Micro-grid intelligent agent control method based on finite time consistency combination

Abstract

The invention relates to a micro-grid intelligent agent control method based on finite time consistency combination. The method comprises the following steps: obtaining adjacent active information and an active normalization coefficient of adjacent intelligent agents; determining a frequency auxiliary variable and an active auxiliary variable of the target agent according to the adjacent active information and the active normalization coefficient; calculating the secondary frequency offset of the target agent according to the frequency auxiliary variable and the active auxiliary variable; determining a target frequency of the target agent based on the secondary frequency offset; and enabling the target agent to operate at the target frequency. According to the method, the relation between the P-f droop secondary control frequency offset and the active sharing and frequency deviation is analyzed, finite time consistency control is carried out on the secondary control frequency adjustment amount, the relation is applied to generation of a consistency protocol, the frequency amount is eliminated, active power proportional accurate sharing and frequency recovery can be achieved within finite time, and the consistency protocol generation efficiency is improved. The control convergence speed is improved, the communication burden is reduced, and the control is simpler.

Description

Technical Field

[0001] The present application relates to the field of microgrid control technology, and in particular to a microgrid intelligent agent control method based on finite-time consistency compounding. Background Art

[0002] A microgrid is a small-scale power system consisting of multiple distributed power sources and loads and connected to a traditional power system. Microgrids have the characteristics of high-proportion integration of renewable energy and distributed control, so advanced control algorithms are needed to achieve stable operation and coordinated management.

[0003] The consistency control algorithm is a decentralized distributed algorithm that is widely used in microgrid control. The algorithm achieves convergence and consistency of system states through information exchange between intelligent agents. However, the traditional consistency control algorithm has the characteristics of asymptotic convergence, infinite convergence time, and cannot meet the real-time requirements. In addition, in order to simultaneously achieve proportional sharing of active power and restore the frequency to the rated value, it is necessary for neighboring intelligent agents to exchange frequency and active power to calculate the active power sharing error and frequency error between neighbors to form a consistency protocol. The frequency interaction process is relatively complex to implement, requiring a large amount of communication and complex protocols, which increases the communication burden, increases the complexity of the system and increases the difficulty of implementation.

[0004] Therefore, a method is needed to perform finite-time consistency control on the secondary control frequency adjustment amount to reduce the communication burden of the microgrid. Summary of the invention

[0005] In order to solve or partially solve the problems existing in the related technology, the present application provides a microgrid intelligent body control method based on finite time consistency compound, which can achieve consistency within a limited time, solve the problems of heavy communication burden and complex algorithm, improve the control speed and simplify the implementation process.

[0006] In a first aspect, the present application provides a microgrid intelligent agent control method based on finite time consistency compounding, which is applied to a target intelligent agent, where the target intelligent agent is any one of a plurality of intelligent agents running in parallel. The microgrid intelligent agent control method comprises the following steps:

[0007] Obtain the adjacent active information and active normalization coefficient of the adjacent intelligent agent;

[0008] Determine the frequency auxiliary variable and active auxiliary variable of the target intelligent agent according to the adjacent active information and the active normalization coefficient;

[0009] Calculate the secondary frequency offset of the target intelligent agent according to the frequency auxiliary variable and the active auxiliary variable;

[0010] Determining a target frequency of the target agent based on the secondary frequency offset;

[0011] Make the target agent run at the target frequency.

[0012] Among them, the secondary frequency offset δ fi Based on the following calculation:

[0013]

[0014] e fi =-m i P i +δ fi

[0015] In the formula, e fi is the frequency deviation of the target agent, m i is the target agent’s active normalization coefficient, P i is the active information of the target agent, u fi is the frequency auxiliary variable, u Pi is the active auxiliary variable, δ fi Depend on Points earned.

[0016] Among them, the frequency auxiliary variable u fi It is expressed as:

[0017]

[0018] Where 0<γ<1, sig(x) k =sgn(x)|x| k (sgn(x) is the sign function), g f is the frequency coupling coefficient.

[0019] Among them, the active auxiliary variable u Pi Based on the finite time consistency algorithm calculation, it is expressed as:

[0020]

[0021] where c P >0, 0<γ<1, sig(x) k =sgn(x)|x| k (sgn(x) is the sign function), g P >0 is the active coupling coefficient,

[0022] Ni is the set of all neighboring agents of the target agent (numbered i), a ij is the weight of information interaction between the target agent and the adjacent agent numbered j, m i is the target agent’s active normalization coefficient, P i is the active information of the target agent, m jis the normalized coefficient of the neighboring agent’s active power, P j It is the active information of neighboring intelligent agents.

[0023] Among them, the target frequency f of the target agent i The calculation formula is:

[0024] f i =f n -mP i +δ fi

[0025] Among them, f n is the rated frequency, m i is the active normalization coefficient, P i is the output active power, δ fi is the secondary frequency offset.

[0026] Among them, the target intelligent agent realizes the integrated control of active power proportional sharing and frequency recovery based on the controller, and the active power proportional sharing and frequency recovery are integrated in the controller.

[0027] Among them, the active normalization coefficient is selected based on the inverse proportion of the rated capacity of the intelligent agent.

[0028] The technical solution provided by this application may have the following beneficial effects:

[0029] The present application provides a microgrid intelligent body control method based on finite-time consistency compound, analyzes the relationship between the Pf droop secondary control frequency offset and the active power sharing and frequency deviation, performs finite-time consistency control on the secondary control frequency adjustment amount, applies the above relationship to the generation of the consistency protocol, eliminates the frequency amount, and can achieve accurate proportional sharing of active power and frequency recovery within a limited time, thereby improving the control convergence speed, reducing the communication burden and making the control simpler.

[0030] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and other objects, features and advantages of the present application will become more apparent through a more detailed description of exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0032] Figure 1 It is a flow chart of a microgrid intelligent body control method shown in an embodiment of the present application;

[0033] Figure 2 It is a schematic diagram of the microgrid structure of the microgrid intelligent body control method shown in the embodiment of the present application;

[0034] Figure 3 It is a simulation schematic diagram of the active power sharing result of the microgrid intelligent body control method shown in the embodiment of the present application;

[0035] Figure 4 It is a microgrid frequency simulation schematic diagram of the microgrid intelligent body control method shown in the embodiment of the present application;

[0036] Figure 5 It is a simulation diagram of the active power sharing result of the traditional method;

[0037] Figure 6 It is a schematic diagram of microgrid frequency simulation using traditional methods. DETAILED DESCRIPTION

[0038] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0039] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0040] In the description of the present application, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0041] Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0042] The technical solution of the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0043] like Figure 1 A microgrid intelligent agent control method based on finite time consistency compound is shown, which is applied to a target intelligent agent, and the target intelligent agent is any one of several intelligent agents running in parallel. The microgrid intelligent agent control method includes the following steps:

[0044] S1. Obtain the adjacent active information and active normalization coefficient of the adjacent intelligent agent;

[0045] S2, determining the frequency auxiliary variable and active auxiliary variable of the target intelligent agent according to the adjacent active information and the active normalization coefficient;

[0046] S3, calculating the secondary frequency offset of the target intelligent agent according to the frequency auxiliary variable and the active auxiliary variable;

[0047] S4. Determine the target frequency of the target agent based on the secondary frequency offset;

[0048] S5. Make the target agent run at the target frequency.

[0049] The microgrid contains multiple distributed generation (DG) based on voltage source converters. DG generally uses LC filters, and its control includes a power sharing loop and a voltage tracking loop. The former calculates the voltage reference value v according to the active and reactive output of each DG. ref , generally adopts droop control; the latter controls the DG output voltage v o Tracking ref , can generally be further divided into a voltage outer loop and a current inner loop. The voltage outer loop is based on v o and v ref The difference is calculated by PI and other methods to calculate the command value i of the current inner loop ref The inner current loop controls the current of the LC filter capacitor (or filter inductor current) to track i ref In order to facilitate the design of control parameters and improve control performance, dq rotating coordinates are generally used for control. This method is very mature and will not be introduced in detail.

[0050] This method can solve the above problems based on the combination of consistency algorithm and droop control.

[0051] The communication network can be represented as a graph G = (V G ,E G ,A G ), where V G ={v1,v2,…,v N} represents N intelligent agents (in the present invention, it refers to the distributed power source of the voltage source converter); E G represents the edge, that is, the connection state (i.e., communication) between agents; the above communication can be further expressed by the adjacency matrix A G =[a ij ]∈R N×N Indicates that the element a ii =0,a ij Represents the edge (v j ,v i ) weight, if v i From v j Get information, then a ij > 0. For v i , all can be v i The agent that provides information consists of v i Neighbors, use N i If a ij =a ji , then G is called an undirected graph, otherwise it becomes a directed graph. The Laplacian matrix of the graph is the N-order matrix L G =[l ij ]∈R N×N , whose diagonal elements are l ii =Σ j∈Ni a ij , the off-diagonal elements are l ij =-a ij .

[0052] In the linear consistency algorithm, each agent follows A G Communicate with neighboring agents, obtain their state information X, and update their own state according to formula (1).

[0053]

[0054] When the graph G is a connected undirected graph or a strongly connected directed graph, 0 is L G The simple eigenvalue of L GAll other eigenvalues ​​have positive real parts. Therefore, formula (1) can ensure that the states of all agents converge to the same value asymptotically in an exponential form. One of the shortcomings of the linear consistency algorithm is that the convergence time is relatively slow. In theory, it takes infinite time to fully converge. This application is based on a finite time consistency algorithm to improve the convergence speed.

[0055] Pf droop usually has the form of equation (2):

[0056] f i =f n -mP i +δ fi (2)

[0057] where f i is the frequency of DG i, f n Rated frequency 50Hz, m i is the reactive power normalization coefficient, which is selected in inverse proportion to the rated capacity of DG i; P i To output active power and eliminate the frequency deviation caused by Pf droop, equation (2) adds a secondary frequency offset δ to the traditional droop control. fi , a simple transformation is performed to obtain formula (3):

[0058] e fi =f i -f n =-m i P i +δ fi (3)

[0059] where e fi is the frequency deviation of DG i, δ fi Obtained by the consistency algorithm. To apply (3), differentiate its two ends to obtain (4):

[0060]

[0061] in is the frequency deviation control quantity, is the active control quantity, u Pi and u fi The calculation of is based on the mutual information between DG i and its neighbor DGs. In order to improve the convergence speed, the present invention adopts finite time consistency control to calculate u Pi , as shown in formula (5):

[0062]

[0063] where c P >0, 0<γ<1, sig(x) k =sgn(x)|x| k(where sgn(x) is the sign function), g P > 0 is the active power coupling coefficient. Traditional finite-time consensus control requires the frequency information of DG i and its neighbors, which increases the communication burden and is relatively complex to implement. To avoid this problem, the present invention proposes to adopt the control strategy shown in Equation (6).

[0064]

[0065] Among them, in the calculation process of (6), the relationship of Equation (3) is utilized to avoid using frequency information. δ is calculated through (3) and (4) fi , and then the frequency f of DG i is calculated using Equation (3) i . Compared with the traditional method, this process only uses the active power information of DG, greatly reducing the communication burden and making the control simpler.

[0066] Next, based on the embodiments, it is proved that the active power can be shared proportionally and the frequency can be restored to the rated value through the above control strategy. Before that, several important lemmas and inequalities are introduced.

[0067] Lemma 1: For a positive definite continuous function V(t), t ≥ 0, if there exist real numbers K > 0 and 0 < α < 1, and

[0068] then V(t) will reach 0 within a finite time t * where t * ≤ V(0) 1-α / (K(1 - α)), and for any t > t * , V(t) = 0.

[0069] Lemma 2: For an undirected graph, Equation (8) holds:

[0070]

[0071] Lemma 3: For the Laplacian matrix L of an undirected graph G , L G all eigenvalues are real numbers. Denote λ2 as the second smallest eigenvalue of L G . If 1 T x = 0, then, x T L G x ≥ λ2(L G )x T x.

[0072] Lemma 4: If a i > 0 and 0 < r < 1, then:

[0073]

[0074] definition:

[0075]

[0076] According to e Pi Definition of, for undirected communication graphs, equations (10) and (11) hold.

[0077]

[0078] Then formula (5) can be further rewritten as formula (12):

[0079]

[0080] Consider the Lyapunov function (13):

[0081]

[0082] where e P =[e P1 e P2 … e PN ] T , e f =[e f1 e f2 … e fN ] T . Define an undirected graph G P , its adjacency matrix is: The corresponding Laplace matrix is ​​denoted as L P .

[0083] Derivative (13) and use it to obtain (14):

[0084]

[0085] According to Lemma 3, we have (15):

[0086]

[0087] The control convergence time is t f , combined with Lemma 1, the upper limit of the convergence time is shown in (16):

[0088]

[0089] In order to verify the effectiveness of the invention, a comparative simulation study was conducted on the consistency control and the finite time integrated consistency control FICC proposed by the present invention. Figure 1 As shown, it includes 4 DGs connected to each other through RL lines; loads 1, 2, and 3 are Y-connected series RL constant impedance loads, and the electrical parameters are shown in Table 1. The present invention requires that the communication topology is an undirected graph, such as Figure 2 The network is shown in grey.

[0090] Step 1: Calculate the frequency auxiliary variable of agent i and active auxiliary transformer quantity

[0091] Step 2: Utilization calculate Further integration gives δ fi .

[0092] Step3: Use f i =f n -mP i +δ fi Calculate the frequency f of agent i i , integrate and multiply by 2π to get the voltage angle θ i =2π∫f i dt; Use the traditional method to calculate the voltage amplitude V of agent i i , the combined command voltage V of agent i i sinθ i If agent i is three-phase, its b and c phase voltage command values ​​are and

[0093] Step 4: Control agent i to track the command voltage in step 3.

[0094] Active and reactive power consistency control use the same communication network, and the adjacency matrix is ​​shown in (17).

[0095]

[0096] Table 1 Electrical parameters of microgrid

[0097]

[0098] The capacity ratio of the four DGs DG1:DG2:DG3:DG4 is 2:1:2:1. The simulation results are as follows Figure 3-Figure 6 In order to compare whether the active power is shared in proportion, the output active power of DG1 to DG4 is multiplied by the inverse ratio of the capacity (i.e., multiplied by 1, 2, 1, and 2 respectively) to obtain the normalized active power. contrast Figure 3 and Figure 5It can be seen that both methods can achieve proportional sharing of active and reactive loads. Comparing 4 and 6, it can be seen that the frequency deviation of both methods is very small and can be quickly restored to the rated value. In summary, the method of the present invention only needs to exchange active power information between DGs, without exchanging frequency information like the traditional method, and can achieve the same active proportional sharing and frequency recovery effect as the traditional method.

[0099] Finally, it should be noted that, in this article, relationships such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms include, include or any other variations are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0100] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0101] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A microgrid intelligent agent control method based on finite time consistency compound, characterized in that: Applied to a target intelligent agent, the target intelligent agent is any one of a number of intelligent agents running in parallel, the microgrid intelligent agent control method comprises the following steps: Obtain the adjacent active information and active normalization coefficient of the adjacent intelligent agent; Determine the frequency auxiliary variable and active auxiliary variable of the target intelligent agent according to the adjacent active information and the active normalization coefficient; Calculate the secondary frequency offset of the target intelligent agent according to the frequency auxiliary variable and the active auxiliary variable; Determining a target frequency of the target agent based on the secondary frequency offset; Make the target agent run at the target frequency.

2. The microgrid intelligent agent control method based on finite time consistency compounding according to claim 1 is characterized in that: Second frequency offset δ fi Based on the following calculation: e fi =-m i P i +d fi In the formula, e fi is the frequency deviation of the target agent, m i is the target agent’s active normalization coefficient, P i is the active information of the target agent, u fi is the frequency auxiliary variable, u Pi is the active auxiliary variable, δ fi Depend on Points earned.

3. The microgrid intelligent agent control method based on finite time consistency compounding according to claim 1 is characterized in that: Frequency auxiliary variable u fi It is expressed as: Where 0<γ<1, sig(x) k =sgn(x)|x| k (sgn(x) is the sign function), g f is the frequency coupling coefficient.

4. The microgrid intelligent agent control method based on finite time consistency compound according to claim 1 is characterized in that the active Auxiliary variable u Pi Based on the finite time consistency algorithm calculation, it is expressed as: where c P >0, 0<γ<1, sig(x) k =sgn(x)|x| k (sgn(x) is the sign function), g P >0 is the active coupling coefficient, Ni is the set of all neighboring agents of the target agent (numbered i), a ij is the weight of information interaction between the target agent and the adjacent agent numbered j, m i is the target agent’s active normalization coefficient, P i is the active information of the target agent, m j is the normalized coefficient of the neighboring agent’s active power, P j It is the active information of neighboring intelligent agents.

5. The microgrid intelligent agent control method based on finite time consistency compounding according to claim 1 is characterized in that: The target frequency f of the target agent i The calculation formula is: f i =f n -mP i +δ fi Among them, f n is the rated frequency, m i is the active normalization coefficient, P i is the output active power, δ fi is the secondary frequency offset.

6. The microgrid intelligent agent control method based on finite time consistency compounding according to claim 1 is characterized in that: The target intelligent agent realizes the integrated control of active power proportional sharing and frequency recovery based on the controller, and the controller integrates active power proportional sharing and frequency recovery.

7. The microgrid intelligent agent control method based on finite time consistency compounding according to claim 1 is characterized in that: The active normalization coefficient is selected based on the inverse proportion of the agent's rated capacity.

Images