Multi-level monitoring method integrating EMS and battery management system

Abstract

The embodiment of the invention provides a multi-level monitoring method integrating an EMS and a battery management system. The multi-level monitoring method comprises the steps that voltage data of all monomers in a battery pack are collected, and the voltage difference degree is calculated; determining an inconsistency grade of the battery pack according to the voltage difference degree; adjusting equalization charging strategy parameters based on the inconsistency level; and transmitting the optimized equalization charging strategy to a battery management system for execution. Through the scheme of the embodiment of the invention, the equalization charging strategy can be regulated and controlled according to the difference degree of the voltage of each single body in the battery pack so as to solve the problem that the capacity fading is too fast due to the inconsistency of the battery cells in a high-density energy storage scene.

Description

Technical Field

[0001] This application relates to the field of power system monitoring technology, specifically to a multi-level monitoring method integrating EMS and battery management system. Background Technology

[0002] In high-density energy storage scenarios, inconsistencies inevitably arise among cells in large-scale battery packs due to factors such as manufacturing processes, operating environments, and cyclic aging. These inconsistencies are mainly manifested in differences in individual cell voltage, internal resistance, and state of charge (SOC). If these inconsistencies are not effectively managed, they will exacerbate energy imbalances between batteries during charging and discharging, leading to overcharging or over-discharging of some cells. This accelerates the overall battery pack's capacity decay, shortens its lifespan, and poses safety hazards. Currently, traditional battery management systems (BMS) mostly adopt passive balancing strategies based on fixed thresholds or simple active balancing strategies. Their balancing judgments often rely solely on voltage as a single parameter, and the strategies are rigid. They fail to comprehensively consider multi-dimensional real-time information such as temperature, load fluctuations, electrochemical impedance, SOC differences, and battery aging evolution. Furthermore, they lack coordinated control with energy management systems (EMS). As a result, in highly dynamic and complex actual operating conditions, the balancing efficiency is low, the response is lagging, and it is difficult to effectively suppress the malignant development of inconsistencies. This fails to meet the core requirements of high-density energy storage systems for long-term maintenance of battery health and safe and economical system operation. Summary of the Invention

[0003] In view of this, the present disclosure provides a multi-level monitoring method integrating EMS and battery management system, which at least partially solves the problems existing in the prior art.

[0004] A multi-level monitoring method integrating EMS and battery management system includes: Collect voltage data of each cell in the battery pack and calculate the voltage difference. The inconsistency level of the battery pack is determined based on the voltage difference. Adjust the equalization charging strategy parameters based on the aforementioned inconsistency level; The optimized equalization charging strategy is transmitted to the battery management system for execution.

[0005] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Get the current cell voltage data V_i (i=1,2,...,n) in the current battery pack; Calculate the average voltage V_avg = (ΣV_i) / n; Calculate the individual unit voltage difference ΔV = max(V_i)-min(V_i); If ΔV > Threshold1, execute balancing strategy A; otherwise, execute balancing strategy B. Here, strategy A is an active balancing strategy and strategy B is a passive balancing strategy. ΔV is the difference between the maximum and minimum voltage values, and Threshold1 is the set threshold for balancing to start, used to determine whether active balancing is required.

[0006] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Collect the battery operating environment temperature T_env; Establish a temperature-dependent dynamic threshold mechanism, i.e., Threshold = Threshold0. (T_env / T_ref); Calculate the individual unit voltage difference ΔV; If ΔV > Threshold, enable fast balancing mode; otherwise, maintain slow balancing mode. Threshold0 is the threshold at the reference temperature, T_ref is the standard temperature, and Threshold is the equilibrium judgment threshold that is dynamically adjusted with temperature to improve adaptability under different operating conditions.

[0007] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Analyze the current fluctuations (I_fluc) generated by the battery pack during discharge; Calculate the inconsistency coefficient IC = (ΔV + α·I_fluc) / V_max; Fuzzy logic is used to evaluate ICs and select corresponding strategies; When IC ≤ IC_min, a no-action policy is executed; when IC > IC_min and ≤ IC_med, a low-intensity equilibrium is executed; when IC > IC_med, a high-intensity equilibrium is executed. α is the current fluctuation weighting coefficient, I_fluc is the current discharge current fluctuation amplitude, and V_max is the maximum single-cell voltage. The equilibrium decision is optimized by combining the changes in charge state.

[0008] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Introduce the remaining capacity estimate SOC_i; According to the monomer SOC difference ΔSOC = max(SOC_i)-min(SOC_i); The design weighting coefficient β = ΔV / ΔSOC; Set strategy priority: If ΔSOC > η, execute the SOC-priority equalization strategy; otherwise, execute the voltage-only equalization strategy. β is a weighting factor for the degree of charge imbalance, and η is a preset SOC inconsistency threshold. The accuracy of the strategy is improved by integrating charge state information.

[0009] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Detect for abnormal signals in the battery management system; Construct a conditional judgment module that includes fault diagnosis; Define the anomaly determination formula: Error = K1·ΔV + K2·ΔT, where K1 and K2 are correction coefficients; If Error > Error_threshold, pause the current load balancing and perform maintenance diagnostics. ΔT is the temperature difference within the battery pack, and K1 and K2 represent the influence coefficients of voltage and temperature difference, respectively. Safety control is achieved through error indicators.

[0010] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: The voltage difference is divided into three categories: low, medium, and high; where low voltage difference is ≤1V, medium voltage difference is >0.1V and ≤0.3V, and high voltage difference is >0.3V. These correspond to three balancing strategies: low intensity, medium intensity, and high frequency. The aging compensation function is set according to the degree of battery aging, such as λ = 1 + k (age / τ); Where age is the number of battery cycles, and τ is the preset aging period, used to dynamically adjust the balancing force. λ represents the balance strength adjustment coefficient for aging, used to adapt to the performance degradation of the battery cell under different life cycles.

[0011] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Considering the impact of system load fluctuations on load balancing, the load fluctuation factor f_Loader = |P_prev-P_curr| / P_avg is calculated. By combining this factor with the voltage difference ΔV, a comprehensive criterion is established: Score = ΔV × f_Loader ×γ, where γ is the equalization efficiency adjustment coefficient; Priority is set based on Score; If Score < Low_thres, the equalization process is skipped to prevent energy waste; if Score ≥ High_thres, equalization control is initiated. γ is a weight parameter coupled with the system's operating state, which enables the equalization control to have dynamic adaptability and is suitable for dynamic power consumption scenarios in high-density energy storage stations.

[0012] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Extracting the electrochemical impedance spectral characteristics Z; The consistency index CI is calculated based on the distribution trend of Z: CI = |(Z_high-Z_low)| / Z_mid. A comprehensive criterion is set based on the voltage difference ΔV; If CI > CI_threshold or ΔV > ΔV_threshold, switch to active balancing mode; otherwise, use passive balancing mode. Z_high and Z_low represent impedance values ​​in the high / low frequency bands, while CI serves as an indicator for evaluating the internal electrochemical consistency of the system, enabling earlier detection of potential cell failure risks.

[0013] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Set the battery pack capacity consistency index CI_cap = |(C_max-C_min)| / C_avg; A composite decision mechanism is established by combining CI_cap and ΔV: if ΔV > V_Thr AND CI_cap > C_Thr, then full equilibrium is forcibly initiated; If CI_cap is not met but ΔV has reached the threshold, then perform targeted load balancing. Otherwise, continue to observe the current running status, where C_max, C_min, and C_avg are the maximum, minimum, and average cell capacities, respectively, used to determine the overall energy storage consistency of the battery pack.

[0014] Preferably, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Analyze the evolution trend of battery inconsistency using historical charge-discharge curves; The trend coefficient ω(t) is established based on the time series as ω(t) = a·ΔV(t) + b·ΔSOC(t); Combine ω(t) for rolling prediction and set a dynamic adjustment mechanism; If ω(t) > ω_Thr, then the adaptive equilibrium control strategy is executed; otherwise, the original scheme is maintained. a and b are weighting parameters, and ΔV(t) and ΔSOC(t) are the voltage and SOC differences at different times, used to predict and prevent future changes in consistency.

[0015] This disclosure provides a multi-level monitoring method integrating an EMS and a battery management system, comprising: collecting voltage data of each cell within a battery pack and calculating the voltage difference; determining the inconsistency level of the battery pack based on the voltage difference; adjusting the equalization charging strategy parameters based on the inconsistency level; and transmitting the optimized equalization charging strategy to the battery management system for execution. The solution provided by this disclosure addresses the problem of rapid capacity decay caused by cell inconsistency in high-density energy storage scenarios by adjusting the equalization charging strategy based on the voltage difference of each cell within the battery pack. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the exemplary embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this disclosure and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a multi-level monitoring method that integrates EMS and battery management systems. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the embodiments of this disclosure will be further described in detail below with reference to the accompanying drawings. The illustrative implementation methods and descriptions of the embodiments of this disclosure are only used to explain the embodiments of this disclosure and are not intended to limit the embodiments of this disclosure.

[0019] This implementation uses a large-scale battery energy storage power station (BESS) as an application scenario. The system consists of multiple battery clusters, each containing a large number of cells connected in series and parallel. Next, refer to... Figure 1 This describes the steps of a multi-level monitoring method integrating an EMS and a battery management system according to the present invention.

[0020] Step 1: Collect voltage data of each cell in the battery pack and calculate the voltage difference.

[0021] In practice, this step involves using a high-precision voltage acquisition circuit (such as a dedicated AFE chip, like TI's BQ79616) deployed within the battery module to acquire the voltage data of each individual cell in the battery pack in real time. The acquisition frequency can be dynamically adjusted according to operating conditions, set to once per second (1Hz) during active charging or discharging periods, and reduced to once every 10 seconds during rest periods to balance data accuracy and system power consumption. The acquired raw voltage data is uploaded to the slave control unit (SBC) of the Battery Management System (BMS) via an isolated CAN bus or daisy-chain communication. The BMS master control unit (BMU) aggregates and preprocesses the data from all slave control units, including filtering and noise reduction, invalid value removal, etc., to form a set of real-time, reliable individual cell voltage datasets {V1, V2, ..., Vn}, where n is the total number of individual cells in the battery pack.

[0022] Subsequently, the BMU calculates the voltage difference (ΔV), a core indicator characterizing battery pack consistency. In this embodiment, ΔV is calculated using the range method, i.e., ΔV = max(V_i) - min(V_i), where max(V_i) and min(V_i) are the maximum and minimum values ​​of all cell voltages within the current sampling period, respectively. This indicator directly reflects the dispersion between the highest and lowest voltage cells in the battery pack and is the most direct parameter for measuring inconsistency.

[0023] Step 2: Determine the inconsistency level of the battery pack based on the voltage difference.

[0024] This step quantifies and classifies the real-time ΔV calculated in Step 1. An inconsistency level mapping table is preset in the EMS (Energy Management System). This mapping table sets multiple threshold ranges based on the battery's electrochemical characteristics, historical operating data, and safety specifications. For example: Level 0 (Normal): ΔV ≤ 50mV. The battery pack is considered to have good consistency and requires no special attention.

[0025] Level 1 (Slight Inconsistency): 50mV < ΔV ≤ 150mV. Initial differentiation is emerging in the battery packs; trends need to be monitored.

[0026] Level 2 (Moderate Inconsistency): 150mV < ΔV ≤ 300mV. The inconsistency is quite significant, and balancing intervention needs to be initiated.

[0027] Level 3 (Severe Inconsistency): ΔV > 300mV. The inconsistency is severe, requiring immediate and forceful equilibration and an early warning.

[0028] The BMU uploads the calculated ΔV value along with a timestamp to the EMS. Upon receiving this data, the monitoring and decision-making module in the EMS immediately queries the aforementioned mapping table, classifies the current ΔV value into the corresponding inconsistency level (e.g., level 2), and attaches a status label to this level.

[0029] Step 3: Adjust the equalization charging strategy parameters based on the inconsistency level.

[0030] This step is used to achieve intelligent control. Based on the inconsistency level determined in step two, the EMS dynamically adjusts the key parameters in the upcoming equilibrium control strategy. These parameters are not fixed but dynamically correlated with the inconsistency level: If it is Level 1, EMS may select the "Observation and Early Warning" mode, with the policy parameters set as follows: the balancing function remains on standby, but the inconsistency monitoring cycle is reduced, and this status is recorded in the log to prompt maintenance personnel to pay attention.

[0031] If it is Level 2, EMS will activate the "Standard Active Balancing" mode. The strategy parameters will be adjusted as follows: activate the active balancing circuit (such as energy transfer balancing based on capacitors or inductors), set the balancing target to reduce ΔV to below 100mV, set the balancing current to 0.5A (Amperes), and tighten the balancing start / stop voltage threshold.

[0032] If it is Level 3, EMS will immediately activate the "Enhanced Balancing and Protection" mode. The strategy parameters are significantly enhanced: the balancing current is increased to 1A or higher (within the allowable range of heat dissipation) to reduce the voltage difference as quickly as possible; at the same time, EMS may adjust the charging and discharging plan of the battery cluster, such as temporarily reducing its maximum allowable charging current, to prevent the risk of overcharging of the cell with the highest voltage before balancing is completed.

[0033] Step 4: Transmit the optimized equalization charging strategy to the battery management system for execution.

[0034] After the EMS generates an optimized equalization strategy containing specific parameter instructions, it sends it as a control command downlink to the corresponding BMS master control unit (BMU) via the power plant's internal high-speed communication network (such as Ethernet or industrial Ethernet). The BMU parses and performs security checks on the command, then converts it into specific hardware drive signals and sends them to the slave control units (SBCs) responsible for executing the equalization operation. Based on the parameters in the command (such as equalization current value, target cell address, etc.), the SBC precisely controls the power MOSFET switches, activates the corresponding equalization resistors or active equalization circuits, and performs discharge or energy transfer operations on the target cell (usually the one with excessive voltage).

[0035] Meanwhile, the BMU continuously monitors the changes in ΔV during execution and feeds back the execution results (such as the rate of ΔV decrease and the current voltage value) to the EMS in real time. Based on the feedback information, the EMS can dynamically fine-tune the strategy parameters or determine whether the target has been achieved and terminate the equalization, forming a closed-loop control loop of "monitoring-evaluation-decision-execution-feedback".

[0036] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Get the current cell voltage data V_i (i=1,2,...,n) in the current battery pack; Calculate the average voltage V_avg = (ΣV_i) / n; Calculate the individual unit voltage difference ΔV = max(V_i)-min(V_i); If ΔV > Threshold1, execute equilibrium strategy A; otherwise, execute equilibrium strategy B, where... ΔV is the difference between the maximum and minimum voltage values, and Threshold1 is the set threshold for balancing to start, used to determine whether active balancing is required.

[0037] Specifically, the BMS accurately obtains the voltage value V_i of each individual cell from the collected real-time cell voltage dataset and calculates its arithmetic mean V_avg, which serves as the benchmark for the current battery pack voltage level. The core step is to calculate the absolute difference ΔV in the cell voltage, i.e., ΔV = max(V_i) - min(V_i). This ΔV value directly quantifies the instantaneous voltage difference between the "best" and "worst" cells within the group.

[0038] The system then compares the ΔV value with a pre-defined, experimentally calibrated equilibrium initiation threshold (Threshold1). This threshold is the key technical boundary distinguishing between "acceptable natural fluctuation" and "imbalance requiring manual intervention." For example, for lithium iron phosphate batteries, Threshold1 might be set at 150mV. The comparison logic is a binary decision: If ΔV > Threshold1: The system determines that the battery pack inconsistency has reached a level requiring intervention, and at this time, balancing strategy A will be executed. Strategy A usually refers to "active balancing" or "strong balancing mode", which is characterized by activating the energy transfer circuit (such as capacitive or inductive type) to transfer charge from high-voltage cells to low-voltage cells or the system bus with a higher balancing current (e.g., 0.5A-1A), aiming to quickly and actively reduce the voltage difference.

[0039] Otherwise (ΔV ≤ Threshold1): The system determines that the current inconsistency is within the normal fluctuation range, and then executes balancing strategy B. Strategy B usually refers to "passive balancing" (bleeder resistor balancing) or "balancing standby / low current balancing" mode. Under passive balancing, only cells with excessively high voltage are discharged through parallel resistors, which is less efficient but has a simpler circuit; in standby mode, balancing may be temporarily turned off to save energy.

[0040] This implementation transforms the judgment of voltage inconsistency from a qualitative level to a quantitative comparison with a fixed threshold, establishing a clear and definite equalization action trigger point (Threshold1). This "if-else" logic is simple, reliable, and easy to stably implement in the BMS microcontroller, ensuring the timeliness and necessity of the system's response to severe voltage deviations. It solves the problems of hesitation and lag in equalization actions caused by the absence of a threshold or ambiguity of the threshold, providing a basic and effective safety protection mechanism for the battery pack, preventing the voltage difference from continuing to expand without intervention, thus preventing the risk of overcharging or over-discharging.

[0041] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Collect the battery operating environment temperature T_env; Establish a temperature-dependent dynamic threshold mechanism, i.e., Threshold = Threshold0. (T_env / T_ref); Calculate the individual unit voltage difference ΔV; If ΔV > Threshold, enable fast balancing mode; otherwise, maintain slow balancing mode. Threshold0 is the threshold at the reference temperature, T_ref is the standard temperature (e.g., 25°C), and Threshold is the equilibrium judgment threshold that is dynamically adjusted with temperature to improve adaptability under different operating conditions.

[0042] Building upon the previous embodiment, this example introduces a dynamic threshold adjustment mechanism to address the significant impact of temperature on battery performance and equalization. This makes the equalization trigger determination more environmentally adaptable. In practice, the system not only collects voltage but also acquires the representative temperature T_env of the battery's operating environment in real time through a network of temperature sensors distributed within the battery module. Simultaneously, the system stores a baseline threshold Threshold0 (e.g., 150mV) calibrated at a standard laboratory temperature (e.g., T_ref = 25°C).

[0043] The key improvement lies in the fact that the system no longer directly uses a fixed Threshold1, but instead dynamically calculates the current effective threshold Threshold based on the real-time temperature. The calculation formula is: Threshold = Threshold0 (T_env / T_ref). This formula reflects the basic principle of temperature compensation: when the ambient temperature T_env is lower than the standard temperature T_ref, the calculated Threshold will decrease accordingly. This is because at low temperatures, the battery's internal resistance increases and polarization becomes more pronounced. Even the same ΔV may indicate more severe electrochemical inconsistencies, thus requiring more sensitive equilibrium triggering conditions. Conversely, at high temperatures, the threshold can be appropriately relaxed to avoid unnecessary frequent equilibrium adjustments.

[0044] The decision-making logic has been upgraded accordingly: If ΔV > the dynamically calculated Threshold, it is determined that immediate and forceful intervention is required, and the system activates the fast balancing mode. This mode may increase the balancing current or adopt a more efficient active balancing topology to quickly suppress inconsistencies exacerbated by temperature effects.

[0045] Otherwise: The system remains in slow balancing mode or in standby. Slow balancing mode may employ low-current passive balancing, used only to eliminate minor voltage drift.

[0046] This embodiment establishes a dynamic correlation between temperature and the equalization threshold, significantly improving the system's adaptability and robustness under different climatic conditions and operating conditions. It overcomes the shortcomings of fixed-threshold schemes, which may react sluggishly at low temperatures and overreact at high temperatures. This allows the equalization strategy to be "sensitive to temperature," preventing inconsistency deterioration in harsh temperature environments and avoiding excessive intervention at suitable temperatures. Consequently, it optimizes the overall performance and lifespan of the battery pack over a wider temperature range, particularly improving the reliability of the energy storage system in complex environments such as outdoor environments and those with large diurnal temperature variations.

[0047] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Analyze the current fluctuations (I_fluc) generated by the battery pack during discharge; Calculate the inconsistency coefficient IC = (ΔV + α·I_fluc) / V_max; Fuzzy logic is used to evaluate ICs and select corresponding strategies; When IC ≤ IC_min, a no-action policy is executed; when IC > IC_min and ≤ IC_med, a low-intensity equilibrium is executed; when IC > IC_med, a high-intensity equilibrium is executed. α is the current fluctuation weighting coefficient, I_fluc is the current discharge current fluctuation amplitude, and V_max is the maximum single-cell voltage. The equilibrium decision is optimized by combining the changes in charge state.

[0048] Specifically, while calculating the voltage range ΔV, the system uses current sensors in the BMS to monitor the instantaneous current value of the battery pack during discharge (or charging) in real time, and calculates its fluctuation amplitude I_fluc within a certain time window (e.g., calculating the standard deviation or peak-to-peak value of the current). Next, the system innovatively constructs an inconsistency coefficient IC, calculated as: IC = (ΔV + α·I_fluc) / V_max. Here, α is a pre-calibrated current fluctuation weighting coefficient (e.g., 0.1) used to adjust the weight of current fluctuation factors in the overall evaluation; V_max is the current maximum single-cell voltage, used for normalization, making IC a dimensionless relative indicator.

[0049] Then, the system uses a fuzzy logic controller to evaluate continuous IC values. The fuzzy logic divides the IC input domain into linguistic variables such as "small," "medium," and "large," and outputs the corresponding equilibrium strategy selection based on an expert knowledge base. To simplify implementation, explicit thresholds can also be set for segmentation. When IC ≤ IC_min (e.g., 0.05): the system considers the inconsistency to be very minor, executes a no-action strategy, and saves energy.

[0050] When IC_min < IC ≤ IC_med (e.g., 0.15): the system considers there to be a certain degree of inconsistency and performs low-intensity equalization, such as low-current passive equalization.

[0051] When IC > IC_med: The system judges that there is a significant inconsistency and performs high-intensity equalization, that is, it starts high-current active equalization.

[0052] This embodiment achieves a leap from "static inconsistency" to "dynamic operating condition inconsistency" by introducing discharge current fluctuation I_fluc and constructing a comprehensive inconsistency coefficient IC. The difference in cell response under load fluctuations is a key driving force leading to state differentiation. This method can capture earlier and more sensitive synergy problems caused by differences in internal resistance or polarization characteristics that may be masked under stable voltage. Combined with fuzzy logic evaluation, it enhances the system's ability to handle complex and nonlinear relationships, making the equilibrium decision closer to the actual electrochemical state, thereby implementing more precise and forward-looking equilibrium intervention and effectively slowing down the rate of capacity decay under dynamic loads.

[0053] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Introduce the remaining capacity estimate SOC_i; According to the monomer SOC difference ΔSOC = max(SOC_i)-min(SOC_i); The design weighting coefficient β = ΔV / ΔSOC; Set strategy priority: If ΔSOC > η, execute the SOC-priority equalization strategy; otherwise, execute the voltage-only equalization strategy. β is a weighting factor for the degree of charge imbalance, and η is a preset SOC inconsistency threshold. The accuracy of the strategy is improved by integrating charge state information.

[0054] Specifically, the process involves the following steps: First, the BMS uses the ampere-hour integral method combined with model estimation to calculate the estimated remaining capacity SOC_i for each cell in real time. Next, the SOC difference between cells is calculated as ΔSOC = max(SOC_i) - min(SOC_i). Then, a key weighting coefficient β = ΔV / ΔSOC is designed. The magnitude of β reveals the degree of correlation between voltage and capacity differences.

[0055] Based on β and ΔSOC, the system performs hierarchical decision-making: First priority judgment: If ΔSOC > η (η is a preset SOC inconsistency threshold, such as 5%), it indicates that there is a significant "energy inconsistency" in the battery pack. In this case, regardless of the voltage difference, the system executes the SOC priority equalization strategy. The goal of this strategy is to balance the state of charge of each cell, and special charging / discharging pulses may be used to calibrate and converge the SOC.

[0056] Second priority judgment: If ΔSOC ≤ η, it indicates that the energy consistency is acceptable, and the system executes the voltage equalization strategy. However, the voltage equalization strategy here is affected by the β value: if the β value is too large, it may mean that the voltage difference is mainly caused by internal resistance or contact resistance, rather than the SOC difference, and the system will adjust the equalization strength accordingly.

[0057] This embodiment, by introducing a State of Charge (SOC) difference ΔSOC and designing a weighting coefficient β, achieves a shift in the balancing objective from "appearance consistency" (consistent voltage) to "essential consistency" (consistent energy). It effectively avoids the ineffective or even harmful balancing operation ("over-balancing") that occurs in traditional methods when voltage differences due to internal cell polarization or contact resistance variations exist, even when the actual SOCs are similar. Conversely, when SOC differences are large but voltages converge due to a plateau, balancing is triggered promptly. This significantly improves the accuracy and inherent effectiveness of the balancing strategy, truly mitigating capacity decay caused by inconsistent charge at its root, and enhancing the battery pack's usable capacity and cycle life.

[0058] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Detect for abnormal signals in the battery management system; Construct a conditional judgment module that includes fault diagnosis; Define the anomaly determination formula: Error = K1·ΔV + K2·ΔT, where K1 and K2 are correction coefficients; If Error > Error_threshold, pause the current load balancing and perform maintenance diagnostics. ΔT is the temperature difference within the battery pack, and K1 and K2 represent the influence coefficients of voltage and temperature difference, respectively. Safety control is achieved through error indicators.

[0059] This implementation deeply integrates safety boundaries and fault diagnosis into the equalization control logic, aiming to ensure that any equalization operation is performed within the system's safety tolerance and to prevent the equalization process itself from causing or exacerbating faults.

[0060] During implementation, the system continuously monitors various abnormal signals within the BMS and the battery itself, including but not limited to: voltage acquisition line breakage, temperature sensor failure, insulation fault, and communication timeout. These signals are input into a dedicated condition judgment module.

[0061] The core of this module is to define a comprehensive anomaly determination formula: Error = K1·ΔV + K2·ΔT. Where ΔT is the maximum temperature difference within the battery pack (max(T_i) - min(T_i)); K1 and K2 are correction coefficients determined based on the battery thermal runaway model and safety specifications, used to quantify the combined contribution of voltage inconsistency and temperature non-uniformity to the overall risk.

[0062] The system compares the real-time calculated Error value with a preset safety threshold, Error_threshold. If Error > Error_threshold: The system immediately determines that the current state is high-risk, suspends all ongoing balancing operations, and jumps to the maintenance diagnostic process. This process may include: recording a fault snapshot, issuing the highest level alarm, forcibly reducing the system's charging and discharging power, and guiding maintenance personnel to conduct a specific inspection.

[0063] Otherwise: the system allows continued execution or activation of the equilibrium strategy determined by the aforementioned steps.

[0064] This implementation adds crucial "safety brakes" and "risk circuit breakers" to the proactive equalization intervention. By constructing a multi-factor error index, Error, which includes voltage and temperature differences, the system can quantitatively assess the overall safety status of the battery pack. When a potential risk is detected (such as excessive temperature difference accompanied by widening voltage difference, which may be a precursor to thermal runaway), safety is prioritized, and equalization actions that might exacerbate the risk are suspended. This greatly enhances the system's self-protection capabilities under abnormal operating conditions, preventing inappropriate equalization strategies from becoming the "last straw" when the battery already has faults or hidden dangers, thus achieving an intelligent trade-off between safety control and performance optimization.

[0065] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: The voltage difference was divided into three categories: low (≤1V), medium (>0.1V and ≤0.3V), and high (>0.3V). These correspond to three equilibrium strategies (low intensity, medium intensity, and high frequency). The aging compensation function is set according to the degree of battery aging, such as λ = 1 + k (age / τ); Where age is the number of battery cycles, and τ is the preset aging period, used to dynamically adjust the balancing force. λ represents the balance strength adjustment coefficient for aging, used to adapt to the performance degradation of the battery cell under different life cycles.

[0066] The implementation is divided into two steps. The first step involves further refining the voltage difference ΔV into three levels: low inconsistency (ΔV ≤ 0.1V), medium inconsistency (0.1V < ΔV ≤ 0.3V), and high inconsistency (ΔV > 0.3V). Each level corresponds to a specific balancing strategy: low-intensity balancing (e.g., balancing only the highest / lowest 1-2 cells), medium-intensity balancing (balancing cells whose voltage deviates from the average value by a certain range), and high-frequency / high-intensity balancing (performing large-scale, periodic balancing of the entire group).

[0067] The second step involves introducing an aging compensation function. The system records the cumulative number of battery cycles (age) and sets a reference aging period τ (e.g., 1000 cycles). The aging adjustment coefficient λ = 1 + k is then calculated. (age / τ). Where k is the gain coefficient (e.g., 0.5). As the number of cycles (age) increases, the value of λ gradually increases from 1. When finally executing the balancing strategy, the balancing strength of the system (such as balancing current, duration) will be multiplied by λ. For example, for the same "medium inconsistency" level, an older battery pack that has undergone 2000 cycles (with a larger λ) will obtain a higher balancing current than a brand new battery pack.

[0068] This embodiment achieves dual optimization. First, the three-level voltage division allows for finer-grained balancing actions, avoiding resource misallocation caused by a "one-size-fits-all" strategy, and enabling balancing energy to be applied more precisely to battery packs with varying degrees of problems. Second, and more forward-lookingly, by introducing an aging compensation coefficient λ, the balancing strategy can dynamically evolve along with the battery's lifecycle. As batteries age, internal resistance increases, self-discharge rate differences intensify, and intrinsic inconsistencies become more severe. The dynamically enhanced balancing force can proactively counteract the negative effects of aging, essentially providing "anti-aging" maintenance treatment for the battery pack, thereby more effectively extending its usable capacity and safety in the later stages of its lifespan and enhancing its overall lifecycle value.

[0069] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Considering the impact of system load fluctuations on load balancing, the load fluctuation factor f_Loader = |P_prev-P_curr| / P_avg is calculated. By combining this factor with the voltage difference ΔV, a comprehensive criterion is established: Score = ΔV × f_Loader ×γ, where γ is the equalization efficiency adjustment coefficient; Priority is set based on Score; If Score < Low_thres, the equalization process is skipped to prevent energy waste; if Score ≥ High_thres, equalization control is initiated. γ is a weight parameter coupled with the system's operating state, which enables the equalization control to have dynamic adaptability and is suitable for dynamic power consumption scenarios in high-density energy storage stations.

[0070] Specifically, the EMS monitors the output (or input) power of the energy storage system in real time, calculates the absolute value of the power change between adjacent sampling periods |P_prev - P_curr|, and combines it with the average power P_avg to calculate the load fluctuation factor f_Load =|P_prev - P_curr| / P_avg. This factor quantifies the severity of instantaneous fluctuations in the system load.

[0071] Subsequently, the system combines the voltage difference ΔV with the load fluctuation factor f_Load to establish a comprehensive criterion, Score: Score = ΔV × f_Load × γ. Here, γ is a balancing efficiency adjustment coefficient related to the system's current operating state (e.g., charging, discharging, or idle). For example, γ may be larger during idle periods, encouraging balancing; while γ may be smaller during high-power discharge, inhibiting balancing.

[0072] The decision-making logic is based on the score: If Score < Low_thres: This means that either ΔV is very small or the load fluctuates drastically (in which case the energy transfer efficiency of the equalization process is low and may interfere with the main circuit). The system decides to skip this equalization process to prevent unnecessary energy loss and potential interference with system stability.

[0073] If Score ≥ High_thres: it means that the inconsistency is significant and the system is in a relatively stable or suitable equilibrium condition, and the system starts equilibrium control.

[0074] This embodiment introduces a load fluctuation factor and a comprehensive criterion, the Score, so that the balancing decision is no longer based solely on the internal state of the battery, but rather takes into full account the needs of the external power grid or load. This effectively solves the problem that traditional balancing may act "inappropriately" during periods of high load fluctuation, thus wasting energy, reducing overall system efficiency, and even affecting power response quality. It makes balancing control an integral part of smart energy management, doing "the right thing" only "at the right time," significantly improving the overall operational economy and stability of high-density energy storage power stations in real-world dynamic power consumption scenarios.

[0075] In one embodiment, Adjusting the equalization charging strategy parameters based on the aforementioned inconsistency level further includes: Extracting the electrochemical impedance spectral characteristics Z; The consistency index CI is calculated based on the distribution trend of Z: CI = |(Z_high-Z_low)| / Z_mid. A comprehensive criterion is set based on the voltage difference ΔV; If CI > CI_threshold or ΔV > ΔV_threshold, switch to active balancing mode; otherwise, use passive balancing mode. Z_high and Z_low represent impedance values ​​in the high / low frequency bands, while CI serves as an indicator for evaluating the internal electrochemical consistency of the system, enabling earlier detection of potential cell failure risks.

[0076] Specifically, the system periodically (e.g., once a day or after each charge-discharge cycle) injects small-amplitude, multi-frequency AC excitation signals into the battery pack using the EIS (electrochemical impedance spectroscopy) excitation and measurement module integrated in the BMS, and measures its response, thereby extracting the electrochemical impedance spectral characteristics Z(f) of each cell or representative cell.

[0077] The analysis focuses on spectral characteristics: the impedance values ​​Z_high (high frequency band, e.g., 1 kHz, reflecting ohmic internal resistance) and Z_low (low frequency band, e.g., 0.1 Hz, reflecting charge transfer impedance) are obtained separately, and the mid-frequency characteristic value Z_mid is calculated. Furthermore, the consistency index CI = |Z_high - Z_low| / Z_mid is calculated. This CI value reflects the degree of impedance dispersion in different polarization processes within the cell; an increase in CI generally indicates differentiation in electrochemical characteristics.

[0078] The system combines CI with the traditional ΔV to form a dual criterion: If CI > CI_threshold or ΔV > ΔV_threshold: If either condition is met, it indicates that the battery pack has serious inconsistencies in its electrochemical nature or external performance, and the system switches to active balancing mode to intervene strongly.

[0079] Otherwise: the system adopts a passive balancing method or only performs maintenance-oriented low-current balancing.

[0080] This embodiment provides a forward-looking and fundamental basis for judging equalization strategies through the electrochemical impedance spectroscopy (CI) index. Changes in impedance often precede significant voltage and capacity decay, providing early warnings of internal degradation such as micro-short circuits, electrolyte drying, and abnormal SEI film growth hundreds or even thousands of cycles in advance. Using CI as one of the equalization trigger conditions allows the system to intervene as soon as inconsistencies begin to emerge at the electrochemical level, achieving early prevention of capacity decay and potential failures. This transforms "post-event remediation" into "pre-event prevention," significantly improving the safety level and lifespan prediction capabilities of battery management.

[0081] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Set the battery pack capacity consistency index CI_cap = |(C_max-C_min)| / C_avg; A composite decision mechanism is established by combining CI_cap and ΔV: if (ΔV > V_Thr AND CI_cap > C_Thr), then full equilibrium is forcibly initiated; If CI_cap is not met but ΔV has reached the threshold, then perform targeted load balancing. Otherwise, continue to observe the current running status, where C_max, C_min, and C_avg are the maximum, minimum, and average cell capacities, respectively, used to determine the overall energy storage consistency of the battery pack.

[0082] Specifically, the system uses historical charge-discharge data (typically a complete full charge-discharge cycle) to estimate the actual usable capacity Ci of each cell using the ampere-hour integration method. Then, it calculates the battery pack capacity consistency index CI_cap = |C_max - C_min| / C_avg, where C_max, C_min, and C_avg represent the maximum, minimum, and average cell capacities, respectively. CI_cap directly reflects the differences in the "capacity" of the cells within the pack.

[0083] The system establishes a composite decision-making mechanism that considers both voltage "appearance" and capacity "physical capacity": Forced full balancing: If (ΔV > V_Thr) AND (CI_cap > C_Thr) are both true, it indicates that the battery pack has both an imbalance in its immediate operating state (voltage) and a significant difference in its inherent energy storage capacity (capacity), indicating a serious problem. The system will force a full balancing process, combined with possible capacity balancing and reconfiguration suggestions.

[0084] Targeted equalization: If CI_cap fails to meet the standard (≤ C_Thr) but ΔV has reached the threshold (> V_Thr), it indicates that the voltage difference may be caused by short-term factors (such as SOC differences or temperature unevenness) rather than permanent capacity loss. The system performs targeted equalization (such as SOC-based equalization) to address immediate inconsistencies.

[0085] Continuous observation: If none of the above conditions are met, the system determines that the current state is good and continues to observe.

[0086] This implementation method achieves a leap from managing the "current state" to managing the "healthy foundation." By introducing the capacity consistency index CI_cap, the system can clearly distinguish whether inconsistencies are temporary and recoverable (only a large voltage difference) or permanent and structural (capacity differentiation). This avoids the ineffective effort of using methods for dealing with temporary misalignments to address permanent capacity degradation. For structural inconsistencies, the system can detect them early and propose higher-level maintenance recommendations (such as battery pack restructuring). Thus, it enables a scientific assessment and tiered maintenance of the overall energy storage capacity and lifespan of the battery pack, ensuring the long-term return on investment of the energy storage system.

[0087] In one embodiment, adjusting the equalization charging strategy parameters based on the inconsistency level further includes: Analyze the evolution trend of battery inconsistency using historical charge-discharge curves; The trend coefficient ω(t) is established based on the time series as ω(t) = a·ΔV(t) + b·ΔSOC(t); Combine ω(t) for rolling prediction and set a dynamic adjustment mechanism; If ω(t) > ω_Thr, then the adaptive equilibrium control strategy is executed; otherwise, the original scheme is maintained. a and b are weighting parameters, and ΔV(t) and ΔSOC(t) are the voltage and SOC differences at different times, used to predict and prevent future changes in consistency.

[0088] Specifically, the system establishes a historical database that continuously records the voltage difference ΔV(t) and SOC difference ΔSOC(t) at each sampling time. Based on this time series data, the system establishes a trend coefficient ω(t) through linear regression or a more advanced algorithm (such as exponential smoothing). A simplified linear model example is: ω(t) = a·ΔV(t) + b·ΔSOC(t), where a and b are weight parameters, and this model fits the slope of the change in difference.

[0089] The system continuously makes rolling predictions: based on historical data from a recent period (such as the past 24 hours), it calculates the current trend coefficient ω(t) and predicts its short-term trend.

[0090] Decision-making logic is based on trend prediction: If ω(t) > ω_Thr (trend deterioration threshold): This means that inconsistency is accelerating, even if the current absolute value may not have reached the emergency threshold yet. The system will execute adaptive equilibrium control strategies, such as initiating equilibrium earlier, increasing the equilibrium strength, or increasing the equilibrium frequency, to proactively curb the deteriorating trend.

[0091] Otherwise: the system maintains the original periodic or trigger-based balancing scheme.

[0092] This implementation analyzes the evolution trend of inconsistencies, enabling the system to identify potential problems that are "not serious now but are rapidly deteriorating," and intervene before they develop into serious failures. This "prevention is better than cure" approach transforms battery health management from a passive, event-driven model to a proactive, predictive model. It can maintain battery pack status more smoothly, avoiding drastic, energy-intensive equalization measures required due to sudden deterioration of inconsistencies. This results in better battery life and reliability management at a lower long-term cost, representing an advanced manifestation of intelligent operation and maintenance.

[0093] This invention effectively solves the problem of rapid capacity decay caused by cell inconsistency in high-density energy storage scenarios by dynamically adjusting the equalization charging strategy based on the voltage differences of individual cells within the battery pack. Traditional static equalization strategies cannot cope with constantly changing battery states, while this invention introduces voltage difference as a core evaluation indicator and combines it with inconsistency levels for graded processing, making equalization charging more precise and flexible. By timely identifying and correcting inconsistencies, it can significantly slow down battery aging, improve the overall lifespan of the system, and ensure the stability and safety of energy storage devices during long-term operation, demonstrating significant application value and economic benefits.

[0094] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this invention, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A multi-level monitoring method integrating EMS and battery management system, characterized in that, include: Collect voltage data of each cell in the battery pack and calculate the voltage difference. The inconsistency level of the battery pack is determined based on the voltage difference. Adjust the equalization charging strategy parameters based on the aforementioned inconsistency level; The optimized equalization charging strategy is transmitted to the battery management system for execution.

2. The multi-level monitoring method integrating EMS and battery management system according to claim 1, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Get the current cell voltage data V_i (i=1,2,...,n) in the current battery pack; Calculate the average voltage V_avg = (ΣV_i) / n; Calculate the individual unit voltage difference ΔV = max(V_i)-min(V_i); If ΔV > Threshold1, execute balancing strategy A; otherwise, execute balancing strategy B. Here, strategy A is an active balancing strategy and strategy B is a passive balancing strategy. ΔV is the difference between the maximum and minimum voltage values, and Threshold1 is the set threshold for balancing to start, used to determine whether active balancing is required.

3. The multi-level monitoring method integrating EMS and battery management system according to claim 2, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Collect the battery operating environment temperature T_env; Establish a temperature-dependent dynamic threshold mechanism, i.e., Threshold = Threshold0. (T_env / T_ref); Calculate the individual unit voltage difference ΔV; If ΔV > Threshold, enable fast balancing mode; otherwise, maintain slow balancing mode. Threshold0 is the threshold at the reference temperature, T_ref is the standard temperature, and Threshold is the equilibrium judgment threshold that is dynamically adjusted with temperature to improve adaptability under different operating conditions.

4. The multi-level monitoring method integrating EMS and battery management system according to claim 3, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Analyze the current fluctuations (I_fluc) generated by the battery pack during discharge; Calculate the inconsistency coefficient IC = (ΔV + α·I_fluc) / V_max; Fuzzy logic is used to evaluate ICs and select corresponding strategies; When IC ≤ IC_min, a no-action policy is executed; when IC > IC_min and ≤ IC_med, a low-intensity equilibrium is executed; when IC > IC_med, a high-intensity equilibrium is executed. α is the current fluctuation weighting coefficient, I_fluc is the current discharge current fluctuation amplitude, and V_max is the maximum single-cell voltage. The equilibrium decision is optimized by combining the changes in charge state.

5. A multi-level monitoring method integrating EMS and battery management system according to claim 4, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Introduce the remaining capacity estimate SOC_i; According to the monomer SOC difference ΔSOC = max(SOC_i)-min(SOC_i); The design weighting coefficient β = ΔV / ΔSOC; Set strategy priority: If ΔSOC > η, execute the SOC-priority equalization strategy; otherwise, execute the voltage-only equalization strategy. β is a weighting factor for the degree of charge imbalance, and η is a preset SOC inconsistency threshold. The accuracy of the strategy is improved by integrating charge state information.

6. A multi-level monitoring method integrating EMS and battery management system according to claim 5, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Detect for abnormal signals in the battery management system; Construct a conditional judgment module that includes fault diagnosis; Define the anomaly determination formula: Error = K1·ΔV + K2·ΔT, where K1 and K2 are correction coefficients; If Error > Error_threshold, pause the current load balancing and perform maintenance diagnostics. ΔT is the temperature difference within the battery pack, and K1 and K2 represent the influence coefficients of voltage and temperature difference, respectively. Safety control is achieved through error indicators.

7. A multi-level monitoring method integrating EMS and battery management system according to claim 6, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: The voltage difference is divided into three categories: low, medium, and high; where low voltage difference is ≤1V, medium voltage difference is >0.1V and ≤0.3V, and high voltage difference is >0.3V. These correspond to three balancing strategies: low intensity, medium intensity, and high frequency. The aging compensation function is set according to the degree of battery aging, such as λ = 1 + k (age / τ); Where age is the number of battery cycles, and τ is the preset aging period, used to dynamically adjust the balancing force. λ represents the balance strength adjustment coefficient for aging, used to adapt to the performance degradation of the battery cell under different life cycles.

8. A multi-level monitoring method integrating EMS and battery management system according to claim 7, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Considering the impact of system load fluctuations on load balancing, calculate the load fluctuation factor f_Loader = |P_prev-P_curr| / P_avg; By combining this factor with the voltage difference ΔV, a comprehensive criterion is established: Score = ΔV × f_Loader × γ, where γ is the equalization efficiency adjustment coefficient; Priority is set based on Score; If Score < Low_thres, the equalization process is skipped to prevent energy waste; if Score ≥ High_thres, equalization control is initiated. γ is a weight parameter coupled with the system's operating state, which enables the equalization control to have dynamic adaptability and is suitable for dynamic power consumption scenarios in high-density energy storage stations.

9. A multi-level monitoring method integrating EMS and battery management system according to claim 8, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Extracting the electrochemical impedance spectral characteristics Z; The consistency index CI is calculated based on the distribution trend of Z: CI = |(Z_high-Z_low)| / Z_mid. A comprehensive criterion is set based on the voltage difference ΔV; If CI > CI_threshold or ΔV > ΔV_threshold, switch to active balancing mode; otherwise, use passive balancing mode. Z_high and Z_low represent impedance values ​​in the high / low frequency bands, while CI serves as an indicator for evaluating the internal electrochemical consistency of the system, enabling earlier detection of potential cell failure risks.

10. A multi-level monitoring method integrating EMS and battery management system according to claim 9, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Set the battery pack capacity consistency index CI_cap = |(C_max-C_min)| / C_avg; A composite decision mechanism is established by combining CI_cap and ΔV: if ΔV > V_Thr AND CI_cap > C_Thr, then full equilibrium is forcibly initiated; If CI_cap is not met but ΔV has reached the threshold, then perform targeted load balancing. Otherwise, continue to observe the current running status, where C_max, C_min, and C_avg are the maximum, minimum, and average cell capacities, respectively, used to determine the overall energy storage consistency of the battery pack.

11. A multi-level monitoring method integrating EMS and battery management system according to claim 10, characterized in that, The adjustment of equalization charging strategy parameters based on the inconsistency level further includes: Analyze the evolution trend of battery inconsistency using historical charge-discharge curves; The trend coefficient ω(t) is established based on the time series as ω(t) = a·ΔV(t) + b·ΔSOC(t); Combine ω(t) for rolling prediction and set a dynamic adjustment mechanism; If ω(t) > ω_Thr, then the adaptive equilibrium control strategy is executed; otherwise, the original scheme is maintained. a and b are weighting parameters, and ΔV(t) and ΔSOC(t) are the voltage and SOC differences at different times, used to predict and prevent future changes in consistency.

Images