Lithium ion battery available capacity estimation method based on SOC

Abstract

The invention relates to the technical field of lithium ion battery available capacity estimation, in particular to an SOC-based lithium ion battery available capacity estimation method, which specifically comprises the following steps of: 1, constructing an SOC estimation result credibility evaluation function according to the change rate of an SOC-OCV curve of a battery by utilizing the phase change characteristics of positive and negative electrode materials in the charging and discharging working conditions of the lithium battery; 2, collecting battery current, voltage and time interval data of the lithium ion battery in a charging and discharging state, establishing a first-order RC equivalent circuit model of the lithium ion battery, and simulating electrochemical characteristics of the lithium ion battery; and 3, setting a battery capacity value as a guess battery capacity Capavai, guess, taking the guess battery capacity Capavai, guess as prior information, and performing equivalent circuit model parameter online identification and SOC online estimation by adopting a recursive least square filtering algorithm. According to the method, the result stability is ensured through the convergence condition, the capacity attenuation is continuously tracked, and the full-life-cycle monitoring requirement is met.

Description

Technical Field

[0001] This invention relates to the field of lithium-ion battery evaluation methods, and in particular to a method for estimating the available capacity of lithium-ion batteries based on SOC. Background Technology

[0002] As the electric vehicle market continues to expand and the proportion of intermittent energy sources such as photovoltaic and wind power connected to the grid continues to increase, the demand for high-performance battery systems is growing exponentially.

[0003] As a core component of energy storage, the performance degradation of lithium-ion batteries directly affects the economy and reliability of the entire energy system. Therefore, establishing a precise battery state of health (SOH) monitoring system, especially for real-time assessment of capacity degradation, has become a key scientific issue that urgently needs to be addressed in the field of battery management technology.

[0004] In many practical applications, it is necessary to estimate the capacity of a single or group of lithium batteries in service, whose health status and SOC are uncertain. However, in the existing technology, the available capacity estimation methods for lithium-ion batteries lack an iterative optimization mechanism, and the deviation between the initial guess capacity and the actual capacity is easily propagated, resulting in poor stability of the capacity estimation results. Summary of the Invention

[0005] To address the technical problem of the lack of an iterative optimization mechanism in the method for estimating the usable capacity of lithium-ion batteries, this invention provides a method for estimating the usable capacity of lithium-ion batteries based on SOC (State of Charge).

[0006] The technical solution adopted in this invention is: a method for estimating the usable capacity of a lithium-ion battery based on SOC, specifically including the following steps:

[0007] Step 1: Utilize the phase transition characteristics of the positive and negative electrode materials during the charging and discharging of lithium batteries, and construct a reliability evaluation function for the SOC estimation results based on the rate of change of the SOC-OCV curve of the battery.

[0008] Step 2: Collect battery current, voltage, and time interval data of lithium-ion batteries under charge and discharge conditions, establish a first-order RC equivalent circuit model of lithium-ion batteries, and simulate the electrochemical characteristics of lithium-ion batteries.

[0009] Step 3: Set the battery capacity value as a guess for the battery capacity Cap. avai,guess This information is used as prior information, and a recursive least squares filtering algorithm is employed to perform online identification of equivalent circuit model parameters and online estimation of SOC.

[0010] Step 4: Based on the SOC estimation results and their reliability evaluation results, the weighted least squares method is used to calculate the optimal estimate of the battery's usable capacity.

[0011] Step 5: Calculate the available capacity estimate results obtained in Step 4. As prior knowledge for the online identification algorithm of equivalent circuit model parameters, a more accurate SOC estimation and available capacity estimation are reimplemented, forming an iterative joint estimation of SOC and available capacity. The process stops when the change in battery capacity estimation results is less than 1%, thus completing the Capacity estimation. avai Estimate.

[0012] In one embodiment, the confidence evaluation function is calculated using the following formula in step one:

[0013]

[0014] Where w(SOC) is the confidence coefficient of the SOC estimation result, and its value ranges from 0 to 1. The slope of the SOC-OCV curve at a certain SOC point represents the rate of change of voltage with respect to SOC; OCV SOC＝100％ OCV is the open-circuit voltage of the battery when it is fully charged, i.e., SOC = 100%. SOC＝0％ This is the open-circuit voltage of the battery when it is in a depleted state, i.e., SOC = 0%.

[0015] In one embodiment, the transfer function calculation formula for the RC equivalent circuit model in step two is as follows:

[0016]

[0017] Among them, U out Where is the battery terminal voltage, OCV is the battery open-circuit voltage, I is the battery current; R0 is the internal resistance in ohms; R p For polarization resistance; C p is the polarization capacitor; s is the complex frequency variable of the Laplace transform; G(s) is the frequency domain transfer function of the battery.

[0018] In one embodiment, in step three, the discrete transfer function of the lithium-ion battery is analyzed using z-transform, and the calculation is as follows:

[0019]

[0020] The recursive least squares filtering algorithm is used to identify the parameters of each model. The parameter identification formula is calculated as follows:

[0021]

[0022] in, y(k) represents the parameter estimate for the k-th iteration; K(k) is the Kalman gain matrix; y(k) is the actual output of the system; φ(k) is the observation vector.

[0023] The real-time state of charge (SOC) of the battery is estimated using the extended Kalman filter algorithm, as follows:

[0024] The formula for calculating the state equation is as follows:

[0025] The calculation formula for the observation equation is as follows: U out =OCV(SOC)-IR0-U p ;

[0026] Where η is the Coulomb efficiency; Cap avai g This represents the currently estimated available capacity; U p This is the polarization voltage.

[0027] In one embodiment, in step four, the usable battery capacity is calculated using weighted least squares, combining the SOC estimate with its confidence weights. Specifically as follows:

[0028] The weighted least squares formula is calculated as follows:

[0029]

[0030] thereby:

[0031]

[0032] Among them, w i ΔQ represents the confidence weight of the i-th sampling point. i ΔSOC represents the cumulative charge / discharge capacity at the i-th sampling point. i SOC represents the change in SOC at the i-th sampling point; SOC0 represents the true SOC value at the initial moment of the operating condition; N represents the total number of sampling points. This represents the optimal estimated available battery capacity to be determined.

[0033] In one embodiment, the available capacity estimate Cap obtained in step four is used as the basis for the calculation. avai As prior knowledge for the online identification algorithm of equivalent circuit model parameters, a more accurate SOC estimation and available capacity estimation are reimplemented, forming an iterative joint estimation of SOC and available capacity. The estimation stops when the change in battery capacity is less than 1%, thus achieving convergence and accurate Capacity estimation. avai The estimation method is as follows:

[0034] The result obtained in step four As the new prior capacity value, steps three and four are re-executed to form an iterative optimization until the capacity estimate changes by less than 1%.

[0035] The convergence condition is calculated as follows:

[0036]

[0037] The execution flow is as follows: using the current Update the model parameters, re-estimate the SOC, and then recalculate the Cap. avai Finally, this process is repeated until the convergence condition is met.

[0038] The beneficial effects of this invention are as follows: Compared with the prior art, in this embodiment, firstly, the reliability evaluation function constructed in step one accurately quantifies the estimation reliability of different SOC intervals, avoiding interference from low-quality data and laying a reliable foundation. Secondly, the first-order RC equivalent circuit model in step two balances accuracy and efficiency, accurately simulating the dynamic characteristics of the battery and adapting to online real-time estimation scenarios. Thirdly, combining recursive least squares and extended Kalman filtering, dynamic parameter updates and real-time SOC estimation are achieved, reducing the impact of factors such as aging and temperature, and lowering the sensitivity to initial values. Fourthly, the weighted least squares algorithm in step four strengthens the contribution of high-reliability data, jointly optimizing capacity and initial SOC to offset random errors. Finally, the iterative optimization in step five forms a closed loop, ensuring result stability through convergence conditions, continuously tracking capacity decay, and meeting the full life cycle monitoring requirements. Attached Figure Description

[0039] Figure 1 This is a flowchart from an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram of the RC equivalent circuit model in an example of the present invention;

[0041] Figure 3 This is a schematic diagram illustrating the post-correction of available capacity based on the credibility evaluation of SOC estimation results in an example of the present invention. Detailed Implementation

[0042] In the description of this invention, it should be noted that the terms "front", "up", "down", "left", "right", "vertical", "horizontal", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0043] refer to Figure 1-3 To address the problems existing in the background technology, this application proposes the following technical solution: a method for estimating the usable capacity of a lithium-ion battery based on SOC, specifically including the following steps:

[0044] Step 1: Utilize the phase transition characteristics of the positive and negative electrode materials during the charging and discharging of lithium batteries, and construct a reliability evaluation function for the SOC estimation results based on the rate of change of the SOC-OCV curve of the battery.

[0045] In step one, the formula for calculating the credibility evaluation function is as follows:

[0046]

[0047] Where w(SOC) is the confidence coefficient of the SOC estimation result, and its value ranges from 0 to 1. The slope of the SOC-OCV curve at a certain SOC point represents the rate of change of voltage with respect to SOC; OCV SOC＝100％ OCV is the open-circuit voltage of the battery when it is fully charged, i.e., SOC = 100%. SOC＝0％ This is the open-circuit voltage of the battery when it is in a depleted state, i.e., SOC = 0%.

[0048] The above technical solution is explained as follows: a reliability evaluation function for SOC estimation results is constructed based on the rate of change of the SOC-OCV curve. This function calculates the ratio of the rate of change of OCV with respect to SOC at a certain SOC point to the difference in OCV between fully charged and depleted states, obtaining a reliability coefficient in the range of 0 to 1, thus quantifying the reliability of the estimation results in different SOC intervals. Firstly, it achieves a quantitative assessment of the reliability of SOC estimation. The slope of the SOC-OCV curve of lithium batteries varies significantly in different intervals. In intervals with a steeper slope (such as the phase transition interval), the OCV is more sensitive to changes in SOC, resulting in smaller estimation errors. The reliability coefficient can accurately identify these high-reliability intervals. Secondly, it provides a scientific weighting basis for subsequent capacity estimation, distinguishing the reliability of different sampling points through the reliability coefficient, avoiding interference from low-reliability data on the overall estimation results. Thirdly, it improves the robustness of capacity estimation by dynamically adjusting the contribution weights of each SOC interval, making the estimation results more dependent on high-reliability data.

[0049] Step 2: Collect battery current, voltage, and time interval data of lithium-ion batteries under charge and discharge conditions, establish a first-order RC equivalent circuit model of lithium-ion batteries, and simulate the electrochemical characteristics of lithium-ion batteries.

[0050] In step two, the transfer function calculation formula for the RC equivalent circuit model is as follows:

[0051]

[0052] Among them, U out This is the battery terminal voltage (output voltage). OCV is the battery open-circuit voltage (related to SOC).

[0053] I is the battery current (positive for charging, negative for discharging). R0 is the internal resistance in ohms (the battery's internal resistance). pThis is the polarization resistance (charge transfer impedance). C p Let be the polarization capacitance (charge transfer capacitive reactance). s is the complex frequency variable of the Laplace transform. G(s) is the frequency domain transfer function of the battery, describing the relationship between the input current and the output voltage.

[0054] Step two involves collecting data on the battery current, voltage, and time intervals during charge and discharge of the lithium-ion battery. A first-order RC equivalent circuit model of the lithium-ion battery is then established to simulate its electrochemical characteristics. Figure 2 As shown:

[0055] Where Uout represents the battery voltage; OCV represents the battery open-circuit voltage; I represents the battery current; R0 represents the battery internal resistance; and Rp and Cp represent the battery polarization impedance.

[0056] The above technical solution is explained as follows: by collecting current, voltage and time interval data of lithium-ion battery under charging and discharging conditions, a first-order RC equivalent circuit model is established to simulate the electrochemical characteristics of the battery. The model transfer function includes key parameters such as ohmic internal resistance, polarization resistance and polarization capacitance. The RC model can effectively simulate the ohmic voltage drop and polarization effect of the battery, and reflect the dynamic response of voltage with current during charging and discharging, which is closer to the actual electrochemical behavior than the simple model.

[0057] Step 3: Set the battery capacity value as a guess for the battery capacity (Cap). avai,guess This information is used as prior information, and a recursive least squares filtering algorithm is employed to perform online identification of equivalent circuit model parameters and online estimation of SOC.

[0058] In step three, the discrete transfer function of the lithium-ion battery is analyzed using the z-transform, and the calculation is as follows:

[0059]

[0060] The recursive least squares filtering algorithm is used to identify the parameters of each model. The parameter identification formula is calculated as follows:

[0061]

[0062] in, The parameter estimates for the k-th iteration (inclusive of R0, R...) p ,τ p (etc.). K(k) is the Kalman gain matrix. y(k) is the actual system output (battery voltage). φ(k) is the observation vector (composed of historical current and voltage data).

[0063] The real-time state of charge (SOC) of the battery is estimated using the extended Kalman filter algorithm, as follows:

[0064] The formula for calculating the state equation is as follows:

[0065] The calculation formula for the observation equation is as follows: U out =OCV(SOC)-IR0-U p ;

[0066] Where η is the coulomb efficiency (usually taken as 1). avai g The currently estimated available capacity (initial value is Cap). avai,guess =0.9 × rated capacity). U p Polarization voltage (U) p =IR p ).

[0067] The above technical solution is explained as follows: A recursive least squares filtering algorithm is used to identify equivalent circuit model parameters (such as ohmic internal resistance and polarization resistance) online, and an extended Kalman filter algorithm is used to estimate the real-time state of charge (SOC) of the battery. The state equation is based on the relationship between current integral and capacity, and the observation equation relates open-circuit voltage, internal resistance voltage drop, and terminal voltage, enabling dynamic parameter updates. The recursive least squares algorithm can track changes in battery parameters in real time, eliminating parameter drift caused by aging, temperature, and other factors, ensuring that the model always matches the actual state of the battery. Furthermore, it improves the real-time performance and accuracy of SOC estimation. The extended Kalman filter effectively handles measurement noise and model errors, maintaining a stable SOC output even under dynamic charge and discharge conditions, reducing accumulated errors compared to the traditional ampere-hour integration method. It also reduces sensitivity to initial conditions, using 0.9 times the rated capacity as the initial guess capacity, providing a reasonable starting point for the algorithm, shortening the convergence time, and avoiding long-term estimation distortion caused by initial value deviations.

[0068] Step 4: Based on the SOC estimation results and their reliability evaluation results, the weighted least squares method is used to calculate the optimal estimate of the battery's usable capacity.

[0069] In step four, the weighted least squares method is used, combined with the SOC estimate and its confidence weight, to calculate the usable battery capacity. Specifically as follows:

[0070] The weighted least squares formula is calculated as follows:

[0071]

[0072] thereby:

[0073]

[0074] Among them, w i The confidence weight for the i-th sampling point is calculated using w(SOC) from step one. ΔQi The cumulative charge / discharge capacity (ΔQ) at the i-th sampling point i =∫Idt); ΔSOC i The change in SOC at the i-th sampling point (ΔSOC) i =SOC i -SOC0). SOC0 is the true SOC value at the initial moment of the operating condition. N is the total number of sampling points. This represents the optimal estimated available battery capacity to be determined.

[0075] The above technical solution is explained as follows: Based on the SOC estimation results and its reliability evaluation, the weighted least squares method is used to calculate the optimal estimate of the battery's available capacity. By constructing a system of equations that includes the cumulative charge and discharge capacity, the change in SOC, and reliability weights, the optimal available capacity and the initial SOC value are obtained. This strengthens the contribution of high-reliability data. The weighted least squares algorithm gives greater weight to sampling points with high reliability (the interval with a large slope of the SOC-OCV curve) in the capacity calculation, reduces the interference of low-reliability data, and increases the stability of the estimation results. It achieves joint optimization of capacity and initial SOC. The system of equations solves for both available capacity and initial SOC simultaneously, avoiding the error propagation caused by treating the initial SOC as a known quantity, and improving the overall estimation accuracy. It can also reduce the impact of random errors. Through statistical optimization of data from multiple sampling points, the weighted average effect is used to offset the random fluctuations of a single measurement, making the capacity estimation results closer to the true value.

[0076] Step 5: Calculate the available capacity estimate results obtained in Step 4. As prior knowledge for the online identification algorithm of equivalent circuit model parameters, a more accurate SOC estimation and available capacity estimation are reimplemented, forming an iterative joint estimation of SOC and available capacity. The process stops when the change in battery capacity estimation results is less than 1%, thus completing the Capacity estimation. avai Estimate.

[0077] The specific method is as follows:

[0078] The result obtained in step four As the new prior capacity value, steps three and four are re-executed to form an iterative optimization until the capacity estimate changes by less than 1%.

[0079] The convergence condition is calculated as follows:

[0080]

[0081] The execution flow is as follows: using the current Cap avai Update model parameters (R0, R p ,τ p Then re-estimate the SOC (Extended Kalman Filter). Then recalculate the Cap. avai(Weighted least squares method). Finally, repeat until the convergence condition is met.

[0082] The above technical solution is explained as follows: Accuracy is improved through iteration. Each iteration updates the model parameters with a more accurate capacity, thereby improving the SOC estimation accuracy. The capacity is then recalculated based on the optimized SOC, forming a closed-loop optimization that gradually reduces the error. A 1% change threshold is set as the convergence condition to avoid excessive iteration and wasting computational resources, while ensuring the consistency of the final capacity estimate and eliminating iteration oscillations. It adapts to long-term operating conditions; the iteration process continuously tracks the slow decay of battery capacity. Even if battery aging causes changes in parameters and capacity, the algorithm can still maintain high-precision estimation through iterative updates.

[0083] In summary, in this embodiment, firstly, the reliability evaluation function constructed in step one accurately quantifies the estimation reliability of different SOC intervals, avoiding interference from low-quality data and laying a reliable foundation. Secondly, the first-order RC equivalent circuit model in step two balances accuracy and efficiency, accurately simulating the dynamic characteristics of the battery and adapting to online real-time estimation scenarios. Thirdly, combining recursive least squares and extended Kalman filtering enables dynamic parameter updates and real-time SOC estimation, reducing the impact of factors such as aging and temperature, and lowering the sensitivity to initial values. Fourthly, the weighted least squares algorithm in step four strengthens the contribution of high-reliability data, jointly optimizing capacity and initial SOC to offset random errors. Finally, the iterative optimization in step five forms a closed loop, ensuring result stability through convergence conditions, continuously tracking capacity decay, and meeting the full lifecycle monitoring requirements.

[0084] Although embodiments of the invention have been shown and described, the scope of the invention will be defined by the appended claims and their equivalents by those skilled in the art.

Claims

1. A method for estimating the usable capacity of a lithium-ion battery based on SOC, characterized in that, Specifically, the following steps are included: Step 1: Utilize the phase transition characteristics of the positive and negative electrode materials during the charging and discharging of lithium batteries, and construct a reliability evaluation function for the SOC estimation results based on the rate of change of the SOC-OCV curve of the battery. Step 2: Collect battery current, voltage, and time interval data of lithium-ion batteries under charge and discharge conditions, establish a first-order RC equivalent circuit model of lithium-ion batteries, and simulate the electrochemical characteristics of lithium-ion batteries. Step 3: Set the battery capacity value as a guess of the battery capacity (Cap). avai,guess This information is used as prior information, and a recursive least squares filtering algorithm is employed to perform online identification of equivalent circuit model parameters and online estimation of SOC. Step 4: Based on the SOC estimation results and their reliability evaluation results, the weighted least squares method is used to calculate the optimal estimate of the battery's usable capacity. Step 5: Calculate the available capacity estimate results obtained in Step 4. As prior knowledge for the online identification algorithm of equivalent circuit model parameters, a more accurate SOC estimation and available capacity estimation are reimplemented, forming an iterative joint estimation of SOC and available capacity. The process stops when the change in battery capacity estimation results is less than 1%, thus completing the Capacity estimation. avai Estimate.

2. The method for estimating the usable capacity of a lithium-ion battery based on SOC according to claim 1, characterized in that, In step one, the formula for calculating the credibility evaluation function is as follows: Where w(SOC) is the confidence coefficient of the SOC estimation result, and its value ranges from 0 to 1. The slope of the SOC-OCV curve at a certain SOC point represents the rate of change of voltage with respect to SOC; OCV SOC＝100％ OCV is the open-circuit voltage of the battery when it is fully charged, i.e., SOC = 100%. SOC＝0％ This is the open-circuit voltage of the battery when it is in a depleted state, i.e., SOC = 0%.

3. The method for estimating the usable capacity of a lithium-ion battery based on SOC according to claim 2, characterized in that, In step two, the transfer function calculation formula for the RC equivalent circuit model is as follows: Among them, U out Where is the battery terminal voltage, OCV is the battery open-circuit voltage, I is the battery current; R0 is the internal resistance in ohms; R p For polarization resistance; C p is the polarization capacitor; s is the complex frequency variable of the Laplace transform; G(s) is the frequency domain transfer function of the battery.

4. The method for estimating the usable capacity of a lithium-ion battery based on SOC according to claim 3, characterized in that, In step three, the discrete transfer function of the lithium-ion battery is analyzed using the z-transform, and the calculation is as follows: The recursive least squares filtering algorithm is used to identify the parameters of each model. The parameter identification formula is calculated as follows: in, y(k) represents the parameter estimate for the k-th iteration; K(k) is the Kalman gain matrix; y(k) is the actual system output; φ(k) is the observation vector. The real-time state of charge (SOC) of the battery is estimated using the extended Kalman filter algorithm, as follows: The formula for calculating the state equation is as follows: The calculation formula for the observation equation is as follows: U out =OCV(SOC)-IR0-U p ; Where η is the Coulomb efficiency; Cap avai g This represents the currently estimated available capacity; U p This is the polarization voltage.

5. The method for estimating the usable capacity of a lithium-ion battery based on SOC according to claim 4, characterized in that, In step four, the weighted least squares method is used, combined with the SOC estimate and its confidence weight, to calculate the usable battery capacity. Specifically as follows: The weighted least squares formula is calculated as follows: thereby: Among them, w i ΔQ represents the confidence weight of the i-th sampling point. i ΔSOC represents the cumulative charge / discharge capacity at the i-th sampling point. i SOC represents the change in SOC at the i-th sampling point; SOC0 represents the true SOC value at the initial moment of the operating condition; N represents the total number of sampling points. This represents the optimal estimated available battery capacity to be determined.

6. The method for estimating the usable capacity of a lithium-ion battery based on SOC according to claim 5, characterized in that, The available capacity estimate Cap obtained in step four avai As prior knowledge for the online identification algorithm of equivalent circuit model parameters, a more accurate SOC estimation and available capacity estimation are reimplemented, forming an iterative joint estimation of SOC and available capacity. The estimation stops when the change in battery capacity is less than 1%, thus achieving convergence and accurate Capacity estimation. avai The estimation method is as follows: The result obtained in step four As the new prior capacity value, steps three and four are re-executed to form an iterative optimization until the capacity estimate changes by less than 1%. The convergence condition is calculated as follows: The execution flow is as follows: using the current Update the model parameters, re-estimate the SOC, and then recalculate the Cap. avai Finally, this process is repeated until the convergence condition is met.

Images