CN109066707B - A Microgrid Energy Management Method Based on NARMA-L2 Model - Google Patents
A Microgrid Energy Management Method Based on NARMA-L2 Model Download PDFInfo
- Publication number
- CN109066707B CN109066707B CN201811054866.5A CN201811054866A CN109066707B CN 109066707 B CN109066707 B CN 109066707B CN 201811054866 A CN201811054866 A CN 201811054866A CN 109066707 B CN109066707 B CN 109066707B
- Authority
- CN
- China
- Prior art keywords
- distributed power
- power supply
- node
- output
- control signal
- 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.)
- Expired - Fee Related
Links
- 238000007726 management method Methods 0.000 title claims abstract description 21
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000013528 artificial neural network Methods 0.000 claims abstract description 13
- 238000010845 search algorithm Methods 0.000 claims abstract description 9
- 238000012886 linear function Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims description 16
- 238000013178 mathematical model Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000010248 power generation Methods 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 5
- 230000008878 coupling Effects 0.000 claims description 5
- 238000010168 coupling process Methods 0.000 claims description 5
- 238000005859 coupling reaction Methods 0.000 claims description 5
- 230000008685 targeting Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 description 9
- 238000005457 optimization Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 3
- 238000011217 control strategy Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- H02J3/12—Circuit arrangements for AC mains or AC distribution networks for adjusting voltage in AC networks by changing a characteristic of the network load
- H02J3/16—Circuit arrangements for AC mains or AC distribution networks for adjusting voltage in AC networks by changing a characteristic of the network load by adjustment of reactive power
-
- 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
-
- 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
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/30—Reactive power compensation
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
本发明公开了一种基于NARMA‑L2模型微电网能量管理方法，包括以下步骤：步骤10）利用NARMA‑L2模型建立分布式电源输出与控制输入之间的关系式；步骤20）基于牵制控制理念，以无功功率均分为目标，针对系统内分布式电源产生控制信号输入与电压输出数据集；步骤30）以步骤20）得到的数据集训练步骤10）中表达式内用于表征复杂非线性函数的神经网络，拟合无功功率均分情形下分布式电源非线性输出动态特性；步骤40）利用梯度下降搜索算法产生参考值，完成微电网无功功率协同控制。该方法利用神经网络拟合输入输出特性中复杂的非线性函数，实现微电网的去模型化控制；同时基于牵制控制理念，实现微电网的协同控制，提升控制响应速度和准确性。
The invention discloses a microgrid energy management method based on the NARMA-L2 model, comprising the following steps: step 10) using the NARMA-L2 model to establish a relational expression between the output of the distributed power supply and the control input; step 20) based on the containment control concept , divide the reactive power into the target equally, and generate control signal input and voltage output data sets for the distributed power supply in the system; step 30) Use the data set obtained in step 20) to train the expression in step 10) to represent complex non-linear The neural network of the linear function is used to fit the nonlinear output dynamic characteristics of the distributed power supply in the case of reactive power equalization; Step 40) Use the gradient descent search algorithm to generate a reference value to complete the microgrid reactive power cooperative control. The method uses neural network to fit the complex nonlinear functions in the input and output characteristics to realize the de-modeling control of the microgrid; at the same time, based on the pinning control concept, the coordinated control of the microgrid is realized, and the control response speed and accuracy are improved.
Description
Technical Field
The invention relates to the technical field of energy management of a microgrid, in particular to a microgrid energy management method based on a NARMA-L2 model.
Background
In recent years, the economic development of China is rapid, the total electricity consumption is rapidly increased, and the contradiction between the increase of energy demand and the shortage of non-renewable energy, the low utilization rate of renewable energy and sustainable development is increasingly highlighted. Therefore, it is imperative to improve the utilization rate of renewable energy, develop new energy utilization techniques, and enhance the utilization of renewable energy.
Distributed power generation is mature day by day, and the increasing new energy power supplies bring new challenges to the stable operation of a power system. The most important problems are: the new energy is often uncontrollable and difficult to control, and has great influence on power supply stability. Since conventional power distribution system architectures and operating strategies cannot accommodate the requirements of large-scale distributed power access, the concept of microgrid was introduced as a solution to reliably utilize new energy. Microgrid versus conventional power distribution systems: the scale is small, the topological structure is easy to change, and the operation can be independent of a large power grid. The aim that a microgrid manager wants to achieve is to complete microgrid control economically, stably and quickly.
The distributed micro-grid cooperative control research is used, the micro-grid is modeled into a complex network by using a containment control concept, voltage cooperative control is converted into the problems of complex network cooperation and neighbor error tracking, and the voltage convergence and the running stability of a system are guaranteed. However, a distributed power supply dynamic mathematical model is adopted, the problem that the mathematical model of the high-frequency inverter is difficult to model is ignored, and control errors can be brought; meanwhile, the containment concept is that a reference value is introduced into a few nodes, and the containment is continuously transmitted among the neighbor nodes to form containment, so that the whole system is coordinated and consistent, and the stability speed is relatively slow. The traditional constraint concept micro-grid cooperative control is based on a distributed power supply dynamic mathematical model, the problem of inaccurate modeling exists, and the constraint control causes the system convergence speed to be slow.
Disclosure of Invention
The invention provides a microgrid energy management method based on an NARMA-L2 model, aiming at solving the problem of modeling errors of a distributed power supply, the NARMA-L2 model is used for replacing a power supply dynamic mathematical model, and based on a containment concept, a data set is generated to train an artificial neural network in the NARMA-L2 model, and a proper nonlinear expression is generated; and aiming at the used model, a gradient descent search algorithm is adopted to give a proper reference value, so that the rapid and accurate model removal control is realized, and the control targets of reactive load equalization and rapid voltage recovery are achieved.
The invention adopts the following technical scheme for solving the technical problems:
the invention provides a microgrid energy management method based on a NARMA-L2 model, which comprises the following steps:
step 10, establishing a relational expression between distributed power supply output and control signal input by utilizing a NARMA-L2 model;
step 20, based on a containment control concept, taking reactive power sharing as a target to obtain a data set of control signal input and voltage output;
step 30, training a neural network used for representing a complex nonlinear function in the relational expression established in the step 10 by using the data set obtained in the step 20, and fitting the nonlinear output dynamic characteristic of the distributed power supply under the condition of uniform reactive power, so as to obtain a relational expression between the trained distributed power supply output and the trained control signal input; constructing a distributed power supply controller based on NARMA-L2 according to a relation between the trained distributed power supply output and the trained control signal input;
and step 40, providing a target output voltage reference value for the distributed power supply controller constructed in the step 30 by using a gradient descent search algorithm, and finishing the reactive power cooperative control of the microgrid.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, in step 10, the NARMA-L2 model is used for replacing a distributed power supply dynamic mathematical model, a complex nonlinear function part in a distributed power supply system is approximately replaced through an artificial neural network, and the relational expression between the output of the distributed power supply and the input of a control signal is as follows:
y(k+d)＝f[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]+g[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]·u(k) (1)
wherein u (k) represents the input of the distributed power supply control signal at the k moment, y (k) represents the output voltage signal of the distributed power supply at the k moment, y (k + d) represents the output voltage signal of the distributed power supply at the k + d moment, and f and g represent complex nonlinear functions of the nonlinear part in the distributed power generation system; collecting the distributed power supply output at the k-n +1 moment in total, wherein y (k-n +1) represents the distributed power supply output at the (k-n +1) th moment; and c, acquiring the control signal input at the k-z +1 moment, wherein u (k-z +1) is the control signal input at the (k-z +1) th moment.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of the step 20 is as follows:
step 201, for the distributed control micro-grid topology, converting the cooperative target into neighbor information error tracking:
in the formula, eiRepresenting the error between the i-th distributed power supply node output voltage and a reference value, yiRepresents the output voltage of the ith distributed power supply node, and j belongs to NiIndicating that the jth node is a neighbor node of the ith node, NiRepresents a set formed by all neighbor nodes of the ith distributed power node, aijRepresenting adjacency matrices in a distributed power system
The value of (i) th row and (j) th column in (1), giRepresenting the holdoff gain, y, of the ith distributed power nodejIndicating the output voltage value, y, of the jth node in communication topology with the ith node0Represents the output voltage value of the pinning node, i.e., the reference value;
step 202, to guarantee error eiConverging to 0, and selecting the control signal input u according to the following formula under the condition that the control target is reactive load sharingi：
In the formula, Q0Representing a reference reactive load, QiIndicating that the ith node outputs reactive power, omegaiIndicating whether the ith node has been pulledIn the preparation method, the raw materials are mixed,
representing the degree of coupling between the nodes;
step 203, replacing the reactive power in the formula (3) with the output voltage according to the droop control characteristic of the distributed power supply, namely the relational expression between the output voltage and the reactive power, wherein the control signal input expression is changed into:
wherein n isQiRepresenting the reactive droop coefficient, y, of the ith node0Representing the value of the output voltage of the holddown node, i.e. the reference value, u0A control signal input representing a holdover node; on the basis of the formula (4), determining corresponding control signal input by giving different distributed power supply outputs, and further forming a data set [ y ] of the control signal input and the voltage outputi,ui]。
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of step 30 is as follows: from the resulting data set y in step 203i,ui]A complex non-linear function f in the relation is established for step 10]、g[*]Training an artificial neural network to replace f [. X [ ]],g[*]And on the basis of the NARMA-L2-based distributed power supply controller; the distributed power controller is a specific implementation of equation (1), and provides control signal inputs u (k) corresponding to different target voltage outputs.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of step 40 is as follows:
step 401, setting a reference value of target output voltage of the distributed power supply as yr(k + d) finding an expression of a target output voltage reference value for targeting reactive power sharing in step 2, and providing a reference value for the distributed power controller constructed in step 30; providing a target voltage output reference value by adopting a gradient descent search algorithmA generator:
wherein j | (i, j) ∈ DGDenotes that the ith, j distributed power node is a neighbor node, DGThe method is a set formed by combining all neighbor nodes in the distributed power system; j (y)i) Represents the average value of the sum of squares of the reactive power differences output by the ith node and the neighboring nodes, QjThe output reactive power of the jth node is represented, and m represents that the ith distributed power supply node has m neighbor nodes;
step 402, according to equation (5), J (y)i) When the minimum value exists, namely the difference between the reactive power output of all the distributed power supplies is minimum, the aim of sharing the reactive power in the step 2 is achieved; to find J (y)i) Minimum time distributed power supply target output voltage reference value yr(k + d) performing the following loop until convergence;
(7) formula (I) is expressed as finding the smallest J (y)i) Value, constantly iterating yiA, denotes the search step size,
represents J (y)i) For yiDerivative, temp, representing the current cycle yiWith the convergence condition set to two cycles yiThe variation is less than or equal to a preset value.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the preset value is 0.001.
Compared with the prior art, the invention adopting the technical scheme has the following technical effects:
according to the method, a NARMA-L2 model is used for replacing a dynamic mathematical model of the distributed power supply, so that the modeling control of the microgrid is realized, and the control speed and precision are improved; and then, by utilizing the control of the drag, generating a data set of an input control signal and a voltage output signal by taking reactive power equalization and voltage rapid recovery as targets, and training an artificial neural network used for representing a complex nonlinear function in an expression so as to achieve a good fitting effect. And finally, giving a reference value generator by adopting a gradient descent search algorithm to finish the quick stabilization of the microgrid.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Fig. 2 is a topology structure diagram of a microgrid according to an embodiment of the present invention.
Fig. 3 is a communication topology of a microgrid in an embodiment of the present invention.
FIG. 4 shows the variation of the power sources with sudden load increase in the system; wherein, (a) is the reactive power output change process of each distributed power supply, and (b) is the voltage change process of each distributed power supply under the condition of sudden load increase.
FIG. 5 shows a variation of the power supply remaining after the power supply is disconnected in the system; wherein, (a) is the voltage change process of the distributed power supply, and (b) is the reactive power output state change process of the distributed power supply.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
The system in the invention refers to a distributed power system, and the nodes refer to distributed power nodes.
As shown in fig. 1, the topology of the stream microgrid according to the embodiment of the present invention is shown in fig. 2. The method comprises the following steps:
step 10, establishing a relational expression between distributed power supply output and control signal input by utilizing a NARMA-L2 model;
step 20, based on a containment control concept, taking reactive power sharing as a target to obtain a data set of control signal input and voltage output;
step 30, training a neural network used for representing a complex nonlinear function in the relational expression established in the step 10 by using the data set obtained in the step 20, and fitting the nonlinear output dynamic characteristic of the distributed power supply under the condition of uniform reactive power, so as to obtain a relational expression between the trained distributed power supply output and the trained control signal input; constructing a distributed power supply controller based on NARMA-L2 according to a relation between the trained distributed power supply output and the trained control signal input;
and step 40, providing a target output voltage reference value for the distributed power supply controller constructed in the step 30 by using a gradient descent search algorithm, and finishing the reactive power cooperative control of the microgrid.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, in step 10, the NARMA-L2 model is used for replacing a distributed power supply dynamic mathematical model, a complex nonlinear function part in a distributed power supply system is approximately replaced through an artificial neural network, and the relational expression between the output of the distributed power supply and the input of a control signal is as follows:
y(k+d)＝f[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]+g[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]·u(k) (1)
wherein u (k) represents the input of the distributed power supply control signal at the k moment, y (k) represents the output voltage signal of the distributed power supply at the k moment, y (k + d) represents the output voltage signal of the distributed power supply at the k + d moment, and f and g represent complex nonlinear functions of the nonlinear part in the distributed power generation system; collecting the distributed power supply output at the k-n +1 moment in total, wherein y (k-n +1) represents the distributed power supply output at the (k-n +1) th moment; and c, acquiring the control signal input at the k-z +1 moment, wherein u (k-z +1) is the control signal input at the (k-z +1) th moment.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of the step 20 is as follows:
step 201, for the distributed control micro-grid topology, converting the cooperative target into neighbor information error tracking:
in the formula, eiRepresenting the error between the i-th distributed power supply node output voltage and a reference value, yiRepresents the output voltage of the ith distributed power supply node, and j belongs to NiIndicating that the jth node is a neighbor node of the ith node, NiRepresents a set formed by all neighbor nodes of the ith distributed power node, aijRepresenting adjacency matrices in a distributed power system
The value of (i) th row and (j) th column in (1), giRepresenting the holdoff gain, y, of the ith distributed power nodejIndicating the output voltage value, y, of the jth node in communication topology with the ith node0Represents the output voltage value of the pinning node, i.e., the reference value;
step 202, to guarantee error eiConverging to 0, and selecting the control signal input u according to the following formula under the condition that the control target is reactive load sharingi：
In the formula, Q0Representing a reference reactive load, QiIndicating that the ith node outputs reactive power, omegaiIndicating whether the ith node has been pinned,
representing the degree of coupling between the nodes;
step 203, replacing the reactive power in the formula (3) with the output voltage according to the droop control characteristic of the distributed power supply, namely the relational expression between the output voltage and the reactive power, wherein the control signal input expression is changed into:
wherein n isQiRepresenting the reactive droop coefficient, y, of the ith node0Representing the value of the output voltage of the holddown node, i.e. the referenceValue u0A control signal input representing a holdover node; on the basis of the formula (4), determining corresponding control signal input by giving different distributed power supply outputs, and further forming a data set [ y ] of the control signal input and the voltage outputi,ui]。
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of step 30 is as follows: from the resulting data set y in step 203i,ui]A complex non-linear function f in the relation is established for step 10]、g[*]Training an artificial neural network to replace f [. X [ ]],g[*]And on the basis of the NARMA-L2-based distributed power supply controller; the distributed power controller is a specific implementation of equation (1), and provides control signal inputs u (k) corresponding to different target voltage outputs.
As a further optimization scheme of the microgrid energy management method based on the NARMA-L2 model, the specific process of step 40 is as follows:
step 401, setting a reference value of target output voltage of the distributed power supply as yr(k + d) finding an expression of a target output voltage reference value for targeting reactive power sharing in step 2, and providing a reference value for the distributed power controller constructed in step 30; a generator for providing a target voltage output reference value by adopting a gradient descent search algorithm:
wherein j | (i, j) ∈ DGDenotes that the ith, j distributed power node is a neighbor node, DGThe method is a set formed by combining all neighbor nodes in the distributed power system; j (y)i) Represents the average value of the sum of squares of the reactive power differences output by the ith node and the neighboring nodes, QjThe output reactive power of the jth node is represented, and m represents that the ith distributed power supply node has m neighbor nodes;
step 402, according to equation (5), J (y)i) When the minimum value exists, all distributed power supplies output reactive powerThe difference between the two is minimum, and the aim of taking reactive power sharing as the step 2 is achieved; to find J (y)i) Minimum time distributed power supply target output voltage reference value yr(k + d) performing the following loop until convergence;
(7) formula (I) is expressed as finding the smallest J (y)i) Value, constantly iterating yiA, denotes the search step size,
represents J (y)i) For yiDerivative, temp, representing the current cycle yiWith the convergence condition set to two cycles yiThe value change is less than or equal to a preset value, which can be 0.001.
According to the method provided by the embodiment of the invention, the NAMRA-L2 model is used for replacing a distributed power supply dynamic mathematical model, so that the model removal control of the micro-grid is realized. And meanwhile, a training data set is generated by using containment control, so that the convergence and stability of the micro-grid system are ensured, and finally, a reference value is generated by using a gradient descent search algorithm based on the generated de-modeling controller, so that the system is quickly converged.
An example is illustrated below.
An independent alternating-current microgrid structure is shown in fig. 2, microgrid energy management is carried out on the microgrid, a communication topology of the microgrid is shown in fig. 3, and relevant parameters are shown in table 1
TABLE 1 System parameters
The application effect of the micro-grid energy management under two conditions is given below
Case 1: sudden increase in load
In this case, the coupling gain c is 4, the microgrid operates in an island mode, the simulation starts at t 0, the control strategy starts at t 0.5s, and the active load of 40kW and the reactive load of 20kVA are added inside the system at t 2 s.
The step length alpha of the gradient descent search is 0.02, the strategy execution interval is 0.1s, and the convergence condition is that the reactive power difference between the adjacent nodes is less than 100 Var.
Fig. 4 (a) shows a reactive power output change process of each distributed power supply. In fig. 4 (a), the system using the NARAMA-L2 controller can achieve reactive sharing, and the system can respond quickly when the reactive load in the system suddenly increases. The balance of reactive power distribution of the system can be ensured, and reactive power circulation is reduced. Obviously, the system can be converged within 1 s.
Fig. 4 (b) shows the voltage change process of each distributed power supply in the case of a sudden load increase, and DG1 is assumed to be a holddown node. It can be seen that even with increased load, the remaining distributed supply voltage in the system does not drop below the reference voltage. That is, the pinning node sets a voltage limit that the voltage output of all DGs will remain greater than. In the power supply range, the system does not have the voltage collapse problem.
Case 2: power supply break in system
In this case, the coupling gain c is 4, the microgrid operates in an island mode, the simulation starts at t 0, and the control strategy starts at t 0.5 s. The DG3 power supply was disconnected at t-2 s, and the rest of the setup was the same as in case 1.
As shown in fig. 5 (a), after the DG3 is disconnected, the internal load of the system is taken by the rest of the distributed power supply. The voltages of the other three power sources should be reduced accordingly according to the droop control characteristics of the power generation. The voltage is stable due to the presence of the pinning node. The rest voltage is changed within a reasonable range and does not exceed the limit value, so that the quality of the electric energy is reduced.
Fig. 5 (b) shows a reactive power output state change process of the distributed power supply, and even if DG3 is disconnected, the remaining distributed power generation still can achieve the objective of reactive power balance, achieve stability of the power supply, and reduce the influence of reactive circulating current.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (3)
1. A microgrid energy management method based on a NARMA-L2 model is characterized by comprising the following steps:
step 10, establishing a relational expression between distributed power supply output and control signal input by utilizing a NARMA-L2 model;
step 20, based on a containment control concept, taking reactive power sharing as a target to obtain a data set of control signal input and voltage output;
step 30, training a neural network used for representing a complex nonlinear function in the relational expression established in the step 10 by using the data set obtained in the step 20, and fitting the nonlinear output dynamic characteristic of the distributed power supply under the condition of uniform reactive power, so as to obtain a relational expression between the trained distributed power supply output and the trained control signal input; constructing a distributed power supply controller based on NARMA-L2 according to a relation between the trained distributed power supply output and the trained control signal input;
step 40, providing a target output voltage reference value for the distributed power supply controller constructed in the step 30 by using a gradient descent search algorithm, and finishing the reactive power cooperative control of the microgrid;
in step 10, replacing a distributed power supply dynamic mathematical model with a NARMA-L2 model, and performing approximate replacement on a complex nonlinear function part in a distributed power supply system through an artificial neural network, wherein a relational expression between distributed power supply output and control signal input is as follows:
y(k+d)＝f[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]+g[y(k),y(k-1),…,y(k-n+1),u(k),u(k-1),u(k-z+1)]·u(k) (1)
wherein u (k) represents the input of the distributed power supply control signal at the k moment, y (k) represents the output voltage signal of the distributed power supply at the k moment, y (k + d) represents the output voltage signal of the distributed power supply at the k + d moment, and f and g represent complex nonlinear functions of the nonlinear part in the distributed power generation system; collecting the distributed power supply output at the k-n +1 moment in total, wherein y (k-n +1) represents the distributed power supply output at the (k-n +1) th moment; control signal input at the moment of k-z +1 is collected together, and u (k-z +1) is the control signal input at the (k-z +1) th moment;
the specific process of step 20 is:
step 201, for the distributed control micro-grid topology, converting the cooperative target into neighbor information error tracking:
in the formula, eiRepresenting the error between the i-th distributed power supply node output voltage and a reference value, yiRepresents the output voltage of the ith distributed power supply node, and j belongs to NiIndicating that the jth node is a neighbor node of the ith node, NiRepresents a set formed by all neighbor nodes of the ith distributed power node, aijRepresenting adjacency matrices in a distributed power system
The value of (i) th row and (j) th column in (1), giRepresenting the holdoff gain, y, of the ith distributed power nodejIndicating the output voltage value, y, of the jth node in communication topology with the ith node0Represents the output voltage value of the pinning node, i.e., the reference value;
step 202, to guarantee error eiConverging to 0, and selecting the control signal input u according to the following formula under the condition that the control target is reactive load sharingi：
In the formula, Q0Representing a reference reactive load, QiIndicating that the ith node outputs reactive power, omegaiIndicating whether the ith node has been pinned,
representing the degree of coupling between the nodes;
step 203, replacing the reactive power in the formula (3) with the output voltage according to the droop control characteristic of the distributed power supply, namely the relational expression between the output voltage and the reactive power, wherein the control signal input expression is changed into:
wherein n isQiRepresenting the reactive droop coefficient, y, of the ith node0Representing the value of the output voltage of the holddown node, i.e. the reference value, u0A control signal input representing a holdover node; on the basis of the formula (4), determining corresponding control signal input by giving different distributed power supply outputs, and further forming a data set [ y ] of the control signal input and the voltage outputi,ui]；
The specific process of step 40 is:
step 401, setting a reference value of target output voltage of the distributed power supply as yr(k + d) finding an expression of a target output voltage reference value for targeting reactive power sharing in step 2, and providing a reference value for the distributed power controller constructed in step 30; a generator for providing a target voltage output reference value by adopting a gradient descent search algorithm:
wherein j | (i, j) ∈ DGDenotes that the ith, j distributed power node is a neighbor node, DGThe method is a set formed by combining all neighbor nodes in the distributed power system; j (y)i) Represents the average value of the sum of squares of the reactive power differences output by the ith node and the neighboring nodes, QjThe output reactive power of the jth node is represented, and m represents that the ith distributed power supply node has m neighbor nodes;
step 402, according to equation (5), J (y)i) Is provided withWhen the reactive power is the minimum value, namely the difference between the reactive powers output by all the distributed power supplies is the minimum value, the aim of sharing the reactive power in the step 2 is achieved; to find J (y)i) Minimum time distributed power supply target output voltage reference value yr(k + d) performing the following loop until convergence;
(7) formula (I) is expressed as finding the smallest J (y)i) Value, constantly iterating yiA, denotes the search step size,
represents J (y)i) For yiDerivative, temp, representing the current cycle yiWith the convergence condition set to two cycles yiThe variation is less than or equal to a preset value.
2. The microgrid energy management method based on the NARMA-L2 model of claim 1, characterized in that the specific process of the step 30 is as follows: from the resulting data set y in step 203i,ui]A complex non-linear function f in the relation is established for step 10]、g[*]Training an artificial neural network to replace f [. X [ ]],g[*]And on the basis of the NARMA-L2-based distributed power supply controller; the distributed power controller is a specific implementation of equation (1), and provides control signal inputs u (k) corresponding to different target voltage outputs.
3. The energy management method for the microgrid based on the NARMA-L2 model is characterized in that the preset value is 0.001.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811054866.5A CN109066707B (en) | 2018-09-11 | 2018-09-11 | A Microgrid Energy Management Method Based on NARMA-L2 Model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811054866.5A CN109066707B (en) | 2018-09-11 | 2018-09-11 | A Microgrid Energy Management Method Based on NARMA-L2 Model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109066707A CN109066707A (en) | 2018-12-21 |
| CN109066707B true CN109066707B (en) | 2021-03-02 |
Family
ID=64761199
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811054866.5A Expired - Fee Related CN109066707B (en) | 2018-09-11 | 2018-09-11 | A Microgrid Energy Management Method Based on NARMA-L2 Model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109066707B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268366A (en) * | 2013-03-06 | 2013-08-28 | 辽宁省电力有限公司电力科学研究院 | A Combined Wind Power Prediction Method Applicable to Distributed Wind Farms |
| CN103488869A (en) * | 2013-08-23 | 2014-01-01 | 上海交通大学 | Wind power generation short-term load forecast method of least squares support vector machine |
| CN104659811A (en) * | 2015-01-28 | 2015-05-27 | 东南大学 | Distributed cooperative control method of micro power grid on basis of holdback |
| CN104992248A (en) * | 2015-07-07 | 2015-10-21 | 中山大学 | Microgrid photovoltaic power station generating capacity combined forecasting method |
| CN107069776A (en) * | 2017-04-12 | 2017-08-18 | 东南大学 | A kind of energy storage prediction distributed control method of smooth microgrid dominant eigenvalues |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10809678B2 (en) * | 2015-03-16 | 2020-10-20 | Board Of Regents, The University Of Texas System | System and method for distributed control of an electrical network |
-
2018
- 2018-09-11 CN CN201811054866.5A patent/CN109066707B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268366A (en) * | 2013-03-06 | 2013-08-28 | 辽宁省电力有限公司电力科学研究院 | A Combined Wind Power Prediction Method Applicable to Distributed Wind Farms |
| CN103488869A (en) * | 2013-08-23 | 2014-01-01 | 上海交通大学 | Wind power generation short-term load forecast method of least squares support vector machine |
| CN104659811A (en) * | 2015-01-28 | 2015-05-27 | 东南大学 | Distributed cooperative control method of micro power grid on basis of holdback |
| CN104992248A (en) * | 2015-07-07 | 2015-10-21 | 中山大学 | Microgrid photovoltaic power station generating capacity combined forecasting method |
| CN107069776A (en) * | 2017-04-12 | 2017-08-18 | 东南大学 | A kind of energy storage prediction distributed control method of smooth microgrid dominant eigenvalues |
Non-Patent Citations (1)
| Title |
|---|
| 基于NARMA_L2控制器的电力系统稳定性分析;薛蕊 等;《现代电子技术》;20170901;第40卷(第17期);第0-4节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109066707A (en) | 2018-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106410808B (en) | Universal micro-capacitance sensor group distributed control method comprising invariable power and droop control | |
| CN108363306B (en) | Micro-grid distributed controller parameter determination method based on linear quadratic optimization | |
| CN111614122A (en) | A Hierarchical Control Strategy for Island Microgrid Based on Event-triggered Mechanism | |
| CN107465211B (en) | Distributed fixed-time coordinated control method for island microgrid | |
| CN108075487A (en) | The hierarchical control method for the isolated island micro-capacitance sensor that adaptive sagging and uniformity is combined | |
| CN104659810B (en) | A kind of micro-capacitance sensor cooperative control method for not knowing communication topology | |
| CN105515010B (en) | A kind of secondary voltage control method for coordinating and system based on cooperative game | |
| CN110190599A (en) | A Control Strategy for Island Microgrid Based on Finite Time Consistency Theory | |
| CN106712085B (en) | A kind of isolated island microgrid voltage/frequency distribution Two-stage control method based on multi-agent system | |
| CN105391094A (en) | Distributed control method for isolated micro grid | |
| CN107834593B (en) | Droop control type island micro-grid static voltage stability probability assessment method | |
| CN110212516A (en) | A kind of direct-current grid stored energy distribution formula control method considering time delay | |
| CN107681693A (en) | Distributed coordinated control method of virtual power supply in low-voltage microgrid based on MAS | |
| CN107565550B (en) | A distribution network partition method and system | |
| CN113541146A (en) | Power flow calculation optimization method for power system considering distributed power supply | |
| Wang et al. | A novel consensus-based optimal control strategy for multi-microgrid systems with battery degradation consideration | |
| CN110120673B (en) | Distributed transmission and distribution cooperative reactive power optimization method and system based on Thevenin equivalent parameter identification | |
| CN105914752B (en) | Pilot bus selection method based on fast search and density peaks cluster | |
| CN109066707B (en) | A Microgrid Energy Management Method Based on NARMA-L2 Model | |
| CN107230982B (en) | A linearized power flow calculation method for microgrid | |
| CN110071503B (en) | Secondary planning model construction method and system for distributed transmission and distribution cooperative reactive power optimization | |
| CN105896613B (en) | A kind of micro-capacitance sensor distribution finite-time control method for considering communication time lag | |
| Chang et al. | Distributed predefined-time secondary control under directed networks for DC microgrids | |
| CN117040003A (en) | Distributed scheduled time sliding mode control method and device of island alternating current micro-grid system based on grid-structured converter | |
| CN117791617A (en) | A power resource aggregation method based on user adjustment willingness to reduce load |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200821 Address after: Four pailou Nanjing Xuanwu District of Jiangsu Province, No. 2 210096 Applicant after: SOUTHEAST University Applicant after: YANGZHOU POWER SUPPLY BRANCH OF STATE GRID JIANGSU ELECTRIC POWER Co.,Ltd. Address before: Four pailou Nanjing Xuanwu District of Jiangsu Province, No. 2 210096 Applicant before: SOUTHEAST University |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210302 |