CN114400387A - Battery equalization management method and system based on multi-agent game - Google Patents

Abstract

The invention provides a battery equalization management method and a system based on a multi-agent game, which comprises the following steps: determining the charge state data of each single battery according to the attenuation data and the current electric quantity data of each single battery in the battery pack; obtaining a charge-discharge strategy of each single battery in the battery pack through a Nash equilibrium method according to the difference value of the charge state data of each single battery and the charge state data of the other single batteries in the battery pack; and carrying out battery balance management according to the charge and discharge strategy. Compared with the prior art, the charge-discharge strategy of each single battery in the battery pack is obtained by the difference value of the charge state data of each single battery and the charge state data of other single batteries through a Nash equilibrium method, the influence of the inconsistency of the single batteries on the available capacity of the whole battery pack is reduced, the service life of the battery is prolonged, and the energy utilization rate of the battery is improved.

Description

Battery balance management method and system based on multi-agent game

technical field

The invention relates to the technical field of optimization algorithms, in particular to a method and system for battery balance management based on multi-agent games.

Background technique

Energy storage batteries and electric vehicle power batteries are the key equipment for the transformation of the energy structure and the promotion of green transportation. Both energy storage batteries and electric vehicle power batteries meet larger capacity requirements and higher voltage requirements by grouping large-scale single cells in series and parallel. Due to the limitation of the "cask principle", the performance of the entire battery pack is limited by each single cell. However, there are inevitably some differences in the manufacturing process, operating temperature, and depth of discharge of single cells. The inconsistency of voltage and state of charge during the charging and discharging process may lead to battery misfire, accelerated aging and other consequences. Seriously affect battery performance. Therefore, the status of a single cell or a battery pack needs to be balanced by the Battery Balancing System (BBS) in the Battery Management System (BMS), so that the battery terminal pressure or state of charge tends to be consistent, thereby improving the Battery performance, prolong battery life.

From the perspective of energy dissipation and transfer, various equalization methods can be divided into two categories: active equalization and passive equalization. Passive equalization is to suppress the rise of the voltage of the single battery through the parallel resistance of the single battery or the shunt of the switching device during the battery charging process. It is a passive energy consumption balancing scheme, the balancing current is small, the balancing time is long, and heat is generated due to energy consumption during the balancing process. However, since the equalization circuit is simple and easy to implement, this equalization scheme can achieve a certain equalization effect when the inconsistency of single cells is relatively light. Active balancing mainly uses capacitors, inductors or transformers as energy storage or energy transmission elements, and transfers energy between single cells and single cells or between single cells and battery packs through switching devices. Ideally, it is a non-energy-consuming equalization scheme, but in practical applications, due to the switching loss of switching devices, a small amount of energy is lost in the equalization circuit during the equalization process.

Patent document CN113394840A discloses an intelligent balance control method and system for energy storage batteries, including: calculating the remaining power of each energy storage battery to obtain the average expected remaining power of all energy storage batteries; Compared with the average expected remaining power, the Nash equilibrium method is used to obtain the optimal charging and discharging strategy adopted by each energy storage battery; the battery energy balance is carried out based on the optimal charging and discharging strategy. However, this method compares the remaining power of each energy storage battery with the average expected remaining power, and fails to solve the problem that the inconsistency of single cells affects the available capacity of the entire battery pack.

SUMMARY OF THE INVENTION

In view of the defects in the prior art, the purpose of the present invention is to provide a battery balance management method and system based on a multi-agent game.

A battery balancing management method based on a multi-agent game provided according to the present invention includes:

Step 1: Determine the state of charge data of each single cell according to the attenuation data and current power data of each single cell in the battery pack;

Step 2: According to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack, the charge and discharge strategy of each single cell in the battery pack is obtained by the Nash equilibrium method;

Step 3: Carry out battery balance management according to the charging and discharging strategy.

Preferably, step 1 includes:

Step 101: Obtain attenuation data and current power data through the current state data and historical state data of each single battery;

Step 102: Determine the state-of-charge data of each single battery according to the attenuation data and the current power data.

Preferably, step 101 further includes:

Step 1011: Input the current state data and the historical state data into the artificial neural network to obtain attenuation data and current electric quantity data output by the artificial neural network.

Preferably, it is characterized in that step 2 includes:

Step 201: Obtain the utility function of each single cell according to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack;

Step 202: Obtain the charge and discharge strategy of each single cell in the battery pack through the Nash equilibrium method and the utility function.

Preferably, step 202 includes:

Step 2021: Obtain a Nash equilibrium solution of the utility function according to a pair of preset constraints by using the Nash equilibrium method;

Step 2022: According to the Nash equilibrium solution, the charging and discharging strategy of each single cell in the battery pack is obtained.

A battery balancing management system based on a multi-agent game provided according to the present invention includes:

Module M1: Determine the state of charge data of each single cell according to the attenuation data and current power data of each single cell in the battery pack;

Module M2: According to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack, the charge and discharge strategy of each single cell in the battery pack is obtained by the Nash equilibrium method;

Module M3: Carry out battery balance management according to the charging and discharging strategy.

Preferably, the module M1 includes:

Sub-module M101: obtain attenuation data and current power data through the current state data and historical state data of each single battery;

Sub-module M102: Determine the state-of-charge data of each single battery according to the attenuation data and the current power data.

Preferably, the submodule M101 further includes:

Unit D1011: Input the current state data and the historical state data into the artificial neural network to obtain attenuation data and current electricity data output by the artificial neural network.

Preferably, the module M2 includes:

Sub-module M201: According to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack, the utility function of each single cell is obtained;

Sub-module M202: Through the Nash equilibrium method and the utility function, the charging and discharging strategy of each single cell in the battery pack is obtained.

Preferably, the submodule M202 includes:

Unit D2021: Obtain the Nash equilibrium solution of the utility function according to the preset constraint condition pair by using the Nash equilibrium method;

Unit D2022: According to the Nash equilibrium solution, the charging and discharging strategy of each single cell in the battery pack is obtained.

Compared with the prior art, the present invention has the following beneficial effects:

1. The present invention solves the utility function through the Nash equilibrium method to obtain the charging and discharging strategy of each single battery, reduces the influence of the inconsistency of the single battery on the available capacity of the entire battery pack, and improves the energy utilization rate.

2. The present invention realizes the effective balance of the battery through the Nash balance method, reduces the number of battery cycles, and thus prolongs the battery life.

3. The present invention realizes the balance of battery power without adding additional power devices and circuits, and can perform online detection of battery power and battery health state, and assist the battery management system to perform battery balance operation.

Description of drawings

Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments with reference to the following drawings:

Fig. 1 is the schematic flow chart of the present invention;

FIG. 2 is a schematic structural diagram of battery state estimation according to the present invention;

FIG. 3a is a schematic diagram of the first battery balancing management of the present invention;

3b is a schematic diagram of the second battery balancing management of the present invention;

FIG. 4 is a schematic structural diagram of the battery balance management system of the present invention.

Detailed ways

The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the inventive concept. These all belong to the protection scope of the present invention.

FIG. 1 is a schematic flowchart of the present invention. As shown in FIG. 1 , the present invention provides a battery balancing management method based on a multi-agent game, including the following steps:

Step 1: Determine the state of charge data of each single cell according to the attenuation data and current power data of each single cell in the battery pack.

Among them, the state of charge (State of Charge, SOC) is used to reflect the remaining capacity of the battery, and its value is defined as the ratio of the remaining capacity to the battery capacity, usually expressed as a percentage. Its value range is 0-1.

In the present invention, the state-of-charge data of each single battery is determined according to the attenuation data and current power data of each single battery.

Among them, the attenuation data represents the current total capacity attenuation of the battery; the current power data represents the actual power of the battery.

Preferably, step 1 includes: step 101: obtaining attenuation data and current power data through the current state data and historical state data of each single cell; step 102: determining each cell according to the attenuation data and current power data State of charge data for the battery.

Among them, the current state data includes data such as voltage, current, temperature, and discharge mode of the battery; historical state data includes data such as charge and discharge capacity, cycle times, and working conditions.

Preferably, step 101 further includes: step 1011 : input the current state data and historical state data into the artificial neural network to obtain attenuation data and current electricity data output by the artificial neural network.

FIG. 2 is a schematic structural diagram of battery state estimation according to the present invention. As shown in FIG. 2 , the current state data of the battery is obtained by measurement, including voltage data, current data and temperature data; the historical state data measured by the battery includes the charge and discharge capacity. , cycle times and working conditions; input the historical state data into the cache unit as battery historical parameters, and input the current state data and battery historical parameters into the artificial neural network, so as to obtain the battery attenuation data and current power data, and at the same time will reflect the current battery power The current power data and the training parameters of the artificial neural network are sent to the cache unit for storage for subsequent acquisition of attenuation data and current power data.

It can be known that obtaining the battery state accurately is the premise of battery balancing management. The invention adopts an algorithm based on artificial intelligence, and comprehensively evaluates the current total capacity attenuation and actual power of the battery by measuring parameters such as battery voltage, current, temperature and discharge mode. Compared with the prior art, the current state data of the battery can be obtained more accurately.

Step 2: According to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack, the charge and discharge strategy of each single cell in the battery pack is obtained by the Nash equilibrium method.

Specifically, according to the difference between the state of charge data of each single cell and the remaining single cells, the difference between each single cell and the remaining single cells is determined, and by minimizing the difference, battery balance management is realized.

Preferably, step 2 includes: step 201: obtaining the utility function of each single cell according to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack; step 202: Through the Nash equilibrium method and the utility function, the charging and discharging strategy of each single cell in the battery pack is obtained.

Specifically, based on multi-agent game theory, battery utility evaluation is performed for each single battery. The utility function of each single cell is defined as minimized as the state difference from adjacent cells. The Nash equilibrium point of the battery utility evaluation problem is analyzed. When all the cell states are the closest, the battery pack reaches the equilibrium state, and each cell also achieves the goal of its minimum utility function.

Preferably, step 202 includes: step 2021 : obtaining a Nash equilibrium solution of the utility function according to a pair of preset constraints by using a Nash equilibrium method; step 2022 : obtaining the charge of each single cell in the battery pack according to the Nash equilibrium solution discharge strategy.

当前电池组需要提供的功率为P，各个单体电池容量为Cⁱ，则对于每个单体电池，其效用函数可以表示为公式(1)：3a is a schematic diagram of the first cell balancing management of the present invention, and FIG. 3b is a schematic diagram of the second cell balancing management of the present invention, as shown in FIGS. 3a and 3b , including a single cell #1 and a single cell #2 , single battery #3, single battery #4, single battery #5, ..., single battery #N, where N is the total number of single batteries in the battery pack. Set the initial SoC of each single cell as

The power that the current battery pack needs to provide is P, and the capacity of each single cell is C ⁱ , then for each single cell, its utility function can be expressed as formula (1):

Among them, i and k are integers; I is an integer; α is a weight coefficient, t is a positive number, indicating time, and p ⁱ is the power provided by the ith elevator battery.

Specifically, the first part in formula (1) represents the inverse of the difference between the state of charge data (SoC) of the ith single cell and the rest of the single cells, and the second part represents the maximum charge of the ith single cell. The opposite of discharge cost.

Further, the utility function needs to satisfy the preset constraints, and the preset constraints can be expressed by formula (2) and formula (3):

∑ _i∈I p ⁱ =P; (2)

That is, the total power provided by all the single cells in the battery pack is P, and the SoC change of each battery is related to the power.

It can be known that the above formula (1), formula (2) and formula (3) constitute a generalized Stauberg game problem, which has a Nash equilibrium solution, that is, each single battery reaches an equilibrium state with the same amount of electricity. By solving it iteratively, the optimal charge and discharge power of each single battery can be obtained, which can be regulated by an active balancing system. At the next moment, the battery SoC is measured, and a new Nash equilibrium point is obtained from the new required power, and then the battery charge and discharge power at each subsequent moment can be obtained in turn to form a stable battery power balance. The final result is shown in Figure 3b.

As can be seen from Figures 3a and 3b, the voltage data of the battery can be divided into 1.5V, 3.0V, 3.6V and 4.4V, wherein the normal working voltage range of the battery is 3.0V-3.6V. The total capacity of the bulk battery varies greatly, which will accelerate the aging of the single battery, and even cause explosion and fire. The working state of the battery, the balanced work of each single cell is within the normal working voltage range of the battery, and the difference in the state of charge data of each single cell is small.

Step 3: Carry out battery balance management according to the charging and discharging strategy.

Specifically, according to the charging and discharging strategy, battery charging and discharging management is performed to achieve active battery balancing. By optimizing the battery charging and discharging sequence, the loop loss in battery balancing is reduced.

The invention provides a battery balance management system based on multi-agent game, including:

Module M1: Determine the state of charge data of each single cell according to the attenuation data and current power data of each single cell in the battery pack.

Preferably, the module M1 includes: a sub-module M101: obtaining attenuation data and current power data through the current state data and historical state data of each single battery; sub-module M102: determining each battery according to the attenuation data and the current power data State of charge data for a single cell.

Preferably, the sub-module M101 further includes: a unit D1011: Input the current state data and the historical state data into the artificial neural network to obtain the attenuation data and the current electric quantity data output by the artificial neural network.

Module M2: According to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack, the charge and discharge strategy of each single cell in the battery pack is obtained through the Nash equilibrium method.

Preferably, the module M2 includes: a sub-module M201: obtaining the utility function of each single cell according to the difference between the state of charge data of each single cell and the state of charge data of the remaining single cells in the battery pack; Sub-module M202: Through the Nash equilibrium method and the utility function, the charging and discharging strategy of each single cell in the battery pack is obtained.

Preferably, the sub-module M202 includes: unit D2021: obtain the Nash equilibrium solution of the utility function according to the preset constraint condition through the Nash equilibrium method; unit D2022: obtain the Nash equilibrium solution of each single cell in the battery pack according to the Nash equilibrium solution charging and discharging strategy.

Module M3: Carry out battery balance management according to the charging and discharging strategy.

FIG. 4 is a schematic structural diagram of the battery balance management system of the present invention. As shown in FIG. 4 , the battery balance management system of multi-agent game includes three parts: battery state estimation, battery utility function evaluation and battery charge and discharge management. Among them, the battery state estimation is used to obtain the state of charge data of the battery. Accurately obtaining the state of charge data of the battery is the premise of battery balance management. The invention adopts an algorithm based on artificial intelligence, and comprehensively evaluates the current total capacity attenuation and actual power of the single battery by measuring parameters such as voltage, current, temperature and discharge mode of the single battery. The historical state data measured from the single battery, including the charge and discharge capacity, cycle times and working conditions are input into the buffer unit as historical parameter information, while the current state data of the battery including voltage, current, and temperature are input into the artificial neural network, thereby obtaining the single battery. The decay data and current charge data of the battery. The battery utility function evaluation is based on the multi-agent game theory framework to model and evaluate the battery utility function of each single battery. The utility function of each cell is defined to minimize the state difference from neighboring cells. Analyzing the Nash equilibrium point of this problem, when the states of all the single cells are the closest, the system reaches the equilibrium state, and each single cell also achieves the goal of its minimum utility function. Battery charge and discharge management, according to the evaluation results of battery utility function, carry out battery charge and discharge management to achieve active battery balance. By optimizing the battery charging and discharging sequence, the loop loss in battery balancing is reduced.

The technical problem solved by the present invention is:

1. The inconsistency of the single cells in the battery pack will have a significant impact on the available capacity of the entire battery pack. And, as the battery pack's usage time accumulates, the inconsistency affects the battery pack more and more.

2. The imbalance of the single cells in the battery pack will increase the number of battery cycles and reduce the battery life.

3. In order to achieve battery balance management, additional power devices and circuits need to be added.

Compared with the prior art, the present invention has the following beneficial effects:

1. The present invention can improve the energy utilization rate and reduce the influence of the inconsistency of the single battery on the usable capacity of the entire battery pack.

2. The present invention can effectively improve the utilization rate of battery energy, prolong the battery life, and can realize less components, reliable structure, easy control, good circuit efficiency and high robustness.

3. The present invention can realize the balance of battery power and can perform on-line detection of battery power and battery health state without adding additional power devices and circuits, and assist the battery management system to perform battery balance operation.

Those skilled in the art know that, in addition to implementing the system, device and its respective modules provided by the present invention in the form of pure computer readable program codes, the system, device and its modules provided by the present invention can be completely implemented by logically programming the method sub-module M. Each module implements the same program in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system, device and each module provided by the present invention can be regarded as a kind of hardware component, and the modules used for realizing various programs included in it can also be regarded as the structure in the hardware component; A module for realizing various functions can be regarded as either a software program for realizing a method or a structure within a hardware component.

The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above-mentioned specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essential content of the present invention. The embodiments of the present application and features in the embodiments may be arbitrarily combined with each other without conflict.

Claims (10)

1. A battery equalization management method based on multi-agent game is characterized by comprising the following steps:

step 1: determining the charge state data of each single battery according to the attenuation data and the current electric quantity data of each single battery in the battery pack;

step 2: obtaining a charge-discharge strategy of each single battery in the battery pack by a Nash equilibrium method according to the difference value of the charge state data of each single battery and the charge state data of the other single batteries in the battery pack;

and step 3: and carrying out battery balance management according to the charge and discharge strategy.

2. The multi-agent gaming-based battery equalization management method of claim 1, wherein said step 1 comprises:

step 101: obtaining the attenuation data and the current electric quantity data through the current state data and the historical state data of each single battery;

step 102: and determining the state of charge data of each single battery according to the attenuation data and the current electric quantity data.

3. The multi-agent gaming based battery equalization management method of claim 2, wherein said step 101 further comprises:

step 1011: and inputting the current state data and the historical state data into an artificial neural network to obtain the attenuation data and the current electric quantity data output by the artificial neural network.

4. The multi-agent gaming-based battery equalization management method of claim 1, wherein said step 2 comprises:

step 201: obtaining a utility function of each single battery according to the difference value of the charge state data of each single battery and the charge state data of other single batteries in the battery pack;

step 202: and obtaining the charge-discharge strategy of each single battery in the battery pack through the Nash equilibrium method and the utility function.

5. The multi-agent gaming based battery equalization management method of claim 1 or 4, wherein said step 202 comprises:

step 2021: obtaining a Nash equilibrium solution of the utility function according to a preset constraint condition pair by the Nash equilibrium method;

step 2022: and obtaining the charge-discharge strategy of each single battery in the battery pack according to the Nash equilibrium solution.

6. A multi-agent gaming-based battery equalization management system, comprising:

module M1: determining the charge state data of each single battery according to the attenuation data and the current electric quantity data of each single battery in the battery pack;

module M2: obtaining a charge-discharge strategy of each single battery in the battery pack by a Nash equilibrium method according to the difference value of the charge state data of each single battery and the charge state data of the other single batteries in the battery pack;

module M3: and carrying out battery balance management according to the charge and discharge strategy.

7. The multi-agent gaming based battery equalization management system of claim 6, wherein said module M1, comprises:

submodule M101: obtaining the attenuation data and the current electric quantity data through the current state data and the historical state data of each single battery;

submodule M102: and determining the state of charge data of each single battery according to the attenuation data and the current electric quantity data.

8. The multi-agent gaming-based battery equalization management system of claim 7, wherein said sub-module M101 further comprises:

unit D1011: and inputting the current state data and the historical state data into an artificial neural network to obtain the attenuation data and the current electric quantity data output by the artificial neural network.

9. The multi-agent gaming based battery equalization management system of claim 6, wherein said module M2, comprises:

submodule M201: obtaining a utility function of each single battery according to the difference value of the charge state data of each single battery and the charge state data of other single batteries in the battery pack;

submodule M202: and obtaining the charge-discharge strategy of each single battery in the battery pack through the Nash equilibrium method and the utility function.

10. The multi-agent gaming-based battery equalization management system according to claim 6 or 9, wherein the sub-module M202 comprises:

unit D2021: obtaining a Nash equilibrium solution of the utility function according to a preset constraint condition pair by the Nash equilibrium method;

unit D2022: and obtaining the charge-discharge strategy of each single battery in the battery pack according to the Nash equilibrium solution.

Images