Lithium battery pack equalization method based on improved Buck-Boost equalization topology and GSSO optimization two-value fuzzy control

Abstract

The invention discloses a lithium battery pack equalization method based on an improved Buck-Boost circuit and GSSO optimization fuzzy control, which is used for solving the problems of inconsistent energy and low equalization efficiency in the long-term use process of a series lithium battery pack. The designed improved Buck-Boost equalization circuit adopts an optimized REB-TCM structure, and by optimizing an energy transmission path, the conversion loss is reduced, and the equalization speed is improved. Aiming at the topological structure, the invention provides a dual-value fuzzy control model (DualFLC) which can dynamically select an equalization target according to the state of charge (SOC) of the battery and dynamically adjust an equalization module through a fuzzy strategy so as to improve the equalization efficiency. On the basis, a swarm ensorption hybrid optimization algorithm (GSSO) is introduced for the first time, global optimization is carried out on node positions of fuzzy control input discourse domains, the limitation that traditional fuzzy control depends on experience and subjective judgment on a membership function is overcome, and the performance of a balance system is further improved. The method effectively improves the equalization speed and efficiency, reduces the energy inconsistency between the batteries, and prolongs the service life of the battery pack.

Description

Technical Field

[0001] The present invention relates to the field of lithium battery equalization technology, and in particular to a lithium battery pack equalization method based on improved Buck-Boost equalization topology and GSSO optimized dual-value fuzzy control. Background Art

[0002] The voltage and capacity of single lithium batteries are low, so in various fields of use, multiple batteries are connected in series and in parallel to increase the overall voltage and capacity. However, the inconsistency between lithium battery cells will affect the life and use of the battery pack. Therefore, a lithium battery balancing control system is very necessary to reduce the inconsistency between single cells and improve the balancing efficiency of the battery pack.

[0003] At present, battery balancing can be divided into two categories: the passive method uses resistors to consume excess energy, which is simple but inefficient; the active method uses energy storage elements to achieve energy transfer, improve energy utilization efficiency and balancing speed, and has become the main research direction. There are various types of active balancing topology circuits: switched capacitor active balancing, inductive energy storage active balancing, combined with switch arrays, and active balancing based on DC / DC converters. Among them, the common balancing circuit has the problem of long balancing paths between batteries that are far apart, resulting in increased balancing time and energy loss. The active balancing circuit based on the switch array type can achieve any single-to-single-cell balancing, with high balancing flexibility, but its logic complexity is high and it will increase the energy loss during the balancing process.

[0004] On the basis of the balancing circuit, a corresponding balancing control strategy needs to be formulated. The voltage-based mean difference strategy can effectively achieve battery balancing, but its efficiency is affected by the difference in the internal resistance of the cells. At present, the battery balancing control method mainly uses the state of charge (SOC) as the balancing target. In traditional fuzzy logic control battery balancing, the determination process of its membership function is highly subjective and empirical, and it is difficult to ensure the optimal control effect. The selection of the balancing target is also crucial. Therefore, it is necessary to select an appropriate balancing control method so that the appropriate balancing current can be output under different battery conditions. Summary of the invention

[0005] In view of the shortcomings of the above problems, the present invention proposes a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control, which optimizes the energy transmission path, reduces energy loss and improves the balancing speed through a closed energy loop system. A fuzzy logic control system that can dynamically select the balancing target according to the battery status is introduced in the balancing strategy. At the same time, the fuzzy logic is optimized through the GSSO algorithm, which effectively overcomes the problem of membership function dependence on experience and subjectivity in traditional fuzzy control, realizes more accurate balancing control, improves the overall balancing efficiency, and reduces the inconsistency between battery packs.

[0006] The proposed balancing method, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control, consists in the following steps:

[0007] S1: Monitor each single cell in the lithium battery pack to obtain its voltage and state of charge (SOC) information;

[0008] S2: Build the REB-TCM balanced topology model and establish the DualFLC fuzzy logic control system model;

[0009] S3: fuzzy logic for setting different balancing targets based on battery state of charge (SOC);

[0010] S4: Introduce GSSO algorithm to optimize and improve the membership function;

[0011] S5: Combine the DualFLC model with different balancing objectives and the REB-TCM balancing topology model to control the balancing of the lithium battery pack.

[0012] The REB-TCM balancing topology model in step S2 of the proposed balancing solution is to introduce an energy conditioning unit between batteries and set energy routing modules at both ends of the battery pack to form a closed energy loop system. The fuzzy strategy based on GSSO optimization controls each balancing submodule to monitor the battery status in real time to adaptively adjust the direction and intensity of energy flow, reduce energy loss, and improve balancing efficiency. The balancing is divided into three stages:

[0013] 1) Discharge process: When B1 is discharged, the balancing system sends a control signal to turn on the M6b switch to form a discharge loop. The conduction time of M6b determines the discharge intensity, and the electrical energy is stored in the inductor L6;

[0014] 2) Charging process: After the discharge is completed, M6c is turned on, and the inductor L6 provides power to charge B6;

[0015] 3) Energy release process: After the charge and discharge process is completed, resistor R6 will consume the energy that has not been completely released in L6 to achieve energy conversion.

[0016] The DualFLC fuzzy logic control model established in step S2 mainly includes the following key contents:

[0017] 1) DualFLC fuzzy system input includes two parts, taking SOC as an example:

[0018] ΔSOC av : The absolute value of the difference between the average SOC of the two batteries and the average SOC of the battery pack, which is used to measure the balance between the battery cell and the overall pack;

[0019] ΔSOC: The absolute value of the difference between the SOCs of the two batteries, used to determine the difference between the battery cells;

[0020] ΔSOC av Using ΔSOC as input can more accurately reflect the differences between single cells and dynamically output the duty cycle to achieve efficient balancing;

[0021] 2) Design of membership function and fuzzy rule base:

[0022] Find the optimal membership function parameters through the GSSO algorithm to improve the accuracy and efficiency of balanced control;

[0023] The fuzzy rule base is adjusted in a targeted manner based on the differences between battery cells to make it more in line with actual balancing needs, further improve control accuracy, and reduce unnecessary rule interventions;

[0024] The fuzzy logic described in step S3 of the balancing strategy is to establish a DualFLC control system model based on voltage and SOC according to the changing characteristics of SOC, and can control battery balancing with different balancing strategies under different SOC states.

[0025] In step S4, the membership function is optimized using the GSSO algorithm, which specifically includes the following steps:

[0026] 1) Parameter encoding design and initialization: Encode the membership node parameters of the DualFLC model input variables (such as SOC difference, voltage difference), and initialize the population and number of iterations;

[0027] 2) Objective function construction: Construct a fitness function with the goal of minimizing the battery SOC difference and evaluate the quality of the membership function node parameters. The formula is as follows:

[0028] f=w1·|SOC dis |+w2·|MeanSOC dis |

[0029] Where w1 and w2 are adjustment coefficients; SOC dis is the SOC difference between the two batteries; MeanSOC dis It is the difference between the average SOC of the two batteries and the average SOC of the battery pack.

[0030] 3) Solution vector update: Update the parameters of the membership node according to the historical optimal position and global optimal position of the solution vector;

[0031] 4) Crossover and mutation operation: cross and mutate the updated solution set to further explore the optimal solution and avoid falling into the local optimum;

[0032] 5) Target update: After each iteration, re-evaluate and update the optimal SOC according to the fitness functiondis and MeanSOC dis , until the population fitness converges;

[0033] 6) Update fuzzy controller: The optimal membership function optimized by GSSO is finally output to update the fuzzy logic controller to achieve charge and discharge balance among battery packs, thereby improving the overall performance.

[0034] The balancing process is to start balancing in step S5 and then select the DualFLC fuzzy control model based on voltage or SOC to control the REB-TCM balancing model according to the battery information collected in step S1 to achieve battery balancing. When the balancing threshold of the battery pack is within the set threshold, the balancing process ends. The balancing threshold uses the variance SOC^2 of the battery pack SOC as the balancing standard, which can accurately reflect the consistency of the SOC between each battery cell, and the initial value of the threshold is set to σ=0.0005.

[0035] The advantages embodied by the present invention are:

[0036] 1) By building the REB-TCM balancing model, the energy balancing path is optimized, which can reduce energy loss and increase the balancing rate. Each balancing module is independent of each other and can be expanded to multiple batteries without increasing circuit complexity.

[0037] 2) The proposed dual-value fuzzy strategy can dynamically select the DualFLC model with different balancing targets according to the SOC state of the battery, and directly control the balancing module through the duty cycle to achieve precise charging and discharging of the single battery, further improving the balancing efficiency.

[0038] 3) The introduced GSSO algorithm optimizes the membership function, improves the accuracy of the equilibrium control system, and has a further improvement effect on the equilibrium time and energy loss.

[0039] 4) The overall balancing scheme based on DualFLC's fuzzy strategy and REB-TCM balancing topology has more significant advantages than traditional balancing schemes in terms of energy loss, balancing rate and complexity. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Balanced circuit topology;

[0041] Figure 2 Discharge equalization principle diagram;

[0042] Figure 3 Charge equalization schematic diagram;

[0043] Figure 4 Input membership function;

[0044] Figure 5 Output membership function;

[0045] Figure 6 Fuzzy rule table;

[0046] Figure 7 Balance control strategy flow chart;

[0047] Figure 8 GSSO optimization flow chart;

[0048] Fig. 9 Optimized input membership function;

[0049] Fig.10 Static balance;

[0050] Fig.11 Charge equalization;

[0051] Fig.12 Discharge balance. DETAILED DESCRIPTION

[0052] The proposed solution is based on an improved Buck-Boost balancing topology, and the topology structure includes a battery module, a SOC range determination module, a main balancing module, and a fuzzy logic control module, such as Figure 1 shown.

[0053] The balancing circuit can operate efficiently in the three stages of battery discharge, charging and demagnetization. Taking the first and last two batteries as an example, when it is detected that the SOC of battery B1 is higher than B6, the balancing process is divided into the following three steps:

[0054] 1) B1 discharge process

[0055] When B1 is discharged, the balancing system sends a control signal, M6b is turned on, and the forward voltage drop diode D6a is turned on to form a discharge loop with the battery B1, switch M6b, and power inductor L6. The discharge balancing process is as follows: Figure 2 As shown in the figure, the loop current increases continuously during the switch conduction process, and the electric energy is stored in the inductor L6.

[0056] 2) B6 charging process

[0057] After the discharge is completed, the diode D6a is turned off, and M6c is turned on through the control signal to charge B6. At this time, the energy for charging B6 is provided by the inductor L6. The charging balancing principle is as follows: Figure 3 When the energy in the inductor L6 cannot maintain the forward conduction of D6d, the charging loop stops.

[0058] 3) Demagnetization process

[0059] After the charging and discharging process is completed, the resistor R6 will consume the energy that has not been completely released in L6, realize energy conversion, and ensure that the circuit operates in a stable state.

[0060] In the balancing control strategy part, a DualFLC model is established to control battery balancing, which mainly includes the following key points: 1) The DualFLC fuzzy system input includes two parts, taking SOC as an example:

[0061] ΔSOC av : The absolute value of the difference between the average SOC of the two batteries and the average SOC of the battery pack, which is used to measure the balance between the battery cell and the overall pack;

[0062] ΔSOC: The absolute value of the difference between the SOCs of the two batteries, used to determine the difference between the battery cells;

[0063] ΔSOC av Using ΔSOC as input can more accurately reflect the differences between single cells and dynamically output the duty cycle to achieve efficient balancing;

[0064] 2) Design of membership function and fuzzy rule base:

[0065] Find the optimal membership function parameters through the GSSO algorithm to improve the accuracy and efficiency of balanced control;

[0066] The fuzzy rule base is adjusted in a targeted manner based on the differences between battery cells to make it more in line with actual balancing needs, further improve control accuracy, and reduce unnecessary rule intervention.

[0067] Two-input ΔSOC of fuzzy control model av The domain range of and ΔSOC is set to (0,10%) and (0,5%) respectively; the two inputs ΔU of the voltage-based fuzzy controller av The basic domains of and ΔU are set to (0, 0.4) and (0, 1.2) respectively; the range of fuzzy logic PWM output is set to (45%, 75%); each domain is divided into 6 fuzzy subsets: very low (VL), low (L), medium low (ML), medium high (MH), high (H), very high (VH). The membership function of input and output is as follows Figure 4 , Figure 5 The established fuzzy rules are shown as Figure 6 shown.

[0068] The dual-value fuzzy control based on GSSO optimization is to establish a fuzzy logic control model based on voltage and SOC according to the changing characteristics of SOC. The range of SOC can be judged according to SOC information. If SOC is between 20% and 80%, the fuzzy logic control strategy based on SOC is adopted. If SOC is outside 20% to 80%, the fuzzy logic based on voltage is adopted. The balanced control process is as follows Figure 7 shown.

[0069] The optimization process of optimizing the membership function using the GSSO algorithm is as follows: Figure 8As shown, the specific steps include:

[0070] 1) Parameter encoding design and initialization: Encode the membership node parameters of the fuzzy model input variables (such as SOC difference, voltage difference), and initialize the population and number of iterations;

[0071] 2) Objective function construction: Construct a fitness function with the goal of minimizing the battery SOC difference and evaluate the quality of the membership function node parameters. The formula is as follows:

[0072] f=w1·|SOC dis |+w2·|MeanSOC dis |

[0073] Where w1 and w2 are adjustment coefficients; SOC dis is the SOC difference between the two batteries; MeanSOC dis It is the difference between the average SOC of the two batteries and the average SOC of the battery pack.

[0074] 3) Solution vector update: Update the parameters of the membership node according to the historical optimal position and global optimal position of the solution vector;

[0075] 4) Crossover and mutation operation: cross and mutate the updated solution set to further explore the optimal solution and avoid falling into the local optimum;

[0076] 5) Target update: After each iteration, re-evaluate and update the optimal SOC according to the fitness function dis and MeanSOC dis , until the population fitness converges;

[0077] 6) Update the fuzzy controller: Finally, the optimal membership function of GSSO optimization is output as Fig. 9 As shown, it is used to update the fuzzy logic controller to achieve charge and discharge balance among battery packs, thereby improving the overall performance.

[0078] The balancing process selects the fuzzy control strategy based on voltage or SOC to control the REB-TCM balancing model to achieve battery balancing according to the obtained battery status information. When the balancing threshold of the battery pack is within the set threshold, the balancing process ends. The balancing threshold uses the variance SOC^2 of the battery pack SOC as the balancing standard, which can accurately reflect the balance between each battery cell. The initial value of the threshold is set to σ=0.0005.

[0079] Example 1

[0080] The lithium battery pack balancing method based on the improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control was modeled and simulated on the Matlab / simulink platform, with a 6-cell series lithium battery pack as the research object. In order to evaluate the effectiveness and advantages of the balancing scheme of the present invention, three balancing schemes were set to conduct experiments and analysis comparisons under three different working conditions: static and charging and discharging. The three balancing schemes are shown in the following table:

[0081]

[0082] (1) Static stage

[0083] By comparing the three equalization schemes, the SOC changes of each single battery are obtained, such as Fig.10 As shown. Comparing the time taken by schemes 1 and 2 to reach equilibrium, it can be seen that the proposed improved balancing topology shortens the balancing time by 31.4% compared with the traditional topology; comparing the time taken by schemes 2 and 3 to reach equilibrium, it can be seen that the balancing time of GSSO optimized fuzzy logic is reduced by 43.9% compared with the traditional fuzzy logic; comparing schemes 1 and 3, the lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control shortens the balancing time by 61.5% compared with the traditional balancing method, and reduces the energy loss by about 0.9%

[0084] (2) Charging stage

[0085] In constant current charging mode, compare the three equalization schemes such as Fig.11 As shown. The balancing time of the improved balancing topology of the comparison scheme 1 and 2 is 870S, which is 28.7% shorter than the balancing time of the traditional balancing circuit; the balancing time of the improved fuzzy control strategy of the comparison scheme 2 and 3 is shortened by 48.3%; the balancing time of the balancing scheme of the comparison scheme 1 and 3 is shortened by 63.1% compared with the traditional scheme, and the overall energy loss is reduced by about 2.6%.

[0086] (3) Discharge stage

[0087] The SOC equilibrium curves of the three schemes in the discharge stage are as follows: Fig.12 As shown. Comparing Schemes 1 and 2, it can be seen that the improved balancing topology shortens the balancing time by 29.2% compared with the traditional topology; comparing Schemes 2 and 3, the improved fuzzy control strategy shortens the balancing time by 46.6%; comparing Schemes 1 and 3, the improved balancing scheme shortens the balancing time by 62.2% compared with the traditional scheme, and the overall energy loss is reduced by about 1.6%.

[0088] In summary, by comparing the balancing time under three different working conditions, the following conclusions are drawn:

[0089] (1) Under the same fuzzy strategy conditions, the improved Buck-Boost balancing topology circuit of the present application has higher balancing efficiency than the traditional Buck-Boost balancing circuit.

[0090] (2) The optimized fuzzy control strategy significantly speeds up the balancing process and is superior to the traditional method.

[0091] (3) This application proposes an overall balancing solution based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control, which not only reduces energy loss, but also significantly improves the balancing efficiency and reduces the inconsistency between batteries.

Claims

1. A lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control, characterized in that The following steps: S1: Monitor each single cell in the lithium battery pack to obtain its voltage and state of charge (SOC) information; S2: Build the REB-TCM balanced topology model and establish the DualFLC fuzzy logic control system model; S3: fuzzy logic for setting different balancing targets based on battery state of charge (SOC); S4: Introduce GSSO algorithm to optimize and improve the membership function; S5: Combine the DualFLC model with different balancing objectives and the REB-TCM balancing topology model to control the balancing of the lithium battery pack.

2. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: In step S1, the voltage and state of charge (SOC) information of each single battery is obtained to provide execution conditions for the subsequent balancing control strategy.

3. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: In step S2, the REB-TCM balancing topology model is built by introducing an energy regulation unit between batteries and setting energy routing modules at both ends of the battery pack to form a closed energy loop system. The fuzzy strategy based on GSSO optimization controls each balancing module to adaptively adjust the direction and intensity of energy flow, reduce energy loss, and improve balancing efficiency. The balancing is divided into three stages: 1) Discharge process: When B1 is discharged, the balancing system sends a control signal to turn on the M6b switch to form a discharge loop, and the electrical energy is stored in the inductor L6; 2) Charging process: After the discharge is completed, M6c is turned on, and the inductor L6 provides power to charge B6; 3) Energy release process: After the charge and discharge process is completed, resistor R6 will consume the energy that has not been completely released in L6 to achieve energy conversion.

4. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: The DualFLC fuzzy logic control model established in step S2 mainly includes the following key contents: 1) DualFLC fuzzy system input includes two parts, taking SOC as an example: ΔSOC av : The absolute value of the difference between the mean SOC of the two batteries and the mean SOC of the battery pack; ΔSOC: the absolute value of the difference between the two battery SOCs; ΔSOC av Using ΔSOC as input can more accurately reflect the differences between single cells and dynamically output the duty cycle to achieve efficient balancing; 2) Design of membership function and fuzzy rule base: Find the optimal membership function parameters through the GSSO algorithm to improve the accuracy and efficiency of balanced control; The fuzzy rule base is adjusted in a targeted manner based on the differences between battery cells to make it more in line with actual balancing needs, further improve control accuracy, and reduce unnecessary rule intervention.

5. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: Step S3 establishes a DualFLC control system model based on voltage and SOC according to the variation characteristics of SOC, and can control battery balancing with different models under different SOC states.

6. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: Step S4 introduces the GSSO algorithm to optimize the membership function to improve the optimal control performance of the battery pack balancing system, which specifically includes the following steps: 1) Parameter encoding design and initialization: Encode the membership node parameters of the DualFLC model input variables (such as SOC difference, voltage difference), and initialize the population and number of iterations; 2) Objective function construction: Construct a fitness function with the goal of minimizing the battery SOC difference. The formula is as follows: f=w1·|SOC dis |+w2·|MeanSOC dis | Where w1 and w2 are adjustment coefficients; SOC dis is the SOC difference between the two batteries; MeanSOC dis is the difference between the mean SOC of the two batteries and the mean SOC of the battery pack; 3) Solution vector update: Update the parameters of the membership node according to the historical optimal position and global optimal position of the solution vector; 4) Crossover and mutation operation: cross and mutate the updated solution set to further explore the optimal solution and avoid falling into the local optimum; 5) Target update: After each iteration, re-evaluate and update the optimal SOC according to the fitness function dis and MeanSOC dis , until the population fitness converges; 6) Update fuzzy controller: The optimal membership function optimized by GSSO is finally output to update the fuzzy logic controller to achieve charge and discharge balance among battery packs, thereby improving the overall performance.

7. According to claim 1, a lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control is characterized in that: After starting the balancing in step S5, the DualFLC control model based on voltage or SOC is selected according to the battery information collected in step S1 to control the REB-TCM balancing model to achieve battery balancing. When the balancing threshold of the battery pack is within the set threshold, the balancing process ends.

8. A lithium battery pack balancing method based on improved Buck-Boost balancing topology and GSSO optimized dual-value fuzzy control according to claim 7, characterized in that: The balance threshold is based on the variance SOC^2 of the battery pack SOC as a balance standard, which can accurately reflect the balance between each battery cell, and the initial value of the threshold is set to σ=0.0005.

Images