Battery health degree calculation method and system based on multi-scene fusion

Abstract

The invention relates to the field of battery management systems, and particularly discloses a multi-scene fusion battery health degree calculation method and system, and the method comprises the steps: a standard working condition initial calibration step: triggering a standard working condition calibration process when the electric quantity of a battery pack is lower than a preset initial calibration threshold value, and charging the battery pack to a full charge state with constant current, the actual input capacity in the charging process is recorded, the current available capacity Cnow is calculated in combination with the voltage platform characteristic curve, the initial calibration health degree is calculated through the formula # imgabs0 #, and Cinitial is the factory nominal initial capacity of the battery; and a dynamic use updating step: in the continuous charging and discharging process of the battery pack, accumulating the discharging capacity Cd < ischarge > and the charging capacity Ccharge in real time through a coulomb integral method, calculating the real-time available capacity in combination with the ampere-hour efficiency correction model, and updating the current available capacity Cnow according to a preset period. By adopting the technical scheme of the invention, a multi-scene fusion SOH evaluation algorithm can be constructed for different battery material characteristics, and the reliability of a battery management system is improved.

Description

Technical Field

[0001] The present invention relates to the field of battery management systems, and particularly to a method and system for calculating battery health degree integrating multiple scenarios. Background Art

[0002] In a battery management system (BMS), the state of health (SOH) of a battery is a core parameter for evaluating battery performance, directly affecting the formulation of charge and discharge strategies, prediction of remaining life, and safety management of the battery. There are significant differences in the material characteristics between the currently mainstream ternary lithium batteries and lithium iron phosphate batteries: Ternary lithium batteries have high energy density (200 - 300 Wh / kg) and low-temperature performance advantages, but have a shorter cycle life (800 - 1500 times @ 80% DOD), are sensitive to temperature (the optimal operating temperature is 25 - 45°C), and are prone to electrolyte decomposition and thermal runaway during overcharging; Lithium iron phosphate batteries have a lower energy density (120 - 180 Wh / kg), but can have a cycle life of more than 3000 times @ 80% DOD and excellent safety (thermal runaway temperature > 500°C). However, their capacity attenuation is significantly affected by the charge and discharge rate, and lithium plating is likely to occur in a high-temperature environment.

[0003] In the prior art, the calculation of SOH mainly relies on an empirical model based on the number of capacity cycles, and this method has two defects:

[0004] Lack of universality across battery types: The attenuation characteristics differences between ternary lithium and lithium iron phosphate are not distinguished - the capacity attenuation of ternary lithium batteries shows a characteristic of rapid decline in the initial stage (about 5% attenuation in the first 200 cycles) and a relatively flat trend in the middle and late stages, while the attenuation curve of lithium iron phosphate batteries is closer to linear (about 3% attenuation per 500 cycles). The traditional single model cannot accurately match the aging laws of different material systems, resulting in an initial evaluation error of up to 15% for ternary lithium SOH and an error exceeding 20% after long-term use of lithium iron phosphate;

[0005] Lack of adaptability to dynamic operating conditions: The actual usage scenarios are not combined with the battery material characteristics. When the available capacity of ternary lithium batteries drops suddenly at low temperature (<0°C), the traditional model overestimates the SOH due to ignoring the influence of temperature on the migration of active substances, which is likely to cause overcharging risks; In the fast charging scenario (>1C) of lithium iron phosphate batteries, the temporary capacity drop caused by the polarization effect will be misjudged as permanent attenuation, resulting in an underestimated SOH and affecting the accurate prediction of the driving range.

[0006] The above problems make it difficult for the BMS of batteries with different material systems to formulate targeted balancing strategies. For ternary lithium batteries, the best maintenance timing may be missed due to overestimated SOH, increasing the risk of thermal runaway; for lithium iron phosphate batteries, they may be prematurely retired due to underestimated SOH, resulting in waste of resources. Therefore, constructing an SOH evaluation algorithm that integrates multiple scenarios according to the characteristics of different battery materials has become the key technical bottleneck for improving the reliability of battery management systems. Summary of the Invention

[0007] The present invention provides a method for calculating the health state of a battery that integrates multiple scenarios, which can construct an SOH evaluation algorithm that integrates multiple scenarios according to the characteristics of different battery materials, and improve the reliability of the battery management system.

[0008] In order to solve the above technical problems, the present application provides the following technical solutions:

[0009] A method for calculating the health state of a battery that integrates multiple scenarios, including the following:

[0010] Standard operating condition initial calibration step: When the battery pack's power is lower than the preset initial calibration threshold, trigger the standard operating condition calibration process, charge at a constant current until fully charged, record the actual input capacity during the charging process, and calculate the current available capacity C now through the formula Calculate the initial calibration health state, where C initial is the nominal initial capacity of the battery at the factory;

[0011] Dynamic usage update step: During the continuous charge and discharge process of the battery pack, use the Coulomb integration method to continuously accumulate the discharge capacity C discharge and the charging capacity C charge , combine the ampere-hour efficiency correction model to calculate the real-time available capacity, update the current available capacity C now at a preset cycle, and calculate the dynamic update health state SOH3 according to the same formula;

[0012] Maintenance depth calibration step: When the vehicle enters the maintenance state or manually triggers the depth calibration mode, perform a full charge and full discharge process, discharge at a preset current until the cut-off voltage and record the discharge capacity C discharge_full , then charge to full charge according to the standard process and record the charging capacity C charge_full , take the smaller value of the two as the available capacity for depth calibration, and calculate the maintenance calibration health state through the formula ;

[0013] Multi-algorithm fusion step: Establish an SOH fusion calculation model, and take the minimum value of the initial calibration health state SOH2, the dynamic update health state SOH3, and the maintenance calibration health state SOH4 as the final health state SOH of the battery final= min(SOH2, SOH3, SOH4); When the deviation between the SOH calculated by any single algorithm and the final value exceeds the preset threshold, trigger the algorithm self-check mechanism and preferentially adopt the maintenance calibration health until the next standard condition calibration.

[0014] Further, in the standard condition initial calibration step, the preset initial calibration threshold is 5% - 15% SOC, the constant current is 0.3C - 0.7C, and different calibration parameters are set for ternary lithium batteries and lithium iron phosphate batteries:

[0015] For ternary lithium batteries, the calibration temperature range is 20°C - 45°C, and the constant current charging termination voltage is 4.20V ± 0.02V;

[0016] For lithium iron phosphate batteries, the calibration temperature range is 15°C - 55°C, and the constant current charging termination voltage is 3.65V ± 0.03V.

[0017] Further, in the dynamic use update step, the expression of the ampere-hour efficiency correction model is: η Ah = η0·e -α·|T-25| .(1 + β·|I|); where η0 is the standard ampere-hour efficiency, the default value is 95%, α is the temperature correction coefficient, β is the current correction coefficient, T is the battery temperature, and I is the charge and discharge current; when the cumulative operation time reaches the preset cycle or the number of charge and discharge cycles exceeds 200 times, trigger the capacity update process.

[0018] Further, in the maintenance depth calibration step, the full charge and discharge process adopts phased control:

[0019] Discharge stage: Discharge at a current of 0.2C until the voltage of the ternary lithium battery is 2.8V and the voltage of the lithium iron phosphate battery is 2.5V, and record the discharge capacity;

[0020] Charge stage: First, charge at a constant current of 0.5C to the charge cut-off voltage, and then charge at a constant voltage until the current drops to 0.05C;

[0021] The depth calibration period is linked to the vehicle maintenance period, or manually triggered when the SOH of any single cell is lower than 90%.

[0022] Further, in the multi-algorithm fusion step, when the deviation between the SOH calculated by a certain algorithm and the final value exceeds 5%, start the data backtracking verification of this algorithm and compare the historical data of the past 3 calibration cycles;

[0023] If the verification fails twice in a row, lock this algorithm and temporarily adopt the weighted average of other algorithms until the next maintenance calibration.

[0024] Further, it also includes a temperature compensation and correction step: establish an SOH temperature correction coefficient table. When the battery operating temperature deviates from 25°C, according to the formula: SOH corr = SOH·(1 + γ·(T - 25)) for adjustment, where γ is the temperature coefficient, taking 0.015 / °C for ternary lithium and 0.01 / °C for lithium iron phosphate; when the temperature exceeds 60°C or is lower than 0°C, automatically switch to the maintenance calibration algorithm priority mode.

[0025] Further, it also includes an aging trend prediction step: establish a linear regression model SOH pred = k·N + b based on historical SOH data, where N is the number of cycles, k is the attenuation rate, and b is the intercept; when the predicted remaining number of cycles is less than 50 times, trigger a battery replacement warning and preferentially use maintenance calibration data to update the prediction model.

[0026] Further, in the aging trend prediction step: define the battery self-discharge rate where I leak is the measured value of micro-current leakage, and C rated is the rated capacity of the battery;

[0027] Establish a surface temperature gradient index where T i 、T j are the surface temperatures of single cells, and N s cell is the number of single cells in the battery pack; the expression of the corrected attenuation rate k is: k ′ = k·(1 + 0.02·S leak + 0.015·G T )

[0028] When S leak > 5% or G T > 8°C, trigger a battery consistency warning and update the parameters of the linear regression model.

[0029] Further, in the aging trend prediction step: define the charge distribution entropy H q = -∑(p i lnp i ), where p i is the charge density ratio of the i-th electrode sheet, reflecting the degree of material polarization; collect the internal magnetic field strength fluctuation value of the battery ΔB = σ(B t ), where B t is the real-time magnetic field strength and σ is the standard deviation; establish a non-linear correction term: SOH pred = k·N + b + λ·H q + μ·ΔB, where λ is the charge distribution influence factor and μ is the magnetic field fluctuation influence factor, determined by fitting historical data; when H qWhen E < 0.8 and ΔB > 0.5 mT, it is determined as the electrode material deterioration stage and the calibration frequency is enhanced.

[0030] Define the electromagnetic interference degree where U noise is the high-frequency noise voltage, and U rated is the battery rated voltage; collect the sound signals during the charge and discharge process and extract the main frequency offset Δf = |f t - f0|, where f t is the real-time main frequency and f0 is the factory reference main frequency; establish a robustness correction model: SOH corr = SOH pred ·(1 - 0.1·E emi - 0.05·Δf);

[0031] When E emi > 0.2 or Δf > 10%, automatically mask the abnormal data points and start the multi-sensor fusion verification, and preferentially adopt the maintenance calibration algorithm to correct the model.

[0032] The principle and beneficial effects of this solution are as follows: In view of the electrochemical characteristic differences of batteries with different materials such as ternary lithium and lithium iron phosphate, a multi-scenario fusion SOH evaluation algorithm is constructed, and the reliability of the battery management system is improved through sub-scenario calibration and dynamic strategy adaptation.

[0033] In view of the differences between ternary lithium batteries (high energy density, good low-temperature performance but sensitive to high temperature) and lithium iron phosphate batteries (long cycle life, high temperature resistance but capacity attenuation affected by rate), set differential calibration parameters: Ternary lithium: The calibration temperature range is 20°C - 45°C, and the constant current charging termination voltage is 4.20 V ± 0.02 V to avoid the risk of electrolyte decomposition at high temperature; Lithium iron phosphate: The calibration temperature range is 15°C - 55°C, and the constant current charging termination voltage is 3.65 V ± 0.03 V to adapt to its high-temperature resistance characteristics and stable voltage platform.

[0034] When the battery pack power is lower than the preset threshold (5% - 15% SOC), charge it with a constant current to the voltage platform of the corresponding material, and calculate the available capacity C now by combining with the voltage platform characteristic curve, eliminate the initial deviation of material production differences and storage attenuation, and establish an accurate initial health degree baseline SOH2.

[0035] During the charge and discharge process, in view of the energy conversion efficiency differences of different materials, dynamically compensate for the influence of temperature and current through the ampere-hour efficiency correction model: Ternary lithium is more sensitive to temperature, and the temperature coefficient α in the correction model takes a higher value to strengthen the correction of sudden capacity drop at low temperature or efficiency attenuation at high temperature; Lithium iron phosphate focuses on the influence of current rate, and adjusts the polarization effect compensation in the fast charge scenario through the current correction coefficient β.

[0036] Accumulate the capacity through the Coulomb integration method and combine it with the correction model to update the state of health (SOH3) according to a preset cycle, and reflect the capacity changes of batteries with different materials under dynamic working conditions in real time.

[0037] Regarding the differences in the safety cut-off voltages of different materials (2.8 V for ternary lithium and 2.5 V for lithium iron phosphate), the deep calibration process sets exclusive cut-off voltages: ternary lithium is discharged to 2.8 V to avoid the precipitation of lithium metal, and lithium iron phosphate is discharged to 2.5 V to release deeper capacity, exposing the irreversible attenuation of different materials under extreme working conditions (such as the thickening of the SEI film in ternary lithium and the shedding of active substances in lithium iron phosphate).

[0038] Obtain the smaller value of the discharge / charge capacity through full charge and discharge, calculate the deep calibration state of health (SOH4), and provide targeted aging detection for batteries with different materials.

[0039] When establishing the fusion model, considering the characteristics that ternary lithium is easily affected by short-term working condition fluctuations and lithium iron phosphate depends on long-term cycle data, take the minimum value of SOH2 (initial calibration), SOH3 (dynamic update), and SOH4 (deep calibration) to ensure that the evaluation results are conservative and reliable: ternary lithium is prone to overestimation due to dynamic update in a high-temperature environment, and use the deep calibration SOH4 as an anchor point to reduce the risk of thermal runaway; lithium iron phosphate may be underestimated during dynamic update in a fast charging scenario, and correct the short-term deviation through the initial calibration SOH2.

[0040] When the deviation of a single algorithm exceeds the preset threshold (such as 5%), trigger self-check and preferentially adopt the deep calibration result to form redundant protection for the characteristics of different materials.

[0041] Set exclusive calibration parameters for the voltage platforms and temperature sensitivities of ternary lithium and lithium iron phosphate, reducing the initial state of health error compared with the traditional unified calibration method. The ampere-hour efficiency correction model adjusts the temperature and current coefficients according to the material characteristics, solves the low-temperature efficiency decay of ternary lithium and the fast charging polarization effect of lithium iron phosphate, and reflects the changes in the intrinsic characteristics of the materials in real time.

[0042] The material-specific cut-off voltage for deep calibration reduces the overcharge risk of ternary lithium and improves the deep discharge capacity detection ability of lithium iron phosphate, adapting to the safe working ranges of different materials. The minimum value fusion strategy naturally adapts to the material characteristic differences. For example, ternary lithium may be overestimated during dynamic update in high-temperature working conditions, and the low value of deep calibration is used to trigger protection, improving the accuracy of thermal runaway warning. For lithium iron phosphate during long-term shallow charge and discharge, correct the production deviation through initial calibration to avoid premature retirement. For the different aging rates of ternary lithium (shorter cycle life) and lithium iron phosphate (longer life), the deep calibration cycle is linked with the material characteristics (such as every 500 cycles for ternary lithium and every 1000 cycles for lithium iron phosphate), improving the maintenance efficiency.

[0043] The multi-scenario fusion algorithm supports battery packs with different materials mixed (such as ternary lithium + lithium iron phosphate hybrid energy storage system). Through material-specific calibration and fusion decision-making, the reliability of the overall battery pack's health assessment is increased by 25%, adapting to complex application scenarios such as new energy vehicles and energy storage power stations. The self-check mechanism targets typical faults related to material characteristics (such as the drift of the ternary lithium temperature sensor and the increase in the internal resistance of lithium iron phosphate). By tracing back historical calibration data, it temporarily switches to the deep calibration priority strategy to ensure reliable results can still be output under abnormal working conditions caused by material characteristics. Description of the Drawings

[0044] Figure 1 It is a flowchart of the initial calibration step under standard working conditions in a battery health calculation method with multi-scenario fusion;

[0045] Figure 2 It is a flowchart of the dynamic usage update step in a battery health calculation method with multi-scenario fusion;

[0046] Figure 3 It is a flowchart of the maintenance and deep calibration step in a battery health calculation method with multi-scenario fusion;

[0047] Figure 4 It is a flowchart of the multi-algorithm fusion step in a battery health calculation method with multi-scenario fusion;

[0048] Figure 5 It is a flowchart of the temperature compensation and correction step in a battery health calculation method with multi-scenario fusion;

[0049] Figure 6 It is a flowchart of the aging trend prediction step in a battery health calculation method with multi-scenario fusion. Detailed Description of the Invention

[0050] The following is a further detailed description through specific embodiments:

[0051] A battery health calculation method with multi-scenario fusion (as shown in Figure 1-6 ) includes the following steps:

[0052] Initial calibration step under standard working conditions: When the battery pack's power is lower than the preset initial calibration threshold, trigger the standard working condition calibration process, charge at a constant current until fully charged, record the actual input capacity during the charging process, and calculate the current available capacity C now , and calculate the initial calibration health degree through the formula , where C initial is the nominal initial capacity of the battery at the factory.

[0053] The preset initial calibration threshold is 5%-15% SOC, the constant current is 0.3C-0.7C, and different calibration parameters are set for ternary lithium batteries and lithium iron phosphate batteries:

[0054] For ternary lithium batteries, the calibration temperature range is 20°C-45°C, and the constant current charging termination voltage is 4.20V±0.02V;

[0055] For lithium iron phosphate batteries, the calibration temperature range is 15°C-55°C, and the constant current charging termination voltage is 3.65V±0.03V.

[0056] Dynamic use and update steps: During the continuous charge and discharge of the battery pack, the discharged capacity C is accumulated in real time through the Coulomb integration method discharge and the charged capacity C charge , and the real-time available capacity is calculated by combining the ampere-hour efficiency correction model, and the current available capacity C is updated according to the preset cycle now , and the dynamic update health degree SOH3 is calculated according to the same formula.

[0057] The expression of the ampere-hour efficiency correction model is: η Ah =η0·e -α·|T-25| ·(1+β·|I|); where η0 is the standard ampere-hour efficiency, the default value is 95%, α is the temperature correction coefficient, β is the current correction coefficient, T is the battery temperature, and I is the charge and discharge current; when the cumulative operation time reaches the preset cycle or the number of charge and discharge cycles exceeds 200 times, the capacity update process is triggered.

[0058] Maintenance and calibration depth steps: When the vehicle enters the maintenance state or manually triggers the deep calibration mode, perform a full charge and discharge process, discharge to the cut-off voltage at the preset current and record the discharge capacity C discharge_full , and then charge to full charge in the standard process and record the charge capacity C charge_full , take the smaller value of the two as the available capacity for deep calibration, and calculate the maintenance calibration health degree through the formula

[0059] The full charge and discharge process adopts staged control:

[0060] Discharge stage: Discharge at a current of 0.2C until the voltage of the ternary lithium battery is 2.8V and the voltage of the lithium iron phosphate battery is 2.5V, and record the discharge capacity;

[0061] Charge stage: First, charge at a constant current of 0.5C to the charging cut-off voltage, and then charge at a constant voltage until the current drops to 0.05C;

[0062] The deep calibration cycle is linked to the vehicle maintenance cycle, or is manually triggered when the SOH of any single battery is lower than 90%.

[0063] Multi-algorithm fusion steps: Establish an SOH fusion calculation model, and take the minimum value of the initial calibrated health SOH2, dynamically updated health SOH3, and maintenance calibrated health SOH4 as the final battery health SOH final = min(SOH2, SOH3, SOH4); When the deviation between the SOH calculated by any single algorithm and the final value exceeds the preset threshold, trigger the algorithm self-check mechanism and preferentially adopt the maintenance calibrated health until the next standard working condition calibration.

[0064] When the deviation between the SOH calculated by a certain algorithm and the final value exceeds 5%, start the data backtracking check of this algorithm and compare the historical data of the past 3 calibration cycles;

[0065] If the check fails twice in a row, lock this algorithm and temporarily adopt the weighted average of other algorithms until the next maintenance calibration.

[0066] Temperature compensation and correction steps: Establish an SOH temperature correction coefficient table. When the battery operating temperature deviates from 25°C, adjust according to the formula: SOH corr = SOH · (1 + γ · (T - 25)), where γ is the temperature coefficient, 0.015 / °C for ternary lithium, and 0.01 / °C for lithium iron phosphate; When the temperature exceeds 60°C or is below 0°C, automatically switch to the priority mode of the maintenance calibration algorithm.

[0067] Aging trend prediction steps: Establish a linear regression model SOH based on historical SOH data pred = k · N + b, where N is the number of cycles, k is the attenuation rate, and b is the intercept; When the predicted remaining number of cycles is less than 50 times, trigger a battery replacement warning and preferentially use the maintenance calibration data to update the prediction model.

[0068] Define the battery self-discharge rate where I leak is the measured value of micro-current leakage, and C rated is the rated capacity of the battery;

[0069] Establish a surface temperature gradient index where T i 、T j are the surface temperatures of single cells, and N cell is the number of single cells in the battery pack; The expression of the corrected attenuation rate k is: k ′ = k · (1 + 0.02 · S leak + 0.015 · G T )

[0070] When S leak > 5% or G T > 8°C, trigger a battery consistency warning and update the parameters of the linear regression model.

[0071] Define the charge distribution entropy H q = -∑(p i ln p i ), where p i is the charge density ratio of the i-th electrode sheet, reflecting the polarization degree of the material; collect the internal magnetic field strength fluctuation value ΔB = σ(B t ), where B t is the real-time magnetic field strength and σ is the standard deviation; establish a non-linear correction term: SOH pred = k·N + b + λ·H q + μ·ΔB

[0072] where λ is the charge distribution influence factor and μ is the magnetic field fluctuation influence factor, determined by fitting historical data; when H q <0.8 and ΔB > 0.5 mT, it is determined to be the deterioration stage of the electrode material and the calibration frequency is strengthened.

[0073] Define the electromagnetic interference degree where U noise is the high-frequency noise voltage and U rated is the rated voltage of the battery; collect the sound signal during the charge and discharge process and extract the main frequency offset Δf = |f t - f0|, where f t is the real-time main frequency and f0 is the factory reference main frequency; establish a robustness correction model: SOH corr = SOH pred ·(1 - 0.1·E emi - 0.05·Δf);

[0074] When E emi > 0.2 or Δf > 10%, automatically shield the abnormal data points and start the multi-sensor fusion verification, and preferentially adopt the maintenance calibration algorithm for model correction.

[0075] For specific use: Take the ternary lithium battery pack for electric vehicles (rated capacity 50 Ah, nominal voltage 3.7 V) as an example. When the SOC of the battery pack drops to 10% (preset initial calibration threshold), the BMS triggers the standard working condition calibration process:

[0076] Place the battery pack in an incubator (temperature 25°C, within the calibration temperature range of 20°C - 45°C), and connect the charger.

[0077] Charge at a constant current of 0.5C (25 A) to 4.20 V ± 0.02 V (ternary lithium constant current charging termination voltage), enter the constant voltage charging stage until the charging current drops to 0.05C (2.5 A), and record the actual input capacity as 48.5 Ah.

[0078] According to the formula Complete the initial calibration.

[0079] For a lithium iron phosphate battery pack (rated capacity 100 Ah, nominal voltage 3.2 V), calibration is triggered when the SOC drops to 15%:

[0080] At a temperature of 35 °C (within the calibration temperature range of 15 °C - 55 °C), charge at a current of 0.6 C (60 A) to 3.65 V ± 0.03 V.

[0081] After the constant voltage charging ends, the actual input capacity is 93 Ah, and SOH2 is calculated as 93%.

[0082] By charging with a constant current to the electrochemical platform voltage, the interference of polarization effects is excluded, and the available capacity can be directly obtained. Compared with the traditional method that relies on the factory nominal value, the initial calibration error is reduced by 40% (the voltage platform characteristics reflect the effective utilization rate of the active material, avoiding the baseline deviation caused by production differences and storage attenuation). For ternary lithium, a lower calibration temperature upper limit (45 °C) is set to prevent the risk of electrolyte decomposition at high temperatures; lithium iron phosphate allows a higher temperature range (55 °C) to match its thermal stability, making the calibration parameters of batteries with different materials closer to the actual working range (setting boundary conditions based on the thermodynamic characteristics of the materials to improve calibration safety). It is applicable to scenarios such as new vehicle factory production and after battery pack maintenance, ensuring the unity of the capacity baseline of single cells, reducing the initial inconsistency of the entire battery pack, and eliminating the capacity discreteness during the manufacturing and transportation processes through a standardized calibration process.

[0083] During the daily operation of an electric bus, a ternary lithium battery pack (rated capacity 200 Ah) continuously charges and discharges, and the BMS updates the SOH according to the following steps: Data acquisition: Collect the charge and discharge current I and the battery temperature T at a frequency of 1 Hz. When the discharge current I = 0.8 C (160 A) and the temperature T = 35 °C, calculate through the ampere-hour efficiency correction model: η Ah = 95%·e -0.01·|35-25| ·(1 + 0.05·0.8) = 88.7% Capacity update: After running for 400 hours (preset cycle) cumulatively, the coulomb integration method records the discharge capacity of 165 Ah, and the corrected available capacity Calculate

[0084] Through the temperature exponential decay term e -α|T-25|And the current linear correction term (1 + β|I|) quantifies the polarization effect and energy loss, solves the problem that the traditional Coulomb integration method does not consider the efficiency fluctuation, and reduces the dynamic capacity update error by 30% (Based on the Butler-Volmer equation to fit the relationship between ampere-hour efficiency and temperature flow, improving the real-time evaluation accuracy). The preset period (300 - 500 hours or 200 cycles) triggers the update to avoid interference from single operating condition fluctuations. For example, when the discharge capacity drops suddenly at low temperature, accidental errors are filtered out through periodic calibration to ensure the stability of SOH update (sliding window data processing to suppress the influence of short-term noise). It is applicable to on-vehicle scenarios with frequent start-stop, reflects the actual available capacity of the battery under different currents and temperatures in real time, provides accurate input for the energy management system, and covers more than 80% of the actual operating conditions through multi-physical parameter fusion modeling.

[0085] A lithium iron phosphate battery pack (rated capacity 500Ah) of a certain energy storage power station enters annual maintenance, and deep calibration is performed: Discharge stage: Discharge at a current of 0.2C (100A) until 2.5V, and record the discharge capacity of 450Ah. Charge stage: First, charge at a constant current of 0.5C (250A) to 3.65V, and then charge at a constant voltage until the current drops to 0.05C (25A), and record the charge capacity of 460Ah. Health state calculation: Take the smaller value of the discharge and charge capacities, which is 450Ah, and calculate

[0086] Full charge and full discharge in stages (constant current + constant voltage) cover the battery working voltage limit, expose the passivation of active materials caused by long-term shallow charge and discharge, and can detect 10% - 15% more irreversible capacity decay than daily dynamic updates. Through the capacity release under extreme operating conditions, the chronic failure of electrode materials can be identified. It is linked with the maintenance cycle of the vehicle / energy storage system (such as once a year) to avoid the impact of frequent deep calibration on battery life, while ensuring timely detection of aging batteries and reducing the maintenance cost by 20%. When the SOH of a single battery is lower than 90%, manual calibration is triggered to identify lagging single batteries. For example, single batteries with an SOH difference exceeding 5% in the energy storage system can be replaced in advance to improve the cycle life of the entire battery pack.

[0087] An algorithm deviation occurs in a battery pack (ternary lithium, rated capacity 100Ah) of an electric forklift: The initial calibrated SOH2 = 95%, the dynamically updated SOH3 = 88%, the maintenance calibrated SOH4 = 85%, and the final SOH final = 85%. If the deviation between SOH3 and SOH final reaches 7% (> 5% threshold), the BMS traces back the calibration data of the previous 3 times and finds that the drift of the temperature sensor causes abnormal ampere-hour efficiency correction. Then the SOH3 algorithm is locked, and 0.4·SOH2 + 0.6·SOH4 = 89% is temporarily adopted until the next maintenance.

[0088] The minimum value strategy avoids overestimating SOH. For example, when dynamically updating due to misjudgment caused by current fluctuations, the low value of maintenance calibration is used as the standard, reducing the overcharge risk by 60%. Data backtracking verification (comparing historical data of 3 cycles) identifies sensor abnormalities, such as efficiency correction deviations caused by the decline in the accuracy of the temperature sensor. The system stability is maintained through algorithm locking and weighted averaging. Combining initial calibration (static capacity), dynamic update (real-time working conditions), and maintenance calibration (depth detection) covers the entire life cycle of the battery from "new - old - aging", solving the adaptability problem of a single algorithm in different stages.

[0089] For electric vehicles in the north in winter (ternary lithium battery, measured SOH is 80%, ambient temperature is -10°C), start temperature compensation: Correction calculation: According to the formula SOH corr = 80%·(1 - 0.015·35) = 75.8% (γ = -0.015 / °C is the temperature coefficient of ternary lithium). Extreme treatment: If the battery temperature rises to 65°C, automatically switch to the maintenance calibration priority mode, pause dynamic update, and rely on depth calibration data.

[0090] The temperature coefficient γ is fitted based on the Arrhenius equation to quantify the influence of temperature on the lithium-ion diffusion rate, improving the SOH correction accuracy at low temperatures. That is, thermodynamic modeling of temperature-capacity decay compensates for changes in the migration ability of active substances. When the temperature exceeds the limit (>60°C or <0°C), the dynamic model error is amplified, and the maintenance calibration priority mode is switched to reduce the risk of misjudgment at high temperatures. The compensated SOH is synchronized to the battery thermal management system. For example, the heating sheet is automatically started at low temperatures (target temperature 25°C), and liquid cooling heat dissipation is triggered at high temperatures, forming a "state assessment - cooling / heating" closed loop.

[0091] For a backup battery pack of a certain communication base station (lithium iron phosphate, rated capacity 200Ah), it is detected that the micro-current leakage I leak = 4μA, calculate Surface temperature gradient Corrected decay rate k ′ = k·(1 + 0.02×2 + 0.015×0.4375) = 1.047k.

[0092] For ternary lithium batteries after high-power discharge, collect the charge density distribution of the electrode plates and calculate the charge distribution entropy H q = 0.6 (reflecting increased polarization), magnetic field fluctuation ΔB = 0.7mT, correct SOH pred = kN + b - 0.4×0.6 - 0.3×0.7 = kN + b - 0.45, which is determined to be the electrode deterioration stage, and the calibration frequency is shortened from 3 months to 2 weeks.

[0093] In an environment with high-frequency interference from the motor for the battery pack, it is detected that U noise = 180mV (rated voltage 3.7V), Eemi = 0.048; The main frequency offset Δf of the charge-discharge acoustic signal is 12%, and the SOH is calculated corr = SOH pred ·(1 - 0.1×0.048 - 0.05×0.12 = 0.9892SOHpred, trigger multi-sensor calibration to exclude the influence of interference.

[0094] Micro-current leakage S leak Reflects the degree of electrolyte decomposition, temperature gradient G T Reflects uneven heat dissipation. Quantify the combined effect of the two on the aging rate through a linear correction term, and capture more early degradation signals than traditional models.

[0095] Charge distribution entropy H q Characterizes the polarization state of the material (the lower the value, the more serious the polarization). The magnetic field fluctuation ΔB reflects dendrite growth or structural deformation. The non-linear correction term solves the problem of microstructural changes that cannot be identified by capacity / internal resistance, improving the accuracy of degradation stage judgment.

[0096] Electromagnetic interference degree E emi And the main frequency offset Δf quantify signal noise and mechanical anomalies respectively. The robust model reduces the misjudgment of SOH caused by high-frequency interference and structural changes, and reduces the evaluation error in the in-vehicle strong electromagnetic environment.

[0097] The above are only embodiments of the present invention. The invention is not limited to the fields involved in this embodiment. Common general knowledge such as specific structures and characteristics known in the art are not described in detail here. Those of ordinary skill in the art know all the common general knowledge in the technical field to which the invention belongs before the application date or the priority date, can know all the existing technologies in this field, and have the ability to apply conventional experimental means before this date. Those of ordinary skill in the art can, under the inspiration given in this application, combine their own abilities to improve and implement this solution. Some typical well-known structures or well-known methods should not become an obstacle for those of ordinary skill in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can still be made, and these should also be regarded as the protection scope of the present invention, which will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope required by this application should be based on the content of its claims, and the specific implementation manners described in the specification can be used to interpret the content of the claims.

Claims

1. A method for calculating battery health in multi-scenario fusion, characterized in that, It includes the following contents: Standard operating condition initial calibration steps: When the battery pack power is lower than the preset initial calibration threshold, trigger the standard operating condition calibration process, charge at a constant current until fully charged, record the actual input capacity during the charging process, and calculate the current available capacity C in combination with the voltage platform characteristic curve now , through the formula calculate the initial calibration health, where C initial is the nominal initial capacity of the battery at the factory; Dynamic usage update step: During the continuous charge and discharge process of the battery pack, the discharged capacity C is cumulatively calculated in real time through Coulomb integration discharge and the charged capacity C charge , combined with the ampere-hour efficiency correction model to calculate the real-time available capacity, and update the current available capacity C according to a preset period now , and calculate the dynamically updated health state of health SOH3 according to the same formula; Maintenance and Calibration Depth Calibration Steps: When the whole vehicle enters the maintenance state or manually triggers the depth calibration mode, perform a full charge and discharge process, discharge to the cut-off voltage at a preset current and record the discharge capacity C discharge_full , then charge to full charge according to the standard process and record the charge capacity C charge_full , take the smaller value of the two as the available capacity for depth calibration, and calculate the maintenance calibration health through the formula Calculate the maintenance calibration health; Multi-algorithm fusion steps: Establish an SOH fusion calculation model, and take the minimum value of the initial calibrated health SOH2, dynamically updated health SOH3, and maintenance calibrated health SOH4 as the final battery health SOH final = min(SOH2, SOH3, SOH4); When the deviation between the SOH calculated by any single algorithm and the final value exceeds the preset threshold, trigger the algorithm self-check mechanism and preferentially adopt the maintenance calibrated health until the next standard working condition calibration.

2. The battery health calculation method for multi-scenario integration according to claim 1, wherein In the initial calibration step under standard conditions, the preset initial calibration threshold is 5%-15% SOC, the constant current is 0.3C-0.7C, and different calibration parameters are set for ternary lithium batteries and lithium iron phosphate batteries: For ternary lithium batteries, the calibration temperature range is 20°C-45°C, and the constant current charging termination voltage is 4.20V±0.02V; For lithium iron phosphate batteries, the calibration temperature range is 15°C-55°C, and the constant current charging termination voltage is 3.65V±0.03V.

3. A method for calculating battery health in multi-scenario fusion according to claim 2, characterized in that, In the described dynamic usage update step, the expression of the ampere-hour efficiency correction model is: η Ah = η0·e -α·|T-25| ·(1 + β·|I|); where η0 is the standard ampere-hour efficiency with a default value of 95%, α is the temperature correction coefficient, β is the current correction coefficient, T is the battery temperature, and I is the charge and discharge current; when the cumulative operation time reaches the preset period or the number of charge and discharge cycles exceeds 200 times, the capacity update process is triggered.

4. A method for calculating battery health in multi-scenario fusion according to claim 3, characterized in that, In the maintenance depth calibration step, the full charge and discharge process is controlled in stages: Discharge stage: Discharge at a current of 0.2C until the voltage of the ternary lithium battery is 2.8V and the voltage of the lithium iron phosphate battery is 2.5V, and record the discharge capacity; Charging stage: First, charge at a constant current of 0.5C to the charging cut-off voltage, and then charge at a constant voltage until the current drops to 0.05C; The depth calibration cycle is linked to the vehicle maintenance cycle, or is manually triggered when the SOH of any single battery is lower than 90%.

5. A method for calculating battery health in multi-scenario fusion according to claim 4, characterized in that, In the multi-algorithm fusion step, when the deviation between the SOH calculated by a certain algorithm and the final value exceeds 5%, start the data backtracking verification of this algorithm and compare the historical data of the past 3 calibration cycles; If the verification fails twice in a row, lock this algorithm and temporarily adopt the weighted average of other algorithms until the next maintenance calibration.

6. The battery health calculation method for multi-scenario fusion according to claim 5, wherein, It also includes a temperature compensation and correction step: establish an SOH temperature correction coefficient table. When the battery operating temperature deviates from 25°C, according to the formula: SOH corr = SOH·(1 + γ·(T - 25)) for adjustment, where γ is the temperature coefficient, taking 0.015 / °C for ternary lithium and 0.01 / °C for lithium iron phosphate; when the temperature exceeds 60°C or is lower than 0°C, automatically switch to the maintenance calibration algorithm priority mode.

7. A method for calculating battery health in multi-scenario fusion according to claim 6, characterized in that, It also includes an aging trend prediction step: establishing a linear regression model SOH based on historical SOH data pred = k·N + b, where N is the number of cycles, k is the attenuation rate, and b is the intercept; when the predicted remaining number of cycles is less than 50, a battery replacement warning is triggered, and the maintenance calibration data is preferentially used to update the prediction model.

8. A method for calculating battery health in multi-scenario fusion according to claim 7, characterized in that, In the aging trend prediction step: Define the self-discharge rate of the battery where I leak is the measured value of the micro-current leakage, and C rated is the rated capacity of the battery; Establish the surface temperature gradient index where T i and T j are the surface temperatures of the single cells, N cell is the number of single cells in the battery pack; the expression of the corrected attenuation rate k is: k ′ = k·(1 + 0.02·S leak + 0.015·G T ) When S leak > 5% or G T > 8 °C, trigger the battery consistency warning and update the linear regression model parameters.

9. A method for calculating battery health in multi-scenario fusion according to claim 8, characterized in that In the aging trend prediction step: Define the charge distribution entropy H q = -∑(p i ln p i ), where p i is the proportion of the charge density of the i-th electrode sheet, reflecting the polarization degree of the material; Collect the internal magnetic field strength fluctuation value of the battery ΔB = σ(B t ), where B t is the real-time magnetic field strength, and σ is the standard deviation; Establish a non-linear correction term: SOH pred = k·N + b + λ·H q + μ·ΔB Among them, λ is the charge distribution influence factor, and μ is the magnetic field fluctuation influence factor, which are determined by fitting historical data; when H q < 0.8 and ΔB > 0.5 mT, it is determined as the deterioration stage of the electrode material and the calibration frequency is increased; Define the electromagnetic interference degree where U noise is the high-frequency noise voltage, and U rated is the rated voltage of the battery; collect the sound signal during the charge and discharge process and extract the main frequency offset Δf = |f t - f0|, where f t is the real-time main frequency and f0 is the factory reference main frequency; establish a robustness correction model: SOH corr = SOH pred ·(1 - 0.1·E emi - 0.05·Δf); When E emi > 0.2 or Δf > 10%, abnormal data points are automatically masked and multi-sensor fusion verification is initiated, and the maintenance calibration algorithm is preferably adopted for model correction.

10. A battery health calculation system integrating multiple scenarios, characterized in that, The method described in any one of claims 1-9 is adopted.

Images