An adaptive method for estimating state of charge of lithium-ion batteries
By combining the second-order RC equivalent circuit model and the adaptive extended Kalman particle filtering algorithm, the model parameters are updated online using HPPC cycle condition experiments and recursive least squares method, which solves the high computational complexity in the state of charge estimation of lithium-ion batteries Problem, efficient estimation under nonlinear and non-Gaussian noise is achieved.
Patent Information
- Application Number
- CN202210212196.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-04
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2042-03-04
AI Technical Summary
The existing lithium-ion battery state-of-charge estimation method is highly complex when dealing with system nonlinearity and noise non-Gaussian conditions, and the adaptive extended Kalman filtering algorithm cannot effectively process non-Gaussian noise. Particle filtering algorithm has particles. The degradation problem leads to low algorithm computing efficiency.
The second-order RC equivalent circuit model is used to combine the adaptive extended Kalman particle filtering algorithm, and the model parameters are identified offline through HPPC cycle condition experiment and exponential fitting method, and the recursive least squares method with forgetting factors is used for online updates to deduce the battery state. Space equations and observation equations realize the estimation of the charge state of lithium-ion batteries.
It improves the accuracy and stability of state-of-charge estimation of lithium-ion batteries, reduces the computational complexity of the algorithm, and ensures efficient computing under nonlinear and non-Gaussian noise.
Smart Images

Figure CN114740385B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of lithium-ion batteries, and in particular relates to an adaptive lithium-ion battery charge state estimation method. Background Art
[0002] Lithium batteries are widely used in new energy vehicles, mobile robots, new energy and other fields due to their high energy density, high operating voltage, low self-discharge rate and long service life. However, lithium batteries also have many disadvantages, such as high internal resistance, which leads to high temperature at high charge and discharge rates. At the same time, overcharging and over-discharging will damage the battery, shorten the battery life, and even cause accidents such as explosions. Therefore, detecting the working status of lithium batteries can significantly improve the performance and life of lithium batteries.
[0003] The state of charge (SOC) of a lithium battery is an important indicator for measuring the state of a lithium battery, and it directly represents the remaining power of the battery. SOC is defined as the percentage of power remaining after a lithium battery is fully charged. Accurate SOC estimation can prevent battery overcharge and overdischarge, improve battery performance, and extend battery life. Due to the nonlinear characteristics of lithium-ion batteries, SOC cannot be directly measured by sensors. Currently, the commonly used methods in the field of SOC estimation include data-driven methods, direct estimation methods, and model-based estimation methods.
[0004] The SOC estimation method based on the equivalent circuit model simulates the battery dynamics by establishing an equivalent circuit model, which has strong robustness and is therefore very suitable for SOC estimation of lithium-ion batteries. In addition, the method based on the equivalent circuit model needs to be combined with a filtering algorithm to estimate the battery SOC. In the invention patent "Power lithium battery SOC estimation method based on adaptive Kalman filtering method" with patent application number CN201910567240.2, the adaptive extended Kalman filter algorithm is used to estimate the battery SOC. Although the adaptive extended Kalman filter algorithm has low computational complexity, the algorithm cannot process non-Gaussian noise, and there is a model linearization error when the system is strongly nonlinear. In the invention patent "State of charge estimation method and battery management system based on adaptive particle filtering" with patent application number CN201910567240.2, the particle filter algorithm is used to estimate the battery SOC. Although the particle filter algorithm can handle system nonlinearity and non-Gaussian noise, there is a problem that particle degradation causes the algorithm calculation complexity to increase. Summary of the invention
[0005] The purpose of the present invention is to provide an adaptive lithium-ion battery state of charge estimation method to solve the problems existing in particle filtering and adaptive extended Kalman filtering: it can handle system nonlinearity and non-Gaussian noise, and can also ensure the computational efficiency of the algorithm.
[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:
[0007] An adaptive lithium-ion battery state of charge estimation method comprises the following steps:
[0008] Step S1, establishing a second-order RC equivalent circuit model of a lithium-ion battery, conducting an HPPC cycle condition experiment, and using an exponential fitting method to perform offline identification of parameters of the equivalent circuit model;
[0009] Step S2, obtaining a correlation curve between the battery open circuit voltage and the state of charge according to the results of the HPPC cycle condition experiment;
[0010] Step S3, inputting the correlation curve of the current, open circuit voltage and state of charge of the battery dynamic stress test and the equivalent circuit model parameters identified offline into the battery equivalent circuit model to obtain the corresponding equivalent circuit model output voltage; comparing the error between the equivalent circuit model output voltage and the actual dynamic stress test voltage to verify the accuracy of the equivalent circuit model;
[0011] Step S4, using a recursive least square method with a forgetting factor to update the parameters of the equivalent circuit model online;
[0012] Step S5, deriving the state space equation and observation equation of the battery according to the mathematical expression of the equivalent circuit model, and estimating the state of charge of the lithium-ion battery by using an adaptive extended Kalman particle filter.
[0013] In a more preferred technical solution, the mathematical expression of the equivalent circuit model in step S1 is:
[0014]
[0015] Among them, U oc is the open circuit voltage of the battery; R0 is the ohmic internal resistance of the battery; R1 and C1 are the resistance and capacitance representing the electrochemical polarization reaction of the battery; R2 and C2 are the resistance and capacitance representing the concentration polarization reaction of the battery; U t is the terminal voltage of the battery.
[0016] In a more optimal technical solution, the specific process of the HPPC experiment in step S1 is: after the battery is allowed to stand for 5 minutes, when the battery voltage is charged to 4.2V with a constant current of 0.5C, it is switched to a constant voltage charging mode; after the battery current is charged to less than 0.05C in the constant voltage charging mode, charging is stopped and the battery is allowed to stand for 2 hours; then the battery is discharged at 3C, and the battery is allowed to stand for 1 hour every time the state of charge decreases by 10%, and the corresponding open circuit voltage is measured, and the above steps are repeated until the battery state of charge is 0%.
[0017] In a more preferred technical solution, the offline identification process of the parameters in step S1 is:
[0018] (1) Based on the terminal voltage V1 of the lithium-ion battery 1 second before the HPPC pulse is loaded, the terminal voltage V2 at the moment of loading the HPPC pulse, the terminal voltage V3 at the moment of loading the HPPC pulse, and the terminal voltage V4 1 second after loading the HPPC pulse, the ohmic resistance R0 in the lithium-ion equivalent circuit is calculated as follows:
[0019]
[0020] (2) The zero input voltage response of the battery during the static process of the HPPC pulse charge and discharge experiment is:
[0021]
[0022] Among them, U1(0) and U2(0) are the terminal voltages of the two RC networks, τ1=R1C1, τ2=R2C2; Use the fitting toolbox of MATLAB to fit the above equation and calculate the specific values of the two time constants τ1 and τ2;
[0023] (3) The zero-state response of the RC network during the HPPC pulse charge and discharge experiment is:
[0024]
[0025] Similarly, the above equation is fitted in the matlab fitting toolbox to obtain the specific values of R1 and R2;
[0026] (4) Based on the relationship τ1=R1C1, τ2=R2C2, and the obtained R1 and R2, the specific values of C1 and C2 are solved.
[0027] In a more optimal technical solution, the specific model of the open circuit voltage and state of charge correlation curve in step S2 is:
[0028] U oc =k1SOC 6 -k2SOC 3 +k3SOC 4 -k4SOC 3 +k5SOC 2 +k6SOC+k7
[0029] Among them, k1, k2...k7 are coefficients to be fitted, which are obtained by fitting the open circuit voltage points under 10 charge states.
[0030] In a more optimal technical solution, the dynamic stress test in step S3 refers to: performing a cyclic test on the battery using a dynamic stress condition until the voltage is less than 3V; wherein, a dynamic stress condition contains multiple charge and discharge pulses, and each dynamic stress condition test takes 360 seconds, and the battery needs to be left to stand for 120 seconds before the next dynamic stress condition test is performed; the shortest duration of a single discharge pulse in the dynamic stress condition is 8 seconds, the longest time can reach 40 seconds, the maximum discharge current is 2C, and the maximum charging current is 0.5C.
[0031] In a more optimal technical solution, step S4 specifically includes:
[0032] (1) Calculate the error between the parameter identification result at the previous moment and the actual battery voltage measured by the sensor:
[0033]
[0034] Among them, y(k) is the actual battery voltage at the current moment, is the measurement vector at the current moment, is the model parameter vector estimated at the last moment, e(k) is the error between the model output voltage obtained by parameter identification at the last moment and the actual battery output voltage;
[0035] (2) Update the gain matrix
[0036]
[0037] Among them, P(k-1) is the covariance matrix of the previous moment, λ is the forgetting factor, and K(k) is the gain matrix of the current moment;
[0038] (3) Update the covariance matrix to calculate the gain matrix at the next moment
[0039]
[0040] (4) Use the gain matrix and error to update the current parameter identification results
[0041]
[0042] In a more preferred technical solution, the state space equation and observation equation of the battery in step S5 are respectively:
[0043]
[0044] U t,k =U oc,k -U 1,k -U 2,k -I k R0
[0045] Where T is the sampling time, k is the discrete time variable, and Q is the rated capacity of the battery.
[0046] In a more optimal technical solution, the particles of the adaptive extended Kalman particle filter described in step S5 are a series of random samples with weights, and the specific steps are:
[0047] (1) Initialize particle set and the particle weights are set to
[0048] (2) Using the adaptive extended Kalman filter as the importance sampling function
[0049] (3) Update the particle weight according to the following formula:
[0050]
[0051] (4) Normalize the weights:
[0052]
[0053] (5) Calculate the effective number of particles according to the following formula:
[0054]
[0055] Among them, N eff is the effective number of particles, if N eff If it is less than 0.7N, resampling is required;
[0056] (6) Obtain the current estimation result:
[0057]
[0058] (7) Return to step (2) and set k = k + 1 until the loop ends.
[0059] The beneficial effects of the present invention are as follows: the present invention establishes a second-order RC equivalent circuit model of a lithium-ion battery, and uses an exponential fitting method to identify the model parameters offline; then the HPPC cycle operating condition experiment is used to fit the OCV and SOC correlation curves. The model parameters are identified online using a recursive least squares method containing a forgetting factor, and based on the online identified model parameters, combined with an adaptive extended Kalman particle filter algorithm, the charge state estimation of the lithium-ion battery is achieved. The present invention combines an adaptive extended Kalman filter algorithm with a particle filter algorithm, and improves the stability and accuracy of the SOC estimation result by outputting a weighted average of multiple particles; the adaptive extended Kalman filter is used as an importance sampling function to improve the computational efficiency of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS
[0060] Figure 1 It is a flow chart of the present invention.
[0061] Figure 2 It is a schematic diagram of the second-order RC equivalent circuit model in the present invention. DETAILED DESCRIPTION
[0062] The present invention is described in detail below in conjunction with an embodiment. This embodiment is based on the technical solution of the present invention, and provides a detailed implementation method and a specific operation process to further explain the technical solution of the present invention.
[0063] like Figure 1 As shown, a charge state estimation method based on an adaptive extended Kalman particle filter of the present invention comprises the following steps:
[0064] Step S1, first establish a second-order RC equivalent circuit model of a lithium-ion battery. The equivalent circuit model is as follows: Figure 2 As shown. R0 represents the ohmic internal resistance; the first RC network, R1 and C1, is used to describe the electrochemical polarization reaction of the battery; the second RC network, R2 and C2, is used to describe the concentration polarization reaction of the battery; U t is the battery terminal voltage; U oc is the open circuit voltage of the battery; I is the battery load current. According to Kirchhoff's current-voltage law, the mathematical expression of the second-order RC equivalent circuit model can be obtained:
[0065]
[0066] In order to identify the initial values of unknown parameters R0, R1, R2, C1, and C2 in the model, the present invention adopts an off-line parameter identification method. The terminal voltage data of the HPPC cycle condition is used in combination with the fitting toolbox of MATLAB to identify the above parameters.
[0067] Before parameter identification, it is necessary to conduct an HPPC pulse charge and discharge experiment. The process is as follows: first, let the battery stand for 5 minutes, then charge the battery voltage to 4.2V at a constant current of 0.5C, and then switch to constant voltage charging mode. After the battery current is charged to less than 0.05C in constant voltage charging mode, stop charging and let the battery stand for 2 hours. After that, discharge at 3C, and let the battery stand for 1 hour for every 10% drop in the state of charge, and measure the corresponding open circuit voltage. Repeat this step until the battery state of charge is 0%.
[0068] After that, the initial values of the model parameters are identified offline. According to the terminal voltage V1 1 second before the HPPC pulse is loaded, the terminal voltage V2 at the moment of the HPPC pulse loading, the terminal voltage V3 at the moment of the HPPC pulse ending, and the terminal voltage V4 1 second after the HPPC pulse ending, the ohmic resistance R0 in the lithium ion equivalent circuit can be calculated. The calculation formula is as follows:
[0069]
[0070] Then, by fitting the zero input voltage response of the battery during the static process of the HPPC pulse charge and discharge experiment, the specific values of the two time constants τ1 and τ2 can be calculated:
[0071]
[0072] By fitting the zero-state response of the RC network, we can get the values of R1 and R2:
[0073]
[0074] Since τ1=R1C1, τ2=R2C2, and R1 and R2 are known at this time, C1 and C2 can be solved, thereby completing the initial value identification of the parameters of the second-order RC equivalent circuit model.
[0075] Step S2, fitting the correlation curve between open circuit voltage and state of charge. First, 10 data points are calibrated using the HPPC cycle condition experiment, namely, the open circuit voltage corresponding to 100% state of charge, the open circuit voltage corresponding to 90% state of charge…the open circuit voltage corresponding to 0% state of charge. Then, a sixth-order polynomial model is used to fit the above 10 data points. The expression of the sixth-order polynomial model is:
[0076] U oc =k1SOC 6 -k2SOC 3 +k3SOC 4 -k4SOC 3 +k5SOC 2 +k6SOC+k7
[0077] After completing the fitting of the open circuit voltage and the state of charge calibration point, the values of k1, k2…k7 can be obtained. This polynomial is also the correlation curve between the open circuit voltage and the state of charge.
[0078] Step S3, use dynamic stress test condition cycle test on the battery to verify the accuracy of the established model. First, a dynamic stress condition test is to be carried out, that is, within 360 seconds, the pulse current applied is of different duration and size. The shortest duration of a single discharge pulse is 8 seconds, the longest time can reach 40 seconds, the maximum discharge current is 2C, and the maximum charging current is 0.5C. Then, after an interval of 120 seconds, the next dynamic stress condition test is carried out until the terminal voltage of the battery is lower than 3V.
[0079] The expression of the output voltage of the battery model is:
[0080] U t =U oc -U1-U2-IR0
[0081] Among them, U oc It can be obtained through the correlation curve of state of charge and open circuit voltage, while U1 and U2 can be calculated through the identified R1, R2, C1, and C2. Therefore, by inputting the current of the battery dynamic stress test into the battery model, the corresponding model output voltage can be obtained. Then, by comparing the error between the model output voltage and the actual dynamic stress test voltage, the accuracy of the model is verified.
[0082] In step S4, the model parameters are identified online using recursive least squares with forgetting factor. The main steps include:
[0083] (1) Calculate the error between the parameter identification result at the previous moment and the actual battery voltage measured by the sensor
[0084]
[0085] Among them, y(k) is the actual battery voltage at the current moment, is the measurement vector at the current moment, is the model parameter vector estimated at the last moment, and e(k) is the error between the model output voltage obtained by parameter identification at the last moment and the actual battery output voltage.
[0086] (2) Update the gain matrix
[0087]
[0088] Among them, P(k-1) is the covariance matrix of the previous moment, λ is the forgetting factor, and K(k) is the gain matrix of the current moment.
[0089] (3) Update the covariance matrix to calculate the gain matrix at the next moment
[0090]
[0091] (4) Use the gain matrix and error to update the current parameter identification results
[0092]
[0093] In step S5, the state of charge of the lithium-ion battery is estimated based on the adaptive extended Kalman particle filter. Before performing the specific filtering algorithm, it is necessary to establish the relevant state equation and measurement equation. For a discrete control system, its state equation and observation equation can be expressed as:
[0094] x k =A k-1 x k-1 +B k-1 u k-1 +w k-1
[0095] y k =C k x k +D k u k +v k
[0096] Among them, x k is the system state vector at the kth moment; y k is the measurement vector at the kth moment; u k is the input vector at the kth moment; w and v are the process noise and measurement noise of the system respectively; A, B, C, D matrices are the parameter matrices of the system. T As the state vector of the system, U t As the output of the system, I is the input of the system. According to the mathematical expression of the second-order RC equivalent circuit model and the definition of SOC, write the state equation and observation equation:
[0097]
[0098] U t,k =U oc,k -U 1,k -U 2,k -I k R0
[0099] Where T is the sampling time, k is the discrete time variable, Q is the rated capacity of the battery, D k =-R0.
[0100] Then, the state of charge of the lithium-ion battery is estimated using an adaptive extended Kalman particle filter, where the particles are a group of random samples with weights. The specific steps are:
[0101] (1) Initialize particle set and the particle weights are set to
[0102] (2) Use adaptive extended Kalman filter to generate importance sampling function The ABCD matrix derived above is needed. Since the adaptive extended Kalman filter is a prior art, it will not be repeatedly described in the present invention;
[0103] (3) Update the particle weight according to the following formula:
[0104]
[0105] (4) Normalize the weights:
[0106]
[0107] (5) Calculate the effective number of particles according to the following formula:
[0108]
[0109] Among them, N eff is the effective number of particles, if N eff If it is less than 0.7N, resampling is required;
[0110] (6) Obtain the current estimation result:
[0111]
[0112] (7) Return to step (2) and set k = k + 1 until the loop ends.
[0113] The above embodiments are preferred embodiments of the present application. Ordinary technicians in this field can also make various changes or improvements on this basis. Without departing from the overall concept of the present application, these changes or improvements should fall within the scope of protection required by the present application.
Claims
1. An adaptive lithium-ion battery state of charge estimation method, characterized in that: The following steps are involved: Step S1, establishing a second-order RC equivalent circuit model of a lithium-ion battery, conducting an HPPC cycle condition experiment, and using an exponential fitting method to perform offline identification of parameters of the equivalent circuit model; Step S2, obtaining a correlation curve between the battery open circuit voltage and the state of charge according to the results of the HPPC cycle condition experiment; Step S3, inputting the correlation curve of the current, open circuit voltage and state of charge of the battery dynamic stress test and the equivalent circuit model parameters identified offline into the battery equivalent circuit model to obtain the corresponding equivalent circuit model output voltage; Compare the error between the output voltage of the equivalent circuit model and the actual dynamic stress test voltage to verify the accuracy of the equivalent circuit model; Step S4, using a recursive least square method with a forgetting factor to update the parameters of the equivalent circuit model online; Step S5, deriving the state space equation and observation equation of the battery according to the mathematical expression of the equivalent circuit model, and estimating the state of charge of the lithium-ion battery by using an adaptive extended Kalman particle filter.
2. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The mathematical expression of the equivalent circuit model in step S1 is: Among them, U oc is the open circuit voltage of the battery; R0 is the ohmic internal resistance of the battery; R1 and C1 are the resistance and capacitance representing the electrochemical polarization reaction of the battery; R2 and C2 are the resistance and capacitance representing the concentration polarization reaction of the battery; U t is the terminal voltage of the battery.
3. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The specific process of the HPPC experiment in step S1 is: after the battery is left to stand for 5 minutes, when the battery voltage is charged to 4.2V with a constant current of 0.5C, it is switched to a constant voltage charging mode; after the battery current is charged to less than 0.05C in the constant voltage charging mode, charging is stopped and the battery is left to stand for 2 hours; then the battery is discharged at 3C, and the battery is left to stand for 1 hour every time the state of charge decreases by 10%, and the corresponding open circuit voltage is measured, and the above steps are repeated until the battery state of charge is 0%.
4. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The offline identification process of the parameters in step S1 is: (1) Based on the terminal voltage V1 of the lithium-ion battery 1 second before the HPPC pulse is loaded, the terminal voltage V2 at the moment of loading the HPPC pulse, the terminal voltage V3 at the moment of loading the HPPC pulse, and the terminal voltage V4 1 second after loading the HPPC pulse, the ohmic resistance R0 in the lithium-ion equivalent circuit is calculated as follows: (2) The zero input voltage response of the battery during the static process of the HPPC pulse charge and discharge experiment is: Among them, U1(0) and U2(0) are the terminal voltages of the two RC networks, τ1=R1C1, τ2=R2C2; Use the fitting toolbox of MATLAB to fit the above equation and calculate the specific values of the two time constants τ1 and τ2; (3) The zero-state response of the RC network during the HPPC pulse charge and discharge experiment is: Similarly, the above equation is fitted in the matlab fitting toolbox to obtain the specific values of R1 and R2; (4) Based on the relationship τ1=R1C1, τ2=R2C2, and the obtained R1 and R2, the specific values of C1 and C2 are solved.
5. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The specific model of the open circuit voltage and state of charge correlation curve in step S2 is: U oc =k1SOC 6 -k2SOC 3 +k3SOC 4 -k4SOC 3 +k5SOC 2 +k6SOC+k7 Among them, k1, k2...k7 are coefficients to be fitted, which are obtained by fitting the open circuit voltage points under 10 charge states.
6. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The dynamic stress test described in step S3 refers to: performing a cyclic test on the battery using a dynamic stress condition until the voltage is less than 3V; wherein, a dynamic stress condition contains multiple charge and discharge pulses, each dynamic stress condition test takes 360 seconds, and the battery needs to be left to stand for 120 seconds before the next dynamic stress condition test; the shortest duration of a single discharge pulse in the dynamic stress condition is 8 seconds, the longest time can reach 40 seconds, the maximum discharge current is 2C, and the maximum charging current is 0.5C.
7. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: Step S4 specifically includes: (1) Calculate the error between the parameter identification result at the previous moment and the actual battery voltage measured by the sensor: Among them, y(k) is the actual battery voltage at the current moment, is the measurement vector at the current moment, is the model parameter vector estimated at the last moment, e(k) is the error between the model output voltage obtained by parameter identification at the last moment and the actual battery output voltage; (2) Update the gain matrix Among them, P(k-1) is the covariance matrix of the previous moment, λ is the forgetting factor, and K(k) is the gain matrix of the current moment; (3) Update the covariance matrix to calculate the gain matrix at the next moment (4) Use the gain matrix and error to update the current parameter identification results 8. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The state space equation and observation equation of the battery described in step S5 are: IN t,k =U oc,k -IN 1,k -IN 2,k -AND k R0 Where T is the sampling time, k is the discrete time variable, and Q is the rated capacity of the battery.
9. The method for estimating the state of charge of a lithium-ion battery according to claim 1, wherein: The particles of the adaptive extended Kalman particle filter described in step S5 are a series of random samples with weights, and the specific steps are: (1) Initialize particle set and the particle weights are set to (2) Using the adaptive extended Kalman filter as the importance sampling function (3) Update the particle weight according to the following formula: (4) Normalize the weights: (5) Calculate the effective number of particles according to the following formula: Among them, N eff is the effective number of particles, if N eff If it is less than 0.7N, resampling is required; (6) Obtain the current estimation result: (7) Return to step (2) and set k = k + 1 until the loop ends.
