CN107147102A - Networked distributed coordinated control method for DC microgrid based on multi-agent - Google Patents

Abstract

The invention discloses a kind of direct-current grid networking distributed and coordinated control method based on multiple agent, the how intelligent distributed coordination control framework of two-stage is built；Design distributed coordination linear quadratic control strategy；Design improved outer shroud droop control device；Design inner ring voltage/current controller；It relies on information network system to carry out decision-making and controls to share with load proportion with voltage uniformity is performed.The distributed and coordinated control that the invention is proposed has taken into full account the network topology switching of information system and the uncertainty of transmission time lag, realizes the dynamic stability control under information physical system integrated environment.

Description

Networked distributed coordinated control method for DC microgrid based on multi-agent

technical field

The invention belongs to the field of micro-grid control under the background of energy Internet, and in particular relates to a multi-agent-based networked distributed coordinated control method of a DC micro-grid.

Background technique

In recent years, the proportion of DC distributed power generation such as photovoltaics, fuel cells and energy storage in distributed energy has been increasing; in addition, DC loads are also on the rise. For example, the expansion of electric vehicle charging piles implies that the DC load of electric vehicles Will be widely connected to the grid. It is also because of the increasing trend of DC power sources and loads that people are interested in the research of DC microgrids. DC microgrids have the following characteristics: (i) power transmission and distribution are more efficient because there is no reactive power transmission in DC microgrids; (ii) compared to AC microgrids, DC microgrids can supply DC loads more efficiently because AC microgrids require two inverter conversions of "AC to DC" and "DC to DC" to power DC loads; while DC microgrids only need one conversion of "DC to DC", thus improving efficiency.

For the control of microgrids, a layered control scheme is generally adopted: that is, the three-layer control is responsible for implementing the optimal dispatch of distributed generation to achieve the best power allocation for distributed generation; the second-layer control is to maintain the grid voltage. Based on the real-time dynamic information of the whole micro-grid, the reference voltage value of each distributed generation is determined and issued in real time; according to the reference voltage or power, the first layer performs local decentralized dynamic adjustment; among them, the control of the third layer and the second layer is based on the communication network Centralized control, while a layer of master control is a typical decentralized control. Corresponding to this hierarchical centralized control scheme, any network failure will cause the corresponding unit information transmission failure, which will inevitably affect the normal operation of other units, and even lead to chain failures such as system overload and system-level instability in some units.

In order to overcome the above problems, distributed coordinated control has attracted attention. If the distributed coordinated control scheme is introduced into the secondary and main control of the microgrid, it will significantly improve the reliability, scalability and reduce communication pressure of the system. effect. In any case, the distributed coordinated control of the DC microgrid must achieve two control objectives, that is, consistent regulation of the entire grid voltage and proportional load sharing. The voltage regulation of the whole network means that the bus voltage of all distributed generation units of the whole microgrid needs to be adjusted consistently according to the reference voltage set by the upper layer; the load proportion sharing is to share the load according to the rated capacity of each distributed generation, thus effectively avoiding the Circulation and overloading between distributed generation. However, the traditional distributed droop control voltage regulation and load sharing capabilities are poor. One of the reasons is that the low-voltage line impedance produces a large voltage drop. The second reason is that the inconsistency of different distributed generation rated voltages leads to poor performance load sharing. At present, the feasible solution to the above problems is to build centralized control based on the network system where any two points of the entire network are connected. However, even though this solution improves the control performance, it reduces the reliability of the control, because any line failure will destroy the control function of the whole system.

In order to overcome the shortcomings of centralized and traditional distributed control, the present invention proposes a networked distributed coordinated control method for DC microgrids based on multi-agents, which not only improves the reliability of control, but also greatly improves the voltage regulation and control of the entire network. Ability to share loads proportionally.

Contents of the invention

The technical problem to be solved by the present invention is to provide a multi-agent-based networked distributed coordinated control method for DC microgrids, which relies on the information network system to make decisions and execute voltage consistency control and load ratio sharing . The distributed coordinated control proposed by the invention fully considers the network topology switching of the information system and the uncertainty of the transmission time lag, and realizes the dynamic stability control under the fusion environment of the information-physical system.

The present invention adopts the following technical solutions to solve the above-mentioned technical problems

A networked distributed coordinated control method for DC microgrid based on multi-agents, which specifically includes the following steps:

Step 1, build a two-level multi-intelligence distributed coordination control architecture;

Step 2, designing a distributed coordinated secondary control strategy;

Step 3, design an improved outer loop droop controller;

Step 4, design the inner loop voltage/current controller.

As a further preferred solution of the multi-agent-based DC microgrid networked distributed coordinated control method of the present invention, the step 1 specifically includes the following steps:

Step 1.1, connect each distributed generation inverter control unit with a first-level unit control agent;

Step 1.2, connect each first-level unit control agent with a second-level distributed coordination control agent;

As a further preferred solution of the multi-agent-based DC microgrid networked distributed coordinated control method of the present invention, the design of the distributed coordinated secondary control strategy described in step 2 specifically includes the following two strategies:

2.1, Voltage consistency control strategy:

Among them, u _jv (t) is the voltage consistency control strategy, j∈{1,2,…,N}; K _j,σ(t) is the voltage control gain; v _ref is the reference voltage for voltage consistency adjustment of the whole network , v _j (t) is the bus voltage of the jth inverter unit system, v _k (t) is the bus voltage of the kth inverter unit system; j∈{1,2,…,N}, v _j (t)=0, if t<0;

2.2, load current proportional sharing control strategy:

Among them, u _jc (t) is the load current proportional sharing control strategy, j∈{1,2,…,N}; For the control gain, i _jpu (t) is the current of the jth inverter unit, and i _kpu (t) is the current of the kth inverter unit; is the network connection matrix coefficient.

As a further preferred solution of the multi-agent-based DC microgrid networked distributed coordinated control method of the present invention, in step 3, the outer loop droop controller is specifically as follows:

Among them, d _j is the gain of the droop controller; u _jref is the rated voltage, which is also the reference voltage of the inner loop controller; I _jrated i _jpu is the product of the rated current of the jth inverter unit and the unit current; k _j is the voltage Transfer controller gain; k∈Ω _j , Ω _j is the set of secondary agents adjacent to the jth distributed coordination agent, the total number is M _j .

As a further preferred solution of the multi-agent-based DC microgrid networked distributed coordinated control method of the present invention, in step 4, the inner loop voltage/current controller is specifically as follows:

Among them, G _j ∈ R ^1×2 is the controller gain, x _j (t) = [Δu _j (t), Δi _tj (t)] ^T is the state variable; for control input.

Compared with the prior art, the present invention adopts the above technical scheme and has the following technical effects:

1. Through each distributed power generation inverter control unit, a two-level multi-intelligence distributed coordination control scheme is proposed, which only uses the interactive information between adjacent two-level distributed coordination multi-agents to effectively implement voltage consistency and load proportion sharing control, which not only ensures the reliability of the control scheme, but also greatly improves the ability of voltage regulation and load proportion sharing of the whole network;

2. The multi-agent control architecture is easy to add, revoke or modify unit function agents according to the "plug and play" operation of the distributed power generation unit. In addition, it can also flexibly reorganize the secondary distributed power generation unit according to the topology changes of the information system. Coordinate and control the interaction mode of agents, so the scheme has strong scalability and compatibility;

3. When designing the secondary control strategy, fully consider the time-varying nature of the information network transmission time-delay and the saturation of the control actuator, and propose a networked voltage consistency and The control strategy of load proportion sharing consistency;

4. When designing and improving the droop controller, a corresponding voltage compensation mechanism is designed based on the droop characteristics of the inverter, and the secondary control strategy of voltage and load current is used to effectively eliminate the line impedance and the rated voltage mismatch of distributed generation. Effects of voltage regulation and load sharing.

Description of drawings

Figure 1 is a two-level distributed coordination control framework based on multi-agents;

Fig. 2 is the distributed coordination control scheme of the jth distributed generation inverter control unit;

Fig. 3 is a circuit structure diagram of the jth distributed generation inverter control unit;

Fig. 4 is a block diagram of the distributed coordinated control strategy execution of the jth distributed generation inverter control unit.

detailed description

Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail:

In order to achieve the above object, the present invention studies the following contents:

Construct a two-level multi-intelligent distributed coordinated control architecture: each distributed generation inverter control unit is connected to a first-level unit control agent, which decides and executes local decentralized control, that is, the main control; and each first-level intelligent body Connecting a second-level distributed coordinated control agent is mainly responsible for the whole network voltage consistency adjustment and load proportion sharing, that is, secondary control; in addition, the second-level agent only interacts with its adjacent second-level multi-intelligence through the information network system.

Design a distributed coordinated secondary control strategy: (1) In order to uniformly adjust the voltage of the entire network according to the reference voltage value given by the upper layer, a voltage regulator is designed in each secondary distributed coordinated control agent to implement the voltage Consistency control strategy for pinning. (2) In order to realize the load proportion sharing of all distributed inverter control units, a current regulator is also designed in each second-level distributed coordinated control agent to implement the control strategy of load current proportion sharing. In addition, when designing the voltage and current consistency control strategy, only the state information of the adjacent second-level distributed coordinated control agents is used, thus reducing the network transmission pressure; at the same time, the change of the network topology of the information system and the transmission time delay are fully considered. The time-varying nature and the saturation of the inverter control actuators realize the secondary control under the fusion of cyber-physical systems.

Design an improved outer-loop droop controller: In each unit control agent, the improved outer-loop droop controller is designed as follows: 1) Using the voltage consistency control strategy from the second-level agent of the unit, inverse In the droop characteristics of the inverter, synthesize a correction item of the inner loop voltage control reference value; 2) use the load current ratio sharing control strategy from the unit's secondary agent to synthesize a voltage shift in the inverter droop characteristics of the unit system Improve the item to further fine-tune the reference value of the inner loop voltage control; 3) use the current proportion sharing control strategy of the second-level agent of the unit, and the current proportion sharing control strategy of all the second-level agents adjacent to the unit, based on The unit system inverter droop characteristic synthesizes a reference value for the inner loop current controller. The above-mentioned improved outer loop droop controller based on the secondary control strategy design can effectively eliminate the voltage drop caused by the line impedance and the poor load sharing caused by the inconsistent rated voltage of the distributed generation through the network distributed coordination , which greatly improves the voltage regulation and load sharing capabilities of distributed droop control.

Design the inner-loop voltage/current controller: In each unit control agent, instead of the traditional PI control, the inner-loop voltage/current controller is designed as an H∞ robust controller, which improves the performance of distributed generation units. Robustness to uncertain factors.

detailed description

The implementation scheme of building a two-level multi-intelligence distributed coordinated control architecture: Taking the physical system of DC microgrid as the research object and supporting by the information system, a two-level multi-agent control architecture is constructed as shown in Figure 1, aiming at the first-level unit control On the basis of the primary master control of the agent, the secondary distributed multi-agent agent realizes the secondary control of the whole network voltage consistency and load proportion sharing under the collaborative interaction mode. The structure and functions of the two-level multi-agent are as follows: (1) Each distributed generation inverter control unit is connected to a first-level unit control agent, which determines and executes local decentralized control, that is, the main control; the unit controls the agent It is designed as a hybrid BDI agent with a reaction layer and a deliberation layer. The reaction layer includes "perception, recognition and execution" modules, which can quickly respond to changes in the operating environment, thus ensuring the adaptability of the unit system to environmental changes ; The deliberation layer includes the "belief, desire and intention" functional module, which can process the state of the distributed power generation unit into knowledge information, and use this to intelligently make decisions and execute the local decentralized dynamic control of the unit system. (2) Each first-level agent is connected to a second-level distributed coordination control agent. Through the information network system, only the second-level multi-intelligence interaction information adjacent to it is mainly responsible for the voltage consistency adjustment and load ratio sharing of the entire network. Secondary control; its structure is designed as a BDI agent composed of "belief, desire and intention" functional modules, its "belief" module filters and screens standardized knowledge information from information systems, useful standardized knowledge information based on "belief", According to the "wish" of the control of voltage consistency regulation and load proportion sharing of the secondary agent, intelligent decision-making and execution of secondary control are made in the "intent" module.

The multi-agent control framework constructed by the present invention adopts the "master-slave" interaction mode between vertical agents, that is, the second-level agent sends a distributed coordination and consistency control strategy to the first-level agent, and the first-level agent receives the strategy. And based on the droop characteristics of the inverter unit, an improved outer loop droop controller is synthesized, and then the reference voltage and reference current are sent to the inner loop voltage/current controller to perform distributed coordinated secondary control and decentralized local control; and Horizontal multi-agents at the same level interact in a "non-master-slave" manner, that is, they have equal interaction rights.

Design the implementation scheme of the distributed coordinated secondary control strategy: corresponding to the two-level multi-agent control architecture shown in Figure 1, taking the jth distributed generation inverter control unit as a research example, the distributed coordinated control scheme is designed as Figure 2 shows. The purpose of the secondary control of the present invention is to realize voltage consistency adjustment and load proportional sharing based on the information network system, so the design methods of voltage consistency and load current proportional sharing control strategies are given below.

(1) Voltage consistency control strategy

In the context of deep integration of cyber-physical systems, considering changes in the network topology of information systems, the switching signal of the information network is defined as a piecewise connection function σ(t), so the information network with switching topology is described as ψ _σ(t) , and the corresponding network before and after each switch has connectivity. In addition, considering the time-varying nature of information system transmission delay, define 0≤τ(t)≤τ, 0≤t≤∞, where τ is the upper limit value, so it is necessary to use the available u _jv (t-τ(t )) instead of u _jv (t). Based on the information network system of topology switching and transmission time-delay uncertainty, the invention proposes a pinning-based voltage consistency control strategy.

The bus voltage dynamics of the jth inverter unit system under voltage consistency control is described as

Among them, v _j (t) is the bus voltage (per unit value) of the jth inverter unit system; η _Δ u _jv (t) is a consistent control strategy subject to actuator saturation. For a positive scalar Δ, the actuator saturation η _Δ :R→R. is described as

in

The pinning-based voltage consistency control strategy is designed as

Among them, j∈{1,2,…,N}; v _j (t)=0, if t<0; K _j,σ(t) voltage control gain; v _ref is the reference voltage for voltage consistency regulation of the whole network.

The difference between the bus voltage of the jth inverter unit and the reference voltage is: Then the whole network voltage dynamics can be described as:

in, Π _σ(t) =K _σ(t) Γ(t)(L _σ(t) +F _σ(t) );

K _σ(t) =diag[k _1,σ(t) ,k _2,σ(t) ,…,k _N,σ(t) ]; Γ(t)=diag[Γ ₁ (t),Γ ₂ (t),...,Γ _N (t)]; F _σ(t) = diag[f _1,σ(t) ,f _2,σ(t) ,...,f _N,σ(t) ].

Theorem 1: Aiming at the dynamic regulation (5) of the whole network voltage, based on the voltage consistency control strategy (4), if there is a matrix Y∈R ^N×N and a symmetric positive definite matrix P,Q,X,Z∈R ^N×N satisfying

with

Among them, Φ ₁₁ = 2PΠ _σ(t) +τX+Y+Y ^T +Q, Φ ₁₂ = Φ ₂₁ = PΠ _σ(t) -Y, Then the voltage consistency regulation lim _t→∞ v _j (t)≡0 can be obtained, where j∈{1,2,…,N}.

According to Theorem 1, when the topology of the information system network is determined, the transmission delay can meet the requirements of the upper limit τ by choosing K _σ(t) reasonably, and then K _σ(t) is determined.

(2) Load current proportional sharing control strategy

Since the secondary control is to ensure voltage consistency regulation and load current ratio sharing at the same time, when designing current ratio sharing, it is still based on the same network topology as voltage consistency regulation, but has a different network weight correlation matrix. At this time, the information The network is described as: and

The current dynamics of the jth inverter unit based on the current proportional sharing control strategy is described as

Where i _jpu (t) is the current of the jth inverter unit; η _Η u _jc (t) is the current proportional sharing control strategy.

Considering the time-delay uncertainty of network transmission, the current proportional sharing control strategy is

where j∈{1,2,…,N}; for the control gain.

For current ratio sharing, Theorem 1 is still satisfied, but in the linear matrix inequality, in, According to Theorem 1, select the appropriate control gain To meet the requirements of the transmission time delay upper limit τ.

Design and improve the implementation of the outer loop droop controller: the first-level main control of the present invention is composed of the outer loop improved droop controller and the inner loop voltage/current controller, the outer loop controller provides voltage and current reference values for the inner loop, The design method of the droop controller with improved outer loop is given below.

Unequal distribution of nominal voltage and line load among DGs can cause currents to deviate significantly from proportional share values. Although this problem can be solved by increasing the droop characteristic gain, a large droop gain brings a large voltage drop. Therefore, the present invention proposes to construct an improved droop controller using a two-level consistency control strategy

Among them, d _j is the droop controller gain; u _{jref is} the rated voltage, which is also the reference voltage of the inner loop controller; k _j is the voltage transfer controller gain; k∈Ω _j , Ω _j is The set of adjacent secondary agents, the total number is M _j .

Compared with the traditional droop characteristic, the improved droop controller (9) has a voltage boost term Also known as a voltage transfer controller, this controller is composed of the per unit value of its inverter unit current and the jth inverter rated current transmitted by all adjacent second-level agents. The transfer controller gain k _j is determined based on the following criteria.

(1) Voltage transfer controller It should be as close as possible to d _j I _jrated i _jpu to ensure that the voltage drop is fully compensated so that the operating voltage is close to or equal to its rated value.

(2) The voltage boost value of the droop characteristic of all distributed generation inverters should be equal to ensure that the voltage transfer does not affect the current sharing, that is

(3) The transfer controller gain should be smaller than the droop gain, namely

k _j ≤ d _{j j} ∈ {1,2,…,N} (11)

Based on the designed voltage transfer controller, when the load increases, the voltage value can be appropriately increased to make the operating voltage close to or equal to its rated value.

According to the improved droop controller, the set value of the inner loop reference voltage is

Substituting equations (1) and (7) to Eq. (12), the reference voltage value of the inner loop is determined as

The inner loop current reference value is

Equations (13) and (14) show that: in the control based on the two-level agent proposed by the present invention, the first-level unit control agent needs to rely on the consistency control strategy sent by all the adjacent agents in the second level to construct an improved droop controller, and then realize the effective compensation of the droop characteristic, on this basis, provide the reference value to the inner loop controller; the inner loop controller performs effective dynamic adjustment according to the voltage and current reference values provided by the outer loop droop controller . This process shows that the voltage and current of the inverter unit are adjusted in real time according to the interactive information of adjacent multi-agents, so it is called distributed coordinated control.

Design the implementation of the inner loop voltage/current controller: The design method of the inner loop voltage/current controller is given below.

In order not to lose generality, it is assumed that the jth DG and the adjacent m-1 DGs pass through the transmission line (R _jk > 0, L _jk > 0, k∈Ω{1,2,,...m }, k≠j) are connected. The circuit structure diagram of the jth distributed generation inverter unit is shown in Figure 3, which supplies power to the internal load connected to the PCC point through an LC filter.

The dynamic equation of the jth distributed generation inverter unit is as follows

Wherein, all parameters and variables of the equation have been marked in Fig. 3; and defined: Δu _j =u _j -u _jref ; Δi _tj =i _tj -i _jref .

Since i _jk is the current between the jth and kth DC buses, that is, di _jk /dt=-di _kj /dt=0, then

i _jk ＝-i _kj ＝(u _k -u _j )/R _jk (16)

Since the voltage consistency of the whole grid is guaranteed by secondary control, this will make Δi _j very small, so Δi _j in equation (15) can be treated as a current disturbance, then the dynamics of the jth distributed generation inverter unit The equation can be described as

Among them, x _j (t) = [Δu _j (t), Δi _tj (t)] ^T is the state variable; is the control input; ω _j (t) = Δi _j disturbance vector; the matrix coefficients are as follows:

The voltage/current controller is designed to

where G _j ∈ R ^1×2 is the controller gain.

The dynamic equation of the jth distributed generation inverter unit under the controller (19) is

in

For the purpose of robust control, the H _∞ control index is given here as

Among them, t _f is the control termination time; Q _j =Q _j ^T >0 and P _j =P _j ^T >0 are the weight matrix; ρ _j is the robust performance index.

Theorem 2: The control unit (20) of the j-th distributed generation inverter is sensitive to any disturbance Has the robust and stable performance described in (21), except that P _j = P _j ^T > 0 satisfies the following linear matrix inequality

The controller design problem of the above Theorem 2 can be transformed into an LMI convex optimization problem:

Subject to P _j ＝P _j ^T ＞0 and (22). (23)

By solving the convex optimization problem of (23), the parameters of the inner loop controller can be obtained.

The execution block diagram of the distributed coordination control strategy of the jth distributed generation inverter control unit based on two-level agents is shown in Figure 4.

It is easy for those skilled in the art to understand that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, etc., All should be included within the protection scope of the present invention.

Claims (5)

1. A multi-agent-based networked distributed coordination control method for a direct current microgrid is characterized by comprising the following steps: the method specifically comprises the following steps:

step 1, constructing a two-stage multi-intelligent distributed coordination control architecture;

step 2, designing a distributed coordination secondary control strategy;

step 3, designing an improved outer ring droop controller;

and 4, designing an inner ring voltage/current controller.

2. The multi-agent based networked distributed coordination control method for the direct current microgrid according to claim 1, characterized in that: the step 1 specifically comprises the following steps:

step 1.1, connecting each distributed generation inverter control unit with a primary unit control intelligent agent respectively;

and 1.2, connecting each primary unit control intelligent agent with a secondary distributed coordination control intelligent agent.

3. The multi-agent based networked distributed coordination control method for the direct current microgrid according to claim 1, characterized in that: the design distributed coordination secondary control strategy described in the step 2 specifically includes the following two strategies:

2.1, voltage consistency control strategy:

wherein u is_jv(t) is a voltage consistency control strategy, j ∈ {1,2, …, N }; K_j,σ(t)Controlling the gain for the voltage; v. of_refReference voltage, v, regulated for full network voltage uniformity_j(t) is a bus voltage of the jth inverter unit system, v_k(t) is a bus voltage of the kth inverter unit system, j ∈ {1,2, …, N }, v_j(t)＝0,if t＜0；

2.2, load current proportion sharing control strategy:

wherein u is_jc(t) load current proportional share control strategy, j ∈ {1,2, …, N };to control the gain, i_jpu(t) is the jth inverter cell current, i_kpu(t) is the kth inverter unit current;are network connection matrix coefficients.

4. The multi-agent based networked distributed coordination control method for the direct current microgrid according to claim 1, characterized in that: in step 3, the outer ring droop controller is specifically as follows:

wherein d is_jIs the droop controller gain; u. of_jrefThe reference voltage is the rated voltage and is also the reference voltage of the inner ring controller; i is_jratedi_jpuIs the product of the rated current of the jth inverter unit and the per unit current, k_jIs the gain of the voltage transfer controller, k ∈ omega_j,Ω_jIs a set of secondary agents adjacent to the jth distributed coordination agent, and the total number is M_j。

5. The multi-agent based networked distributed coordination control method for the direct current microgrid according to claim 1, characterized in that: in step 4, the inner loop voltage/current controller is specifically as follows:

wherein G is_j∈R^1×2Is the controller gain, x_j(t)＝[Δu_j(t),Δi_tj(t)]^TIs a state variable;is a control input.