Lithium battery cycle life prediction method and system

Abstract

The invention relates to the technical field of lithium battery life prediction, and discloses a lithium battery cycle life prediction method and system. The method comprises the following steps: collecting charging voltage state data and discharging voltage drop data of an nth cycle in a lithium battery, and calculating charging internal resistance and discharging internal resistance; subtracting the charging internal resistance from the discharging internal resistance to obtain a first internal resistance difference value, and calculating a difference evolution rate and a corresponding difference evolution acceleration based on the first internal resistance difference value; calculating the capacity retention ratio of the nth cycle based on the first internal resistance difference value, the difference evolution rate and the difference evolution acceleration; and superposing the product of the difference evolution rate and the to-be-predicted cycle index to the first internal resistance difference value for linear extrapolation to obtain a second internal resistance difference value, and solving the residual available cycle index based on the second internal resistance difference value and the capacity retention ratio. According to the invention, the cycle life of the battery is predicted by using the evolution law of asymmetry of the lithium ion transmission path in the charging and discharging process, and the accuracy of lithium battery cycle life prediction is improved.

Description

Technical Field

[0001] This invention relates to the field of lithium battery life prediction technology, and in particular to a method and system for predicting the cycle life of lithium batteries. Background Technology

[0002] As a core piece of equipment in intelligent logistics systems, industrial AGVs' power lithium batteries directly impact operating costs and operational continuity. With increasingly complex AGV application scenarios, frequent start-stop cycles and high-rate discharges place high demands on battery health. Existing lithium battery life prediction schemes rely on capacity decay curve fitting or single equivalent internal resistance measurement. Current capacity calibration requires a complete charge-discharge cycle, which is time-consuming and disrupts AGV operation. While the equivalent internal resistance method is convenient, it only reflects the battery's impedance characteristics under specific operating conditions and cannot capture the asymmetric evolution of lithium-ion transport paths during charging and discharging, resulting in low accuracy in lithium battery cycle life prediction. Summary of the Invention

[0003] The main objective of this invention is to provide a method and system for predicting the cycle life of lithium batteries. This invention utilizes the evolution law of lithium-ion transport path asymmetry during charging and discharging to predict battery cycle life, thereby improving the accuracy of lithium battery cycle life prediction.

[0004] To achieve the above objectives, the present invention provides a method for predicting the cycle life of lithium batteries, comprising the following steps: Collect charging voltage state data and discharging voltage drop data for the nth cycle of the lithium battery, and calculate the charging internal resistance and discharging internal resistance. The first internal resistance difference is obtained by subtracting the charging internal resistance from the discharge internal resistance, and the differential evolution rate and the corresponding differential evolution acceleration are calculated based on the first internal resistance difference. The capacity retention rate of the nth cycle is calculated based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration. The product of the differential evolution rate and the number of cycles to be predicted is superimposed on the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value. The remaining number of available cycles is then calculated based on the second internal resistance difference value and the capacity retention rate.

[0005] Optionally, in a first implementation of the first aspect of the present invention, the step of collecting charging voltage state data and discharging voltage drop data of the nth cycle in the lithium battery, and calculating the charging internal resistance and discharging internal resistance, includes: During constant rate charging, the voltage and state of charge of the lithium battery within a preset state of charge range are recorded in real time to obtain a sampling sequence. The sampling sequence is then filtered and abnormal points where the current fluctuation exceeds a first preset threshold are removed to obtain charging voltage state data. When the AGV starts from a stationary state and is suddenly subjected to a load of a preset ratio, the stationary open circuit voltage and the transient voltage value within a preset time period after the load are continuously collected to obtain a voltage sequence. The voltage sequence is then filtered and abnormal points where the current fluctuation exceeds the second preset threshold are removed to obtain discharge voltage drop data. The charging internal resistance is calculated based on the charging voltage state data, and the discharging internal resistance is calculated based on the discharging voltage drop data.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the step of calculating the charging internal resistance based on the charging voltage state data and simultaneously calculating the discharging internal resistance based on the discharging voltage drop data includes: Linear interpolation is performed on the charging voltage state data to locate the first voltage value corresponding to the preset first state of charge and the second voltage value corresponding to the preset second state of charge. Divide the difference between the second voltage value and the first voltage value by the charging current to obtain the charging internal resistance; The open-circuit voltage under static conditions is extracted from the discharge voltage drop data as the static open-circuit voltage. The voltage value corresponding to a preset time after a sudden load is applied is extracted as the steady-state voltage. The difference between the static open-circuit voltage and the steady-state voltage is divided by the discharge current to obtain the discharge internal resistance.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the step of subtracting the charging internal resistance from the discharging internal resistance to obtain a first internal resistance difference, and calculating the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference, includes: Subtracting the charging internal resistance from the discharge internal resistance yields the first internal resistance difference for the nth cycle. Retrieve the internal resistance difference value of the nth cycle from the historical data cache, perform a difference operation between the first internal resistance difference value of the nth cycle and the internal resistance difference value of the nth cycle, and divide by m to obtain the differential evolution rate. The differential evolution rate of the nmth cycle is retrieved from the historical data cache. The differential evolution rate is then divided by m after performing a difference operation between the differential evolution rate of the nmth cycle and the differential evolution rate of the nmth cycle to obtain the differential evolution acceleration.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the step of retrieving the differential evolution rate of the nmth cycle from the historical data cache, performing a difference operation between the differential evolution rate and the differential evolution rate of the nmth cycle, and then dividing by m to obtain the differential evolution acceleration includes: Extract the differential evolution rate corresponding to the nmth cycle from the historical data cache by indexing the cycle number; The rate increment is obtained by subtracting the rate of differential evolution corresponding to the nth cycle from the rate of differential evolution in the nth cycle. The rate increment is then divided by the cycle interval m to obtain the differential evolution acceleration, where m is 10.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the step of calculating the capacity retention rate of the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration includes: Divide the first internal resistance difference by the initial reference value and then subtract 1 to obtain the growth factor; The first capacity loss component is obtained by multiplying the growth factor by the first weighting coefficient, the second capacity loss component is obtained by multiplying the product of the differential evolution rate and the number of cycles n by the second weighting coefficient, and the third capacity loss component is obtained by multiplying the product of the differential evolution acceleration and the square of the number of cycles n by the third weighting coefficient. The capacity retention rate for the nth cycle is obtained by subtracting the sum of the first capacity loss component, the second capacity loss component, and the third capacity loss component from the initial capacity retention rate.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the step of superimposing the product of the differential evolution rate and the number of cycles to be predicted onto the first internal resistance difference for linear extrapolation to obtain a second internal resistance difference, and solving for the remaining available number of cycles based on the second internal resistance difference and the capacity retention rate, includes: The product of the differential evolution rate and the number of cycles to be predicted is added to the first internal resistance difference to obtain the second internal resistance difference. Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to the preset target value, and solve for the positive real root to obtain the remaining number of available cycles.

[0011] Optionally, in a seventh implementation of the first aspect of the present invention, substituting the second internal resistance difference into the prediction equation for the capacity retention rate decaying to a preset target value, and solving for the positive real root to obtain the remaining usable number of cycles includes: Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to a preset target value. In the prediction equation, the coefficient of the square term of the number of cycles to be predicted is the quadratic coefficient, the coefficient of the linear term of the number of cycles to be predicted is the linear coefficient, and the sum of all terms excluding the number of cycles to be predicted is the constant term. The discriminant and two root values ​​are calculated using the root-finding formula for the prediction equation, and the positive real root is selected as the remaining number of available iterations.

[0012] Optionally, in an eighth implementation of the first aspect of the present invention, the lithium battery cycle life prediction method further includes: When the number of iterations in the nth iteration is an integer multiple of h, the internal resistance difference values ​​from the nh-1th iteration to the nth iteration (a total of h iterations) are extracted from the historical data cache to obtain the internal resistance difference value sequence, where the value of h is 50. Sort the h values ​​in the internal resistance difference sequence in ascending order, and extract the value at the j-th position after sorting as the updated initial reference value, where j is 5. Replace the currently used initial baseline value with the updated initial baseline value, recalculate the growth factor and capacity retention rate based on the updated initial baseline value, and resolve to obtain the updated remaining available cycle count.

[0013] The present invention also provides a lithium battery cycle life prediction system, comprising: The data acquisition module is used to collect charging voltage state data and discharging voltage drop data of the lithium battery during the nth cycle, and to calculate the charging internal resistance and discharging internal resistance. The first calculation module is used to subtract the charging internal resistance from the discharging internal resistance to obtain a first internal resistance difference, and to calculate the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference. The second calculation module is used to calculate the capacity retention rate of the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration. The solution module is used to superimpose the product of the differential evolution rate and the number of cycles to be predicted onto the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value, and to solve for the remaining available number of cycles based on the second internal resistance difference value and the capacity retention rate.

[0014] In summary, this invention establishes a differentiated extraction method for charging and discharging internal resistance, and uses the evolution law of lithium-ion transport path asymmetry during charging and discharging to predict battery cycle life. This invention constructs a three-level characteristic parameter system of internal resistance difference, differential evolution rate, and differential evolution acceleration. The charging internal resistance reflects the impedance of the lithium-ion insertion process, and the discharging internal resistance reflects the impedance of the deintercalation process. The difference between the two quantifies the differential influence of SEI film asymmetric growth and concentration polarization effect. The evolution rate obtained by stepwise differential calculation characterizes the dynamic trend of degradation, and the evolution acceleration captures the nonlinear accelerated degradation characteristics. The three parameters synergistically describe the static, dynamic, and accelerated characteristics of battery health state. The established capacity degradation correlation model comprehensively considers the relative growth of internal resistance difference, the cumulative effect of evolution rate, and the nonlinear contribution of acceleration. It uses a weighted summation method to quantify the influence of each factor on capacity loss. The model achieves remaining lifetime prediction through linear extrapolation and quadratic equation solving, which can more accurately capture the signal of the battery entering the accelerated degradation stage compared with traditional linear models. The dynamic correction mechanism that periodically extracts the percentile of the internal resistance difference sequence to update the health benchmark value can adapt to changes in the actual battery usage environment and maintain long-term prediction accuracy and stability. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the steps of a lithium battery cycle life prediction method in one embodiment of the present invention; Figure 2 This is a block diagram of a lithium battery cycle life prediction system according to an embodiment of the present invention.

[0016] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Reference Figure 1 This embodiment provides a method for predicting the cycle life of a lithium battery, including the following steps: S1, collect the charging voltage state data and discharging voltage drop data of the lithium battery in the nth cycle, and calculate the charging internal resistance and discharging internal resistance. In this process, lithium batteries are monitored in real time within the operating environment of industrial AGVs. When the battery enters the constant current charging stage at a constant rate, the voltage and corresponding SOC values ​​within a preset state of charge (SOC) range are continuously recorded. The preset range is set to an SOC between 40% and 50%, which, being within the linear region of the charging curve, effectively avoids interference from polarization voltage in impedance calculations. Uniform sampling at a frequency of 100Hz is used to obtain a high-resolution voltage-SOC sampling sequence. The original sampling sequence is then subjected to median filtering to eliminate spike interference, with the filtering window set to 5 sampling points. Simultaneously, the fluctuation level of the current signal within the recording range is identified, and abnormal points with current fluctuations exceeding a first preset threshold are removed to obtain charging voltage state data. When the AGV undergoes a startup process involving a sudden increase in high-rate load from a stationary state, the control system performs a rest period of at least 30 minutes before the load is applied to obtain the true open-circuit voltage. Then, within the first 5 seconds after the load is applied, the battery voltage change curve is continuously acquired using high-speed sampling. The voltage sequence includes dynamic response information after the startup current surge, with the recording focusing on the steady-state voltage in the first second after the load is applied. Median filtering is applied to the voltage sequence, and all sampling points with current fluctuations exceeding a second preset threshold are excluded to extract discharge voltage drop data. After obtaining the charging voltage state data, the voltage values ​​at SOC=40% and SOC=50% are extracted using linear interpolation. The voltage difference ΔUc is calculated and divided by the constant current charging current value to obtain the charging internal resistance Rc(n) for the nth cycle. Using the voltage difference ΔUd from the discharge stage, and substituting the temperature-corrected discharge current value, the discharge internal resistance Rd(n) for the nth cycle is calculated.

[0019] S2, subtract the charging internal resistance from the discharge internal resistance to obtain the first internal resistance difference, and calculate the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference. Specifically, in the nth cycle, the first internal resistance difference ΔR(n) of this cycle is calculated by performing a difference operation on the discharge internal resistance Rd(n) and charging internal resistance Rc(n) extracted from the lithium battery under actual charge and discharge conditions, i.e., ΔR(n) = Rd(n) - Rc(n), which reflects the electrochemical impedance asymmetry of the lithium-ion transport path during high-rate discharge and low-rate charging. The internal resistance difference ΔR(nm) corresponding to the nmth cycle n times before the current cycle is retrieved from the historical data cache, and the current ΔR(n) and the historical ΔR(nm) are differentially calculated, i.e., [ΔR(n) - ΔR(nm)] is calculated, and then divided by the cycle interval m to obtain the rate of change of internal resistance difference per unit number of cycles, i.e., the rate of difference evolution, which reflects the rate of increase of impedance asymmetry caused by factors such as SEI film growth and active material loss within the m-cycle span. To detect whether there is accelerated evolution behavior in the battery degradation process and to assess the changing trend of the differential evolution rate itself, the differential evolution rate vΔR(nm) recorded at the time of the nmth cycle is retrieved from the cache. The currently calculated vΔR(n) is then differentially calculated with the historical rate vΔR(nm) and divided by m to obtain the differential evolution acceleration aΔR(n) = [vΔR(n) - vΔR(nm)] / m, which reflects whether the growth of the asymmetric internal impedance of the battery exhibits nonlinear positive feedback characteristics. When aΔR(n) is significantly greater than zero, it means that the battery has entered a rapid degradation stage.

[0020] S3, calculate the capacity retention rate of the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration; It should be noted that dividing the first internal resistance difference ΔR(n) of the current nth cycle by ΔR(0) and subtracting 1 calculates the growth factor of the current impedance difference relative to the initial state. The growth factor reflects the degree of structural degradation caused by the thickening of the SEI film and enhanced polarization due to impedance asymmetry. Multiplying the growth factor by the preset first weighting coefficient k1 yields the first capacity loss component caused by the expansion of the absolute value of the internal resistance structural asymmetry, quantifying the active capacity decay caused by the limitation of the reversible migration path of lithium ions. Multiplying the difference evolution rate vΔR(n) by the current cycle number n represents the cumulative effect caused by the continuous growth of impedance difference over n cycles. Multiplying this result by the second weighting coefficient k2 yields the second capacity loss component, reflecting the total influence of the degradation rate over the entire cycle and demonstrating the cumulative effect of the kinetic degradation rate. Multiplying the difference evolution acceleration aΔR(n) by the square of the cycle number n amplifies the influence of acceleration on long-term capacity evolution, and then multiplying it by the third weighting coefficient k3 yields the third capacity loss component. The three capacity loss components are summed up and then subtracted sequentially from the initial capacity retention rate of 100% to obtain the capacity retention rate for the current nth cycle.

[0021] S4. The product of the differential evolution rate and the number of cycles to be predicted is superimposed on the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value. The remaining number of available cycles is then calculated based on the second internal resistance difference value and the capacity retention rate.

[0022] Specifically, based on the known first internal resistance difference ΔR(n) and the corresponding difference evolution rate vΔR(n) in the current nth cycle, a future cycle increment m is set to be predicted. It is assumed that the impedance evolution rate remains constant within the prediction period, i.e., vΔR(n) is considered locally constant, thus extrapolating the future internal resistance evolution with a linear trend. The results of multiplying the current first internal resistance difference ΔR(n) by the difference evolution rate vΔR(n) by m are accumulated to obtain the second internal resistance difference ΔR(n+m) = ΔR(n) + vΔR(n) × m in the (n+m)th cycle at the prediction time, which is the estimated state after the impedance difference increases in the next m cycles. Substituting the extrapolated second internal resistance difference ΔR(n+m) into the capacity retention prediction model, the model is defined as SOH_pred(n+m)=100%-k1×[ΔR(n+m) / ΔR(0)-1]-k2×vΔR(n)×(n+m)-k3×aΔR(n)×(n+m) 2 The formula is given in the form of ΔR(0) is the initial reference internal resistance difference, k1, k2, and k3 are capacity decay weighting coefficients obtained through experimental fitting, and SOH_pred(n+m) represents the predicted capacity retention rate at the (n+m)th cycle. When the battery capacity decreases to the end-of-life threshold, for example, set to 80%, the above formula is transformed into a quadratic equation, with m as the unknown, and the equation is in the form A×m. 2 +B×m+C=0, where A, B, and C represent the combination coefficients of differential evolution acceleration, rate, and current state, respectively. By solving the quadratic equation in the real domain and selecting the smallest positive real solution that satisfies the physical constraints, the remaining number of cycles m from the end of the lifetime in the current state is obtained, which is the desired remaining number of usable cycles.

[0023] In one example, charging voltage state data and discharging voltage drop data are collected for the nth cycle of a lithium battery, and the charging internal resistance and discharging internal resistance are calculated, including: During constant rate charging, the voltage and state of charge of the lithium battery within a preset state of charge range are recorded in real time to obtain a sampling sequence. The sampling sequence is then filtered and abnormal points where the current fluctuation exceeds the first preset threshold are removed to obtain charging voltage state data. When the AGV starts from a stationary state and is suddenly subjected to a load of a preset ratio, the stationary open circuit voltage and the transient voltage value within a preset time period after the load are continuously collected to obtain a voltage sequence. The voltage sequence is then filtered and abnormal points where the current fluctuation exceeds the second preset threshold are removed to obtain discharge voltage drop data. The charging internal resistance is calculated based on the charging voltage state data, and the discharging internal resistance is calculated based on the discharging voltage drop data.

[0024] In this example, charging voltage state data and discharging voltage drop data of the lithium battery are collected under two typical operating conditions: constant rate charging and sudden load discharging. During the charging phase, when the AGV vehicle battery system enters constant current charging, the target SOC range is identified and locked, for example, the range of 40% to 50% SOC is selected. This range is located in the linear region of the voltage-SOC curve, which has the advantages of minimizing polarization interference and enhancing impedance estimation stability. Within this range, the battery voltage and SOC value are synchronously recorded at a fixed sampling frequency (e.g., 100Hz) to obtain a continuous raw sampling sequence. Median filtering is performed on the sampling sequence, with the filtering window set to 5 points, to remove instantaneous spike interference that occurs during charging. At the same time, stability analysis is performed on the charging current to filter out all abnormal points where the current fluctuation exceeds a first preset threshold (e.g., ±5%) within the target SOC range. Voltage recording points under unstable operating conditions are also removed to obtain charging voltage state data. During the discharge phase, when the AGV starts up after being in a stable state for more than 30 minutes and then suddenly subjected to a high-rate load (such as 2C rate), the open-circuit voltage before the load is applied is recorded. Within the first 5 seconds after the load is applied, the transient voltage response sequence of the battery is acquired using high-frequency continuous sampling. The steady-state voltage at the first second after the load is applied is then monitored. Median filtering is used to denoise the voltage sequence, and sampling points with fluctuations exceeding a second preset threshold (such as ±5%) are removed based on the real-time current signal acquired through the current channel. This ensures that the extracted steady-state voltage accurately reflects the true internal resistance characteristics at the initial stage of discharge, resulting in effective discharge voltage drop data. Internal resistance calculation is performed based on charging voltage state data and discharging voltage drop data: For the extraction of charging internal resistance, the voltage values ​​at two times, SOC=40% and SOC=50%, are located using linear interpolation. The voltage difference ΔUc is calculated and divided by the constant current charging current Ic to obtain the charging internal resistance Rc. For the extraction of discharging internal resistance, the voltage difference ΔUd between the open circuit voltage and the load steady-state voltage is calculated, and the discharging internal resistance Rd is obtained by combining it with the temperature-compensated discharge current.

[0025] In one example, the charging internal resistance is calculated based on charging voltage state data, and the discharging internal resistance is calculated based on discharging voltage drop data, including: Linear interpolation is performed on the charging voltage state data to locate the first voltage value corresponding to the preset first state of charge and the second voltage value corresponding to the preset second state of charge. Divide the difference between the second voltage value and the first voltage value by the charging current to obtain the charging internal resistance; The open-circuit voltage under static conditions is extracted from the discharge voltage drop data as the static open-circuit voltage. The voltage value corresponding to a preset time after a sudden load is extracted as the steady-state voltage. The difference between the static open-circuit voltage and the steady-state voltage is divided by the discharge current to obtain the discharge internal resistance.

[0026] In this example, linear interpolation is performed on the charging voltage state data to accurately locate the voltage values ​​corresponding to two key state of charge (SOC) points. During constant-rate charging, the SOC-voltage curve of the lithium battery exhibits a good linear relationship; therefore, SOC=40% and SOC=50% are selected as the preset first and second SOC values. Using a linear interpolation algorithm, two neighboring sampling points located before and after the target SOC value are found in the original sampling points. The voltage values ​​are then weighted and averaged based on the SOC difference to calculate the precise first voltage value U1 and the second voltage value U2 corresponding to SOC=40% and SOC=50%, respectively. The difference between these two voltage values, ΔUc=U2, is used as the basis for the calculation. Using U1 as the numerator and the actual charging current Ic applied at a constant rate as the denominator, a division operation is performed to obtain the charging internal resistance Rc = ΔUc / Ic, which reflects the overall resistivity faced by lithium ions during the process of detaching from the positive electrode and migrating through the electrolyte to the negative electrode for insertion under low-rate stable charging conditions. Two key voltage indicators are extracted from the voltage drop data during the discharge process: the open-circuit voltage Urest measured by the AGV system after a resting period of more than 30 minutes before discharge start-up, which is highly representative and eliminates the influence of polarization potential; and the steady-state voltage Uload measured at the first second after a sudden application of a high-rate discharge load (such as 2C rate), at which point the lithium battery voltage has stabilized from its initial instantaneous drop, effectively reflecting the conductivity and reaction kinetics after the load impact. The difference between the open-circuit voltage Urest and the steady-state voltage Uload is the voltage drop ΔUd = Urest. Uload, then divide the voltage drop value by the actual discharge current after temperature compensation correction, to obtain the discharge internal resistance, which reflects the accumulated impedance of lithium ions during the entire process of lithium ions being extracted from the negative electrode graphite, passing through the SEI film and electrolyte to the positive electrode under high-rate, high-current rapid discharge conditions.

[0027] In one example, the first internal resistance difference is obtained by subtracting the charging internal resistance from the discharging internal resistance, and the differential evolution rate and the corresponding differential evolution acceleration are calculated based on the first internal resistance difference, including: Subtracting the charging internal resistance from the discharging internal resistance yields the first internal resistance difference for the nth cycle. Retrieve the internal resistance difference value of the nmth cycle from the historical data cache, perform a difference operation between the first internal resistance difference value of the nth cycle and the internal resistance difference value of the nmth cycle, and divide by m to obtain the differential evolution rate. Retrieve the differential evolution rate of the nmth cycle from the historical data cache, perform a difference operation between the differential evolution rate and the differential evolution rate of the nmth cycle, and divide by m to obtain the differential evolution acceleration.

[0028] In this example, the first internal resistance difference for the nth cycle is obtained by calculating the difference between the discharge internal resistance and the charging internal resistance. This difference is denoted by ΔR(n), and its calculation formula is ΔR(n) = Rd(n). Rc(n), where Rd(n) is the discharge internal resistance reflected by the steady-state voltage drop after a sudden high-rate load in the nth cycle, and Rc(n) is the charging resistance obtained by linear interpolation under constant current charging conditions. The physical significance of the first internal resistance difference lies in quantitatively revealing whether there is asymmetry in the transmission path impedance encountered by lithium ions from the positive electrode to the negative electrode and from the negative electrode back to the positive electrode in the same complete charge-discharge cycle, thereby reflecting internal degradation characteristics such as changes in SEI film thickness, electrode structure loss, or enhanced polarization effect. To analyze the dynamic evolution trend of the first internal resistance difference with the number of cycles, the record of the nth cycle is retrieved from the record stored in the system historical data cache. The internal resistance difference ΔR(n) after m cycles m), and compare the current ΔR(n) with the historical ΔR(n) m) Perform the difference operation to obtain ΔR(n). ΔR(n m), then divide the difference by the time interval m to obtain the average increase in the internal resistance difference per cycle, i.e., the rate of difference evolution vΔR(n) = [ΔR(n)]. ΔR(n [m)] / m reflects the rate characteristics of impedance structure degradation. If vΔR(n) is positive and continuously increases, it means that the SEI film on the negative electrode surface is continuously thickening or the electron migration path is obstructed due to structural collapse of the positive electrode active material, thus causing the impedance to show an increasing trend. Assess the changing trend of the differential evolution rate itself. Retrieve the nth [value] from the historical cache. The differential evolution rate vΔR(n) at m cycles m), and for the current vΔR(n) and vΔR(n) By subtracting m, we obtain vΔR(n). vΔR(n Divide the result by the same time span m to obtain the differential evolution acceleration aΔR(n) = [vΔR(n)]. vΔR(n The differential evolution acceleration reflects whether the impedance asymmetry deteriorates rapidly in the short term. When aΔR(n) is significantly greater than zero and remains at a high value, it indicates that the lithium battery has entered the degradation feedback range, that is, the impedance change begins to show nonlinear amplification characteristics. At this time, maintenance measures should be intervened as soon as possible or its service life should be reassessed to avoid sudden performance failure due to system function degradation or insufficient energy density.

[0029] In one example, the differential evolution rate of the nm-th cycle is retrieved from the historical data cache. The differential evolution rate is then subtracted from the differential evolution rate of the nm-th cycle and divided by m to obtain the differential evolution acceleration, which includes: Extract the differential evolution rate corresponding to the nmth cycle from the historical data cache by indexing the cycle number; The rate increment is obtained by subtracting the rate of differential evolution corresponding to the nth cycle from the rate of differential evolution in the nth cycle. The rate increment is then divided by the cycle interval m to obtain the differential evolution acceleration, where m is 10.

[0030] In this example, when the current iteration reaches the nth iteration, the nth iteration is retrieved from the historical data cache using the iteration number index. The differential evolution rate vΔR(n) corresponding to m cycles m), where m is a fixed time span. To improve the stability and sensitivity of evolutionary trend identification, the value of m is 10, meaning that every 10 cycles constitute one evolutionary analysis cycle. Indexing operations rely on the system's structured storage of internal resistance evolution data for each cycle, ensuring vΔR(n) m) can be directly obtained from historical records without recalculation. After extraction, the differential evolution rate vΔR(n) of the current cycle n is compared with that of the historical cycle n. The rate vΔR(n) of 10 10) Perform the difference operation to obtain the rate increment Δv = vΔR(n). vΔR(n 10) The rate increment represents the degree of change in the unit rate over the past 10 cycles, used to assess the intensification or mitigation of impedance asymmetry. Dividing the rate increment Δv by a fixed cycle interval m=10 yields the rate of change per unit number of cycles, i.e., the differential evolution acceleration aΔR(n)=Δv / 10. This acceleration parameter reflects the nonlinear trend of the degradation process of the battery's internal structure. When aΔR(n) is significantly greater than zero, it means that the SEI film growth is exponential or the active region of the electrode material is rapidly shrinking, leading to a rapid increase in the overall transmission impedance asymmetry, thereby accelerating capacity decay.

[0031] In one example, the capacity retention rate for the nth cycle is calculated based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration, including: Divide the first internal resistance difference by the initial reference value and then subtract 1 to obtain the growth factor; The first capacity loss component is obtained by multiplying the growth factor by the first weighting coefficient, the second capacity loss component is obtained by multiplying the product of the differential evolution rate and the number of cycles n by the second weighting coefficient, and the third capacity loss component is obtained by multiplying the product of the differential evolution acceleration and the square of the number of cycles n by the third weighting coefficient. The capacity retention rate for the nth cycle is obtained by subtracting the sum of the first, second, and third capacity loss components from the initial capacity retention rate.

[0032] In this example, given that the first internal resistance difference ΔR(n) and the initial reference internal resistance difference ΔR(0) corresponding to the current cycle have been obtained, a normalization operation is performed, that is, ΔR(n) is divided by ΔR(0) and then 1 is subtracted to obtain the growth factor representing the current internal resistance growth relative to the initial state. The growth factor reflects the degree of accumulation of impedance asymmetry in the charging and discharging path inside the lithium battery. Multiplying the growth factor by the first weighting coefficient k1 yields the first capacity loss component, characterizing the capacity decay effect caused by the relative increase in internal resistance difference. The difference evolution rate vΔR(n) measures the rate of increase in impedance asymmetry over time. Multiplying this by the current cycle number n and then by the second weighting coefficient k2 yields the second capacity loss component. This component represents the linear cumulative effect of impedance change rate throughout the cycle and reveals the capacity loss trend of the battery within the stable degradation range. Simultaneously, to reflect the nonlinear accelerated degradation characteristics during impedance evolution, the difference evolution acceleration aΔR(n) is multiplied by the square of the cycle number n and then multiplied by the third weighting coefficient k3 to calculate the third capacity loss component, capturing the sudden capacity drop after the battery enters the degradation feedback stage. These three capacity loss components are summed and subtracted from the initial capacity retention rate (set to 100%) to obtain the capacity retention rate SOH_pred(n) corresponding to the nth cycle, where SOH_pred(n) = 100%. k1×(ΔR(n) / ΔR(0) 1) k2×vΔR(n)×n k3×aΔR(n)×n 2 .

[0033] In one example, the product of the differential evolution rate and the number of cycles to be predicted is linearly extrapolated to the first internal resistance difference to obtain the second internal resistance difference. The remaining available number of cycles is then calculated based on the second internal resistance difference and the capacity retention rate, including: The product of the differential evolution rate and the number of cycles to be predicted is added to the first internal resistance difference to obtain the second internal resistance difference. Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to the preset target value, and solve for the positive real root to obtain the remaining number of available cycles.

[0034] In this example, the first internal resistance difference ΔR(n) obtained from the current nth cycle and the predicted future cycle number m are used as input variables. The predicted internal resistance difference ΔR(n+m) = ΔR(n) + vΔR(n) × m is constructed by multiplying the differential evolution rate vΔR(n) by m and then summing the results to ΔR(n). The second internal resistance difference reflects the level of internal resistance asymmetry reached by the battery in the (n+m)th future cycle under the assumption that the current degradation trend remains unchanged. Substituting ΔR(n+m) into the capacity retention prediction function built based on the evolution model, the function has the form SOH_pred(n+m) = 100%. k1×(ΔR(n+m) / ΔR(0) 1) k2×vΔR(n)×(n+m) k3×aΔR(n)×(n+m) 2 Where k1, k2, and k3 are capacity decay sensitivity coefficients obtained by fitting actual sample data, ΔR(0) is the initial reference benchmark for the internal resistance difference, and aΔR(n) is the difference evolution acceleration at the current moment. The prediction function is used to solve for the number of cycles m that makes SOH_pred(n+m) equal to the preset capacity termination threshold SOH_end. Setting SOH_end to 80% indicates that the battery has reached an unacceptable lower capacity limit. Substituting 80% and rearranging the formula, a quadratic equation in m is obtained, in the form A×m. 2 +B×m+C=0, where A, B, and C are composite coefficients calculated from existing parameters. The positive real root of the equation is the additional number of cycles required to reduce the capacity retention rate to the target threshold. Using the standard root-finding formula m=[ B+√(B 2 Solve using 4AC)] / (2A), retaining only positive real solutions as valid solutions. Use the obtained value of m as the remaining available iterations in the current loop state.

[0035] In one example, the second internal resistance difference is substituted into the prediction equation for the capacity retention rate decaying to a preset target value, and the positive real root is solved to obtain the remaining available number of cycles, including: Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to the preset target value. In the prediction equation, the coefficient of the square term of the number of cycles to be predicted is the quadratic term coefficient, the coefficient of the linear term of the number of cycles to be predicted is the linear term coefficient, and the sum of all terms excluding the number of cycles to be predicted is the constant term. The discriminant and two root values ​​are calculated using the quadratic formula for the prediction equation, and the positive real root is selected as the remaining number of available iterations.

[0036] In this example, based on the current state of the nth cycle, the second internal resistance difference ΔR(n+m) = ΔR(n) + vΔR(n) × m is obtained by multiplying the differential evolution rate vΔR(n) by the remaining number of cycles to be predicted m and adding it to the current first internal resistance difference ΔR(n). Substituting ΔR(n+m) into the capacity prediction equation SOH_pred(n+m) = 100% k1×(ΔR(n+m) / ΔR(0) 1) k2×vΔR(n)×(n+m) k3× aΔR(n)×(n+m) 2 Where k1, k2, and k3 represent the sensitivity weighting coefficients of capacity to internal resistance growth, evolution rate, and evolution acceleration, respectively; ΔR(0) is the initial baseline internal resistance difference; and aΔR(n) is the current differential evolution acceleration. Substituting both sides of the equation into the capacity termination threshold SOH_end (e.g., 80%) and rearranging the terms, the prediction model is transformed into a standard quadratic equation, i.e., A×m 2 +B×m+C=0, where the coefficients A of the quadratic term include k3×aΔR(n) and k1×vΔR(n). 2 The product terms, such as / ΔR(0), are used to form the linear term coefficient B, which is composed of k2×vΔR(n), k1×vΔR(n) / ΔR(0), and other linear terms related to n. The constant term C includes k1×ΔR(n) / ΔR(0), k2×vΔR(n)×n, and k3×aΔR(n)×n. 2 The sum of all terms unrelated to m, including the difference between the initial capacity and SOH_end. After forming the standard quadratic equation, use the quadratic formula m=[ B±√(B 2 Analytically solve 4AC)] / (2A) and calculate the discriminant D=B. 2 4AC is used to determine whether there is a real root. If D is less than zero, it means that the current state can no longer be maintained to the target capacity level, and the life is determined to end. Let m be 0. If D is greater than or equal to zero, then continue to calculate two root values ​​and select the positive real root that satisfies the physical meaning as the remaining number of available cycles.

[0037] In one example, the lithium battery cycle life prediction method also includes: When the number of iterations in the nth iteration is an integer multiple of h, the internal resistance difference values ​​from the nh-1th iteration to the nth iteration (a total of h iterations) are extracted from the historical data cache to obtain the internal resistance difference value sequence, where the value of h is 50. Sort the h values ​​in the internal resistance difference sequence from smallest to largest, and extract the value at the j-th position after sorting as the updated initial reference value, where j is 5. Replace the currently used initial baseline value with the updated initial baseline value, recalculate the growth factor and capacity retention rate based on the updated initial baseline value, and resolve to obtain the updated remaining available cycle count.

[0038] In this example, after each nth iteration, it checks if n is an integer multiple of h. If the condition n mod h = 0 is met, the baseline update process is triggered, where h is set to 50, indicating that a baseline calibration operation is performed every 50 complete iterations. At this point, the baseline data is retrieved from the historical data cache from the nth iteration. h The internal resistance difference ΔR(n) corresponding to h cycles from the 1st to the nth cycle. h 1) ΔR(n) h), ..., ΔR(n) form an internal resistance difference sequence of length h. The internal resistance difference sequence is sorted in ascending order, i.e., the values ​​are reorganized in ascending order. After sorting, the value at the j-th position is selected as the new initial baseline value, where j is set to 5, indicating that the 5th smallest internal resistance difference value is selected as the approximate lower limit representing the health status within the interval. After the baseline update, the previously used initial baseline value ΔR(0) is replaced with the updated initial baseline value, and the growth factor of the current n-th cycle is recalculated using this as the denominator. The updated growth factor is substituted into the capacity retention rate prediction model, and together with the differential evolution rate vΔR(n) and acceleration aΔR(n), forms the SOH prediction expression, thereby calculating the new capacity retention rate SOH_pred(n). Based on the new SOH_pred(n) value and the updated initial baseline value, the prediction equation is reconstructed and the capacity termination threshold SOH_end=80% is substituted. The quadratic equation is then re-expanded and the root-finding operation is performed to obtain the updated remaining number of available loops.

[0039] The calculation process includes calculating the charging internal resistance based on charging voltage state data and the discharging internal resistance based on discharging voltage drop data. Following this, a temperature-rate coupling compensation correction step is performed on the charging and discharging internal resistances. This involves simultaneously acquiring the real-time temperature and actual charge / discharge rate of the industrial AGV lithium battery during the calculation of charging and discharging internal resistances to obtain a temperature-rate operating condition parameter pair. A baseline charging internal resistance and a baseline discharging internal resistance under standard operating conditions are extracted from a pre-established temperature-rate-internal resistance three-dimensional mapping database. The ratio of the charging internal resistance to the baseline charging internal resistance is multiplied by a normalization coefficient for the charging internal resistance at a standard temperature of 25 degrees Celsius and a standard charging rate to obtain the temperature-compensated charging internal resistance. The ratio of the discharging internal resistance to the baseline discharging internal resistance is then calculated. The ratio of the standard value is multiplied by the normalization coefficient of the discharge internal resistance at the standard temperature of 25 degrees Celsius and the standard discharge rate to obtain the temperature-compensated discharge internal resistance. The charging internal resistance after temperature compensation is corrected by multiplying the difference between the current actual charging rate and the standard charging rate by the charging rate correction coefficient and then adding it to the temperature-compensated charging internal resistance to obtain the temperature-rate coupled compensated charging internal resistance. The discharge internal resistance after temperature compensation is corrected by multiplying the difference between the current actual discharge rate and the standard discharge rate by the discharge rate correction coefficient and then adding it to the temperature-compensated discharge internal resistance to obtain the temperature-rate coupled compensated discharge internal resistance. The temperature-rate coupled compensated charging internal resistance and the temperature-rate coupled compensated discharge internal resistance replace the original charging internal resistance and the original discharge internal resistance, respectively.

[0040] The process includes calculating the capacity retention rate of the nth cycle based on the first internal resistance difference, the rate of differential evolution, and the acceleration of differential evolution. It also includes a step of classifying and warning about battery health status based on multi-dimensional evolution characteristics: establishing a health status classification threshold matrix. This matrix includes first-level, second-level, and third-level thresholds for the relative growth factor of the internal resistance difference; first-level, second-level, and third-level thresholds for the rate of differential evolution; and first-level, second-level, and third-level acceleration thresholds for the acceleration of differential evolution. The first-level threshold corresponds to a good health status, the second-level threshold corresponds to a health status warning, and the third-level threshold corresponds to a severely degraded health status. The relative growth factor of the internal resistance difference, the rate of differential evolution, and the acceleration of differential evolution in the current nth cycle are compared with the corresponding thresholds in the threshold matrix. When the relative growth factor of the internal resistance difference exceeds the third-level threshold, or the rate of differential evolution exceeds the threshold... When either the third rate threshold or the differential evolution acceleration exceeds the third acceleration threshold is met, the battery health status is determined to be at the severe degradation level. When the relative growth factor of the internal resistance difference is between the second and third thresholds, and the differential evolution rate exceeds the second rate threshold, and the differential evolution acceleration exceeds the second acceleration threshold, the battery health status is determined to be at the warning level. Based on the determined health status level, a corresponding warning signal is generated. When the health status level is severe degradation, a high-priority warning signal is output and an immediate battery replacement plan is recommended. When the health status level is warning, a medium-priority warning signal is output and preventive maintenance is recommended within a preset number of cycles. When the health status level is good, normal monitoring continues. The health status level and the corresponding warning signal, along with the capacity retention rate and remaining available cycle count of the nth cycle, are output to the battery management system of the industrial AGV.

[0041] The process involves calculating the discriminant and two root values ​​using the root-finding formula for the prediction equation, selecting the positive real root as the remaining usable cycle count, and then performing prediction correction based on historical evolution trajectory similarity matching. This includes extracting reference battery samples that have completed full lifecycle testing from the historical degradation database of lithium batteries used in industrial AGVs from the same batch. Each reference battery sample contains an internal resistance difference evolution sequence, a difference evolution rate sequence, and the actual total number of cycles from the initial stage to the end of the lifecycle. The process also involves extracting the internal resistance difference evolution sequence and the difference evolution rate sequence of the battery to be predicted from the initial stage to the nth cycle, arranging the internal resistance difference evolution sequence in chronological order to form the current evolution trajectory vector, and calculating the Euclidean distance between the current evolution trajectory vector and the evolution trajectory vector of each reference battery sample corresponding to the cycle count segment. Finally, the calculated Euclidean distances are sorted, and the k reference battery samples with the smallest Euclidean distances are selected as the similarity sample set, where k ranges from 3 to 10. Calculate the similarity weight coefficient for each similar sample. The similarity weight coefficient is the ratio of the reciprocal of the Euclidean distance of the similar sample to the sum of the reciprocals of the Euclidean distances of all similar samples. Extract the actual remaining number of cycles for each similar sample in the similar sample set at the current nth cycle time. Multiply the actual remaining number of cycles for each similar sample by the corresponding similarity weight coefficient and sum them to obtain the reference remaining number of cycles based on historical trajectory matching. Perform a weighted fusion of the remaining available number of cycles and the reference remaining number of cycles. The fusion weight is dynamically adjusted according to the proportion of the current battery cycle count to the expected total cycle count. When the cycle count is less than 30%, the remaining available number of cycles is given a higher weight. When the cycle count is greater than 70%, the reference remaining number of cycles is given a higher weight, resulting in the corrected final remaining available number of cycles. Store the final remaining available number of cycles in the historical data cache and output it to the battery management system of the industrial AGV.

[0042] The process includes subtracting the charging internal resistance from the discharging internal resistance to obtain the first internal resistance difference value for the nth cycle, followed by a step of dynamically correcting the internal resistance difference based on the temperature-state-of-charge coupling effect. This involves simultaneously collecting the measured temperature value of the industrial AGV lithium battery, the average state-of-charge value during the charging phase, and the state-of-charge value at the discharging moment when calculating the charging and discharging internal resistances, thus obtaining a temperature-state-of-charge parameter set. The charging internal resistance temperature correction coefficient and the discharging internal resistance temperature correction coefficient, corresponding to the measured temperature values, are extracted from a pre-established temperature-internal resistance correlation database. The charging internal resistance temperature correction coefficient characterizes the influence of a temperature deviation of 25 degrees Celsius from the standard temperature on the measured charging internal resistance value, and the discharging internal resistance temperature correction coefficient characterizes the influence of a temperature deviation from the standard temperature on the measured discharging internal resistance value. The charging internal resistance is multiplied by the charging internal resistance temperature correction coefficient to obtain the temperature-normalized charging internal resistance, and the discharging internal resistance is multiplied by the discharging internal resistance temperature correction coefficient to obtain the temperature-normalized discharging internal resistance. Finally, the process involves extracting the nth-th internal resistance difference value from the historical data cache. The depth of discharge (DHC) value of the first cycle and the state of charge (SOC) value at the end of discharge are used to calculate the historical influence factor of the discharge process in the (n-1)th cycle on the internal resistance measurement in the nth cycle. The historical influence factor is equal to the product of the DHC value of the (n-1)th cycle and the historical influence weight of the internal resistance. The temperature-normalized internal resistance is subtracted from the historical influence factor to obtain the historically corrected internal resistance. The difference between the average SOC value during the charging phase and the SOC value at the discharge moment is calculated as the charge-discharge electrical state span. The SOC coupling correction coefficient corresponding to the charge-discharge electrical state span is extracted from the SOC-internal resistance difference association database. The SOC coupling correction coefficient characterizes the degree of influence of the charging and discharging process on the internal resistance difference when measured in different SOC intervals. The temperature-normalized discharge internal resistance is subtracted from the historically corrected charging internal resistance to obtain the preliminary internal resistance difference value. The preliminary internal resistance difference value is multiplied by the SOC coupling correction coefficient to obtain the corrected internal resistance difference value. The corrected internal resistance difference value replaces the first internal resistance difference value.

[0043] Reference Figure 2 This embodiment provides a lithium battery cycle life prediction system, including: The acquisition module 1 is used to acquire the charging voltage state data and discharging voltage drop data of the lithium battery during the nth cycle, and to calculate the charging internal resistance and discharging internal resistance. The first calculation module 2 is used to subtract the charging internal resistance from the discharge internal resistance to obtain the first internal resistance difference, and to calculate the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference. The second calculation module 3 is used to calculate the capacity retention rate of the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration. The solver module 4 is used to superimpose the product of the differential evolution rate and the number of cycles to be predicted onto the first internal resistance difference for linear extrapolation to obtain the second internal resistance difference, and to solve for the remaining number of cycles based on the second internal resistance difference and the capacity retention rate.

[0044] In this embodiment, the specific implementation of each unit in the above system embodiment is described in the above method embodiment, and will not be repeated here.

[0045] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, system, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, system, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, system, article, or method that includes that element.

[0046] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for predicting the cycle life of a lithium battery, characterized in that, include: Collect charging voltage state data and discharging voltage drop data for the nth cycle of the lithium battery, and calculate the charging internal resistance and discharging internal resistance. The first internal resistance difference is obtained by subtracting the charging internal resistance from the discharge internal resistance, and the differential evolution rate and the corresponding differential evolution acceleration are calculated based on the first internal resistance difference. The capacity retention rate of the nth cycle is calculated based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration. The product of the differential evolution rate and the number of cycles to be predicted is superimposed on the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value. The remaining number of available cycles is then calculated based on the second internal resistance difference value and the capacity retention rate.

2. The lithium battery cycle life prediction method according to claim 1, characterized in that, The process of collecting charging voltage state data and discharging voltage drop data for the nth cycle of the lithium battery, and calculating charging internal resistance and discharging internal resistance, includes: During constant rate charging, the voltage and state of charge of the lithium battery within a preset state of charge range are recorded in real time to obtain a sampling sequence. The sampling sequence is then filtered and abnormal points where the current fluctuation exceeds a first preset threshold are removed to obtain charging voltage state data. When the AGV starts from a stationary state and is suddenly subjected to a load of a preset ratio, the stationary open circuit voltage and the transient voltage value within a preset time period after the load are continuously collected to obtain a voltage sequence. The voltage sequence is then filtered and abnormal points where the current fluctuation exceeds the second preset threshold are removed to obtain discharge voltage drop data. The charging internal resistance is calculated based on the charging voltage state data, and the discharging internal resistance is calculated based on the discharging voltage drop data.

3. The lithium battery cycle life prediction method according to claim 2, characterized in that, The calculation of charging internal resistance based on the charging voltage state data and the calculation of discharging internal resistance based on the discharging voltage drop data include: Linear interpolation is performed on the charging voltage state data to locate the first voltage value corresponding to the preset first state of charge and the second voltage value corresponding to the preset second state of charge. Divide the difference between the second voltage value and the first voltage value by the charging current to obtain the charging internal resistance; The open-circuit voltage under static conditions is extracted from the discharge voltage drop data as the static open-circuit voltage. The voltage value corresponding to a preset time after a sudden load is applied is extracted as the steady-state voltage. The difference between the static open-circuit voltage and the steady-state voltage is divided by the discharge current to obtain the discharge internal resistance.

4. The method for predicting the cycle life of a lithium battery according to claim 1, characterized in that, The step of subtracting the charging internal resistance from the discharging internal resistance to obtain a first internal resistance difference, and calculating the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference, includes: Subtracting the charging internal resistance from the discharge internal resistance yields the first internal resistance difference for the nth cycle. Retrieve the internal resistance difference value of the nth cycle from the historical data cache, perform a difference operation between the first internal resistance difference value of the nth cycle and the internal resistance difference value of the nth cycle, and divide by m to obtain the differential evolution rate. The differential evolution rate of the nmth cycle is retrieved from the historical data cache. The differential evolution rate is then divided by m after performing a difference operation between the differential evolution rate of the nmth cycle and the differential evolution rate of the nmth cycle to obtain the differential evolution acceleration.

5. The lithium battery cycle life prediction method according to claim 4, characterized in that, The step of retrieving the differential evolution rate of the nmth cycle from the historical data cache, performing a difference operation between the differential evolution rate and the differential evolution rate of the nmth cycle, and then dividing by m to obtain the differential evolution acceleration includes: Extract the differential evolution rate corresponding to the nmth cycle from the historical data cache by indexing the cycle number; The rate increment is obtained by subtracting the rate of differential evolution corresponding to the nth cycle from the rate of differential evolution in the nth cycle. The rate increment is then divided by the cycle interval m to obtain the differential evolution acceleration, where m is 10.

6. The method for predicting the cycle life of a lithium battery according to claim 1, characterized in that, The calculation of the capacity retention rate for the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration includes: Divide the first internal resistance difference by the initial reference value and then subtract 1 to obtain the growth factor; The first capacity loss component is obtained by multiplying the growth factor by the first weighting coefficient, the second capacity loss component is obtained by multiplying the product of the differential evolution rate and the number of cycles n by the second weighting coefficient, and the third capacity loss component is obtained by multiplying the product of the differential evolution acceleration and the square of the number of cycles n by the third weighting coefficient. The capacity retention rate for the nth cycle is obtained by subtracting the sum of the first capacity loss component, the second capacity loss component, and the third capacity loss component from the initial capacity retention rate.

7. The method for predicting the cycle life of a lithium battery according to claim 1, characterized in that, The step of superimposing the product of the differential evolution rate and the number of cycles to be predicted onto the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value, and calculating the remaining available number of cycles based on the second internal resistance difference value and the capacity retention rate, includes: The product of the differential evolution rate and the number of cycles to be predicted is added to the first internal resistance difference to obtain the second internal resistance difference. Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to the preset target value, and solve for the positive real root to obtain the remaining number of available cycles.

8. The method for predicting the cycle life of a lithium battery according to claim 7, characterized in that, The step of substituting the second internal resistance difference into the prediction equation for the capacity retention rate decaying to a preset target value, and solving for the positive real root to obtain the remaining usable cycle count includes: Substitute the second internal resistance difference into the prediction equation for the capacity retention rate decaying to a preset target value. In the prediction equation, the coefficient of the square term of the number of cycles to be predicted is the quadratic coefficient, the coefficient of the linear term of the number of cycles to be predicted is the linear coefficient, and the sum of all terms excluding the number of cycles to be predicted is the constant term. The discriminant and two root values ​​are calculated using the quadratic formula for the prediction equation, and the positive real root is selected as the remaining number of available iterations.

9. The method for predicting the cycle life of a lithium battery according to claim 1, characterized in that, The lithium battery cycle life prediction method also includes: When the number of iterations in the nth iteration is an integer multiple of h, the internal resistance difference values ​​from the nh-1th iteration to the nth iteration (a total of h iterations) are extracted from the historical data cache to obtain the internal resistance difference value sequence, where the value of h is 50. Sort the h values ​​in the internal resistance difference sequence in ascending order, and extract the value at the j-th position after sorting as the updated initial reference value, where j is 5. Replace the currently used initial baseline value with the updated initial baseline value, recalculate the growth factor and capacity retention rate based on the updated initial baseline value, and resolve to obtain the updated remaining available cycle count.

10. A lithium battery cycle life prediction system, characterized in that, The steps for implementing the lithium battery cycle life prediction method according to any one of claims 1 to 9 include: The data acquisition module is used to collect charging voltage state data and discharging voltage drop data of the lithium battery during the nth cycle, and to calculate the charging internal resistance and discharging internal resistance. The first calculation module is used to subtract the charging internal resistance from the discharging internal resistance to obtain a first internal resistance difference, and to calculate the differential evolution rate and the corresponding differential evolution acceleration based on the first internal resistance difference. The second calculation module is used to calculate the capacity retention rate of the nth cycle based on the first internal resistance difference, the differential evolution rate, and the differential evolution acceleration. The solution module is used to superimpose the product of the differential evolution rate and the number of cycles to be predicted onto the first internal resistance difference value for linear extrapolation to obtain the second internal resistance difference value, and to solve for the remaining available number of cycles based on the second internal resistance difference value and the capacity retention rate.

Images