BMS control method for battery prediction

Abstract

The invention relates to the field of battery management, and discloses a battery management system (BMS) control method for battery prediction, which comprises the following steps of: analyzing the difference condition of single batteries; constructing a conventional hybrid balanced switching strategy; constructing a differential hybrid balanced switching strategy; forming different equalization adjustment strategies by the conventional hybrid equalization switching strategy and the difference hybrid equalization switching strategy; determining an executed balance adjustment strategy according to the difference condition of the single batteries; active equalization and passive equalization are carried out separately in a hardware isolation mode and a logic interlocking mode; and constructing a battery state prediction model, and predicting the battery state according to the battery state prediction model. According to the method, the battery equalization adaptability is optimized, the dual goals of improving the equalization effect and reducing the equalization loss are achieved at the same time, the battery state prediction process is doubly optimized through the equalization reference parameters and the conventional operation data, the influence of the equalization process is considered by the battery state prediction model, and the battery state prediction is more accurate.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically to a BMS control method for battery prediction. Background Technology

[0002] Battery Management Systems (BMS) address battery management issues in electric vehicles through precise monitoring, intelligent control, and proactive protection. BMS is crucial for ensuring the safety of electric vehicle batteries; however, existing BMS systems have several shortcomings: First, single balancing technology has adaptability defects and cannot cope with the dynamic deterioration of battery cell consistency caused by fluctuations in charge / discharge rate and temperature changes. Currently, most BMS adopt a single solution of active or passive balancing: In the early stage of fast charging, the voltage of battery cells rises rapidly and the consistency difference is easily amplified, requiring the high-speed adjustment capability of active balancing. However, single passive balancing is slow in adjustment speed, which not only fails to level the voltage difference in time, but also causes some cells to remain in a high voltage state for a long time. In the float charging stage, only slight voltage deviations need to be corrected. If single active balancing is used, the excessive energy consumption will increase the energy loss of the battery system. The adaptability defects of active and passive balancing described above make it impossible for the balancing process to be dynamically adjusted according to the operating conditions, which not only fails to guarantee the balancing effect, but also generates energy redundancy loss.

[0003] Second, the adaptability defects of equalization technology can lead to uncontrolled equalization parameters, which may cause a decrease in the accuracy of battery state prediction. During the equalization process, equalization parameters such as the number of equalization cycles and equalization current will affect the prediction of battery state. For example, insufficient equalization cycles will lead to an increase in the voltage difference between individual cells, resulting in an overestimation of the predicted state of battery (SOH). Over-equalization will interfere with the cycle life prediction, shortening the predicted life. Insufficient equalization time can easily create a consistency illusion, overestimating capacity and range. Excessive equalization time will cause over-repair loss, resulting in an underestimation of the remaining life. Current conventional battery state prediction does not take into account the impact of the equalization process on the battery prediction results.

[0004] Therefore, the current limitations of single or non-single balancing technologies not only result in poor balancing effects and redundant balancing energy consumption, but also cause balancing parameters to become out of control, leading to low accuracy in battery state prediction. Summary of the Invention

[0005] Therefore, this invention provides a BMS control method for battery prediction, which effectively solves the technical problem that the adaptability defects of single or non-single equalization technology in the prior art not only result in poor equalization effect and redundant equalization energy consumption, but also cause the equalization parameters to run out of control, thus leading to low accuracy of battery state prediction.

[0006] To solve the above-mentioned technical problems, the present invention specifically provides the following technical solution: a BMS control method for battery prediction, comprising the following steps: Step 1: Obtain the state parameters of each battery cell in the battery pack, and analyze the differences between battery cells based on the state parameters of the battery cells; Step 2: Construct a hierarchical-domain hybrid balancing topology architecture, and build a conventional hybrid balancing switching strategy based on the hierarchical-domain hybrid balancing topology architecture; A differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of individual battery cells; Different equilibrium adjustment strategies are composed of conventional hybrid equilibrium switching strategies and differential hybrid equilibrium switching strategies; Step 3: Determine the balancing adjustment strategy to be implemented based on the differences between individual battery cells; Step 4: Block energy interference between active and passive balancing by hardware isolation, and establish the operational boundaries of active and passive balancing by logical interlocking, so that active and passive balancing are carried out separately. Step 5: Construct a battery state prediction model based on normal operating data and equalization benchmark parameters, and predict the battery state based on the battery state prediction model.

[0007] Further, in step 1, the state parameters of each battery cell in the battery pack are obtained, and the differences between battery cells are analyzed based on the state parameters of the battery cells, including the following steps: Step 11: Conduct identical charge and discharge test records for the battery pack; When the battery pack is charged and discharged under the same charging and discharging conditions for the same period of time, the detection is carried out and the data is recorded; Step 12: Analyze the voltage fluctuation curve based on the data obtained from multiple charge and discharge processes, and calculate the voltage difference parameters; During multiple charge and discharge cycles of the battery pack, voltage data of each battery cell is collected at fixed time intervals, voltage fluctuation curves are plotted, the collected voltage fluctuation curves are analyzed, voltage data collected at equal time intervals under conditions without equalization are obtained, and the average voltage of all battery cells is calculated. The voltage difference is calculated by subtracting the voltage data from the average voltage. The average voltage difference is calculated based on the voltage difference at all time points, and the average voltage difference is used as the voltage difference parameter of the battery cells. Step 13: Calculate the capacity difference parameters of individual battery cells based on the data obtained from multiple charge and discharge processes; Based on the charge and discharge capacity data recorded during multiple charge and discharge cycles of the battery pack, combined with the charge and discharge characteristics of the battery, the actual capacity of each battery cell at equal time intervals is obtained and the average actual capacity of all battery cells is calculated. The capacity difference is calculated by comparing the actual capacity of the battery cells with the average actual capacity. The average capacity difference is calculated based on the capacity difference at all time points and used as the capacity difference parameter of the battery cells. Step 14: Calculate the internal resistance difference parameter of the battery cells based on the data obtained from multiple charge and discharge processes; During the charging and discharging process of the battery pack, the internal resistance value of each battery cell is monitored and recorded in real time. The internal resistance value of each battery cell at equal time intervals is obtained and the average internal resistance value of all battery cells is calculated. The difference between the internal resistance value data and the average internal resistance value is obtained. The average resistance difference is calculated based on the internal resistance difference measured at all time points. The average internal resistance difference is used as the internal resistance difference parameter of the battery cells. Step 15: Calculate the individual unit difference parameters based on voltage difference parameters, capacity difference parameters, and internal resistance difference parameters; Based on voltage difference parameters, capacity difference parameters, and internal resistance difference parameters, the individual cell difference parameters are calculated using a preset weighting algorithm.

[0008] Further, in step 15, the individual cell difference parameters of the battery cells are calculated using a preset weight allocation algorithm, specifically including the following steps: Step 151: Determine the basic weight range; Step 152: Adjust the weights according to the battery type; Step 153: Adjust the weights a second time based on the application scenario; Step 154: Based on the adjusted weights above, the specific calculation steps for the individual cell difference parameters of the battery cells are as follows: Obtain the voltage difference parameters, capacity difference parameters, and internal resistance difference parameters of individual battery cells; The voltage difference parameter, capacity difference parameter, and internal resistance difference parameter are all normalized. The individual cell difference parameters are obtained by weighting the normalized voltage difference parameters, capacity difference parameters, and internal resistance difference parameters.

[0009] Further, in step 2, a hierarchical-domain hybrid load balancer topology is constructed, and a conventional hybrid load balancer switching strategy is built based on the hierarchical-domain hybrid load balancer topology architecture, specifically including the following steps: Step 21: Establish a hierarchical balanced topology between and within groups; The battery pack is divided into battery sub-groups, and inter-group active balancing units are deployed between different battery sub-groups, while intra-group passive balancing units are deployed between battery cells in the same battery sub-group. Step 22: Construct a domain-balanced topology; Divide the temperature zones, deploy temperature monitoring units, and formulate a temperature zone regional balance switching strategy; Step 23: Construct a fuzzy control algorithm to adaptively switch the equalization mode based on voltage difference amplitude, temperature range, and charging / discharging stage, so as to formulate a conventional hybrid equalization switching strategy.

[0010] Furthermore, in step 23, a fuzzy control algorithm is constructed to adaptively switch the equalization mode based on the voltage difference amplitude, temperature range, and charging / discharging stage, in order to formulate a conventional hybrid equalization switching strategy, including the following sub-steps: Step 231, divide the charging and discharging phase scenarios; Based on the charging and discharging current and SOC value of the battery pack, the working process of the battery pack is divided into different charging and discharging stages: fast charging stage, float charging stage, discharging stage, and resting stage. Step 232: Collect operating parameters in real time and make judgments; The system collects operating parameters in real time and transmits them to the BMS main control unit. The BMS main control unit processes and analyzes the operating parameters in real time to determine the current charging and discharging stage of the battery pack. Step 233: Combining the hierarchical equalization topology and the temperature-regional equalization switching strategy, formulate the equalization mode switching logic of the conventional hybrid equalization switching strategy.

[0011] Furthermore, in step 2, a differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of the battery cells. This strategy comprises different equalization adjustment strategies, including the following steps: Step 24 involves constructing a differentiated hybrid equalization switching strategy based on the aging level and operating conditions of individual battery cells. This strategy includes the following sub-steps: Step 241: Analyze and determine the degree of battery aging; Adjust the phased equalization switching strategy according to the degree of battery aging, formulate the extreme operating condition switching strategy for fluctuations in charge and discharge rates, and formulate the fault warning and emergency switching strategy based on system safety requirements. Furthermore, in step 2, a differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of the battery cells. This strategy comprises different equalization adjustment strategies, consisting of a conventional hybrid equalization switching strategy and a differentiated hybrid equalization switching strategy. The steps also include: Step 25: Construct a priority strategy for the load balancing switching strategy, which includes the following sub-steps: Step 251: Divide the priority layers of the balanced switching strategy; Based on the battery pack's operational requirements, the equalization switching strategy is divided into three levels of priority from high to low: First priority: emergency switching strategy for fault warning; Second priority: switching strategy for extreme operating conditions; Third priority: phased equalization switching strategy and regular hybrid equalization switching strategy. Step 252: Develop priority conflict handling rules.

[0012] Further, step 3, determining the balancing adjustment strategy to be implemented based on the differences in individual battery cells, includes the following steps: Step 31: Preset the threshold corresponding to the individual cell difference parameter and the threshold for the number of abnormal battery cells; Step 32: Compare the calculated individual cell difference parameters with the preset corresponding thresholds, and count the number of abnormal battery cells whose individual cell difference parameters exceed the thresholds. If the number of abnormal battery cells exceeds the corresponding threshold, a differential hybrid equalization switching strategy will be initiated. If the number of abnormal battery cells does not exceed the quantity threshold, a regular hybrid equalization switching strategy will be initiated.

[0013] Furthermore, in step 4, energy interference between active and passive balancing is blocked by hardware isolation, and the operational boundaries of active and passive balancing are established by logical interlocking, so that active and passive balancing are performed separately. This includes the following steps: Step 41: Connect two independent switches in series at the positive and negative terminals of each battery cell, corresponding to the passive balancing circuit and the active balancing circuit, respectively. Step 42, achieving logical interlocking between active and passive equilibrium through state interlocking and trigger condition exclusion, includes the following two sub-steps: Step 421: Establish interlocking logic between individual battery cells; Set an equalization status flag for each battery cell and define interlocking rules; Step 422: Priority scheduling of passive and active balancing based on sudden changes in complex operating conditions.

[0014] Furthermore, in step 5, a battery state prediction model is constructed based on conventional operating data and equalization benchmark parameters. The battery state is then predicted according to this model, specifically including the following steps: Step 51: Multi-dimensional raw data collection and synchronization; Real-time acquisition of battery operation data and balancing parameter data, ensuring timestamp alignment, includes the following sub-steps: Step 511: Determine the data collection range and frequency; Set the data collection frequency and clearly define the types of data to be collected, dividing them into two categories: routine operating data and balanced baseline parameter data; Step 512: Collect routine operating data; The system collects and stores data on the voltage, current, and temperature of individual battery cells using current sensors, voltage sensors, and temperature sensors. Step 513: Collect equalization benchmark parameter data; Collect the baseline parameters for each balancing process and match them with the regular operation data by timestamp; Step 514: Perform preliminary verification and storage of routine operating data and equilibrium benchmark parameters; Verify and remove abnormal data, and store the verified data in layers; Step 52: Quantize and preprocess the equilibrium benchmark parameters; Step 53: Preprocess the routine operating data; Step 54: Construct a balanced-regular joint feature set; By integrating the preprocessed equilibrium benchmark parameters with regular operating characteristics, effective features are selected to form a joint equilibrium-regular feature set. Step 55: Perform feature filtering on the balanced-conventional joint feature set; Calculate the correlation coefficients between each feature in the balanced-conventional joint feature set and the SOH and RUL values, and remove redundant features whose absolute values ​​of the correlation coefficients are lower than the corresponding thresholds; Step 56: Train and optimize the battery state prediction model using the balanced-conventional joint feature set as input; Step 57: Predict and output the SOH and RUL values ​​based on the battery state prediction model.

[0015] Compared with the prior art, the present invention has the following advantages: This invention addresses the problems of poor balance adaptation, uncontrolled prediction parameters, and low battery prediction accuracy by optimizing the balance switching strategy and integrating balance into the positive correlation system of prediction. This achieves adjustments to the adaptability balance switching strategy and optimization of battery prediction accuracy. (i) To address the problem of dynamic deterioration in battery cell consistency caused by the adaptability defects of a single equalization technology, this invention constructs a multi-level, multi-dimensional equalization switching strategy system, achieving deep adaptation of the equalization technology to different operating conditions: To test the consistency of individual battery cells, when the consistency of individual battery cells is good, a conventional hybrid equalization switching strategy is implemented based on the voltage difference magnitude, temperature range, and charging / discharging stage. When the consistency of individual battery cells is poor and there are no other safety warnings or extreme operating conditions, a corresponding differential hybrid equalization switching strategy is implemented based on the aging stage. Based on the consistency of individual battery cells, the equalization technology is adapted, thereby optimizing the equalization adaptability. (II) Regarding the impact of balancing parameters on battery state prediction results, this invention uses conventional operating data and balancing reference parameters as a benchmark to predict battery state, achieving a positive correlation between balancing reference parameters and battery state: By using equalization benchmark parameters such as equalization number, equalization current, and equalization time as inputs to the battery state prediction model, and addressing the limitation of single input features in the battery state prediction model, this method integrates preprocessed equalization benchmark parameters with regular operating data, eliminates redundant features, retains key features, and establishes a correlation between equalization benchmark parameters and battery state prediction results through the battery state prediction model. This effectively captures the implicit impact of the equalization process on battery health and optimizes the accuracy of battery state prediction.

[0016] This invention optimizes battery balancing adaptability, achieving the dual goals of improving balancing effect and reducing balancing loss. It also provides a more stable data source for the battery state prediction process, optimizing the battery state prediction process with both balancing benchmark parameters and regular operating data to solve the problem of balancing parameter runaway. At the same time, the battery state prediction model takes into account the impact of the balancing process, making battery state prediction more accurate. Attached Figure Description

[0017] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0018] Figure 1 A flowchart of a BMS control method for battery prediction provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the calculation of monomer difference parameters in this invention; Figure 3 This is a flowchart illustrating the conventional hybrid equalization switching strategy in this invention; Figure 4 This is a flowchart of the equalization mode switching logic of the conventional hybrid equalization switching strategy in this invention; Figure 5 This is a schematic diagram illustrating the corresponding balancing switch based on aging stage, extreme operating conditions, and system safety in this invention. Figure 6 This is a flowchart illustrating the equalization adjustment strategy determined based on the differences between individual battery cells in this invention. Figure 7 This is a flowchart of the training and optimization process for the battery state prediction model in this invention; Figure 8A basic weighting interval table for different battery types; Figure 9 A table of criteria for different battery types; Figure 10 This is a program algorithm diagram of the equalization mode switching logic of the conventional hybrid equalization switching strategy in this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, the present invention provides a BMS control method for battery prediction, comprising the following steps: Step 1: Obtain the state parameters of each battery cell in the battery pack, and analyze the differences between battery cells based on the state parameters of the battery cells; Step 2: Construct a hierarchical-domain hybrid balancing topology architecture, and build a conventional hybrid balancing switching strategy based on the hierarchical-domain hybrid balancing topology architecture; A differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of individual battery cells; Different equilibrium adjustment strategies are composed of conventional hybrid equilibrium switching strategies and differential hybrid equilibrium switching strategies; Step 3: Determine the balancing adjustment strategy to be implemented based on the differences between individual battery cells; Step 4: Block energy interference between active and passive balancing by hardware isolation, and establish the operational boundaries of active and passive balancing by logical interlocking, so that active and passive balancing are carried out separately. Step 5: Construct a battery state prediction model based on normal operating data and equalization benchmark parameters, and predict the battery state based on the battery state prediction model.

[0021] Step 1 involves systematic testing and data analysis of the battery pack to accurately grasp the differences between individual battery cells, providing data support for selecting an appropriate balancing strategy. This step includes the following sub-steps: Step 11: Conduct identical charge and discharge test records for the battery pack; When the battery pack is charged and discharged under the same charging and discharging conditions for the same period of time, the data (operating parameters, etc.) are detected and recorded. By testing through multiple charge-discharge cycles, random errors that may occur in the data of a single charge-discharge process can be eliminated, providing a more comprehensive reflection of the battery pack's performance in actual use and laying the foundation for obtaining accurate difference parameters in the future.

[0022] Step 12: Analyze the voltage fluctuation curve based on the data obtained from multiple charge and discharge processes, and calculate the voltage difference parameters; During multiple charge and discharge cycles of the battery pack, voltage data for each individual battery cell is collected at fixed time intervals to form a complete voltage fluctuation curve. The collected voltage fluctuation curves are analyzed to obtain voltage data collected at equal time intervals under conditions without equilibration and to calculate the average voltage of all battery cells. The voltage difference is calculated by subtracting the voltage data from the average voltage. The average voltage difference is then used as the voltage difference parameter for each battery cell. This voltage difference parameter can intuitively reflect the degree of inconsistency in voltage changes of battery cells during charge and discharge.

[0023] Step 13: Calculate the capacity difference parameters of individual battery cells based on the data obtained from multiple charge and discharge processes; Based on the charge and discharge capacity data recorded during multiple charge and discharge cycles of the battery pack, and combined with the battery's charge and discharge characteristics, the actual capacity of each battery cell at equal time intervals is obtained, and the average actual capacity of all battery cells is calculated. The capacity difference is calculated by comparing the actual capacity of each battery cell with the average actual capacity. The average capacity difference is then used as the capacity difference parameter of the battery cells. The capacity difference parameter is a key indicator for evaluating the differences in the energy storage capacity of battery cells and directly affects the overall range performance of the battery pack.

[0024] Step 14: Calculate the internal resistance difference parameter of the battery cells based on the data obtained from multiple charge and discharge processes; During the charging and discharging process of the battery pack, the internal resistance value of each battery cell is monitored and recorded in real time. The internal resistance value of each battery cell at equal time intervals is obtained, and the average internal resistance value of all battery cells is calculated. The difference between the internal resistance data and the average internal resistance value is obtained. The average resistance difference is calculated using the internal resistance difference measured at all time points. The average internal resistance difference is used as the internal resistance difference parameter of the battery cells. The internal resistance difference parameter can reflect the difference in the internal conductivity of the battery cells. Battery cells with excessive internal resistance are more likely to generate heat during charging and discharging, which affects the safety and service life of the battery pack.

[0025] Step 15: Calculate the individual unit difference parameters based on voltage difference parameters, capacity difference parameters, and internal resistance difference parameters; Based on the voltage difference parameters obtained in step 12, the capacity difference parameters obtained in step 13, and the internal resistance difference parameters obtained in step 14, the final individual battery cell difference parameters are calculated using a preset weight allocation algorithm (the weights can be adjusted according to factors such as battery type and application scenario). like Figure 2 As shown, in step 15, the individual cell difference parameters of the battery cells are calculated using a preset weight allocation algorithm, specifically including the following steps: Step 151: Determine the basic weight range; Step 152: Adjust the weights according to the battery type; Specific values ​​are adjusted within the basic weight range based on the characteristics of different battery types, such as... Figure 8 As shown.

[0026] Step 153: Adjust the weights a second time based on the application scenario; Based on the weighting of battery type, the weights are further fine-tuned according to the core needs of specific application scenarios to ensure that they match the pain points of the scenario, as shown in the following example: Scenario 1: Electric vehicle fast charging system (ternary lithium battery) During fast charging (above 1.5C), the voltage difference increases rapidly, so overcharging risk needs to be controlled first. Therefore, the voltage weight is increased. The adjusted weights are: w_V=0.6, w_C=0.25, w_R=0.15. During fast charging, the voltage rise rate reaches 0.02-0.03V / min. The voltage difference directly determines whether overvoltage protection is triggered. Therefore, the voltage weight is increased to the highest level. The capacity difference has a relatively small impact on a single fast charge, so the capacity weight is appropriately reduced.

[0027] Scenario 2: Low-temperature power battery (lithium iron phosphate battery, -20℃ operating condition) At low temperatures, the difference in internal resistance increases by 3-5 times, affecting discharge efficiency. It is necessary to increase the weight of internal resistance. The adjusted weights are: w_V=0.45, w_C=0.35, w_R=0.2. The difference in internal resistance at low temperatures leads to inconsistent voltage drops. It is necessary to increase the weight of internal resistance while retaining a relatively high voltage weight to avoid over-discharge at low temperatures.

[0028] Based on the adjusted weights above, the specific calculation steps for the individual cell difference parameters of the battery cells are as follows: First, obtain the voltage difference parameters, capacity difference parameters, and internal resistance difference parameters of the individual battery cells; Secondly, the voltage difference parameter, capacity difference parameter, and internal resistance difference parameter are normalized—the parameters with different dimensions are converted into dimensionless values ​​between 0 and 1 to avoid calculation deviations caused by differences in battery units. (1) Voltage difference parameter normalization, normalization formula: ΔV'=ΔV / ΔV_max, where ΔV is the voltage difference parameter before normalization, ΔV' is the voltage difference parameter after normalization, and ΔV_max is the maximum allowable voltage difference of the battery cell of this battery type. (2) Normalization of capacity difference parameters. Normalization formula: ΔC'=ΔC / C_rated, where ΔC is the capacity difference parameter before normalization, ΔC' is the capacity difference parameter after normalization, and C_rated is the rated capacity of the battery cell.

[0029] (3) Normalization of internal resistance difference parameter. Normalization formula: ΔR'=ΔR / R_max, where ΔR is the internal resistance difference parameter before normalization, ΔC' is the internal resistance difference parameter after normalization, and R_max is the maximum allowable internal resistance of the battery cell of this battery type (e.g., the internal resistance of a new ternary lithium battery is ≤30mΩ, and the maximum allowable internal resistance after aging is ≤60mΩ, so take R_max=60mΩ).

[0030] Finally, the normalized voltage difference parameter, capacity difference parameter, and internal resistance difference parameter are weighted and calculated to obtain the single-cell difference parameter D, with the formula: D=w_V×ΔV'+w_C×ΔC'+w_R×ΔR'; The D value ranges from 0 to 1. The closer the value is to 1, the greater the difference between the individual battery cells and other battery cells in the battery pack. The closer the D value is to 0, the better the consistency between individual battery cells.

[0031] Taking the fast charging scenario of electric vehicles - ternary lithium batteries as an example, the scenario weights are: w_V=0.6, w_C=0.25, w_R=0.15, and the normalized parameters are: ΔV'=0.333, ΔC'=0.05, ΔR'=0.083. The weighted calculation is D=0.6×0.333+0.25×0.05+0.15×0.083≈0.22475. The individual battery cell difference parameter is about 0.225, which is at a low level, indicating that the consistency of the individual battery cell in the entire battery pack is good.

[0032] In addition to adjusting the voltage weight, capacity weight, and internal resistance weight for different battery types and application scenarios, adjustments to these weights are also necessary for certain special scenarios, such as: 1. High-aging batteries (internal resistance change rate > 18%): The difference in internal resistance of high-aging batteries has a significant impact on heat generation and discharge efficiency, and the weight of internal resistance needs to be increased.

[0033] 2. In ultra-high temperature environments (>50℃), the risk of heating due to differences in internal resistance is aggravated at high temperatures, and voltage stability decreases at the same time. It is necessary to increase the weighting of voltage and internal resistance simultaneously.

[0034] Through the detailed weight allocation and calculation process described above, it can be ensured that the individual battery cell difference parameters can objectively and accurately reflect the actual consistency level of the individual battery cells in the battery pack, providing reliable data support for subsequent equalization switching strategy selection and other processes, and avoiding problems of insufficient or excessive equalization caused by parameter misjudgment.

[0035] Step 2: Construct a hierarchical-domain hybrid balancing topology, and based on the hierarchical-domain hybrid balancing topology architecture, construct a conventional hybrid balancing switching strategy, and construct a differential hybrid balancing switching strategy based on the aging degree and operating conditions of individual battery cells. The conventional hybrid balancing switching strategy and the differential hybrid balancing switching strategy constitute different balancing adjustment strategies.

[0036] Step 2 addresses the varying requirements of battery packs under different operating conditions by designing a scientifically sound and reasonable balancing topology architecture. This allows active and passive balancing to fully leverage their respective advantages, and includes the following sub-steps: Step 21: Establish a hierarchical balanced topology between and within groups; Step 211: Divide the battery pack into battery subgroups; Based on the total number of battery cells in the battery pack and the actual application requirements, the battery cells in the battery pack are evenly divided into several battery subgroups to ensure that the number of battery cells in each battery subgroup is consistent.

[0037] Taking the 48-cell ternary lithium battery commonly used in electric vehicle power battery systems as an example, it can be divided into 8 battery subgroups, each containing 6 battery cells. This grouping method is not only easy to manage, but also effectively balances the equilibrium pressure between and within groups.

[0038] Step 212: Deploy inter-group active balancing units between different battery sub-groups; By deploying an active balancing unit between different battery sub-groups, in the case of 48 ternary lithium batteries divided into 8 battery sub-groups, an inductor-based bidirectional DC / DC converter is deployed between the 8 battery sub-groups. The active balancing unit can realize the rapid transfer of energy between different battery sub-groups.

[0039] Step 213: Deploy intra-group passive balancing units among the individual cells of the same battery sub-group; An intra-group passive balancing unit is constructed between the battery cells in the same battery sub-group. For each group of 6 battery cells, a shunt resistor controlled by a MOSFET is connected in series. The resistance value is matched according to the individual cell capacity. When the voltage difference between the battery cells in the battery sub-group is small, the passive balancing unit can dissipate a small amount of excess energy from the high-voltage battery cells, thereby achieving voltage balancing of the battery cells in the battery sub-group with low power consumption and avoiding energy waste.

[0040] Step 22: Construct a domain-specific (high temperature region + low temperature region) balanced topology; Step 221: Divide the temperature zones; Based on the battery's operating characteristics and the impact of temperature on battery performance, the battery pack's operating temperature range is divided into two regions: "low temperature range (<10℃)" and "normal temperature range (≥10℃)". The battery's internal resistance, charge and discharge efficiency, and other performance parameters vary significantly in different temperature regions, requiring targeted matching and equalization switching strategies.

[0041] Step 222: Deploy the temperature monitoring unit; Temperature sensors are strategically placed within the battery pack to ensure accurate and real-time collection of temperature data from various locations within the pack. This data is then transmitted to the BMS main control unit, which determines the current temperature range of the battery pack based on the received temperature data.

[0042] Step 223: Develop a temperature zone balancing switching strategy; Low-temperature domain equalization switching strategy: When the temperature sensor detects that the battery pack temperature is <10℃, the active equalization priority mode is automatically activated.

[0043] Taking the lithium iron phosphate battery operating at -20℃ as an example, the internal resistance deviation of the battery cell can reach 20%. If only passive balancing is used to correct the 0.1V deviation, it will take 4 hours. However, by using active balancing (balancing current 5A), the correction can be completed within 30 minutes. At the same time, the efficiency of active balancing in the low temperature range can still reach 80%, which is 30% higher than that of passive balancing. It can effectively reduce energy loss at low temperatures and ensure the normal operation performance of the battery pack in low temperature environments. Normal temperature range equalization switching strategy: When the temperature sensor detects that the battery pack temperature is ≥10℃, it switches to passive equalization priority mode.

[0044] Under normal temperature conditions, passive balancing is sufficient to handle deviations of 0.05-0.1V (correction time is about 1.5 hours), and its power consumption is only 1 / 3 of that of active balancing (active balancing consumes about 3W at normal temperature, while passive balancing consumes only 1W). This avoids the waste problem of high energy consumption for low deviations and improves the energy utilization efficiency of the battery pack.

[0045] Step 23: Construct a fuzzy control algorithm to adaptively switch the equalization mode based on voltage difference amplitude, temperature range, and charging / discharging stage, so as to formulate a conventional hybrid equalization switching strategy; like Figure 3 As shown, a fuzzy control algorithm is constructed using "voltage difference amplitude," "temperature range," and "charging / discharging stage" as core input parameters. The algorithm includes the following steps: Step 231: Divide the charging and discharging phase scenarios; Based on the battery pack's charging and discharging current and SOC value, the battery pack's operation process is divided into different charging and discharging stages, as follows: Fast charging phase: SOC value is in the range of 0-80%, and charging current value is >0.5C; Float charging stage: The SOC value is in the range of 80-100%, and the charging current value is <0.1C; Discharge phase: Discharge current value > 0.2C; During the resting phase: current value < 0.05C; The voltage change patterns and balancing requirements of battery packs differ significantly at different charging and discharging stages. Dividing the stages into scenarios provides a basis for accurately switching balancing modes.

[0046] Step 232: Collect operating parameters in real time and make judgments; The system collects operating parameters such as individual battery cell voltage, individual cell temperature, and charge / discharge current in real time. The collected operating parameters are transmitted to the BMS main control unit, which processes and analyzes the operating parameters in real time to determine the current charge / discharge stage of the battery pack.

[0047] Step 233: Combining the hierarchical equalization topology and the temperature-regional equalization switching strategy, formulate the equalization mode switching logic of the conventional hybrid equalization switching strategy; like Figure 4 and Figure 10 As shown, the equalization mode switching logic is as follows: During the fast charging phase (0-80% SOC), when the voltage difference (ΔV) between battery sub-groups is detected to be greater than 0.05V, an active balancing mode between sub-groups is adopted to quickly suppress the expansion of the difference and ensure that the voltage of each battery cell remains consistent during the fast charging process, thus avoiding overcharging of some battery cells.

[0048] During the float charging phase (80-100% SOC), if the voltage difference ΔV between individual cells in the battery sub-group is less than 0.05V, a passive balancing mode is adopted to correct slight deviations with low power consumption and improve energy utilization efficiency.

[0049] During the low-temperature discharge stage (temperature below -10℃), when the voltage difference ΔV between battery sub-groups is detected to be greater than 0.08V, an active balancing mode is adopted to overcome the influence of increased internal resistance at low temperatures and ensure the discharge capacity of the battery pack.

[0050] During the room temperature resting stage (temperature between 10-30℃), if the voltage difference ΔV between individual cells in the battery sub-group is less than 0.08V, the passive balancing mode is used to extend the battery cycle life by reducing energy consumption.

[0051] Step 24: Construct a differentiated hybrid equalization switching strategy based on the aging degree and operating conditions of individual battery cells; Step 24: Based on the battery's aging level and different extreme operating conditions, formulate a targeted balancing and switching strategy to ensure that the battery pack can achieve efficient and safe balancing throughout its entire life cycle and under complex operating conditions. This includes the following sub-steps: Step 241: Analyze and determine the degree of battery aging; Step 2411: Determine the criteria and quantification thresholds for assessing the degree of aging; Using "cycle life," "capacity decay rate," and "internal resistance change rate" as the three main evaluation indicators, and combining them with battery type (such as ternary lithium batteries and lithium iron phosphate batteries), quantitative thresholds for each aging stage are established. Figure 9 As shown, this provides a standard basis for accurately determining the degree of battery aging.

[0052] Step 2412: Determine the equivalent cycle number of a single battery cell based on the charge-discharge cycle process; Record the equivalent number of cycles for each charge-discharge cycle (full charge and discharge is one cycle, partial charge and discharge is converted into "energy integral", such as charging 50% and then discharging 50%, which is converted into 0.5 cycles), and accumulate the equivalent number of cycles (cycle life) in real time. Considering the impact of the battery charging and discharging stages, if the battery cell is in a shallow charge and discharge state (such as SOC 30-70% cycle) for a period of time exceeding the corresponding threshold, the equivalent number of cycles needs to be corrected according to the life extension factor (the actual life of shallow charge and discharge is about 1.5-2 times that of standard cycle) so that the estimation results are more in line with the actual aging of the battery.

[0053] Step 2413: Determine the capacity decay rate based on energy integral and temperature correction; During battery charging, the current actual capacity of the battery is estimated by integrating the amount of charge [for example, if the SOC is 20% before charging and 3kWh of charge is added when the battery reaches 80%, the current actual capacity = 3kWh / (80%-20%) = 5kWh]. The ratio of the current actual capacity to the battery's rated capacity is then calculated to obtain the capacity decay rate. To ensure the accuracy of the calculation results, the influence of charging efficiency needs to be eliminated. For example, the charging efficiency is low in low temperature environments, and the charged amount may be lower than the actual capacity. In this case, the charged amount needs to be adjusted according to the temperature correction coefficient. That is, the charged amount is divided by the temperature correction coefficient before calculating the actual capacity. Under normal circumstances, the capacity decay rate is automatically calculated once after each complete charge and discharge cycle.

[0054] Step 2414: Calculate the rate of change of internal resistance of a single battery cell; Real-time measurement of battery internal resistance (e.g., applying a short-duration 1.5C pulse current during battery discharge, recording the voltage change, and calculating the internal resistance based on Ohm's law), and calculation of the rate of change of internal resistance based on the battery internal resistance; For example, the initial internal resistance of a certain vehicle ternary lithium battery is 30mΩ. After 2 years of use, the internal resistance is measured to be 35mΩ. The internal resistance change rate is calculated as: (35-30) / 30×100%≈16.7%.

[0055] Step 2415: Perform auxiliary verification and correction of the aging degree of individual battery cells based on consistency; The consistency differences between individual battery cells are used to help verify the degree of battery aging. The principle is that the higher the degree of battery aging, the worse the consistency (capacity, internal resistance) between individual battery cells. Calculate the capacity deviation of each battery cell [(actual capacity - average actual capacity) / average actual capacity × 100%) and the internal resistance deviation [(internal resistance data - average internal resistance) / average internal resistance × 100%), and perform auxiliary verification according to the following standards: New battery / low aging stage: capacity deviation <5%, internal resistance deviation <8%; Medium aging stage: Capacity deviation is within the range of 5-15%, and internal resistance deviation is within the range of 8-18%. High aging stage: Capacity deviation > 15%, internal resistance deviation > 18%.

[0056] Based on the consistency verification results, the aging degree determined by core indicators is corrected to improve the accuracy of the determination.

[0057] Step 2416: Determine the degree of aging based on the equivalent cycle number, capacity decay rate, and internal resistance change rate of the battery cells.

[0058] Step 2417: Analyze and correct the degree of aging based on extreme operating conditions; Under extreme operating conditions (such as prolonged low temperatures, overcharging, and over-discharging), a single aging indicator may exhibit "abnormal fluctuations," requiring cross-validation using multiple indicators to avoid misjudging the aging stage. Long-term low-temperature use scenario: First, heat the individual battery cells to room temperature (25℃), then test the actual capacity of the individual battery cells and compare the low-temperature capacity recovery rate (the ratio of the difference between the actual capacity at low temperature and the actual capacity at room temperature to the actual capacity at room temperature). Among them, for new batteries / low aging stage, the capacity recovery rate after heating is >95%; for moderate aging stage, the recovery rate is in the range of 90-95%; for high aging stage, the recovery rate is <90%. The aging degree judgment is corrected based on the low-temperature capacity recovery rate results.

[0059] Overcharge and over-discharge fault scenarios: Test the actual capacity of the charging and discharging process three times in a row. If the measured parameters are stable (fluctuation <3%), the aging stage is determined according to the current parameters. If the parameters continue to drop and the drop exceeds 1%, it is directly determined to be in the high aging stage, and the battery needs to be replaced.

[0060] Step 242: Adjust the phased equalization switching strategy according to the degree of battery aging, such as... Figure 5 As shown, Step 2421: Develop a balanced switching strategy for new batteries / low aging stages; On-demand balancing is performed only when the voltage difference between individual battery cells exceeds a set threshold, thus avoiding excessive balancing that leads to energy waste.

[0061] The specific strategy is as follows: when the voltage difference ΔV > 0.06V, active balancing is initiated to quickly level the difference; when ΔV < 0.04V, passive balancing is switched to correct slight deviations; if ΔV remains stable within the range of 0.02-0.04V for three consecutive charge-discharge cycles, balancing operation can be paused (only voltage monitoring is maintained) to further reduce energy consumption and extend battery life.

[0062] Step 2422: Develop a balanced transition strategy for the moderate aging stage; The balancing frequency needs to be increased to avoid the accumulation of differences between individual battery cells, which could lead to overcharging or over-discharging of some cells.

[0063] The specific strategy is as follows: maintain active balancing priority mode throughout the entire battery charging and discharging stage; when ΔV > 0.05V is detected, active balancing is immediately initiated; even if ΔV < 0.05V, if the capacity decay rate of a certain battery cell is detected to be more than 1.5 times the average value, active balancing still needs to be maintained (but the balancing current is reduced to 0.5-1A) to slow down the decay rate of the battery cell through energy transfer; only during the resting stage (current < 0.05C) and when ΔV < 0.03V, switch to passive balancing to reduce energy consumption.

[0064] Step 2423: Develop a balanced switching strategy for the high-aging phase; The core requirement for balancing is to avoid the impact of strong current balancing on weaker battery cells, while frequent correction of differences is necessary to ensure the safety of the battery system.

[0065] The specific strategy is as follows: disable high-current active balancing (limit the maximum balancing current to within 0.8A); when ΔV > 0.07V is detected, adopt a low-current active balancing and passive balancing co-mode (active balancing transfers some energy, passive balancing assists in discharge correction); when ΔV < 0.06V, switch completely to passive balancing (reduce the balancing current to 0.2-0.3A) to prevent weak battery cells from being further damaged by the energy impact of active balancing; if the voltage of a certain battery cell is lower than the set threshold at the end of discharge (e.g., the voltage threshold at the end of discharge for ternary lithium batteries is 2.5V), immediately start passive balancing to discharge other high-voltage cells to avoid over-discharge of the low-voltage cell and ensure the safety of the battery system.

[0066] Step 243: Develop extreme operating condition switching strategies to address charge / discharge rate fluctuations, such as... Figure 5 As shown; Step 2431: Develop a balanced switching strategy for ultra-fast charging scenarios (charging current > 1.5C); In ultra-fast charging scenarios, the voltage rise rate of a single battery cell can reach 0.02-0.03V / min. Traditional equalization modes struggle to keep up with the rate of difference expansion, which can easily lead to some battery cells reaching the charging cutoff voltage first.

[0067] The switching logic for this scenario is as follows: Forcefully activate high-power active balancing (increase the balancing current to 2.5-3A), while shortening the voltage sampling interval (from the usual 500ms to 100ms) to track voltage difference changes in real time. If the voltage of a battery cell is detected to be close to the charging cut-off voltage, passive balancing must be activated immediately to force discharge the battery cell, and active balancing must be used to pull its voltage back to a safe range (such as below 4.1V). Once the charging current drops below 1C (usually corresponding to about 60% SOC), the balancing mode switching logic under non-extreme operating conditions is restored to avoid energy waste caused by continuous high-power balancing.

[0068] Step 2432: Develop a balanced switching strategy for high-rate discharge scenarios (discharge current > 2C); During high-rate discharge, the voltage of individual battery cells drops rapidly, and differences in internal resistance can lead to inconsistent voltage drops. Battery cells with higher internal resistance experience faster voltage drops, which can easily result in false low voltages. If this is mistakenly interpreted as a genuine difference and balancing is initiated, it will waste battery energy.

[0069] The switching logic for this scenario is as follows: First, passive equalization is disabled (to avoid further exacerbating voltage drops during the passive equalization discharge process), and only active equalization is enabled with the equalization current reduced to 0.5-1A (to reduce the impact of energy transfer on individual battery cells). When a voltage difference ΔV > 0.1V is detected, the equalization operation is paused and monitored for 30 seconds. If ΔV remains > 0.08V after 30 seconds (excluding false voltage interference), active equalization is then started to transfer energy, leveling out the voltage differences between individual battery cells. Once the discharge current drops below 1C, the equalization mode switching logic under non-extreme operating conditions is restored to balance the equalization effect and energy loss.

[0070] Step 2433: Develop a balancing switching strategy for pulse charging and discharging scenarios (such as the start-stop operation of hybrid vehicles); In pulse charging and discharging scenarios, the charging and discharging currents alternate frequently (such as 10 seconds of charging and 5 seconds of discharging), and the voltage fluctuates frequently. If the balancing mode is switched frequently, it will easily lead to increased losses of balancing devices (such as MOSFETs and DC / DC converters) and shorten their service life.

[0071] The switching logic for this scenario is as follows: First, set a switching delay threshold (e.g., 30 seconds). If the interval between charging and discharging modes is less than 30 seconds, keep the current balancing mode unchanged to avoid repeatedly starting and stopping the balancing module in a short period of time. If the interval is greater than or equal to 30 seconds, perform a regular balancing mode switch based on the voltage difference, temperature range, charging and discharging stage, and / or aging stage to reduce the start and stop frequency of active balancing. For example, after active balancing is started, it needs to run continuously for at least 1 minute before stopping to avoid frequent start and stop of active balancing due to short-term voltage fluctuations, reduce device losses, and extend the service life of the balancing system.

[0072] Step 244: Develop a fault warning and emergency switchover strategy based on system security requirements, such as... Figure 5 As shown; When the battery pack presents safety risks (such as individual cell overvoltage, module overheating, or insulation failure), the balancing mode needs to shift from "optimizing performance" to "ensuring safety," adopting a low-risk balancing method, which includes the following steps: Step 2441: Develop a battery cell overvoltage warning (voltage > 4.18V, ternary lithium battery) equalization switching strategy; When an overvoltage is detected in a battery cell, the core requirement is to quickly reduce the overvoltage of the battery cell to avoid triggering the overvoltage protection (OVP) and causing the system to shut down. The emergency switching logic is as follows: immediately initiate passive balancing to force discharge the overvoltage battery cell, accelerating the voltage drop of the overvoltage battery cell. During this process, active balancing is disabled (to prevent energy from being transferred from the overvoltage battery cell to other battery cells, which could lead to new overvoltage risks). Once the overvoltage battery cell voltage drops below 4.1V, the normal balancing mode is restored, balancing safety and balancing efficiency.

[0073] Step 2442: Develop a balanced switching strategy for module overheating warning (temperature > 55℃); When the temperature sensor detects that the temperature of one of the modules is greater than 55°C (overheat warning threshold), the core requirement is to reduce the heat generated by the equalization module to prevent the temperature from rising further and causing the risk of thermal runaway.

[0074] The emergency switching logic is as follows: immediately switch from the current balancing mode to passive balancing (the heat generated by the active balancing module is usually 2-3 times that of passive balancing), and reduce the passive balancing current to 0.1-0.2A (to reduce the heat generated by the shunt resistor). If the temperature continues to rise to 60℃, all balancing operations are suspended, and only temperature and voltage monitoring is maintained. The battery temperature is reduced first through the heat dissipation system (such as fan, liquid cooling). Once the module temperature drops below 50℃, the normal balancing mode is restored to balance safety and balancing requirements.

[0075] Step 2443: Develop an insulation fault early warning (insulation resistance < 500Ω / V) equalization switching strategy; When the insulation resistance of the battery pack is detected to be <500Ω / V (insulation fault warning threshold) by the insulation monitoring module, the core requirement is to avoid the balancing current from exacerbating the insulation problem.

[0076] The emergency switching logic is as follows: disable active balancing (active balancing energy transfer process is prone to generating common-mode voltage, increasing the risk of leakage current), enable only passive balancing and reduce the balancing current to below 0.1A (minimize the impact of balancing current on the insulation system); at the same time, shorten the insulation resistance monitoring interval (from the usual 1 minute to 10 seconds) to track changes in insulation status in real time. If the insulation resistance further drops below 300Ω / V, immediately suspend all balancing operations and trigger a fault alarm to prompt the user to inspect the insulation system and ensure the safety of the battery system.

[0077] Step 25: Construct a priority strategy for comprehensive switching logic to achieve accurate adaptation under complex operating conditions; Constructing a priority-based switching system ensures that the balanced mode can adapt to performance requirements while guaranteeing system security under complex operating conditions. This involves the following steps: Step 251: Divide the equalization switching strategy priority layers. Based on the operating requirements of the battery system, divide the equalization switching strategy priority into three levels from high to low: First priority (safety first): emergency switching strategy for fault warning (such as battery cell overvoltage, module overheating, insulation fault warning equalization switching strategy). This level of strategy is not affected by other operating parameters. Once a safety warning is triggered, the corresponding emergency equalization switching strategy is executed immediately to prioritize the safety of the battery system.

[0078] Second priority (operating condition adaptation): extreme operating condition switching strategy (such as ultra-fast charging, high-rate discharge, and pulse charging and discharging scenario balanced switching strategy). This level of strategy is executed under the premise of no safety warning. If it conflicts with the fault warning emergency switching strategy, the fault warning emergency switching strategy shall prevail.

[0079] The third priority (aging and normal parameters): phased equalization switching strategy and normal hybrid equalization switching strategy. This level of strategy is executed under the premise of no safety warning and non-extreme operating conditions, to balance efficiency, energy consumption and battery life.

[0080] Step 252: Establish priority conflict handling rules; To avoid conflicts between different levels of strategies, the following conflict resolution rules are defined: When the first priority policy is triggered, regardless of the current operating condition, the second and third priority policies are immediately terminated, and the emergency switching policy of the first priority is executed. After the safety warning is lifted, the corresponding policies are restored in the order of "second priority → third priority".

[0081] When the second priority strategy is triggered, if the third priority strategy (such as ambient temperature passive balancing) is currently being executed, the third priority strategy is immediately terminated and the second priority strategy is executed; if the first priority strategy is triggered at this time, the strategy is immediately switched to the first priority strategy.

[0082] The third priority strategy is executed only when the first and second priority strategies are not triggered. If the number of abnormal battery cells does not exceed the threshold, the regular hybrid equalization switching strategy is executed. If the number of abnormal battery cells exceeds the threshold, the differential hybrid equalization switching strategy is executed (consisting of the phased equalization switching strategy, the extreme operating condition switching strategy, and the fault warning emergency switching strategy. When the first and second priorities are not triggered, the extreme operating condition switching strategy and the fault warning emergency switching strategy will not be triggered, only the phased equalization switching strategy will be triggered), prioritizing the correction of significant differences in battery cells.

[0083] Step 3: Determine the balancing adjustment strategy to be implemented based on the differences in individual battery cells, such as... Figure 6 As shown.

[0084] Step 3 can determine different balancing adjustment strategies to be implemented based on the differences in individual battery cells.

[0085] Step 31: Preset the threshold corresponding to the individual cell difference parameter and the threshold for the number of abnormal battery cells; Step 32: Compare the calculated individual cell difference parameters with the preset corresponding thresholds, and count the number of abnormal battery cells whose individual cell difference parameters exceed the thresholds. If the number of abnormal battery cells exceeds the corresponding threshold, start the difference-based hybrid equalization switching strategy. If the number of abnormal battery cells does not exceed the number threshold, start the regular hybrid equalization switching strategy.

[0086] Step 4: Block energy interference between active and passive balancing by hardware isolation, and establish the operational boundaries of active and passive balancing by logical interlocking, so that active and passive balancing are carried out separately.

[0087] The isolation mechanism between active and passive balancing is only not applicable to the balancing switching strategy in the high aging stage and the balancing switching strategy in the ultra-fast charging scenario. It is applicable to all other balancing scenarios. In the high aging stage and the ultra-fast charging scenario, active and passive balancing need to work together to achieve better overall battery control. In other stages, in order to avoid interference between active and passive balancing, active and passive balancing are generally carried out independently through the isolation mechanism.

[0088] Step 4 eliminates the mutual interference between active and passive balancing through the dual collaboration of hardware and software, from both physical circuits and logical rules, providing a stable balancing environment for battery prediction, ensuring the accuracy of prediction data, avoiding energy waste, and extending battery life.

[0089] Step 41: For the intra-group balancing scenario, two independent switches (S1, S2) are connected in series with the positive and negative terminals of each battery cell, corresponding to the passive balancing circuit and the active balancing circuit, respectively. The specific operation is as follows: When passive balancing needs to be initiated, the switch S1 of the corresponding battery cell is turned on to form an independent discharge circuit of "positive terminal of battery cell → S1 → shunt resistor → negative terminal of battery cell". When active balancing is running, S1 is forcibly turned off to prevent the shunt resistor from consuming the energy transferred by active balancing.

[0090] When active balancing needs to be initiated, the switch S2 of the corresponding battery cell is turned on to establish an energy transfer channel of "battery cell positive terminal → S2 → active balancing energy bus → battery cell negative terminal"; when passive balancing is running, S2 is forcibly turned off to prevent the active balancing circuit from being pulled down by the passive balancing resistor.

[0091] For inter-group balancing scenarios, a group-level energy isolation bus is set up in the active balancing unit between multiple battery subgroups, which is separated from the passive balancing circuit through an independent DC / DC converter.

[0092] Step 41 physically disconnects the energy interaction path between active and passive equalization to prevent problems such as current crosstalk and voltage interference, providing a hardware foundation for the independent operation of the two equalization technologies and avoiding equalization failure or energy loss caused by loop crossing.

[0093] Step 42: Logical interlocking between active and passive equilibrium is achieved through state interlocking and trigger condition exclusion.

[0094] Step 42 ensures that only one equalization technology is operating for the same battery cell / battery sub-pack at any given time, avoiding conflicts caused by the simultaneous activation of two equalization technologies. This step includes the following two sub-steps: Step 421: Establish interlocking logic between individual battery cells; Set a balancing status flag for each battery cell (0: no balancing, 1: passive balancing, 2: active balancing) and define interlocking rules, as follows: Passive balancing trigger and interlock: When a battery cell triggers passive balancing, if the flag bit is 0, passive balancing is started and the flag bit is switched to 1. At this time, active balancing is disabled.

[0095] Active balancing triggering and interlocking: When a battery cell triggers active balancing, if the flag bit is 0, active balancing is started and the flag bit is switched to 2. At this time, passive balancing is disabled.

[0096] Equalization switching control: If the operating parameters change during the equalization process (such as a sudden increase in the inter-group deviation during passive equalization), the current equalization status flag is first reset to 0, the corresponding switch is turned off (such as S1 being turned off), and a new equalization process is started after a 50ms delay (such as active equalization, with S2 turned on) to avoid simultaneous operation at the moment of switching.

[0097] Additional Functionality: By resetting the flag and delaying the switching, the system resolves the balance conflict problem during sudden changes in operating conditions, prevents energy interference caused by the brief conduction of the two balance circuits during switching, and ensures the smoothness and safety of the switching process.

[0098] Step 422: Priority scheduling of passive and active balancing based on sudden changes in complex operating conditions; When the same battery pack simultaneously meets the active and passive balancing trigger conditions, the execution order is determined through priority scheduling to avoid logical conflicts. The specific operation is as follows: Priority setting: Set active balancing to high priority and passive balancing to low priority. Passive balancing can only be activated when active balancing is not triggered.

[0099] Conflict handling: If active balancing is started, passive balancing will be paused immediately, and passive balancing trigger conditions will not be checked during active balancing to avoid frequent switching.

[0100] Step 5 incorporates battery balancing parameters (number of balancing attempts, balancing current, balancing time) as benchmark parameters into battery state prediction, eliminating the interference of balancing behavior on SOH (State of Health) and RUL (Relative Lifetime) predictions, thus improving prediction accuracy. Figure 7 As shown, the specific steps include: Step 51: Multi-dimensional raw data acquisition and synchronization. Real-time acquisition of battery operation data and equalization parameter data, ensuring timestamp alignment (sampling frequency ≥ 1Hz), to provide basic data for subsequent feature processing. This step includes the following sub-steps: Step 511: Determine the data acquisition range and frequency, set the data acquisition frequency to cover the complete charge and discharge cycle of the battery, and clarify that the data types to be acquired are divided into two categories: regular operating data and equalization benchmark parameter data, to avoid data omission.

[0101] Step 512: Collect routine operating data. Collect and store the voltage, current, and temperature data of individual battery cells using current sensors, voltage sensors, and temperature sensors.

[0102] Step 513: Collect equilibrium benchmark parameter data. Based on step 3, synchronously collect equilibrium benchmark parameter data for each equilibrium process and match it with the regular operation data by timestamp.

[0103] Step 514: Perform preliminary verification and storage of routine operating data and equalization benchmark parameters, remove abnormal data (such as current surges or invalid values ​​such as voltage zero caused by sensor failure), and store the verified data in layers according to the number of cycles and timestamps to support subsequent periodic retrieval.

[0104] Step 52: Quantize and preprocess the equalization reference parameters (equalization times, equalization current, equalization time); Step 53: Preprocess the routine operating data (filtering, noise reduction, etc.); Step 54: Construct a balanced-regular joint feature set, integrate the preprocessed balanced baseline parameters with regular operation features, filter effective features, and form a balanced-regular joint feature set; Step 55: Perform feature filtering on the balanced-conventional joint feature set; Calculate the correlation coefficients between each feature in the balanced-conventional joint feature set and the SOH and RUL values, and remove redundant features with an absolute value of correlation coefficient <0.3 (such as ambient temperature without fluctuations).

[0105] Step 56: Train and optimize the battery state prediction model using the balanced-conventional joint feature set as input; Step 57: Predict and output the SOH and RUL values ​​based on the battery state prediction model.

[0106] Based on the above technical solutions, the present invention provides the following embodiments: Example 1: 1.8C ultra-fast charging scenario for new energy passenger vehicles; In winter low-temperature scenarios (temperature < 0℃), when the voltage difference ΔV between battery sub-groups is greater than 0.08V, an active balancing mode is adopted to overcome the influence of increased internal resistance at low temperatures. When the electric vehicle is stationary, when the voltage difference ΔV between individual battery cells in the battery sub-group is less than 0.08V, a passive balancing mode is used.

[0107] Data from ultra-fast charging scenarios was collected for 7 consecutive days, including voltage, current, and temperature data every 100ms, and equalization counts and equalization current every minute, totaling over 100,000 data points. A battery state prediction model was built, and the above data was input into the model, with outputs being SOH and remaining charging time. After 500 rounds of iterative training, the loss function was reduced from 0.08 to 0.012. Ten ultra-fast charging tests at different temperatures (-5℃, 25℃, and 35℃) were selected, and the model parameters were adjusted. Ultimately, the SOH prediction deviation was less than 2%, and the remaining charging time error was less than ±3 minutes.

[0108] Example 2: High-aging scenario for low-temperature logistics vehicles; This embodiment is designed for aging logistics vehicles of cold chain logistics companies. The vehicles deliver goods from suburban warehouses to 10 fresh food stores in the city every day, with a one-way distance of 50km and two round trips per day. The refrigerated compartment is set to a temperature of -10℃ and the refrigeration unit is powered by batteries. By performing three consecutive charge-discharge cycles, the capacity deviation and internal resistance deviation of individual battery cells are tested. Some individual battery cells are identified as being in the high aging stage. When the battery pack is charged and discharged, the ΔV of the individual battery cells is tested. When ΔV > 0.07V, a low-current active balancing and passive balancing co-mode is adopted. When ΔV < 0.06V, passive balancing is completely switched to reduce the balancing current to 0.25A. Training of prediction model for high-aging battery cells: Historical data of 6 high-aging battery cells were collected to obtain 1200 sets of samples. A state prediction model for high-aging battery cells was constructed. The historical data was used as the input of the model and the output was the number of remaining cycles. Two high-aging battery cells were selected for charge and discharge tests. The model predicted the number of remaining cycles to be 180 and 210, respectively. The actual number of cycles to failure was 175 and 205, respectively. The error was less than 3%, which meets the operation and maintenance requirements.

[0109] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.

Claims

1. A BMS control method for battery prediction, characterized in that, Includes the following steps: Step 1: Obtain the state parameters of each battery cell in the battery pack, and analyze the differences between battery cells based on the state parameters of the battery cells; Step 2: Construct a hierarchical-domain hybrid balancing topology architecture, and build a conventional hybrid balancing switching strategy based on the hierarchical-domain hybrid balancing topology architecture; A differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of individual battery cells; Different equilibrium adjustment strategies are composed of conventional hybrid equilibrium switching strategies and differential hybrid equilibrium switching strategies; Step 3: Determine the balancing adjustment strategy to be implemented based on the differences between individual battery cells; Step 4: Block energy interference between active and passive balancing by hardware isolation, and establish the operational boundaries of active and passive balancing by logical interlocking, so that active and passive balancing are carried out separately. Step 5: Construct a battery state prediction model based on normal operating data and equalization benchmark parameters, and predict the battery state based on the battery state prediction model.

2. The BMS control method for battery prediction according to claim 1, characterized in that, Step 1 involves obtaining the state parameters of each individual battery cell in the battery pack, and analyzing the differences between the individual cells based on these state parameters. This includes the following steps: Step 11: Conduct identical charge and discharge test records for the battery pack; When the battery pack is charged and discharged under the same charging and discharging conditions for the same period of time, the detection is carried out and the data is recorded; Step 12: Analyze the voltage fluctuation curve based on the data obtained from multiple charge and discharge processes, and calculate the voltage difference parameters; During multiple charge and discharge cycles of the battery pack, voltage data of each battery cell is collected at fixed time intervals, voltage fluctuation curves are plotted, the collected voltage fluctuation curves are analyzed, voltage data collected at equal time intervals under conditions without equalization are obtained, and the average voltage of all battery cells is calculated. The voltage difference is calculated by subtracting the voltage data from the average voltage. The average voltage difference is calculated based on the voltage difference at all time points, and the average voltage difference is used as the voltage difference parameter of the battery cells. Step 13: Calculate the capacity difference parameters of individual battery cells based on the data obtained from multiple charge and discharge processes; Based on the charge and discharge capacity data recorded during multiple charge and discharge cycles of the battery pack, combined with the charge and discharge characteristics of the battery, the actual capacity of each battery cell at equal time intervals is obtained and the average actual capacity of all battery cells is calculated. The capacity difference is calculated by comparing the actual capacity of the battery cells with the average actual capacity. The average capacity difference is calculated based on the capacity difference at all time points and used as the capacity difference parameter of the battery cells. Step 14: Calculate the internal resistance difference parameter of the battery cells based on the data obtained from multiple charge and discharge processes; During the charging and discharging process of the battery pack, the internal resistance value of each battery cell is monitored and recorded in real time. The internal resistance value of each battery cell at equal time intervals is obtained and the average internal resistance value of all battery cells is calculated. The difference between the internal resistance value data and the average internal resistance value is obtained. The average resistance difference is calculated based on the internal resistance difference measured at all time points. The average internal resistance difference is used as the internal resistance difference parameter of the battery cells. Step 15: Calculate the individual unit difference parameters based on voltage difference parameters, capacity difference parameters, and internal resistance difference parameters; Based on voltage difference parameters, capacity difference parameters, and internal resistance difference parameters, the individual cell difference parameters are calculated using a preset weighting algorithm.

3. The BMS control method for battery prediction according to claim 2, characterized in that, In step 15, the individual cell difference parameters of the battery cells are calculated using a preset weight allocation algorithm, specifically including the following steps: Step 151: Determine the basic weight range; Step 152: Adjust the weights according to the battery type; Step 153: Adjust the weights a second time based on the application scenario; Step 154: Based on the adjusted weights above, the specific calculation steps for the individual cell difference parameters of the battery cells are as follows: Obtain the voltage difference parameters, capacity difference parameters, and internal resistance difference parameters of individual battery cells; The voltage difference parameter, capacity difference parameter, and internal resistance difference parameter are all normalized. The individual cell difference parameters are obtained by weighting the normalized voltage difference parameters, capacity difference parameters, and internal resistance difference parameters.

4. The BMS control method for battery prediction according to claim 3, characterized in that, In step 2, a hierarchical-domain hybrid load balancer topology is constructed, and a conventional hybrid load balancer switching strategy is built based on the hierarchical-domain hybrid load balancer topology architecture. This specifically includes the following steps: Step 21: Establish a hierarchical balanced topology between and within groups; The battery pack is divided into battery sub-groups, and inter-group active balancing units are deployed between different battery sub-groups, while intra-group passive balancing units are deployed between battery cells in the same battery sub-group. Step 22: Construct a domain-balanced topology; Divide the temperature zones, deploy temperature monitoring units, and formulate a temperature zone regional balance switching strategy; Step 23: Construct a fuzzy control algorithm to adaptively switch the equalization mode based on voltage difference amplitude, temperature range, and charging / discharging stage, so as to formulate a conventional hybrid equalization switching strategy.

5. The BMS control method for battery prediction according to claim 4, characterized in that, In step 23, a fuzzy control algorithm is constructed to adaptively switch the equalization mode based on the voltage difference amplitude, temperature range, and charging / discharging stage, in order to formulate a conventional hybrid equalization switching strategy, including the following sub-steps: Step 231, divide the charging and discharging phase scenarios; Based on the charging and discharging current and SOC value of the battery pack, the working process of the battery pack is divided into different charging and discharging stages: fast charging stage, float charging stage, discharging stage, and resting stage. Step 232: Collect operating parameters in real time and make judgments; The system collects operating parameters in real time and transmits them to the BMS main control unit. The BMS main control unit processes and analyzes the operating parameters in real time to determine the current charging and discharging stage of the battery pack. Step 233: Combining the hierarchical equalization topology and temperature-regional equalization switching strategy, formulate the equalization mode switching logic for the conventional hybrid equalization switching strategy: During the fast charging phase, when a voltage difference greater than 0.05V is detected between battery sub-groups, an active balancing mode between sub-groups is adopted. During the float charging phase, when the voltage difference between individual cells in the battery sub-group is detected to be less than 0.05V, a passive balancing mode is adopted. During the low-temperature discharge stage, when the voltage difference between battery sub-groups is detected to be greater than 0.08V, an active balancing mode is adopted. During the room temperature resting stage, when the voltage difference between battery sub-groups is less than 0.08V, a passive balancing mode is adopted.

6. The BMS control method for battery prediction according to claim 5, characterized in that, In step 2, a differentiated hybrid equalization switching strategy is constructed based on the aging degree and operating conditions of individual battery cells. This strategy consists of a conventional hybrid equalization switching strategy and a differentiated hybrid equalization switching strategy, forming different equalization adjustment strategies. The steps include: Step 24 involves constructing a differentiated hybrid equalization switching strategy based on the aging level and operating conditions of individual battery cells. This strategy includes the following sub-steps: Step 241: Analyze and determine the degree of battery aging; Step 242: Adjust the phased equalization switching strategy according to the degree of battery aging; The phased equalization switching strategy includes a new battery / low aging phase equalization switching strategy, a medium aging phase equalization switching strategy, and a high aging phase equalization switching strategy. Step 243: Develop a switching strategy for extreme operating conditions to address fluctuations in charge / discharge rates; The extreme operating condition switching strategies include an ultra-fast charging scenario balanced switching strategy, a high-rate discharge scenario balanced switching strategy, and a pulse charging and discharging scenario balanced switching strategy. Step 244: Develop a fault warning and emergency switching strategy based on system security requirements; The fault warning emergency switching strategy includes a battery cell overvoltage warning equalization switching strategy, a module overheat warning equalization switching strategy, and an insulation fault warning equalization switching strategy.

7. The BMS control method for battery prediction according to claim 6, characterized in that, Step 2 involves constructing differentiated hybrid equalization switching strategies based on the aging level and operating conditions of individual battery cells. These strategies consist of a conventional hybrid equalization switching strategy and a differentiated hybrid equalization switching strategy, forming different equalization adjustment strategies. The steps also include: Step 25: Construct a priority strategy for the load balancing switching strategy, which includes the following sub-steps: Step 251: Divide the priority layers of the balanced switching strategy; Based on the battery pack's operational requirements, the balancing switching strategy is divided into three levels of priority from high to low: First priority: Fault warning and emergency switchover strategy; Second priority: Extreme operating condition switching strategy; Third priority: phased equilibrium switching strategy, regular hybrid equilibrium switching strategy; Step 252, define priority conflict handling rules: When the first priority policy is triggered, the second and third priority policies are immediately terminated, and the first priority balanced switching policy is executed. When the second priority policy is triggered, if the third priority policy is currently being executed, the third priority policy is immediately terminated and the second priority policy is executed. If the first priority policy is triggered at this time, the policy is immediately switched to the first priority policy. The third priority strategy is only executed when the first and second priority strategies are not triggered. If the number of abnormal battery cells does not exceed the threshold, the regular hybrid equalization switching strategy is executed; if the number of abnormal battery cells exceeds the threshold, the differential hybrid equalization switching strategy is executed.

8. The BMS control method for battery prediction according to claim 7, characterized in that, Step 3: Determine the balancing adjustment strategy to be implemented based on the differences between individual battery cells, including the following steps: Step 31: Preset the threshold corresponding to the individual cell difference parameter and the threshold for the number of abnormal battery cells; Step 32: Compare the calculated individual cell difference parameters with the preset corresponding thresholds, and count the number of abnormal battery cells whose individual cell difference parameters exceed the thresholds. If the number of abnormal battery cells exceeds the corresponding threshold, a differential hybrid equalization switching strategy will be initiated. If the number of abnormal battery cells does not exceed the quantity threshold, a regular hybrid equalization switching strategy will be initiated.

9. The BMS control method for battery prediction according to claim 8, characterized in that, In step 4, energy interference between active and passive balancing is blocked by hardware isolation, and the operational boundaries of active and passive balancing are established by logical interlocking, so that active and passive balancing are performed separately. This includes the following steps: Step 41: Connect two independent switches in series at the positive and negative terminals of each battery cell, corresponding to the passive balancing circuit and the active balancing circuit, respectively. Step 42, achieving logical interlocking between active and passive equilibrium through state interlocking and trigger condition exclusion, includes the following two sub-steps: Step 421: Establish interlocking logic between individual battery cells; Set an equalization status flag for each battery cell and define interlocking rules; Step 422: Priority scheduling of passive and active balancing based on sudden changes in complex operating conditions.

10. The BMS control method for battery prediction according to claim 9, characterized in that, In step 5, a battery state prediction model is constructed based on normal operating data and equalization benchmark parameters. The battery state is then predicted based on the battery state prediction model, specifically including the following steps: Step 51: Multi-dimensional raw data collection and synchronization; Real-time acquisition of battery operation data and balancing parameter data, ensuring timestamp alignment, includes the following sub-steps: Step 511: Determine the data collection range and frequency; Set the data collection frequency and clearly define the types of data to be collected, dividing them into two categories: routine operating data and balanced baseline parameter data; Step 512: Collect routine operating data; The system collects and stores data on the voltage, current, and temperature of individual battery cells using current sensors, voltage sensors, and temperature sensors. Step 513: Collect equalization benchmark parameter data; Collect the baseline parameters for each balancing process and match them with the regular operation data by timestamp; Step 514: Perform preliminary verification and storage of routine operating data and equilibrium benchmark parameters; Verify and remove abnormal data, and store the verified data in layers; Step 52: Quantize and preprocess the equilibrium benchmark parameters; Step 53: Preprocess the routine operating data; Step 54: Construct a balanced-regular joint feature set; By integrating the preprocessed equilibrium benchmark parameters with regular operating characteristics, effective features are selected to form a joint equilibrium-regular feature set. Step 55: Perform feature filtering on the balanced-conventional joint feature set; Calculate the correlation coefficients between each feature in the balanced-conventional joint feature set and the SOH and RUL values, and remove redundant features whose absolute values ​​of the correlation coefficients are lower than the corresponding thresholds; Step 56: Train and optimize the battery state prediction model using the balanced-conventional joint feature set as input; Step 57: Predict and output the SOH and RUL values ​​based on the battery state prediction model.

Images