A multi-agent based hierarchical battery equalization device and control method

Abstract

The application relates to a multi-agent-based layered battery equalization device and a control method, and belongs to the technical field of battery equalization.The layered battery equalization device is composed of two layers, wherein the upper layer is composed of a plurality of battery equalization units to realize distributed communication without a center, and the lower layer is composed of the battery equalization units and battery monomers to be equalized to realize a master-slave topology with a center; the main function of the upper layer is to realize necessary information exchange between modules; the main function of the lower layer is to calculate a current equalization strategy according to state information by adopting an improved consistency algorithm and the thought of average value control in the battery equalization unit, and to control the charging and discharging current of the battery monomers to be equalized according to the equalization strategy; and the layered battery equalization device can realize efficient equalization of a large number of batteries, maintain the performance of a battery pack and prolong the service life.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery equalization, and in particular to a layered battery equalization device and control method based on multiple agents. BACKGROUND

[0002] New energy vehicles and energy storage technologies are developing in scale, and lithium batteries have been widely used due to their excellent energy density characteristics. In engineering practice, multiple battery monomers of the same model are connected in series and parallel to form a battery pack to meet the use requirements. In the process of battery operation, due to the differences in working conditions and parameters, there is inconsistency between battery monomers, which will cause overcharging, overdischarging and life attenuation of single battery. Therefore, the consistency equalization and maintenance of lithium batteries have become a hot research topic.

[0003] Equalization can be divided into energy consumption type equalization and non-energy consumption type equalization. The current design of battery equalization device limits the number of battery monomers that can be equalized at the same time, and the battery equalization device can only simultaneously equalize a limited number of battery units. When the number of battery monomers that need to be equalized is large, a single equalization device cannot complete the task in one operation. If multiple devices are used for simultaneous equalization or a single device is used for multiple operations, the result may be that the batteries in each batch operation will reach different equalization states. This not only leads to a doubling of the time required for re-equalization of the batteries, but also increases the complexity of the operation. For example, taking 10 batteries as an example, if equalization is divided into two times, the first 5 batteries after the first equalization may reach an average of 5.5V, and the other 5 batteries after the second equalization may reach an average of 6V. This obviously fails to achieve the goal of making the 10 batteries as a whole. Such inconsistency not only affects the performance of the battery pack, but also shortens the service life of the battery, increases maintenance costs, and reduces the overall system efficiency. Therefore, developing a solution that can simultaneously and efficiently equalize a large number of battery units is crucial for maintaining the performance of a battery pack composed of numerous battery monomers and prolonging its service life.

[0004] In addition, due to the idea based on the consistency algorithm, the traditional consistency algorithm has the problem of slower and slower consistency process. As the consistency process proceeds, the deviation between the control targets gradually decreases, and according to the principle of the consistency algorithm, the reduction of the deviation will cause the overall consistency of the controlled targets to converge gradually slower, so how to use a more efficient algorithm to speed up the battery equalization process is also a key problem. SUMMARY

[0005] In view of the defects in the prior art, the purpose of the present application is to provide a layered battery equalization device and control method based on multiple agents, the upper layer of the system is a distributed multi-agent system without a center formed by multiple battery equalization units through information exchange, and the lower layer of the system is a battery monomer connected by each battery equalization unit; the control method is a layered collaborative equalization algorithm based on an improved consensus algorithm between battery equalization units and an average value control idea in the battery equalization device.

[0006] To achieve the above purpose, the technical solution adopted by the present application is:

[0007] A layered battery equalization device based on multiple agents,

[0008] The battery equalization device comprises an upper-layer multi-agent system and a lower-layer multi-agent system;

[0009] The upper-layer nodes of the upper-layer multi-agent system correspond to battery equalization units, and the lower-layer nodes of the lower-layer multi-agent system correspond to battery monomers to be equalized;

[0010] The upper-layer multi-agent system is a distributed topology without a center, and a distributed communication without a center is established between the set battery equalization units, each battery equalization unit is an upper-layer node in the upper-layer multi-agent structure, the communication relationship between nodes is equivalent to an edge of the upper-layer multi-agent system, the upper-layer multi-agent system exchanges information between modules, and forms a distributed multi-agent system without a center;

[0011] The operation of "establishing a distributed communication without a center between the set battery equalization units" is in accordance with the principle of giving priority to the nearest, that is, each unit is set to preferentially establish communication with the nearest other battery equalization unit, and to establish communication with as many battery equalization units as possible.

[0012] Each battery monomer to be equalized is a lower-layer node of the lower-layer multi-agent system; each node in this layer is controlled by the connected upper-layer node, the lower-layer multi-agent nodes do not communicate with each other, and each lower-layer node is completely controlled by the connected upper-layer node; the lower-layer nodes are connected with only one upper-layer node of the upper layer, and each upper-layer node is connected with multiple lower-layer nodes.

[0013] Preferably,

[0014] Each battery equalization unit comprises the following modules:

[0015] A voltage detection module: collecting the voltage of the battery monomer connected to the battery equalization unit at each sampling time;

[0016] A communication module: establishing communication with the nearby battery equalization units and exchanging average battery state information;

[0017] The calculation module calculates the average state information of the battery; and calculates the balancing strategy of each battery monomer according to the improved consistency algorithm and the average voltage of all battery monomers, based on the local battery voltage state information and the average voltage state information of the battery obtained by the communication module of the adjacent battery balancing unit;

[0018] The charge and discharge control module controls the charge and discharge current of each battery monomer according to the obtained battery balancing strategy.

[0019] A control method of layered battery balancing based on multi-agent, using the layered battery balancing device of any one of the above, comprising the following steps:

[0020] S1: The upper layer node of the upper layer multi-agent system collects the state of the lower layer node connected thereto, and all battery balancing units sample the voltage state information of each battery monomer connected thereto at each sampling time, and calculate the average voltage of all battery monomers to be balanced connected thereto at the current time according to the obtained voltage state information;

[0021] S2: The upper layer nodes of the upper layer multi-agent system perform non-central distributed communication, and each battery balancing unit sends the average voltage information obtained in S1 to the adjacent battery balancing unit having communication therewith according to the set agreed communication relationship (wherein adjacent refers to adjacent in the communication topology, that is, the communication relationship is established between two battery balancing units), and at the same time, each battery balancing unit receives the average voltage information of the adjacent battery balancing unit;

[0022] S3: According to the average voltage of the battery monomer to be balanced collected in S1 and the average voltage of the adjacent battery balancing unit obtained in S2, the battery balancing unit of the upper layer multi-agent system performs battery balancing strategy calculation, and each battery balancing unit calculates the balancing strategy for each battery monomer connected thereto according to the layered collaborative balancing algorithm based on the improved consistency algorithm;

[0023] S4: Each upper layer agent node controls the lower layer node according to the balancing strategy obtained in S3, that is, controls the charge and discharge current of the battery monomer corresponding to the lower layer node at the time;

[0024] S5: Repeat steps S1-S4 until the voltages of the battery monomers controlled by all battery balancing units reach the same level.

[0025] Preferably,

[0026] The layered collaborative balancing algorithm of the improved consistency algorithm in S3 is as follows:

[0027] Each of the battery balancing units calculates a balancing strategy of the battery monomer according to the collected voltage information of the connected battery monomer and the received average voltage information of the battery monomer of the adjacent battery balancing unit, defines a Laplace matrix L describing the topology of the upper multi-agent system, and the element of the i-th row and j-th column of the Laplace matrix L is

[0028]

[0029] Wherein,

[0030]

[0031] N i represents a set of other battery balancing units in communication connection with the i-th battery balancing unit; m i is the number of other battery balancing units in communication connection with the i-th battery balancing unit; the communication topology structure of the upper multi-agent system is represented by an undirected connected graph, the Laplace matrix is semi-positive definite, the smallest non-zero eigenvalue of the Laplace matrix is denoted as μ; for each battery monomer, the hierarchical collaborative balancing algorithm based on the improved consensus algorithm is:

[0032]

[0033] In the formula, T represents a preset time parameter, the value of which will affect the convergence speed of the whole system; k represents a discrete time; V i,j,k represents the voltage state of the j-th battery monomer connected to the i-th battery balancing unit detected by the i-th battery balancing unit at time k; n i represents the total number of battery monomers connected to the i-th battery balancing unit; is the average state information of all battery monomers connected to the i-th battery balancing unit; l∈N i represents that l is a battery balancing unit in N i ; λ1, λ2 are preset balancing coefficients; u i,j,k is the balancing strategy of the j-th battery monomer connected to the i-th battery balancing unit obtained by the i-th battery balancing unit at time k.

[0034] Preferably,

[0035] The charge and discharge current calculation formula of the battery monomer at this moment in S4 is:

[0036] I i,j,k =μ·u i,j,k (6)

[0037] Wherein, μ is a preset proportional coefficient; I i,j,k is the balancing current of the j-th battery monomer connected to the i-th battery balancing unit at time k, and correspondingly, the SOC change of each battery monomer is:

[0038]

[0039] Where Δt represents the unit step size of the discrete time interval, ΔS i,j,k Represents the equilibrium current I at time k i,j,k The resulting change in SOC; C i,j This is the maximum capacity of the battery.

[0040] S i,j,k+1 =S i,j,k +ΔS i,j,k (8)

[0041] Among them, S i,j,k This represents the SOC of the j-th battery cell connected to the i-th battery equalization unit at time k; the voltage change is:

[0042] V i,j,k+1 =V i,j,k +ΔV i,j,k (9)

[0043] Among them, V i,j,k ΔV represents the voltage of the j-th battery cell connected to the i-th battery equalization unit at time k. i,j,k The voltage change of the j-th battery cell connected to the i-th battery equalization unit at time k.

[0044] Preferred,

[0045] The agreed communication relationship set in S2 is that, provided that the communication capability allows, each upper-level node communicates with the nearest few adjacent upper-level nodes of the battery equalization unit, or each upper-level node communicates with any other upper-level node of the battery equalization unit.

[0046] The beneficial effects of this invention are:

[0047] 1. Each battery balancing unit can synchronize its connected battery cells to the same balancing level while simultaneously balancing all battery cells connected to different battery balancing units to the same level.

[0048] 2. This invention, based on an improved consensus algorithm and the concept of intra-module averaging, ensures that at any given time, the amount of electricity released by the discharging battery in the entire battery balancing device is equal to the amount of electricity received by the charging battery, without consuming external power during overall operation. 3. This invention improves the step size based on the consensus algorithm. As balancing progresses, the algorithm's step size adaptively increases, thereby reducing the slowdown in convergence speed and accelerating the battery balancing process.

[0049] 4. The advantages of the decentralized distributed system of the present invention

[0050] Scalability: The centerless topology is usually more flexible and easy to expand. New nodes (new battery equalization units) can easily join the network without the need for coordination by the center node;

[0051] Fault tolerance: The present application adopts a centerless topology, and when a single point (a single equalization unit) fails, it does not affect the normal equalization function of other units. Because information transmission is distributed, even if some nodes in the network fail, other nodes can continue to communicate and work, improving the stability and reliability of the network; because the communication transmission data volume between units is small, the communication requirement is low;

[0052] Load balancing: Information processing is distributed on each node (battery equalization unit), so the centerless topology can achieve load balancing, and compared with the traditional center topology, it can avoid overloading of the center node;

[0053] Autonomy: Each node (single equalization unit) can make independent decisions and operations without the instruction of the center node, which enables each part of the network to make a quick response based on local information. BRIEF DESCRIPTION OF DRAWINGS

[0054] The present application has the following drawings:

[0055] Figure 1 A schematic diagram of the present application for multiple battery equalization units to communicate with each other

[0056] Figure 2 A schematic diagram of the present application for a battery equalization unit

[0057] Figure 3 A graph theory equivalent diagram of the equalization topology of the present application

[0058] Figure 4 A flowchart of the equalization strategy execution of the present application

[0059] Figure 5 A schematic diagram of a single battery equalization unit according to the algorithm of the present application to control the equalization current DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0061] The application provides a layered battery equalization device based on multiple agents, which is composed of two layers, including an upper-layer multiple agent system and a lower-layer multiple agent system, an upper-layer node of the upper-layer multiple agent system corresponds to a battery equalization unit, and a lower-layer node of the lower-layer multiple agent system corresponds to a battery monomer to be equalized.

[0062] In order to solve the inconsistency problem of the average value of the battery monomers between different battery equalization units, based on the layered battery equalization device based on multiple agents, the application further provides a control method for layered battery equalization based on multiple agents, which includes the idea of using an improved consistency algorithm between battery equalization units and using average value control in the battery equalization unit, each battery equalization unit obtains the state information of the connected battery monomers through sampling, and the average state information of the connected battery monomers is calculated through average value; the average state information of the connected battery monomers of the similar battery equalization unit is obtained through communication, then the equalization strategy of each battery connected to the unit is updated according to the layered collaborative equalization algorithm based on the improved consistency algorithm, and the corresponding battery monomer equalization current is adjusted according to the strategy.

[0063] The method combines the ideas of the improved consistency algorithm and the state averaging of the battery monomers connected to the unit, so that each battery equalization unit can make reasonable charge and discharge control on each battery monomer connected to it according to the collected state information of the battery monomers and the average state information received from other battery equalization units, and a large number of battery monomers connected to different battery equalization units can reach the same level of equalization through one equalization operation, the iteration step of the algorithm is changed, and the method has higher efficiency.

[0064] A layered battery equalization device based on multiple agents, as shown in Figure 1 The application is a schematic diagram of communication between four battery equalization units, each battery equalization unit has the following functions:

[0065] 1) Each battery equalization unit can connect multiple battery monomers, the battery equalization unit can collect the voltage state information of the connected battery monomers through a voltage detection module, calculate the average voltage, and perform charge and discharge operation on each battery monomer according to the strategy;

[0066] 2) Each battery equalization unit is equivalent to an agent, and the agents can communicate with each other to form a multiple agent system.

[0067] As shown in Figure 2 Each battery equalization unit is composed of the following modules:

[0068] Voltage detection module: collect the voltage of the battery monomers connected to the battery equalization unit at each sampling time.

[0069] Communication module: Establish communication with similar battery equalization units, exchange average battery state information.

[0070] Calculation module: Calculate the average battery state information; according to the local battery state information and the battery average state information of the similar battery equalization units received by the communication module, calculate the equalization strategy of each battery by combining the algorithm provided by the application.

[0071] Charge and discharge control module: Control the charge and discharge current of each battery monomer according to the obtained battery equalization strategy.

[0072] Each battery equalization unit in the application has all the functions of the above modules.

[0073] The application provides a layered battery equalization device based on multi-agent,

[0074] The upper layer of the system is composed of a plurality of battery equalization units, and a distributed communication without a center is established between similar battery equalization units, which is regarded as a multi-agent system, each battery equalization unit is regarded as a node in the multi-agent system, and the communication relationship between the nodes is equivalent to an edge of the multi-agent system, the main function of the upper layer is to realize the necessary information exchange between the modules, and a distributed multi-agent system without a center is formed through information exchange; the lower layer of the system is composed of battery monomers to be equalized connected by the battery equalization unit, and each battery monomer in the lower layer multi-agent system is a lower layer node; each lower layer node in this layer is controlled by the connected upper layer node, the lower layer multi-agent nodes do not communicate with each other, and each lower layer node is completely controlled by the connected upper layer node; each lower layer node is connected with only one node in the upper layer, and each upper layer node is connected with a plurality of lower layer nodes.

[0075] The multi-agent system can be described by graph theory to describe its topological relationship, as shown in Figure 3 , which is a graph theory display of the layered distributed topology corresponding to the system. Figure 1

[0076] The application provides a layered battery equalization method based on multi-agent, and a flow chart is shown in Figure 4 When the battery monomer equalization process starts, each battery equalization unit executes the following equalization control method:

[0077] Step 1: The upper layer multi-agent system node collects the current state information of the connected lower layer node, all battery equalization units sample the voltage state information of each battery monomer connected by the battery equalization unit at each sampling time, and calculate the average voltage of the connected battery monomer at the current time according to the data obtained by sampling.

[0078] ​Step Two: The upper-layer multi-agent system nodes establish decentralized distributed communication. Each battery equalization unit sends its average voltage information to other battery equalization units according to a pre-agreed communication relationship. Simultaneously, each battery equalization unit receives voltage information from other battery equalization units. This "pre-agreed communication relationship" is pre-set by the operator and offers high flexibility. For example, provided communication capabilities allow, each node can communicate with its nearest neighboring agent nodes or any other node. Battery equalization units with communication relationships are considered topologically adjacent and are called adjacent battery equalization units.

[0079] Step 3: The battery balancing unit of the upper-layer multi-agent system performs policy calculations. Each node calculates the control policy for each slave node (battery cell). Specifically, each battery balancing unit calculates the balancing policy for each connected battery cell according to the hierarchical collaborative balancing algorithm based on the improved consensus algorithm.

[0080] The calculation process for each central node in this process is illustrated in the diagram below. Figure 5 As shown.

[0081] The hierarchical collaborative equilibrium algorithm based on the improved consensus algorithm is as follows:

[0082] Battery equalization control principle:

[0083] During the battery equalization process, the discretized battery SOC change model is represented as follows:

[0084] S k+1 =S k +ΔS k (1)

[0085]

[0086] Where k represents discrete time; S k Represents the battery's SOC; ΔS k Represents the equilibrium current I at time k k The resulting change in SOC; C is the battery's maximum capacity.

[0087] This invention uses the voltage of individual battery cells to reflect the difference in battery SOC. Each balancing module calculates the balancing strategy for each battery cell based on the collected voltage and the average voltage information received from other modules.

[0088] Algorithm parameter settings (determine algorithm parameter μ based on the upper-level system topology):

[0089] Define the Laplace matrix L of the upper-level multi-agent system, where the element in the i-th row and j-th column is...

[0090]

[0091] wherein,

[0092] N i represents other battery balancing units that establish a communication connection with the i-th battery balancing unit; m i is the number of other battery balancing units that establish a communication connection with the i-th battery balancing unit.

[0093] Since the communication topology of the upper multi-agent system can be represented by an undirected connected graph, the Laplacian matrix thereof is semi-positive definite, and the minimum non-zero eigenvalue of the Laplacian matrix is denoted as μ; for each battery monomer, the hierarchical collaborative balancing algorithm based on the improved consensus algorithm is:

[0094]

[0095] In the formula, T represents a preset time parameter, and the value of T will affect the convergence speed of the whole system; k represents a discrete time; V i,j,k represents the voltage state of the j-th battery monomer connected to the i-th battery balancing unit detected by the i-th battery balancing unit at the k-th time; n i represents the total number of battery monomers connected to the i-th battery balancing unit; that is, the average state information of all battery monomers connected to the i-th battery balancing unit; l is an element in N i represents that l is an element in N i ; λ1 and λ2 are preset balancing coefficients; u i,j,k that is, the balancing strategy of the j-th battery monomer connected to the i-th battery balancing unit by the i-th battery balancing unit at the k-th time.

[0096] The improved consensus algorithm adopted by the application can accelerate the speed of the controlled object tending to be balanced, and the average voltage state of the upper node can reach a consensus in a preset time T.

[0097] Step four: the upper node controls the charging and discharging of the lower node. Each upper node determines and controls the charging and discharging current of the corresponding lower node, i.e., the battery monomer at the moment, according to the balancing strategy obtained in step three, and the specific calculation formula is:

[0098] I i,j,k = μ·u i,j,k (6)

[0099] wherein, μ is a preset proportional coefficient; I i,j,k is the balancing current of the j-th battery monomer connected to the i-th battery balancing unit by the i-th battery balancing unit at the k-th time.

[0100] Correspondingly, the SOC change of each battery cell is:

[0101]

[0102] where Δt represents the unit step length of discrete time, ΔS i,j,k represents the SOC change caused by the balancing current I i,j,k at k time; C i,j is the maximum capacity of the battery. S i,j,k+1 = S i,j,k + ΔS i,j,k (8)

[0103] where S i,j,k represents the SOC of the jth battery cell connected to the ith battery balancing unit at k time;

[0104] The voltage change is:

[0105] V i,j,k+1 = V i,j,k + ΔV i,j,k (9)

[0106] where V i,j,k represents the voltage of the jth battery cell connected to the ith battery balancing unit at k time, ΔV i,j,k is the voltage change value of the jth battery cell connected to the ith battery balancing unit at k time.

[0107] With the progress of the balancing process, the SOC and voltage of each battery cell tend to be consistent.

[0108] Step five: repeat steps one to four until the battery cell voltage of all battery balancing units reaches the same level.

[0109] The contents not described in detail in the specification belong to the prior art known to those skilled in the art.

[0110] Finally, it should be pointed out that the above description is only a preferred example of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements to some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A hierarchical battery equalization control method based on multi-agent systems, characterized in that, Includes the following steps: S1: The upper-level nodes of the upper-level multi-agent system collect the state of the lower-level nodes connected to them. All battery equalization units sample the voltage state information of each battery cell connected to them at each sampling time, and calculate the average voltage of all battery cells to be equalized connected to them at the current time based on the obtained voltage state information. S2: The upper-layer nodes of the upper-layer multi-agent system conduct decentralized distributed communication. Each battery equalization unit sends the average voltage information obtained in S1 to the adjacent battery equalization unit with which it communicates according to the established communication relationship. At the same time, each battery equalization unit receives the average voltage information of the adjacent battery equalization unit. S3: Based on the average voltage of the battery cell to be balanced collected in S1 and the average voltage of the adjacent battery balancing unit obtained in S2, the battery balancing unit of the upper-layer multi-agent system performs battery balancing strategy calculation. Each battery balancing unit calculates the balancing strategy for the battery cell connected to each battery balancing unit according to the hierarchical collaborative balancing algorithm based on the improved consensus algorithm. S4: Each upper-layer intelligent agent node controls the lower-layer node according to the equalization strategy obtained in S3, that is, controls the charging and discharging current of the battery cell corresponding to the lower-layer node at that moment. S5: Repeat steps S1 to S4 until the voltage of all battery cells controlled by the battery equalization unit reaches the same level. The specific details of the hierarchical collaborative equilibrium algorithm of the improved consensus algorithm in S3 are as follows: Each battery balancing unit calculates a battery balancing strategy based on the collected voltage information of the connected battery cells and the average voltage information of the battery cells from neighboring battery balancing units, and defines a Laplace matrix describing the topology of the upper-layer multi-agent system. , its first i OK j Column elements are ; in, ; Representative and the i The set of other battery equalization units that establish communication connections with each battery equalization unit; In order to be with the first i The number of other battery equalization units that establish communication connections with each battery equalization unit; the communication topology of the upper-layer multi-agent system is represented by an undirected connected graph, the Laplace matrix is ​​positive semi-definite, and the smallest non-zero eigenvalue of the Laplace matrix is ​​denoted as... For each battery cell, the hierarchical collaborative equalization algorithm based on the improved consensus algorithm is as follows: ; In the formula, This represents a preset time parameter, the value of which affects the convergence speed of the entire system. k Represents discrete moments; represent k Time of the first i The battery equalization unit detected the first battery unit it was connected to. j The voltage state of an individual battery cell; Representing the i The total number of battery cells connected to each battery equalization unit. That is, the first i The average state information of all battery cells connected to each battery equalization unit; represent l for One of the battery equalization units; These are the preset balance coefficients; That is k Time of the first i The battery equalization unit's result on its connected first battery equalization unit j Balancing strategy for individual battery cells.

2. The hierarchical battery equalization control method based on multi-agent systems according to claim 1, characterized in that, The formula for calculating the charging and discharging current of a single battery cell at that moment in S4 is: ; in, This is a preset proportionality coefficient; That is k Time of the first i The battery equalization unit connects to the first... j The equalization current of each battery cell, and correspondingly, the SOC change of each battery cell is as follows: ; in, The unit step size represents the discrete time step. Represents the equalization current at time k The resulting change in SOC; This is the maximum capacity of the battery; ; in, represent k Time of the first i The battery equalization unit connected to the first j The state of charge (SOC) of each individual battery cell; voltage change is: ; in, represent k Time of the first i The battery equalization unit connected to the first j The voltage of each individual battery cell, Time of the first i The battery equalization unit connected to the first j The voltage change value of each battery cell.

3. The hierarchical battery equalization control method based on multi-agent systems according to claim 1, characterized in that, The agreed communication relationship set in S2 is that, provided that the communication capability allows, each upper-level node communicates with the nearest few adjacent upper-level nodes of the battery equalization unit, or each upper-level node communicates with any other upper-level node of the battery equalization unit.

4. The hierarchical battery equalization control method based on multi-agent systems according to claim 1, characterized in that, The battery balancing device includes an upper-level multi-agent system and a lower-level multi-agent system; The upper-level nodes of the upper-level multi-agent system correspond to the battery balancing unit, and the lower-level nodes of the lower-level multi-agent system correspond to the individual battery cells to be balanced. The upper-layer multi-agent system is a decentralized distributed topology structure. Decentralized distributed communication is established between the set battery balancing units. Each battery balancing unit is an upper-layer node in the upper-layer multi-agent structure. The communication relationship between nodes is equivalent to an edge of the upper-layer multi-agent system. The upper-layer multi-agent system exchanges information between modules to form a decentralized distributed multi-agent system. Each battery cell to be balanced is a lower-level node of the lower-level multi-agent system; Each node in this layer is controlled by the upper-level node it is connected to. The lower-level multi-agent nodes do not communicate with each other. Each lower-level node is completely controlled by the upper-level node it is connected to. Each lower-level node is connected to a single upper-level node, and each upper-level node is connected to multiple lower-level nodes.

5. The hierarchical battery equalization control method based on multi-agent systems according to claim 1, characterized in that, Each of the battery equalization units includes the following modules: Voltage detection module: Collects the voltage of the individual battery cells connected to the battery equalization unit at each sampling time; Communication module: Establishes communication with nearby battery equalization units to exchange average battery status information; Calculation module: Calculates the average state information of the battery; Based on the local battery voltage state information and the average voltage state information of the battery equalization units of similar batteries obtained by the communication module, and combined with the improved consensus algorithm and the average state voltage of all battery cells, calculates the equalization strategy for each battery cell. Charge / discharge control module: controls the charge / discharge current of each battery cell according to the obtained battery balancing strategy.

Images