the distributed power supply, and the like, and the constraint ranges of the voltages of the grid-connected points of different distributed power supplies can be different. The variation range of the injection power range of the grid-connected point of the distributed power supply is determined according to the capacity of the voltage inverter adopted by the grid-connected point of the distributed power supply, the requirement on the injection power, the communication requirements of the distributed power supply generation equipment and a power grid system at the grid-connected point of the distributed power supply and other conditions, and the constraint ranges of the injection power of different grid-connected points of the distributed power supply can be different.
Wherein, the optimization objective function can be expressed as:
wherein i represents the serial number of the distributed power supply grid-connected point,
a voltage reference value representing a grid-connected point of the distributed power supply,
the voltage measured value of the grid-connected point of the distributed power supply is shown,
representing the corresponding relation between the injection power and the voltage of the grid-connected point of the distributed power supply,
and the increment function represents the corresponding relation between the injection power and the voltage of the grid-connected point of the distributed power supply. In the constraint condition, the number of the optical fiber,
represents the minimum value of the grid-connected point voltage of the distributed power supply,
representing the maximum allowable value, Q, of the grid-connected point voltage of the distributed power supplyDGiRepresenting reactive power, P, injected into the grid system by the grid-connected point of the distributed power supplyDGiThe active power, delta Q, injected into a power grid system by a grid-connected point of a distributed power supply is representedDGiExpressing the reactive power increment, delta P, injected into the power grid system by the grid-connected point of the distributed power supplyDGiThe active power increment injected into the power grid system by the grid-connected point of the distributed power supply is represented,
representing the maximum capacity of the distributed power grid-connected inverter.
In the above embodiment, since there is a specific correspondence between the voltage and the injected power, the communication constraint condition is defined according to the actual condition of the distributed power grid-connected point.
In one embodiment, as shown in fig. 4, the training of the initial convolutional neural network according to the measured value of the voltage of the distributed power supply point of connection and the measured value of the injection power in a first preset period to obtain a first corresponding relationship between the voltage of the distributed power supply point of connection and the injection power includes:
specifically, as shown in fig. 5, the initial convolutional neural network is composed of a three-layer convolutional structure and a two-layer fully-connected structure. Each convolution structure comprises a convolution layer and a batch regularization layer, the convolution layer can extract key features among input data, and the batch regularization layer can improve the convergence efficiency and stability of the convolution neural network model. The convolution kernel sizes of the three convolution layers may be 2 x 9, 2 x 7 and 2 x 7, respectively, with an activation function behind each batch of regularization layers and fully-connected layers. Specifically, on the basis of the initial convolutional neural network structure, the injection power measured value in a first preset period is input into the initial convolutional neural network, and a training result of the initial neural network is obtained.
And 2044, obtaining a loss function according to the training result and the voltage measured value of the distributed power supply grid-connected point.
The training result in step 2042 is an intermediate training result performed on the initial convolutional neural network, and the acquisition of the final training result requires repeated training iteration on the initial convolutional neural network. In order to train the initial convolutional neural network repeatedly, a loss function is constructed according to the training result in step 2042 and the voltage measured value of the grid-connected point of the distributed power supply. Further, the loss function is a mean square error between the training result output voltage and the voltage measured value in step 2042. Specifically, for an electric power system with n observable nodes and m key nodes, the mean square error between the output voltage and the voltage measured value of the fitting function model is taken as the loss function of the model, and the loss function L can be expressed as:
wherein U represents a training result output voltage set,
representing a set of voltage measurements, UiRepresents the training result output voltage of the distributed power supply grid-connected point i,
and the voltage measured value of the grid-connected point i of the distributed power supply is shown.
The preset condition is to set different distribution requirements for the loss function according to different precision requirements for the convolutional neural network, and the preset condition for the loss function is determined according to actual requirements and is not specifically limited herein. Specifically, it is determined whether the loss function satisfies a preset condition during the training process, and if not, the step of inputting the measured injection power value in the first preset period to the initial convolutional neural network is executed. If the first correspondence relationship is satisfied, the initial convolutional neural network can obtain the correspondence relationship between the grid-connected point voltage and the injection power of the distributed power supply, and the correspondence relationship between the grid-connected point voltage and the injection power of the distributed power supply is the first correspondence relationship.
In the above embodiment, the measured injection power value in the first preset period is input to the initial convolutional neural network to obtain a training result, and the loss function is constructed according to the training result. And repeatedly training the initial neural network by using the loss function and the preset condition thereof to obtain a first corresponding relation. By adopting the method in the embodiment, different requirements can be set on the loss function preset condition according to different requirements, so that the first corresponding relation meeting the preset precision requirement can be obtained.
In one embodiment, as shown in fig. 6, the obtaining a second corresponding relationship between the first-time injection power and the second-time injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relationship, and the measured value of the injection power of the grid-connected point of the distributed power supply in a second preset period includes:
According to steps 2022 to 2026, the dimension of the optimization objective function may be reduced, and further, solving the optimization objective function may be understood as solving the optimization objective function for the voltage or the injection power under the communication constraint condition. Specifically, after the first corresponding relation is substituted into the optimization objective function, the intermediate variable voltage can be eliminated, and the dimension-reduced optimization objective function is obtained. The dimension reduction optimization objective function is related to the injection power, and further the second corresponding relation can be obtained by solving the optimization objective function relative to the injection power within the communication constraint condition range of the optimization objective function.
In one embodiment, as shown in fig. 7, the obtaining the second corresponding relationship according to the dimension reduction optimization objective function and the measured value of the injection power of the grid-connected point of the distributed power supply in the second preset period includes:
The dimension reduction optimization objective function is obtained by eliminating variable voltage dimension reduction through a first corresponding relation, and the first corresponding relation is obtained by repeatedly training a convolutional neural network. Therefore, the corresponding relation between the dimensionality reduction optimization objective function and the injection power of the distributed power supply grid-connected point comprises the corresponding relation of a convolutional neural network, and the gradient of the dimensionality reduction optimization objective function to the injection power cannot be obtained through explicit derivation. Therefore, the gradient of the optimization objective function to the injection power needs to be obtained through the measured value of the injection power of the grid-connected point of the distributed power supply and the dimension reduction optimization objective function in the second preset period. Further, a second corresponding relationship is obtained.
In one embodiment, as shown in fig. 8, the controlling the grid voltage according to the adjusted first time injection power, the second corresponding relationship, and the first corresponding relationship includes:
Specifically, as shown in fig. 9, the gradient of the injection power is optimized by calculating the optimization objective functionUpdating Q by gradient descent methodDGiAnd PDGi. Further, according to the first corresponding relationship, performing gradient pass-back on the loss function L, and obtaining the loss function L, which is the reactive power Q injected into the grid-connected point of the distributed power supplyDGiAnd gradient of active power PDGi. And copying the loss function gradient feedback model to be applied to the optimization objective function, and transmitting the gradient of the injection power of the optimization objective function on the basis of the loss function gradient feedback model to obtain a corresponding relation between the injection power at the first moment and the injection power at the second moment as a second corresponding relation. The specific second correspondence may be expressed as:
wherein the content of the first and second substances,
the reactive power of a grid-connected point of the distributed power supply can be observed at the moment k,
shows that the active power, lambda, of a grid connection point of the distributed power supply can be observed at the moment k1And λ2Indicating the update step size of the injected power.
And 2086, controlling the power grid voltage at the second moment by using the voltage at the second moment.
In the above embodiment, the control of the first-time injection power to the second-time injection power, the control of the second-time injection power to the second-time voltage, and the control of the second-time voltage to the grid voltage may be gradually achieved according to the adjusted first-time injection power, the adjusted second correspondence relationship, and the adjusted first correspondence relationship. And furthermore, the control of the power grid voltage by adjusting the power is realized.
In one embodiment, as shown in fig. 10, the injected power includes reactive power and active power.
Wherein the voltage is controlled by adjusting the grid-connected point power of the distributed power supply from step 202 to step 208. Because the inverter of the distributed power supply usually reserves 5% -10% of extra capacity, a large amount of reactive power regulation capacity can be released only by adjusting the control strategy of the inverter at the moment, extra cost is not needed, and the method is an economical and effective voltage control means. Therefore, when the injected power is adjusted, the reactive power of the grid-connected point of the distributed power supply is preferentially adjusted, and when the reactive power has no adjustable capacity, the voltage of the power grid is controlled by adjusting the active power.
In the above embodiment, in order to ensure the consumption of new energy, the voltage of the power grid is controlled by adjusting the injection power of the grid-connected point of the distributed power supply, and a control strategy of firstly performing reactive power and then performing active power is adopted. Namely, when the adjustable reactive power capacity exists in the grid-connected point of the distributed power supply, the reactive power capacity of the grid-connected point of the distributed power supply is preferentially adjusted to control the voltage of the power grid. And when the reactive power capacity of the grid-connected point of the distributed power supply is used up, the voltage of the power grid is controlled by means of active reduction and the like.
The power grid system with the distributed power supply part accessed to the control center is subjected to voltage sensing by the convolutional neural network, and then voltage control is carried out based on the trained convolutional neural network. Therefore, accurate modeling of the system is not needed, and all nodes are not needed to be accessed to the control center.
In one embodiment, there is provided a grid voltage control apparatus, the apparatus comprising:
the optimization objective function building module is used for building an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply;
the first corresponding relation determining module is used for training the initial convolutional neural network according to the voltage measured value and the injection power measured value of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point;
the second corresponding relation determining module is used for obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period;
and the power grid voltage control module is used for controlling the power grid voltage according to the adjusted first moment injection power, the adjusted second corresponding relation and the adjusted first corresponding relation.
For specific limitations of the grid voltage control device, reference may be made to the above limitations of the grid voltage control method, which are not described herein again. The modules in the above-mentioned network voltage control device may be implemented wholly or partially by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 11. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing grid voltage control data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a grid voltage control method.
Those skilled in the art will appreciate that the architecture shown in fig. 11 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In an embodiment, a computer device is provided, comprising a memory storing a computer program and a processor implementing the steps of the method of any one of claims 1 to 7 when the processor executes the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
In one embodiment, as shown in FIG. 12, an IEEE33 (Institute of Electrical and Electronics Engineers, IEEE) node algorithm system incorporating a distributed power supply is illustrated. 8 distributed photovoltaics are added into the standard 33-node calculation system, and the capacity of each distributed power supply is 4 MW. Reference voltage V of systembaseAdjusted to 35 kV. In this embodiment, 4 and 12 load nodes in the distributed power supply are not connected to the scheduling master station, and belong to an invisible state, which is indicated by a dotted line in the figure. The analysis is carried out according to 17 days of data in historical operation data of the power grid system, and 24480 data points are included, wherein 20000 points serve as training data, and 4480 points serve as test data. For the system shown in fig. 12, for the convolutional neural network, active power and reactive power injection power flow of the nodes of the access scheduling master station are used as model inputs, a 2 x 17-dimensional power flow matrix is used as an input, and voltages of the appreciable controllable nodes 2, 14, 20 and 27 are used as outputs.
On the basis of test data, the fitting error of the convolutional neural network is smaller than 0.01 through operation, and the voltage control requirement is met. After the distributed power supply is connected, the voltage fluctuation range of the system is obviously enlarged, and the node voltage fluctuation range at the tail end of the power grid is larger. After the voltage of the power grid is controlled, the voltage fluctuation curve is obviously improved. As shown in fig. 13, compared with the grid voltage under the condition of no intervention of a voltage control means and the grid voltage under the condition of control by other control means, the voltage curve obtained by the grid voltage control method provided by the application is distributed more intensively near the reference value, and the defect that the grid voltage after the distributed power supply is connected to the grid is difficult to control is overcome.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of grid voltage control, the method comprising:
constructing an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply; wherein the communication constraint condition is that the distributed power supply grid-connected point is an observable controllable node;
training an initial convolutional neural network according to the voltage measured value and the injection power measured value of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point;
obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period;
and controlling the voltage of the power grid according to the adjusted first moment injection power, the adjusted second corresponding relation and the adjusted first corresponding relation.
2. The method of claim 1, wherein constructing an optimization objective function according to the voltage, injection power and communication constraints of the grid-connected point of the distributed power supply comprises:
constructing an initial objective function according to the voltage of the grid-connected point of the distributed power supply and the injection power of the grid-connected point of the distributed power supply;
determining the communication constraint condition according to the voltage constraint condition of the grid-connected point of the distributed power supply and the inverter capacity of the grid-connected point of the distributed power supply;
and determining the optimized objective function according to the communication constraint condition and the initial objective function.
3. The method according to claim 1, wherein the training of the initial convolutional neural network according to the measured values of the voltage and the injection power of the grid-connected point of the distributed power supply in a first preset period to obtain a first corresponding relationship between the voltage and the injection power of the grid-connected point of the distributed power supply comprises:
inputting the measured injection power value in the first preset period into the initial convolutional neural network to obtain a training result;
obtaining a loss function according to the training result and the voltage measured value of the distributed power supply grid-connected point;
if the loss function does not meet the preset condition, returning to the step of inputting the measured injection power value in the first preset period into the initial convolutional neural network;
and if the loss function meets a preset condition, obtaining the first corresponding relation according to the initial convolutional neural network.
4. The method according to claim 1, wherein obtaining a second corresponding relationship between a first-time injection power and a second-time injection power of a grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relationship, and an actual measured value of the injection power of the grid-connected point of the distributed power supply in a second preset period comprises:
substituting the first corresponding relation into the optimization objective function to obtain a dimension reduction optimization objective function, wherein the dimension reduction optimization objective function is related to the injection power;
and obtaining the second corresponding relation according to the dimension reduction optimization objective function and the measured value of the grid-connected point injection power of the distributed power supply in the second preset period.
5. The method according to claim 4, wherein the obtaining the second corresponding relationship according to the dimension reduction optimization objective function and the measured value of the grid-connected point injection power of the distributed power supply in the second preset period comprises:
obtaining the gradient of the dimension reduction optimization objective function to the injection power according to the dimension reduction optimization objective function and the measured value of the injection power of the grid-connected point of the distributed power supply in the second preset period;
and obtaining the second corresponding relation according to the dimension reduction optimization objective function and the gradient.
6. The method of claim 1, wherein controlling the grid voltage according to the adjusted first time injected power, the second correspondence, and the first correspondence comprises:
determining the injection power at the second moment according to the adjusted injection power at the first moment and the second corresponding relation;
obtaining the voltage at the second moment according to the injection power at the second moment and the first corresponding relation;
and controlling the power grid voltage at the second moment by using the voltage at the second moment.
7. The method of any of claims 1-6, wherein the injected power comprises reactive power, active power;
if the adjustable reactive power capacity exists in the grid-connected point of the distributed power supply, the reactive power capacity of the grid-connected point of the distributed power supply is adjusted;
controlling the voltage of the power grid according to the adjusted reactive power capacity of the grid-connected point of the distributed power supply;
if the grid-connected point of the distributed power supply has no adjustable reactive power capacity, adjusting the active power capacity of the grid-connected point of the distributed power supply;
and controlling the voltage of the power grid according to the adjusted active power capacity of the grid-connected point of the distributed power supply.
8. A grid voltage control apparatus, the apparatus comprising:
the optimization objective function building module is used for building an optimization objective function according to the voltage, the injection power and the communication constraint conditions of the grid-connected point of the distributed power supply; wherein the communication constraint condition is that the distributed power supply grid-connected point is an observable controllable node;
the first corresponding relation determining module is used for training the initial convolutional neural network according to the voltage measured value and the injection power measured value of the distributed power supply grid-connected point in a first preset period to obtain a first corresponding relation between the voltage and the injection power of the distributed power supply grid-connected point;
the second corresponding relation determining module is used for obtaining a second corresponding relation between the first-moment injection power and the second-moment injection power of the grid-connected point of the distributed power supply according to the optimization objective function, the first corresponding relation and the measured injection power value of the grid-connected point of the distributed power supply in a second preset period;
and the power grid voltage control module is used for controlling the power grid voltage according to the adjusted first moment injection power, the second corresponding relation and the first corresponding relation.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811641746.5A CN109638883B (en) | 2018-12-29 | 2018-12-29 | Power grid voltage control method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811641746.5A CN109638883B (en) | 2018-12-29 | 2018-12-29 | Power grid voltage control method and device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109638883A CN109638883A (en) | 2019-04-16 |
| CN109638883B true CN109638883B (en) | 2020-12-22 |
Family
ID=66054871
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811641746.5A Active CN109638883B (en) | 2018-12-29 | 2018-12-29 | Power grid voltage control method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109638883B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110556874B (en) * | 2019-09-30 | 2021-08-10 | 西安特锐德领充新能源科技有限公司 | Power control method, device, electronic equipment and storage medium |
| CN112467752A (en) * | 2020-11-25 | 2021-03-09 | 青岛创新奇智科技集团有限公司 | Voltage regulating method and device for distributed energy distribution system |
| CN113300379B (en) * | 2021-05-08 | 2022-04-29 | 武汉大学 | Electric power system reactive voltage control method and system based on deep learning |
| CN113746111B (en) * | 2021-09-07 | 2023-10-20 | 广东电网有限责任公司 | Power distribution network reactive power optimization method and device based on data driving |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104111373A (en) * | 2014-07-23 | 2014-10-22 | 国家电网公司 | Metering method used for digital electric energy of intelligent substation |
| CN108183481A (en) * | 2018-01-29 | 2018-06-19 | 中国电力科学研究院有限公司 | One kind quickly sentences steady method and system based on deep learning power grid |
| CN106532762B (en) * | 2015-09-14 | 2018-12-14 | 南京理工大学 | A kind of direct current Probabilistic Load Flow method of classification processing load and wind speed correlation |
-
2018
- 2018-12-29 CN CN201811641746.5A patent/CN109638883B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104111373A (en) * | 2014-07-23 | 2014-10-22 | 国家电网公司 | Metering method used for digital electric energy of intelligent substation |
| CN106532762B (en) * | 2015-09-14 | 2018-12-14 | 南京理工大学 | A kind of direct current Probabilistic Load Flow method of classification processing load and wind speed correlation |
| CN108183481A (en) * | 2018-01-29 | 2018-06-19 | 中国电力科学研究院有限公司 | One kind quickly sentences steady method and system based on deep learning power grid |
Non-Patent Citations (2)
| Title |
|---|
| Power Analysis of a General Convolution Algorithm Mapped on a Linear Processor Array;Rashindra Manniesing, Richard Kleihorst, André van der Avoird &;《Journal of VLSI signal processing systems for signal, image and video technology》;20040501;全文 * |
| 基于实测数据的配网理论网损计算方法;王刚军等;《电网技术》;20021231;第26卷(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109638883A (en) | 2019-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109638883B (en) | Power grid voltage control method and device, computer equipment and storage medium | |
| CN110738344B (en) | Distributed reactive power optimization method and device for load prediction of power system | |
| CN107436971B (en) | Improved Latin hypercube sampling method suitable for non-positive stereotype correlation control | |
| CN107846024B (en) | Island micro-grid probability load flow analysis calculation method | |
| Chen et al. | Input convex neural networks for optimal voltage regulation | |
| CN110518591B (en) | Load flow calculation method for uncertain power system | |
| CN109088407B (en) | Power distribution network state estimation method based on deep belief network pseudo-measurement modeling | |
| CN112383065A (en) | Distributed MPC-based power distribution network dynamic voltage control method | |
| CN112418496B (en) | Power distribution station energy storage configuration method based on deep learning | |
| CN111697625A (en) | Island micro-grid opportunity constraint energy scheduling method based on distributed robust optimization | |
| CN111181164B (en) | Improved master-slave split transmission and distribution cooperative power flow calculation method and system | |
| CN111600309B (en) | Voltage control method, device, equipment, computer equipment and storage medium | |
| Yang et al. | Optimization of static voltage stability margin considering uncertainties of wind power generation | |
| CN108551177B (en) | Sensitivity analysis-based transient load shedding control optimization method for direct current receiving end system | |
| Xie et al. | Discrete/continuous-time online algorithm application for time-varying optimal power flow in active distribution networks | |
| CN111507591B (en) | Power system state determination method, device, computer medium and storage medium | |
| Dubyago | Mathematical description of interaction of the elements in the power network system | |
| CN105207220B (en) | A kind of tapping voltage regulation and control method based on progressive learning | |
| CN112016702A (en) | Medical data processing method, device, equipment and medium based on transfer learning | |
| Li et al. | A Region-Based Approach for the Operational Reliability Evaluation of Power Systems with Renewable Energy Integration | |
| CN112488224B (en) | Model training method, device, computer equipment and storage medium | |
| CN115579955A (en) | Neural network prediction-based power distribution network photovoltaic consumption capacity evaluation method | |
| CN115549137A (en) | Distributed power grid regulation and control system and regulation and control method | |
| CN113240072A (en) | Deep learning-based prediction method for direct-current micro-grid cluster | |
| CN107086603A (en) | A kind of Random-fuzzy Continuation power flow of power system containing DFIG |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |