Off-line rapid identification method for lithium battery parameters

Abstract

The invention provides an off-line rapid identification method for lithium battery parameters, which belongs to the technical field of lithium batteries and is characterized in that model simplification is realized through quantification of transient change characteristics on the basis of a first-order equivalent circuit model, and in a single step transient response stage, a time domain transient change rule is extracted by using measurement data of a plurality of sampling points, so that the identification accuracy of the parameters of the lithium battery is improved. The battery polarization is ignored on a working condition point where the initial value of the transient voltage is 0, and on the premise that the polarization resistance and the polarization capacitance are kept constant within the calculation time, the internal relation among the R0, the Rp and the TVR is obtained according to the dynamic equation; according to the method, calculation of R0 and Rp is simplified into solving of a linear equation set composed of two equations, the sampling point number condition, the sampling period condition and the TVR condition of rapid measurement of the lithium battery parameters are analyzed, response data of the lithium battery are collected, and the lithium battery parameters and the lithium battery terminal voltage are calculated through a deduced calculation formula.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium batteries, and in particular, to an offline fast identification method for lithium battery parameters. Background Art

[0002] When lithium batteries leave the factory and are retired, they need to be capacity-determined and classified, which is not only related to the performance of lithium batteries, but also involves environmental protection and the effectiveness of resource recovery. Due to the immaturity of related technologies, lithium batteries face many safety hazards during operation, such as thermal runaway, short circuit, and overcharging. Therefore, accurate identification and detection of lithium battery-related parameters have become the key technologies for their application and development. However, in scenarios such as capacity determination, classification, and maintenance of lithium batteries, the detection objects are mostly offline lithium batteries or lithium battery packs that are not connected to loads or charging devices. Therefore, the offline method for quickly and accurately identifying lithium battery parameters has gradually become an important engineering requirement.

[0003] Short detection time, few original test data, and unknown lithium battery status are the main characteristics of the fast identification application scenario of lithium batteries. The offline identification method requires experimental data obtained in advance, and based on the fitting of the lithium battery equivalent model and actual data or other calculation conditions to obtain lithium battery parameter information. Its identification results are usually fixed values, mappings, or functional relationships. The offline identification methods mainly include the current step method and the electrochemical impedance spectroscopy method. The current step method and the electrochemical impedance spectroscopy method obtain lithium battery parameters based on specific calculation and analysis conditions. The current step method calculates the ohmic internal resistance through the voltage transient generated by the current step and can quickly obtain the ohmic internal resistance parameter of the lithium battery. The electrochemical impedance spectroscopy method can directly give the impedance of the lithium battery at different working frequency points according to the electrochemical impedance model of the lithium battery, with strong pertinence, and is recognized as a method that can accurately reflect the impedance of the lithium battery. The online identification method and the offline method rely on the same basic model and can realize online real-time identification of parameters. Typical online algorithms include the algorithm based on Kalman filter KF, the algorithm based on LS, the co-evolutionary particle swarm optimization CPSO algorithm based on PSO, and the data-driven algorithm. Among them, the data-driven method is a parameter identification method based on the data-driven model, and can directly obtain the corresponding lithium battery equivalent parameters by measuring the external characteristics of the lithium battery. Due to the limited data volume that can be provided in the fast identification scenario, the application of data-driven algorithms that require a large amount of historical data for training is restricted. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention provides an offline fast identification method for lithium battery parameters. According to the measured values of a small number of sampling points in a single-step transient process, a quantization result TVR of the time-domain transient change rule is obtained, so that the change rule of the step transient is quantified into a single-value TVR, simplifying the calculation of the ohmic internal resistance and the polarization internal resistance into the solution of a linear equation system composed of two equations, thereby realizing fast parameter identification, greatly simplifying the traditional calculation process, and making the evaluation of the performance of lithium batteries more efficient in practical applications.

[0005] The present invention provides an offline fast identification method for lithium battery parameters, including the following steps: Step 1: Construct an offline measurement system for lithium batteries and design the working conditions of lithium batteries according to the actual requirements of fast measurement.

[0006] Step 2: Under the designed working conditions of lithium batteries, use the offline measurement system for lithium batteries to perform charge and discharge tests on the lithium battery to be measured, and obtain the transient response data of the lithium battery after the current mutation.

[0007] Step 3: Based on the first-order RC equivalent circuit model, deduce the equivalent circuit state equation and output equation of the lithium battery, and analyze the transient process of the lithium battery terminal voltage under the condition of current step change to determine the equivalent circuit state equation and output equation of the lithium battery in the step transient process and the terminal voltage change formula of sampling points at different times. Furthermore, analyze the change rule of the lithium battery in the step transient process to obtain the TVR calculation formula for the quantization result of the time-domain transient change rule.

[0008] Step 4: On the premise that the resistance and polarization capacitance remain constant during the calculation period and the initial value of the transient voltage is ignored, deduce the relationship between the lithium battery parameters and the quantization result TVR of the time-domain transient change rule according to the equivalent circuit state equation and output equation of the lithium battery in the step transient process and the terminal voltage change formula of sampling points at different times, so as to use the transient response data of the lithium battery at several sampling points for continuous state analysis of the lithium battery and fast calculation of the lithium battery parameters. Among them, the lithium battery parameters include ohmic internal resistance, polarization resistance, and polarization capacitance.

[0009] Step 5: Analyze the influence of the number of sampling points of sampling points at different times from the starting point on the quantization result TVR and the measurement result of lithium battery parameters, and obtain the sampling point number condition for fast measurement of lithium battery parameters.

[0010] Step 6: Analyze the influence of different sampling periods on the quantization result TVR and the measurement result of lithium battery parameters, and obtain the sampling period condition for fast measurement of lithium battery parameters.

[0011] Step 7: Combine Step 5 and Step 6, compare the method with the current step method and the measurement results of lithium battery parameters under the full SOC region of the co-evolutionary particle swarm optimization algorithm, and obtain the fast parameter measurement conditions that can effectively balance the calculation efficiency and accuracy.

[0012] Step 8: Preset the sampling point conditions and sampling period conditions for fast measurement of lithium battery parameters in the lithium battery offline measurement system, and collect the lithium battery response data through the set lithium battery parameter offline measurement system, and calculate the lithium battery parameters and the lithium battery terminal voltage according to the calculation formulas deduced in Steps 3 and 4.

[0013] Step 9: Verify the accuracy of the lithium battery parameters according to the lithium battery parameter calculation results and the corresponding lithium battery terminal voltage calculation results, and analyze the factors affecting the accuracy of the offline fast measurement of lithium battery parameters.

[0014] Optionally, in Step 3, based on the first-order RC equivalent circuit model, deduce the equivalent circuit state equation and output equation of the lithium battery, including: Use the first-order RC equivalent circuit model to simulate the internal characteristics of the lithium battery, u OC is the voltage of the ideal voltage source, used to simulate the open-circuit voltage of the lithium battery, u L is the terminal voltage, R0 is the ohmic resistance, R p is the polarization internal resistance, C p is the polarization capacitance, u p is the polarization voltage, i L is the load current. When a constant current is input, the current reference direction for charging is set as positive. The transient component of the terminal voltage is generated under the action of the current source, and a step current will be generated at the end of charging, forming a current difference i d , i d determines the form of the transient response of the lithium battery.

[0015] According to Kirchhoff's law, Equations (1) and (2) are the continuous state equation and output equation of the equivalent circuit respectively: (1).

[0016] (2).

[0017] Equations (3) and (4) are the discrete state equation and output equation of the equivalent circuit respectively: (3).

[0018] (4).

[0019] In the formula, τ = R p C p , is the time constant, and Ts is the sampling period.

[0020] Furthermore, in step 3, under the condition of current step change, analyze the transient process of the terminal voltage of the lithium battery to determine the equivalent circuit state equation and output equation of the lithium battery in the step transient process, as well as the terminal voltage change formula at different sampling points, and then analyze the change law of the lithium battery in the step transient process to obtain the quantization result TVR calculation formula of the change law in the time domain transient, including: Analyze the step transient process of the lithium battery based on the current step change condition to obtain i d The step transient process curve when it is positive. Take the 1 sampling point before the current transient as the starting point and the 1 sampling point before the next current transient as the ending point to intercept a step transient load step, and subtract the voltage and current data at the starting point from the overall voltage and current.

[0021] Take the starting point as a constant reference point, and change the sampling period Ts in formula (3) to a time-varying factor t. The t of the constant reference point is 0. The circuit continuous state and output equations of the step dynamic model are as follows: (5).

[0022] (6).

[0023] In the formula, u p0 is the initial value of the polarization voltage, u pd is the change amount of the polarization voltage, u d is the change amount of the terminal voltage, i d is the change amount of the current, u OCd is the change amount of the open-circuit voltage.

[0024] Formula (6) includes the change law of voltage in the transient process , and the state before the current transient is concentrated in u p0 . Transform the Ohm internal resistance calculation formula (7) of the current step method into formula (8): (7).

[0025] (8).

[0026] Assume that within nTs, R p , C p are fixed values, then the following formulas hold at the three sampling points of t = nTs, (n - 1)Ts, and (n - 2)Ts: (9).

[0027] (10).

[0028] The difference in the output voltage between two adjacent sampling points determined by Equation (10) is proportional to E. The quantization result TVR of the variation law of the time-domain transient is derived from Equation (10): (11).

[0029] Among them, n>2, which is the number of points of the sampling point from the starting point, defined as the distance points of the sampling point, and nTs is a certain sampling time.

[0030] Furthermore, in step 4, according to the equivalent circuit state equation and output equation of the lithium battery during the step transient process and the terminal voltage change formula of the sampling points at different times, the relationship between the lithium battery parameters and the quantization result TVR of the variation law of the time-domain transient is derived, including: As can be seen from Equation (9), the state at the latter sampling time is related to the state at the previous sampling time, and E is the key parameter. Derived from Equation (10), we get:[[]] (12).

[0031] (13).

[0032] Analyze the step transient process curve when i d is positive and R0 and u determined by Equations (5) and (6) p0 jointly cause the instantaneous voltage change. Combining Equation (12) and Equation (13), we get:[[]] (14).

[0033] Equation (14) is the calculation formula for the voltage transient variable at the 0 moment of the current step. In the lithium battery off-line detection scenario, the lithium battery mostly undergoes long-term static treatment, and u p0 can be regarded as 0. When u p0 is not 0, the separation of the ohmic internal resistance and the polarization internal resistance is realized, then the calculation formulas of R0, R p , C p can be expressed as follows:[[]] (15).

[0034] (16).

[0035] (17).

[0036] After adopting the above technical solutions, the present invention has at least the following beneficial effects: (1) Based on the measured values of a small number of sampling points in a single-step transient process, the present invention obtains the quantization result TVR of the time-domain transient change law, thereby quantizing the change law of the step transient into a single-valued TVR, simplifying the calculation of the ohmic internal resistance and the polarization internal resistance into solving a linear equation system composed of two equations, and thus realizing fast parameter identification. This method greatly simplifies the traditional calculation process and makes the evaluation of the performance of lithium batteries more efficient in practical applications.

[0037] (2) Different from calculating the ohmic internal resistance by the current step method, the off-line fast parameter extraction studied in the present invention can effectively separate the ohmic internal resistance and the polarization internal resistance. Among them, the key to separating the polarization internal resistance is: ignoring the polarization of the lithium battery at the operating point where the initial value of the transient voltage is 0, extracting the change amount of the transient voltage at different stages of the transient process, and the polarization internal resistance can be directly obtained through the change amount of the transient voltage and the TVR.

[0038] (3) The present invention gives clear algorithm conditions, taking three factors of TVR, the number of sampling points, and the sampling period as important factors affecting the parameter identification result, enabling the key data information in the dynamic process to be accurately reflected in the identification result. In the shortest possible detection time, using a custom working condition and the method of the present invention to identify parameters, ensuring that the identification result of the ohmic internal resistance parameter of the lithium battery has high accuracy and stability in engineering applications. In addition, the factors affecting the characteristics of TVR are analyzed, the relationship between TVR and other parameters is obtained, and the parameter identification process of the lithium battery is further optimized. Brief Description of the Drawings

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0040] Figure 1 is a first-order RC equivalent circuit.

[0041] Figure 2 is a scatter plot of the comparison between the predicted value and the actual value of the model paleo-elevation.

[0042] Figure 3 is i d is the step transient process curve when it is positive.

[0043] Figure 4 is a derivation process diagram for separating the parameters of the lithium battery in the step transient process.

[0044] Figure 5For the dynamic test and load cycle results, (a) shows the overall operating condition test results of 13 groups of discharge processes, and (b) shows 26 load steps in a single load condition cycle.

[0045] Figure 6 For Ts = 0.1 s and n = 3, the ohmic internal resistance obtained by the method of the present invention and the ohmic internal resistance curve obtained by the current step method, and the TVR curve obtained by the method of the present invention. (a) shows the ohmic internal resistance R obtained by the method of the present invention d and the ohmic internal resistance R3 curve obtained by the current step method. (b) shows the TVR curve obtained by the method of the present invention.

[0046] Figure 7 For Ts = 0.1 s and different n values, the ohmic internal resistance obtained by the method of the present invention and the ohmic internal resistance curve obtained by the current step method, and the TVR curve obtained by the method of the present invention. (a) shows R d and the R6 curve. (b) shows the TVR curve for n = 6. (c) shows R d and R 10 curve. (d) shows the TVR curve for n = 10. (e) shows R d and R 20 curve. (f) shows the TVR curve for n = 20. (g) shows R d and R max curve. (h) shows the TVR curve for n = MAX.

[0047] Figure 8 For n = 10 and different Ts values, the ohmic internal resistance obtained by the method of the present invention and the ohmic internal resistance curve obtained by the current step method, and the TVR curve obtained by the method of the present invention. (a) shows R d and the R0 curve when Ts = 0.05 s. (b) shows R d and the R0 curve when Ts = 0.1 s. (c) shows R d and the R0 curve when Ts = 0.2 s. (d) shows R d and the R0 curve when Ts = 0.4 s. (e) shows the ohmic internal resistance curve obtained by the method of the present invention under different step load conditions when Ts = 0.1 s. (f) shows the TVR curve obtained by the method of the present invention.

[0048] Figure 9 For n = 10 and Ts = 0.1 s, the R p curve and C p curve obtained by the method of the present invention. (a) shows the R p curve. (b) shows the C p curve.

[0049] Figure 10For n = 10, the curves of the ohmic internal resistance obtained by the method of the present invention at different Ts and the ohmic internal resistance obtained by the current step method. (a) is the ohmic internal resistance R obtained by the method of the present invention d curve, and (b) is the curve of the ohmic internal resistance R0 obtained by the current step method.

[0050] Figure 11 are the CPSO fitting results and errors. (a) is the fitting result at high SOC, (b) is the fitting error at high SOC, (c) is the fitting result at medium SOC, (d) is the fitting error at medium SOC, (e) is the fitting result at low SOC, and (f) is the fitting error at low SOC.

[0051] Figure 12 are the results of R0 and R in the medium SOC region p Results. (a) are the R0 results in the medium SOC region of the three methods of CPSO, the method of the present invention, and the current step method. (b) are the R p results.

[0052] Figure 13 are the calculation results and errors of the terminal voltage of the single-load curve. (a) are the calculation results of the terminal voltage, (b) are the calculation errors of the terminal voltage, and (c) are the specific values of the three errors of MEAN, ME, and RMSE.

[0053] Figure 14 are the calculation results and errors of the terminal voltage in the first 1 s of the step transient of the single-load curve. (a) are the calculation results of the terminal voltage, (b) are the calculation errors of the terminal voltage, and (c) are the specific values of the three errors of MEAN, ME, and RMSE. Detailed implementation manners

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0055] Within a limited observation time, the acquisition of lithium battery dynamic information is severely restricted, which directly affects the accurate extraction of lithium battery parameters. The higher the effective utilization degree of the recognition algorithm for real lithium battery dynamic data, the more effective the accurate recognition of model parameters. However, a large amount of data will lead to challenges in storage and calculation. Therefore, lightweight algorithms become a better choice in engineering applications. Although iterative algorithms do not need to record continuous historical data, they still directly or indirectly rely on a section of historical state information of the system before the current time. Usually, such algorithms need to run continuously for a period of time to ensure that the state information can be effectively used for model optimization and finally achieve convergence. In real-time and online applications, the immediate update of raw data is crucial, which can ensure that the algorithm timely adjusts model parameters and state information. Therefore, these algorithms are suitable for online parameter identification scenarios, but in the field of rapid identification, too little data will pose new challenges to the effectiveness of the algorithm.

[0056] Methods to improve the speed of the recognition algorithm include algorithms such as using simplified equivalent models, parallel computing, iterative algorithms, and lightweight optimization. Currently, simplified equivalent models, lightweight optimization algorithms, and iterative algorithms have become the mainstream solutions. With the diversification of application requirements such as the rapid evaluation, classification, combination, and sorting of retired lithium batteries in lithium battery production, and the continuous improvement of accuracy requirements, rapid recognition algorithms for specific application scenarios need to be customized. In lithium battery production, such as lithium battery classification and rapid sorting, usually in pursuit of simplicity and efficiency, the Ohmic internal resistance of the lithium battery is calculated by using the ratio of the voltage change rate to the current change rate under specific working conditions, and the measured value of the lithium battery terminal voltage after long-term static placement is regarded as the open-circuit voltage of the lithium battery. This method simplifies the calculation of the internal resistance and the measurement of the open-circuit voltage to the greatest extent. Although the ratio of voltage to current change can reflect the Ohmic internal resistance to a certain extent, due to the impedance change caused by the polarization phenomenon, it is difficult to effectively present this method. The current step method obtains the internal resistance of the lithium battery by calculating the ratio of the voltage difference to the current difference between two sampling points before and after the current step, and has the characteristics of simplicity and speed, and is widely used in actual engineering. However, the Ohmic internal resistance causes instantaneous voltage changes. Limited by equipment accuracy and rough operation, the calculation results of the current step method cannot subdivide the Ohmic internal resistance and the polarization internal resistance, weakening its application value.

[0057] The inventor proposed a new method for separating the Ohmic internal resistance and the polarization internal resistance by analyzing the variation law of the time-domain transient of the lithium battery. Therefore, the equivalent parameter extraction of the simplified model analysis of the lithium battery transient process is a very good starting point to improve the rapidity and accuracy of the rapid identification method.

[0058] The technical solutions of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0059] An embodiment of the present disclosure provides an off-line fast identification method for lithium battery parameters.

[0060] Step 1: Construct an off-line measurement system for lithium batteries, and design the working conditions of lithium batteries according to the actual requirements of fast measurement.

[0061] Step 2: Under the designed working conditions of the lithium battery, use the off-line measurement system for lithium batteries to perform charge and discharge tests on the lithium battery to be measured, and obtain the transient response data of the lithium battery after the current mutation.

[0062] Step 3: Based on the first-order RC equivalent circuit model, deduce the equivalent circuit state equation and output equation of the lithium battery, and analyze the transient process of the lithium battery terminal voltage under the condition of current step change to determine the equivalent circuit state equation and output equation of the lithium battery during the step transient process and the terminal voltage change formula at different time sampling points. Furthermore, analyze the change law of the lithium battery during the step transient process to obtain the quantization result TVR calculation formula of the change law in the time domain transient.

[0063] As Figure 1 shown, use the first-order RC equivalent circuit model to simulate the internal characteristics of the lithium battery. u OC is the voltage of the ideal voltage source, which is used to simulate the open-circuit voltage of the lithium battery. u L is the terminal voltage, R0 is the ohmic resistance, R p is the polarization internal resistance, C p is the polarization capacitance, u p is the polarization voltage, i L is the load current. When a constant current is input, the current reference direction for charging is set as positive. The transient component of the terminal voltage is generated under the action of the current source. At the end of charging, a step current will be generated, forming a current difference i d , i d determines the form of the transient response of the lithium battery.

[0064] According to Kirchhoff's law, equations (1) and (2) are respectively the continuous state equation and output equation of the equivalent circuit: (1).

[0065] (2).

[0066] Equations (3) and (4) are respectively the discrete state equation and output equation of the equivalent circuit: (3).

[0067] (4).

[0068] In the formula, τ = R p C p , is the time constant, and Ts is the sampling period.

[0069] Based on the analysis of the step transient process of a lithium battery under the condition of a step change in current, the step transient process curve when i is positive is obtained. Taking the 1 sampling point before the current transient as the starting point and the 1 sampling point before the next current transient as the ending point, a step transient load step is intercepted, and the overall voltage and current are subtracted from the voltage and current data at the starting point. d Taking the starting point as a constant reference point, and changing the sampling period Ts in Equation (3) to a time-varying factor t, with t = 0 at the constant reference point, the circuit continuous state and output equations of the step dynamic model are as follows:

[0070] Taking the starting point as a constant reference point, and changing the sampling period Ts in Equation (3) to a time-varying factor t, with t = 0 at the constant reference point, the circuit continuous state and output equations of the step dynamic model are as follows: (5).

[0071] (6).

[0072] In the formula, u p0 is the initial value of the polarization voltage, u pd is the change in the polarization voltage, u d is the change in the terminal voltage, i d is the change in the current, u OCd is the change in the open-circuit voltage. Since the duration of a step transient is short, the change in the open-circuit voltage in a step transient is small and can be ignored. OCd In a step transient, the change in the open-circuit voltage is small and can be ignored.

[0073] When the load changes instantaneously from one value to another, the response of the system (such as voltage, current, or other parameters) will experience a transient stage. Figure 2 is the terminal voltage transient response of the lithium battery under constant current. The shown is the terminal voltage transient response characteristic of the LIB under constant current excitation, establishing a connection between the equivalent model theory and the excitation current response curve. In the stage from A to B, it can be observed that the terminal voltage shows a significant step characteristic, which is due to the constant characteristic of the internal ohmic resistance of the battery. It should be noted that the ohmic impedance shows an approximately constant time-invariant characteristic under the test conditions. In the process from B to F, the terminal voltage of the battery gradually decreases until it reaches the initial voltage u0, and then the battery is left standing for a period of time.

[0074] In this process, the step response characteristic of the system reflects its transient impedance behavior. In particular, it is worth noting that there is a significant non-linear attenuation trend after the current transient point B. This phenomenon is related to the polarization and diffusion processes. The time-domain transient process evolution of the polarization voltage attenuation rate shows a logarithmic attenuation trend, and the change is relatively slow and the signal is weak during the descent. Therefore, by intercepting a single step transient response stage, effectively quantifying the change law of the transient response after the current transient can improve the accuracy and reliability of the model, and then accurately and quickly identify the battery parameters.

[0075] Obtain the voltage change rates k1 of two points at the start of the step and k2 of two points when the voltage tends to be stable after the nth point respectively, as Figure 2 shown. k1 is significantly greater than k2, mainly because the sudden change in the charging current causes the terminal voltage to drop rapidly. Since the polarization voltage u p does not change suddenly, it will discharge slowly through the polarization resistance R p , and as time goes by, until the polarization capacitor C p is fully discharged, at this time the terminal voltage is equal to the open-circuit voltage u OC .

[0076] Usually, the ratio of the voltage drop Δu0 between A and B and the step change in current Δi d is used to calculate the ohmic internal resistance. The calculation of the ohmic internal resistance by the current step method can be expressed as: (7).

[0077] Equation (6) includes the voltage change law during the transient process , and the state before the current transient is concentrated as u p0 . Transform the calculation formula (7) of the ohmic internal resistance by the current step method into formula (8): (8).

[0078] In addition to the ohmic internal resistance R0, the calculation result of formula (8) also includes R p , C p , u p0 , i d , and the influence of the t factor. It is a dynamic internal resistance R d that changes with time. If R d is directly used to replace R0, a large error will occur. Since the device cannot achieve the accuracy of instantaneous sampling, if the instantaneous voltage fluctuation value of the current step can be obtained through analysis and the influence of factors unrelated to R0 can be reduced, a more accurate ohmic internal resistance can be obtained.

[0079] Assume that R p , C p are fixed values within nTs. Then the following formulas hold at the three sampling points of t = nTs, (n - 1)Ts, and (n - 2)Ts: (9).

[0080] (10).

[0081] It is determined from formula (10) that the difference in the output voltage between two adjacent sampling points is proportional to E. The quantization result TVR of the change law of the time-domain transient is derived from formula (10): (11).

[0082] Among them, n > 2, which is the number of points of the sampling point from the starting point, defined as the distance points of the sampling point. nTs is a certain sampling moment, and the TVR value can be calculated through the data of 3 to 4 sampling points.

[0083] Step 4: On the premise that the resistance and polarization capacitance remain constant during the calculation period and the initial value of the transient voltage is ignored, according to the equivalent circuit state equation and output equation of the lithium battery in the step transient process and the terminal voltage change formula of the sampling points at different moments, deduce the relationship between the lithium battery parameters and the quantization result TVR of the change law of the time-domain transient, so as to use the transient response data of several sampling points of the lithium battery for continuous state analysis of the lithium battery and rapid calculation of the lithium battery parameters. Among them, the lithium battery parameters include ohmic internal resistance, polarization resistance, and polarization capacitance.

[0084] As can be seen from Equation (9), the state at the latter sampling moment is related to the state at the previous sampling moment, and E is a key parameter. Derived from Equation (10), we get: (12).

[0085] (13).

[0086] It can be seen that since there are three unknowns and only two effective formulas, it is impossible to accurately obtain the calculation results of R p , R0 and u p0 through Equations (12) and (13).

[0087] Such as Figure 3 shown, u R0 is the voltage transient variable at time 0 caused by the ohmic internal resistance R0. u p01 , u p02 are the true voltage transient variables at time 0 when they are in the same direction and opposite direction to i p0 respectively. u d . u d is the measured curve of voltage fluctuation in the step dynamic process. u dp01 , u dp02 are the true voltage fluctuation curves in the step dynamic process when they are in the same direction and opposite direction to i p0 respectively. d

[0088] Analysis Figure 3 and Equations (5) and (6), when the current steps, R0 causes a fixed voltage transient variable u1. Since the polarization state of the lithium battery cannot change instantaneously, the initial value u p0 of the polarization voltage of the lithium battery will also cause voltage transient. Therefore, R0 and u p0 jointly cause instantaneous voltage change. Combining Equations (12) and (13), we get: (14).

[0089] Equation (14) is the calculation formula for the voltage transient variable at the 0 moment of the current step. In the lithium battery offline detection scenario, most lithium batteries are subjected to a long-term static treatment, and u p0 can be regarded as 0. When u p0 is not 0, the separation of the ohmic internal resistance and the polarization internal resistance is realized. The derivation process is as Figure 4 shown. Then, the calculation formulas for R0, R p , and C p can be expressed as follows: (15).

[0090] (16).

[0091] (17).

[0092] Step 5: Analyze the influence of the number of points of the sampling points at different moments from the starting point on the quantization result TVR and the measurement result of the lithium battery parameters, and obtain the sampling point number conditions for the rapid measurement of the lithium battery parameters.

[0093] Step 6: Analyze the influence of different sampling periods on the quantization result TVR and the measurement result of the lithium battery parameters, and obtain the sampling period conditions for the rapid measurement of the lithium battery parameters.

[0094] Step 7: Combine Step 5 and Step 6, compare the measurement results of the lithium battery parameters of this method with those of the current step method and the co-evolutionary particle swarm optimization algorithm in the full SOC region, and obtain the rapid parameter measurement conditions that can effectively balance the calculation efficiency and accuracy.

[0095] Step 8: Preset the sampling point number conditions and sampling period conditions for the rapid measurement of the lithium battery parameters in the lithium battery offline measurement system, and collect the lithium battery response data through the set lithium battery parameter offline measurement system and calculate the lithium battery parameters and the lithium battery terminal voltage according to the calculation formulas deduced in Steps 3 and 4.

[0096] Step 9: Verify the accuracy of the lithium battery parameters according to the calculation results of the lithium battery parameters and the corresponding calculation results of the lithium battery terminal voltage, and analyze the factors affecting the accuracy of the offline rapid measurement of the lithium battery parameters.

[0097] Combined with the above embodiments, the following specific examples are proposed. It can be understood that the following specific examples only exemplarily elaborate on the specific implementation of the above embodiments, and do not limit the technical solutions of the above embodiments.

[0098] 1. Lithium battery offline measurement The lithium battery used for testing is a ternary lithium battery, and its detailed parameters are shown in Table 1. The system uses a main controller of model STM32F407ZGT6. The load condition is defined through the 12-bit DA converter integrated in the MCU. The programmable LoadProfile is provided to the lithium battery by controlling the charge and discharge current through a power amplifier. ADS1274 is used to collect the terminal voltage and current of the lithium battery. The collected data is preprocessed by the MCU and then sent to the PC side (Intel Core i7-7500U CPU, 2.70GHz, 12GB, 64bit, Windows 10) to save the experimental data. After the original experimental data test is completed and stored, MATLAB is used to verify the method of the present invention.

[0099] Table 1 Lithium battery parameters

[0100] 2. Customized condition To simulate the short-term load condition in the scenario of rapid identification of lithium battery parameters, the short-term step condition is one of the easiest-to-implement short-term transient conditions of lithium batteries. It has only one current step, and the duration after the current step is controlled to complete the lithium battery condition test in a short time. To explore the influence of charging and discharging on the lithium battery parameter identification results, the customized load current of the short-term step condition should include various step transients, such as continuous charging, continuous discharging, alternating charging and discharging, and charging and discharging starting from rest. Therefore, a customized experimental load configuration file as shown in Table 2 is designed.

[0101] Table 2 Customized experimental load configuration file

[0102] The load configuration file shown in Table 2 can be used to simulate multiple polarization states of the lithium battery when transient current occurs. The load current and the change of the current can be from large to small and from positive to negative. There are 26 load steps in the load configuration file used in the experiment, and the duration (T L ) of each load step ranges from 4 s to 10 s, and the total duration is 200 s. The current change amount (i d ) has 12 change values from -3200 mA to 3200 mA, and the average current is -176 mA.

[0103] Figure 5 A complete load current cycle is plotted in (a) below. After the experimental system runs the load curve for 3 consecutive cycles, the lithium battery is discharged at a current of -400 mA for 30 minutes, and then this process is repeated until the lithium battery voltage reaches the discharge cut-off voltage. The measured terminal voltage of the lithium battery is as shown in (a) below. Figure 5 as shown in (a) below.Figure 5 The 13 dynamic test results numbered from A to M in (b) include Figure 5 The 39 load profile cycles shown in (b) are numbered in the order of occurrence. Since the data in Group A was collected when the lithium battery was fully charged, under the charging condition, the lithium battery voltage data in Group A repeatedly saturates at the charging cut-off voltage, which is not conducive to comparison with other data groups. Therefore, the data in Group A is not used, and the last two load curves in Groups B - M are used to verify and discuss the accuracy and rationality of the method of the present invention.

[0104] 3. Analysis and Discussion Several different comparison schemes are set up to discuss the influence of the number of sampling points n from the starting point and the sampling period Ts participating in the calculation on the method of the present invention. To verify the accuracy of the method, the ohmic internal resistance R0 obtained by different methods is compared. Through comparison, we can identify the performance differences of the methods under various sampling conditions. In addition, to provide a more comprehensive analysis, the ohmic internal resistance R0, polarization internal resistance R p , polarization capacitance C p parameters are used to calculate the terminal voltage. This process aims to verify the effectiveness of the method of the present invention in the state assessment of lithium batteries and its feasibility in practical applications.

[0105] (1) The number of sampling points from the starting point affects the parameter identification accuracy The fewer the sampling points, the less time-consuming the algorithm for the ohmic internal resistance, and the better the rapid performance of the method. As shown in Equation (11), the adjacent two sampling points affect TVR. The more sampling points, the longer the calculation time, and the smoother TVR. Increasing the distance points can reduce the calculation error of the method. When the distance points take an appropriate value, both the accuracy and rapidity of the method can be better guaranteed.

[0106] Combined with Equation (11) and it can be known that when the distance points n = 3 (Time = 0.3s), the rapidity of the method of the present invention is the best. Let Ts = 0.1s, Figure 6 The curves of R0 and TVR with n = 3 are given in Figure 6 In (a), the ohmic internal resistance R d obtained by the method of the present invention in the full SOC region with n = 3 and the ohmic internal resistance R3 obtained by the current step method Figure 6 In (b), the TVR curve diagram in the full SOC region with n = 3 is shown. Figure 6 In (c) and (d), the local enlarged views in the same SOC region are shown. It can be seen that due to the low accuracy of the first-order RC circuit and measurement errors, etc., the TVR fluctuates greatly when n = 3, resulting in many mutation points in the calculation results of R0. However, in the local enlargement Figure 6As can be seen from (c) and (d) in the figure, the ohmic internal resistances R3 and R measured by the method of the present invention and the current step method respectively d are not much different. The overall trend of the R0 curve is flat in the full SOC region, and is consistent with the overall trend of the R d curve of the current step method, and has the value of further research.

[0107] Increasing the number of distance points n in formula (11) can reduce the fluctuation of TVR. At the same time, it will also lead to the deterioration of the rapidity of the method. On the basis of fixing R p and C p , the influence of different n on the calculation results of TVR and R0 is discussed. Table 3 gives the corresponding relationships of each parameter when SOC = 52%.

[0108] Table 3 Corresponding relationship table of parameters when SOC = 52%

[0109] It is known that Figure 6 the R3 and E3 curves are given. As Figure 7 shown are the TVR and R0 curves corresponding to other n values in Table 3. Among them, R d is the calculation result of the current step method with Ts = 0.1 s. As Figure 7 shown in (b), (d), (f), and (h) in the figure, all TVR values are distributed in the range of 0-1. This phenomenon is consistent with the theoretical derivation of the normalization characteristics of the time constant on the voltage response in the first-order RC equivalent circuit, which not only verifies the physical meaning of TVR representing the transient attenuation of the polarization voltage, but also verifies the rationality of the model construction.

[0110] Comparing the TVR curves under different n values, it can be seen that with the increase of n, the fluctuation of TVR in the full SOC range is significantly reduced. Although increasing the number of distance points n can effectively suppress the fluctuation of TVR, it will also lead to the deterioration of the accuracy and rapidity of the method. With the increase of the n value, the time consumed by the algorithm calculation also increases, but the information reflected in the battery phase change process is more, covering a longer time response, so as to average out the transient fluctuation and make the TVR curve smoother. In addition, increasing the number of distance points can effectively reduce the calculation error of the method, while reducing the number of sampling points can significantly improve the calculation efficiency. Therefore, selecting an appropriate n value plays a key role in balancing the identification accuracy and calculation efficiency.

[0111] As Figure 6 shown in (b) in the figure and Figure 7 shown in (b) in the figure, the fluctuation amplitude of TVR is significantly larger when n = 3 and n = 6 than other n values, especially in the low SOC region. Figure 7 As shown in (h) in the figure, when n = max, the TVR curve tends to be completely smooth. This smoothness directly determines the stability of the R0 calculation result. AsFigure 7 As shown, when n = 6, the TVR curve still fluctuates greatly in the low SOC range. When n = 10 or above, the TVR trend is flatter and the calculation result of R0 is closer to that of the current step method. Therefore, the improvement of TVR stability can reduce the mutation of the R0 curve. When n = 10 and the corresponding calculation time Time = 1s, the R0 in the full SOC region obtained by the method of the present invention has no significant change and the overall trend is flat. However, too large an n value will cause loss of polarization information and reduction of effective information. Therefore, n = 10 is the preferred parameter considering both calculation efficiency and accuracy.

[0112] In the low SOC region where the polarization effect is significant, the traditional method is prone to cause calculation deviation of R0 due to the enhanced nonlinearity of the polarization voltage. The method of the present invention realizes the consistency of the R0 calculation results in the high and low polarization regions and the stable middle SOC region through the adjustment of TVR. Therefore, the ohmic internal resistance of the battery is relatively stable within a full SOC cycle and shows a slow upward trend from the high SOC to the low SOC interval. Figure 7 As shown in (c), when Ts = 0.1s and n = 10 (Time = 1s), relatively stable and accurate TVR and R0 calculation results can be obtained, and at the same time, the method of the present invention has good rapidity.

[0113] (2) A smaller sampling period is beneficial to the accuracy of R0 The method of the present invention can realize the stable calculation of the ohmic internal resistance R0 under different sampling periods. Research shows that as the sampling period decreases, with the same number of distance points n maintained, the algorithm time required for ohmic internal resistance identification is significantly reduced, thereby improving the rapidity of the method. This means the importance of a smaller sampling period for improving the identification speed, providing greater flexibility and convenience for practical applications. In addition, using a smaller sampling period can not only shorten the identification time but also significantly reduce the loss of effective information. When the sampling time is short, the changes in the state of the lithium battery can be captured more frequently. This high-frequency data acquisition greatly enhances the accuracy and stability of the algorithm in obtaining lithium battery parameters. Just because of this, when evaluating the performance and state of the lithium battery, the method of the present invention shows stronger advantages, making the identification results more reliable.

[0114] Based on the three sampling period conditions defined in Table 4, with the number of points n = 10 from the starting point, Table 4 shows the corresponding algorithm identification results at SOC = 52%. These results further verify the significant influence of different sampling periods on the parameter identification effect, revealing how to optimize the performance of parameter identification by selecting an appropriate sampling period in practical applications. These characteristics provide a solid foundation for future lithium battery performance monitoring and management, contributing to the further development and application of lithium battery technology.

[0115] Table 4 Table of parameter relationships corresponding to different sampling periods when SOC = 52%

[0116] The sampling period has an important influence on the calculation result of the ohmic internal resistance. By using the method of the present invention and the current step method, the calculation results of the ohmic internal resistance are compared under three sampling periods Ts, and the effective algorithm conditions in parameter identification of the method of the present invention can be obtained.

[0117] As Figure 8 shown, in the medium SOC region where polarization is relatively stable, as Ts increases, the consistency of the identification results of the two methods improves, but with the reduction of the effective feature information amount and the significant increase of the calculation time. When Ts = 0.1 s, the TVR has a more stable fluctuation trend in the low SOC region than other sampling periods. In Figure 8 (a), (b), (c) and (d) in it reflect that the ohmic internal resistance in the polarization region is more stable when Ts = 0.1 s, and the identification results of the two methods have good consistency. This trend further verifies the influence of TVR on the parameter identification accuracy, that is, the polarization interference in the low SOC region is strong. If Ts is too large, it is easy to miss the key dynamic information, resulting in a significant increase in errors. When Ts = 0.05 s, although the overall trend of the ohmic internal resistance is flat and the robustness is good, the consistency with the identification result of the current step method is poor. In the TVR calculation formula, the denominator is the change amount of the voltage at two moments in the initial stage of sampling, which reveals the transient response characteristics in the stage where polarization is not saturated from the physical meaning. Therefore, by selecting the sampling data in the initial stage of the step, that is, at a smaller Ts, the transient characteristics where the polarization effect is not fully established can be effectively captured, which not only helps to extract the intrinsic response characteristics of the ohmic resistance, but also is beneficial to suppressing the polarization interference caused by the relaxation of the RC network, thereby improving the identification accuracy and stability.

[0118] In summary, choosing an appropriate Ts has a significant influence on the parameter identification accuracy. A smaller Ts has high calculation efficiency and can better retain dynamic characteristics, but if it is too small, it may increase the influence of measurement noise; a larger Ts can improve the consistency of the identification results to a certain extent, but the influence on information loss and calculation complexity needs to be carefully weighed. Therefore, the identification accuracy, dynamic response capture ability and algorithm efficiency should be comprehensively considered to determine the optimal sampling period.

[0119] From Figure 8 (e) in it, it can be seen that when Ts = 0.1 s, the method can maintain a stable R0 output characteristic under different load condition cycles. As Figure 9 shown, compared with the current step method, the R p and C pThe fluctuation amplitude within the entire SOC range (especially in the high and low SOC regions) is significantly reduced. The method of the present invention effectively avoids the polarization interference accumulated during the long-time relaxation process in the current step method. In addition, the smaller sampling period enhances the ability to capture dynamic characteristics, making the robustness of the parameter identification results under extreme SOC conditions significantly improved. Therefore, through the collaborative optimization of simplifying the model and quantifying TVR, the method of the present invention realizes the stable identification of polarization parameters within the entire SOC range while ensuring the calculation efficiency, providing a more reliable data basis for battery modeling.

[0120] As Figure 10 shown, on the premise of n = 10, three Ts with better consistency in the identification results of the two methods are selected. The Ohmic resistance calculation results of the method of the present invention and the current step method are compared at Ts = 0.1 s, Ts = 0.2 s, and Ts = 0.4 s respectively. In the range from Ts = 0.1 s to 0.4 s, compared with the traditional current step method, the method of the present invention exhibits better parameter consistency and robustness characteristics. The research shows that the selection of the sampling period directly affects the characterization accuracy of the dynamic attenuation process of the polarization voltage. When the value of Ts gradually increases, due to the polarization voltage accumulation effect and excessive loss of effective information in the low SOC region of the current step method, the calculated result of R0 is too high and the fluctuation intensifies. However, the method of the present invention effectively suppresses the interference of the transient component of the polarization voltage by dynamically adjusting TVR, making the fluctuation amplitude of R0 smaller than that of the traditional method and the calculation result more stable.

[0121] In addition, when Ts = 0.1 s, the stability of R0 within the entire SOC range is significantly better than the calculation results of other sampling periods, verifying the theoretical hypothesis that a smaller Ts value is beneficial to enhancing the integrity of feature information extraction and calculation speed. Therefore, the optimized parameter settings (Ts = 0.1 s, n = 10) are beneficial to the extraction of effective information and the enhancement of the stability of the parameter identification results while ensuring the calculation efficiency.

[0122] (3) Comparative verification of the method of the present invention and other R0 extraction methods There are commonalities between the method of the present invention, the current step method, and the CPSO algorithm, which enables effective comparative analysis during the internal resistance identification process of lithium batteries. By comparing the results of different algorithms during the internal resistance identification process of lithium batteries, the rationality and effectiveness of these algorithms can be effectively verified. Specifically, the method of the present invention combines the current step method with the equivalent circuit model to form an innovative internal resistance identification method. To verify the accuracy and consistency of the method of the present invention, in this subsection, a comparative analysis of the Ohmic resistance R0 curves obtained by the current step method, the equivalent circuit model method, and the method of the present invention will be carried out. This comparison will not only show the performance differences of each method in internal resistance identification but also further consolidate the superiority of the method of the present invention.

[0123] The co-evolutionary particle swarm optimization (CPSO) algorithm is selected to obtain the ohmic internal resistance R0 through a first-order RC circuit. The CPSO algorithm has excellent global search ability and local optimization characteristics, enabling it to online identify R0 in the full SOC region. The CPSO parameter table is set as follows: Table 5 CPSO algorithm parameter table

[0124] Let the sampling period Ts = 0.1 s. Using the co-evolutionary particle swarm CPSO algorithm, the ohmic internal resistance R0 is identified in different SOC regions. Respectively in Figure 5 In (b) of [reference], in the third load curve of groups B, H, and M, the CPSO algorithm is used to fit the terminal voltage. The fitting results of the CPSO algorithm are as Figure 11 shown, where MEAN is the mean absolute error, ME is the maximum absolute error, and RMSE is the root mean square error. From the analysis of the fitting results, the fitting error of the CPSO algorithm is relatively large in the high and low SOC regions, which is related to the low accuracy of the first-order RC circuit and the severe polarization of the lithium battery in these two regions. In the middle SOC region, the mean absolute error of the CPSO algorithm is within 1 mV, indicating that its R0 identification results in this region have high accuracy and reliability.

[0125] As Figure 12 shown, comparing the results of the ohmic internal resistance R0 obtained by three methods in the middle SOC region, where R 0_cpso and R p_cpso are the R0 and R p curves obtained by the CPSO algorithm through the first-order RC equivalent circuit respectively, while R d and R0 are the ohmic internal resistance curves obtained by the current step method and the method of the present invention respectively, and R p is the polarization internal resistance curve obtained by the current step method.

[0126] Analysis shows that: (1) The ohmic internal resistance curves obtained by the three methods all fluctuate around 40 mΩ, showing certain commonalities; (2) The fluctuation range of the ohmic internal resistance R0 curve of the method of the present invention is the smallest and the most stable; (3) Due to the limitations of the first-order RC circuit model and the optimization interval, the R p curve of the CPSO algorithm fluctuates greatly, but it shows commonalities with the method of the present invention in the stable region.

[0127] Therefore, the commonalities of the method of the present invention, the current step method, and the CPSO algorithm verify their rationality and accuracy. In addition, the R0 and R p results obtained by the method of the present invention are significantly more stable, verifying the superiority of the method.

[0128] (4)Verification by comparing terminal voltage fitting

[0129] Systematically analyze the calculation error of the terminal voltage. The overall error is controlled within about ±5 mV, showing high accuracy. To further improve the fitting accuracy, the time period with fixed assumed values of R p and C p can be shortened to improve the reliability and consistency of the calculation results.

[0130] To verify the accuracy of R0, R p and C p obtained by the method of the present invention, assume that in a single step transient, the polarization resistance R p and the polarization capacitance C p remain unchanged, and calculate the terminal voltage u L on the third load curve of group H. In Figure 13 , the calculation results and error conditions of the terminal voltage are presented. The results show that the overall calculation error of the terminal voltage is controlled within about ±5 mV, with good accuracy. To deeply understand the cause of the error, by comparing (c) and (d) in Figure 11 , the mean absolute error MEAN and the root mean square error RMSE of the terminal voltage calculation result u c are relatively large. This phenomenon is mainly due to the accuracy limitation of the first-order RC circuit model, and in a single step transient, the values of the polarization resistance R p and the polarization capacitance C p are not fixed, resulting in obvious fluctuations in the calculation results.

[0131] To improve the calculation efficiency and accuracy, the calculation time of this method is only 1 s. Therefore, in a single step transient, only calculate the terminal voltage in the first 1 s, and shorten the time period with fixed assumed values of R p and C p . The specific fitting results are as shown in Figure 14 . The results show that shortening the time period with fixed assumed values of R p and C p can significantly improve the accuracy of the terminal voltage calculation result u c .

[0132] 4. Conclusion (1) Effectively simplify the transient process model by quantifying TVR. Ignore battery polarization at the operating point where the initial value of the transient voltage is 0, and on the premise that the polarization resistance and the polarization capacitance remain constant during the calculation time, obtain the internal relationship between R0, R p and TVR according to the dynamic equation. The equivalent parameters can be directly obtained through the change amount and TVR. This method simplifies the calculation process and makes the evaluation of battery performance more efficient in practical applications.

[0133] (2) The computational time consumed by the algorithm affects the stability and consistency of the parameter identification results. The number of points n of the sampling points from the starting point directly affects the subsequent calculation of R0. As the number of distance points n increases, the stability of TVR is higher, and the mutation points in the R0 calculation are fewer and more stable. In addition, under different sampling periods, the R0 calculation results of this method have good consistency and stability. By changing the sampling period and n, the algorithm can obtain the original time-domain information of different durations and different transients. The shorter the sampling period under the same n, the less loss of effective information, which is more conducive to the rapidity and accuracy of the method.

[0134] (3) On the premise of effectively separating R0 and R p , appropriate algorithm conditions are given. The R0 calculation results of the method of the present invention have good consistency in the full SOC region of the battery. Its R0 curve has commonalities with the CPSO algorithm and the current step method, and is more stable. Selecting the algorithm conditions of n = 10 and Ts = 0.1s can not only ensure that the algorithm accurately separates R0 and R p while fully capturing the effective information in the relaxation process, but also guarantee the calculation efficiency.

[0135] (4) Based on the parameters obtained by the method of the present invention, the terminal voltage is calculated, and its average absolute error is within 2 mV, with good accuracy. In addition, shortening the time when the assumed values of R p and C p are fixed helps to improve the accuracy of the method of the present invention.

[0136] So far in the embodiments of the present invention, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. An offline fast identification method for lithium battery parameters, characterized in that, It includes the following steps: Step 1: Construct a lithium battery off-line measurement system and design the working conditions of the lithium battery according to the actual requirements of rapid measurement; Step 2: Under the designed working conditions of the lithium battery, use the lithium battery off-line measurement system to perform charge and discharge tests on the lithium battery to be measured, and obtain the transient response data of the lithium battery after the current mutation; Step 3: Based on the first-order RC equivalent circuit model, deduce the equivalent circuit state equation and output equation of the lithium battery, and analyze the transient process of the lithium battery terminal voltage under the condition of current step change to determine the equivalent circuit state equation and output equation of the lithium battery during the step transient process and the terminal voltage change formula at different sampling points, and then analyze the change law of the lithium battery during the step transient process to obtain the quantization result TVR calculation formula of the change law in the time-domain transient; Step 4: On the premise that the resistance and polarization capacitance remain constant during the calculation period and the initial value of the transient voltage is ignored, deduce the relationship between the lithium battery parameters and the quantization result TVR of the change law in the time-domain transient according to the equivalent circuit state equation and output equation of the lithium battery during the step transient process and the terminal voltage change formula at different sampling points, so as to use the transient response data of the lithium battery at several sampling points for continuous state analysis of the lithium battery and rapid calculation of the lithium battery parameters. Among them, the lithium battery parameters include ohmic internal resistance, polarization resistance, and polarization capacitance; Step 5: Analyze the influence of the number of points of the sampling points at different times from the starting point on the quantization result TVR and the measurement result of the lithium battery parameters, and obtain the sampling point number condition for rapid measurement of the lithium battery parameters; Step 6: Analyze the influence of different sampling periods on the quantization result TVR and the measurement result of the lithium battery parameters, and obtain the sampling period condition for rapid measurement of the lithium battery parameters; Step 7: Combine Step 5 and Step 6, compare this method with the measurement results of the lithium battery parameters under the full SOC region of the current step method and the co-evolutionary particle swarm optimization algorithm, and obtain the rapid measurement condition of the parameters that can effectively balance the calculation efficiency and accuracy; Step 8: Preset the sampling point number condition and sampling period condition for rapid measurement of the lithium battery parameters in the lithium battery off-line measurement system, and collect the lithium battery response data through the set lithium battery parameter off-line measurement system and calculate the lithium battery parameters and the lithium battery terminal voltage according to the calculation formulas deduced in Steps 3 and 4; Step 9: Verify the accuracy of the lithium battery parameters according to the lithium battery parameter calculation results and the corresponding lithium battery terminal voltage calculation results, and analyze the factors affecting the accuracy of the off-line rapid measurement of the lithium battery parameters.

2. The offline fast identification method for lithium battery parameters according to claim 1, characterized in that In Step 3, deducing the equivalent circuit state equation and output equation of the lithium battery based on the first-order RC equivalent circuit model includes: The internal characteristics of a lithium battery are simulated using a first-order RC equivalent circuit model, where u OC is the voltage of an ideal voltage source, used to simulate the open-circuit voltage of the lithium battery, and u L is the terminal voltage, R0 is the ohmic resistance, and R p is the polarization internal resistance, and C p is the polarization capacitance, and u p is the polarization voltage, and i L is the load current. When a constant current is input, the current reference direction for charging is set as positive. The transient component of the terminal voltage is generated under the action of the current source, and a step current will be generated at the end of charging, forming a current difference i d , and i d determines the form of the transient response of the lithium battery; According to Kirchhoff's law, Equation (1) and Equation (2) are respectively the continuous state equation and output equation of the equivalent circuit: （1） （2） Equation (3) and (4) are respectively the discrete state equation and output equation of the equivalent circuit: （3） （4） where τ = R p C p , is the time constant, and Ts is the sampling period.

3. The offline fast identification method for lithium battery parameters according to claim 2, wherein In step 3, under the condition of step change in current, analyze the transient process of the terminal voltage of the lithium battery to determine the equivalent circuit state equation and output equation of the lithium battery during the step transient process, as well as the terminal voltage change formula at different sampling points, and then analyze the change law of the lithium battery during the step transient process to obtain the quantization result TVR calculation formula of the change law in the time-domain transient, including: Analyze the step transient process of a lithium battery based on the condition of a step change in current to obtain the step transient process curve when i d is positive. Take the first sampling point before the current transient as the starting point and the first sampling point before the next current transient as the ending point to intercept a step transient load step, and subtract the voltage and current data at the starting point from the overall voltage and current; Taking the starting point as a constant reference point, and changing the sampling period Ts in formula (3) to a time-varying factor t, with t = 0 at the constant reference point, the circuit continuous state and output equation of the step dynamic model are as follows: （5） （6） where u p0 is the initial value of the polarization voltage, u pd is the change in the polarization voltage, u d is the change in the terminal voltage, i d is the change in the current, u OCd is the change in the open-circuit voltage; Equation (6) contains the variation law of voltage during the transient process , and the state before the current transient is concentrated as u p0 , and the Ohm's internal resistance calculation formula (7) of the current step method is transformed into formula (8): （7） （8） Assume that R p and C p are fixed values. Then the following formulas hold at the three sampling points t = nTs, (n - 1)Ts, and (n - 2)Ts: （9） （10） It is determined from formula (10) that the difference in output voltage between two adjacent sampling points is proportional to E, and the quantization result TVR of the change law in the time-domain transient is derived from formula (10): （11） where n > 2, which is the number of points of the sampling point from the starting point, defined as the distance number of the sampling point, and nTs is a certain sampling moment.

4. The offline rapid measurement method for lithium battery parameters according to claim 3, wherein In step 4, based on the equivalent circuit state equation and output equation of the lithium battery during the step transient process, as well as the terminal voltage change formula at different sampling points, derive the relationship between the lithium battery parameters and the quantization result TVR of the change law in the time-domain transient, including: It can be seen from formula (9) that the state at the latter sampling moment is related to the state at the previous sampling moment, and E is the key parameter. Derived from formula (10), we get: （12） （13） Analyze the i d When it is positive, the step transient process curve and equations (5) and (6) determine R0 and u p0 Together, they cause the instantaneous voltage change. Combining equations (12) and (13), we get: （14） Equation (14) is the calculation formula for the voltage transient at the 0 moment of the current step. In the off-line detection scenario of lithium batteries, most lithium batteries have undergone long-term static treatment, and u p0 can be regarded as 0. When u p0 is not 0, the separation of the ohmic internal resistance and the polarization internal resistance is achieved. Then the calculation formulas for R0, R p , and C p can be expressed as follows: （15） （16） （17）。

Images