Lithium ion battery SOC estimation method based on coulomb counting and UKF joint algorithm

Abstract

The invention relates to a lithium ion battery SOC (State of Charge) estimation method based on coulomb counting and a UKF (Unscented Kalman Filter) joint algorithm. The method comprises the following steps: establishing a first-order equivalent circuit model of a lithium ion battery; constructing an SOC estimation equation based on coulomb counting and a UKF joint algorithm; updating the SOC estimation value; gaussian noise is introduced into an input voltage parameter, a novel SOC estimation equation and an estimation process are established, and simulation verification is performed in different battery degradation periods. The SOC estimation method has the beneficial effects that the coulomb counting method and the unscented Kalman filtering algorithm are combined, so that the accumulative error can be reduced, and the SOC estimation precision can be improved; moreover, a first-order equivalent circuit model is adopted to simplify battery dynamic characteristic modeling, and through joint calibration of a UKF state equation and equivalent circuit parameters, calculation resource consumption is reduced while precision is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium-ion batteries, and more specifically, it relates to a method for estimating the state of charge (SOC) of a lithium-ion battery based on a combined algorithm of Coulomb counting and UKF. Background Art

[0002] Under different operating conditions, the charge-discharge performance and internal characteristics of lithium-ion batteries will change. By using accurate and effective circuit models and battery models, designers can predict and optimize the battery's operating time, real-time state of charge (SOC), and circuit performance. If the SOC value of the battery cannot be effectively predicted, overcharging or over-discharging of the battery voltage, current, and capacity may occur, leading to permanent damage to a single lithium-ion battery. The charge-discharge process of lithium-ion batteries is a non-linear dynamic system, and traditional SOC estimation methods such as the extended Kalman filter algorithm (EKF) cannot simulate non-linear systems and have an error accumulation effect. Summary of the Invention

[0003] The object of the present invention is to address the deficiencies of the existing technology and propose a method for estimating the SOC of a lithium-ion battery based on a combined algorithm of Coulomb counting and UKF.

[0004] In a first aspect, there is provided a method for estimating the SOC of a lithium-ion battery based on a combined algorithm of Coulomb counting and UKF, including:

[0005] Step 1: Establish a first-order equivalent circuit model of the lithium-ion battery;

[0006] Step 2: Based on the combined algorithm of Coulomb counting and UKF, construct an SOC estimation equation;

[0007] Step 3: Update the SOC estimated value;

[0008] Step 4: Introduce Gaussian noise into the input voltage parameters, establish a new SOC estimation equation and estimation process, and perform simulation verification at different battery degradation cycles.

[0009] Preferably, in Step 1, the first-order equivalent circuit model includes a parallel RC network, an ohmic resistor in series with the RC network, and an ideal voltage source, where the RC network is composed of a polarization resistor and a polarization capacitor.

[0010] Preferably, in Step 1, it further includes: combining Kirchhoff's law and the first-order equivalent circuit model to establish a voltage-current relationship equation.

[0011] Preferably, Step 2 includes:

[0012] Step 2.1: Based on the Coulomb counting method, establish an SOC estimation equation;

[0013] Step 2.2: Based on the parameters of the first-order equivalent circuit model, establish the state equation and observation equation of the UKF algorithm;

[0014] Step 2.3: Use the parameters of the first-order equivalent circuit model and the parameters of the Coulomb counting method to calculate the coefficient parameters of the state equation and the observation equation.

[0015] Preferably, step 3 includes:

[0016] Step 3.1: Parameter initialization, including: initializing state variables, covariance matrices, Sigma points, mean weight parameters, and covariance weight parameters;

[0017] Step 3.2: Perform time updates including state estimation, error covariance, and output estimation;

[0018] Step 3.3: Update the covariance matrix;

[0019] Step 3.4: Perform state estimation updates and error covariance updates.

[0020] In a second aspect, a lithium-ion battery SOC estimation system based on a combined Coulomb counting and UKF algorithm is provided for performing any of the methods described in the first aspect, including:

[0021] A building module for building a first-order equivalent circuit model of a lithium-ion battery;

[0022] A construction module for constructing an SOC estimation equation based on a combined Coulomb counting and UKF algorithm;

[0023] An update module for updating the SOC estimated value;

[0024] A verification module for introducing Gaussian noise into the input voltage parameters, establishing a new SOC estimation equation and estimation process, and performing simulation verification at different battery degradation cycles.

[0025] In a third aspect, a computer storage medium is provided, in which a computer program is stored; when the computer program runs on a computer, the computer is enabled to execute any of the methods described in the first aspect.

[0026] In a fourth aspect, an electronic device is provided, including:

[0027] A memory for storing a computer program;

[0028] A processor for executing the computer program to implement any of the methods described in the first aspect.

[0029] The beneficial effects of the present invention are:

[0030] 1. The present invention combines the Coulomb Counter (CC) method with the Unscented Kalman Filter (UKF) algorithm to reduce the cumulative error and improve the accuracy of SOC estimation.

[0031] 2. The present invention simplifies the modeling of battery dynamic characteristics using a first-order equivalent circuit model. By jointly calibrating the UKF state equation and the equivalent circuit parameters, it reduces the consumption of computing resources while ensuring accuracy.

[0032] 3. The present invention introduces Gaussian white noise to simulate battery capacity degradation and verifies the applicability of the CC-UKF joint algorithm throughout the entire battery life cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a schematic diagram of the first-order equivalent circuit model provided by this application;

[0034] Figure 2a is a schematic diagram of the change in the rated capacity of the battery for 1 degradation cycle provided by this application;

[0035] Figure 2b is a schematic diagram of the change in the rated capacity of the battery for 3 degradation cycles provided by this application;

[0036] Figure 2c is a schematic diagram of the change in the rated capacity of the battery for 5 degradation cycles provided by this application;

[0037] Figure 2d is a schematic diagram of the change in the rated capacity of the battery for 7 degradation cycles provided by this application;

[0038] Figure 3 is a schematic diagram of the SOC estimated value and the actual measured SOC value based on CC-UKF provided by this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The present invention will be further described below in conjunction with embodiments. The description of the following embodiments is only for helping to understand the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several modifications can still be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

[0040] Embodiment 1:

[0041] To solve the problems of the existing technology, Embodiment 1 of the present application provides a method for estimating the state of charge (SOC) of a lithium-ion battery based on a combined Coulomb counting and unscented Kalman filter (UKF) algorithm. Meanwhile, a first-order equivalent circuit model is established, and the combined algorithm is applied to the accurate estimation process of the SOC of the lithium-ion battery. To overcome the shortcomings of the extended Kalman filter (EKF), the UKF generates more Sigma points near the current state sampling point according to the covariance. The generated Sigma points are propagated in the non-linear mapping to accurately estimate the mean and covariance of the SOC. The unscented UKF algorithm has good performance for non-linear systems. At the same time, Coulomb counting can overcome the shortcoming of the capacity degradation of lithium-ion batteries caused by the increase in the degradation cycle.

[0042] Embodiment 1 of the present application provides a method for estimating the SOC of a lithium-ion battery based on a combined Coulomb counting and UKF algorithm. The input parameters of the UKF algorithm are the temperature value, the voltage value (with Gaussian noise added), and the capacity value of the lithium-ion battery estimated by the CC algorithm. Specifically, the method includes:

[0043] Step 1: Establish a first-order equivalent circuit model of the lithium-ion battery.

[0044] Step 1 includes:

[0045] Step 1.1: Establish the equivalent circuit model.

[0046] In the process of SOC estimation, lithium-ion battery modeling plays an important role. Battery modeling is also called the mathematical description of the battery based on the internal structure and charge-discharge characteristics of the battery. Battery models are divided into empirical models, equivalent circuit models, electrochemical models, and artificial intelligence-based models. Among them, the equivalent circuit model (ECM) has a simple circuit structure, good parameter identification characteristics, low computational complexity, and good dynamic performance, and is often used in the lithium-ion battery models of electric vehicles.

[0047] In Step 1.1, the Thevenin first-order equivalent circuit model is adopted in the embodiment of the present application. As Figure 1 shown, the first-order equivalent circuit model consists of a parallel RC network, and the voltage of the parallel RC network is . The parallel RC network is connected in series with the ideal voltage source and the ohmic resistance . is the polarization resistance, is the polarization capacitance, is the output terminal voltage, is the output current.

[0048] Step 1.2, Establishment of equivalent circuit relationship equations. Combining Kirchhoff's law with the Thevenin first-order equivalent circuit model, the voltage and current relationship equations are established as follows:

[0049] (1)

[0050] (2)

[0051] In the formula, is the output terminal voltage, is the voltage value of the RC parallel network at the Kth moment, is the voltage value of the RC parallel network at the (K - 1)th moment. is the current value at the (K - 1)th moment. is the time constant, satisfying . is the sampling time interval.

[0052] Step 2, Construct the SOC estimation equation based on the combined Coulomb counting and UKF algorithms.

[0053] In Step 2, first, combine the first-order equivalent circuit model and the CC algorithm to establish the SOC estimation equation. Second, combine the UKF state equation and observation equation with the equivalent circuit model parameters to establish the state equation and observation equation of the CC-UKF combined algorithm. Finally, calculate the coefficient parameters A to D of the UKF estimation equation using the equivalent circuit model parameters and CC algorithm parameters.

[0054] Specifically, Step 2 includes:

[0055] Step 2.1, Establish the SOC estimation equation based on the Coulomb counting method:

[0056] (3)

[0057] In the formula, is the initial SOC value, is the SOC value at the current time t, is the battery charge and discharge efficiency.

[0058] Step 2.2, Establish the state equation and observation equation of the UKF algorithm based on the parameters of the first-order equivalent circuit model.

[0059] Specifically, the state space equation of the lithium-ion battery is as shown in formula (4). Assume is the state variable, is the observation variable, and K is the sampling time. Then the state parameters are and . Establish the observation equation as shown in formula (5), and then the observation parameter is . , , and are coefficient variables.

[0060] (4)

[0061] (5)

[0062] In the formula, is an observation parameter, is model noise, is Gaussian noise, and and are uncorrelated with each other.

[0063] Step 2.3: Calculate the coefficient parameters of the state equation and the observation equation by using the parameters of the first-order equivalent circuit model and the parameters of the Coulomb counting method.

[0064] Specifically, combining formulas (1)-(5), the coefficient parameter expression is calculated as:

[0065] (6)

[0066] (7)

[0067] (8)

[0068] (9)

[0069] In the formula, SOC is the SOC value of the lithium-ion battery, represents taking the derivative.

[0070] Step 3: Update the SOC estimated value.

[0071] Step 4: Introduce Gaussian noise into the input voltage parameter, establish a new SOC estimation equation and estimation process, and perform simulation verification at different battery degradation cycles.

[0072] Example 2:

[0073] Based on Example 1, Example 2 of the present application provides a more specific method for estimating the SOC of a lithium-ion battery based on the combined algorithm of Coulomb counting and UKF, including:

[0074] Step 1: Establish a first-order equivalent circuit model of the lithium-ion battery.

[0075] Step 2: Construct an SOC estimation equation based on the combined algorithm of Coulomb counting and UKF.

[0076] Step 3: Update the SOC estimated value.

[0077] Step 3 includes:

[0078] Step 3.1, parameter initialization, including: initializing state variables, covariance matrices, Sigma points, mean weight parameters, and covariance weight parameters.

[0079] Specifically, the initialization of state variables and covariance matrices is expressed as:

[0080] (10)

[0081] (11)

[0082] In the formula, is the initial value of the state variable, is the initial value of the state covariance, represents the formula for calculating the expected value.

[0083] The initialization of Sigma points is expressed as:

[0084] (12)

[0085] In the formula, and are the mean and covariance values of, is the correction value, is the number of sampling points, and the subscript represents the i-th Sigma point.

[0086] The initialization of mean weight parameters is expressed as:

[0087] (13)

[0088] (14)

[0089] In the formula, is the initial value of the mean weight value factor, is the i-th mean weight value factor, is defined as the distance between the mean point and the sampling point as the correction value, and the calculation formula is,

[0090] (15)

[0091] In the formula, represents the distribution parameter of Sigma points around the current estimated value, and its value range is from 0 to 1.

[0092] The initialization of covariance weight parameters is expressed as:

[0093] (16)

[0094] (17)

[0095] Wherein, is the initial value of the covariance weight value factor, is the i-th covariance weight value factor, represents the prior parameter.

[0096] Step 3.2: Perform time updates including state estimation, error covariance, and output estimation.

[0097] In Step 3.2, the generated Sigma points are converted into a nonlinear function using Equation (18) is the conversion function, and the mean value of Y is calculated using Equation (19) and the covariance value .

[0098] (18)

[0099] (19)

[0100] Specifically, the time update of state estimation is expressed as:

[0101] (20)

[0102] Wherein, the subscript represents the k-th sample value estimated based on the (k - 1)-th sample value, is the i-th state value, is the state update value calculated based on all i state values.

[0103] The time update of error covariance is expressed as:

[0104] (21)

[0105] Wherein, is 's mean value, is the covariance update value based on the (k - 1)-th covariance value (for the state variable ).

[0106] The time update of output estimation is expressed as:

[0107] (22)

[0108] Wherein, the i-th output variable value, The updated value of the output variable estimate calculated based on all i output variables.

[0109] Step 3.3: Perform covariance matrix update, expressed as:

[0110] (23)

[0111] (24)

[0112] In the formula, represents the covariance matrix of the average values of the k-th sampled values x and y. represents the covariance matrix of the k-th sampled value y.

[0113] Step 3.4: Perform state estimate update and error covariance update.

[0114] State estimate update, expressed as:

[0115] (25)

[0116] In the formula, is the gain matrix, satisfying . is the updated state, is the state before update.

[0117] Error covariance update, expressed as:

[0118] (26)

[0119] In the formula, is the error covariance matrix of the (k - 1)-th sampled value, is the gain matrix transpose matrix.

[0120] Step 4: Introduce Gaussian noise into the input voltage parameters, establish a new SOC estimation equation and estimation process, and perform simulation verification at different battery degradation cycles.

[0121] The present invention adds Gaussian noise to the input voltage parameters to prove the estimation performance of the CC-UKF joint algorithm. In the simulation of the lithium-ion battery degradation cycle, Gaussian white noise is added to the input voltage parameters , then the rated capacity of a single lithium-ion battery is expressed as:

[0122] (27)

[0123] In the formula, is the rated capacity of the lithium-ion battery for the k-th sampled value, is the rated capacity of the lithium-ion battery for the (k + 1)-th sampling value. Calculate the maximum value of the SOC change and the maximum value of the input voltage change as shown in Formula (28) and Formula (29):

[0124] (28)

[0125] (29)

[0126] (30)

[0127] In the formula, is the total charging or discharging duration (unit: hour), is the sampling time interval (unit: second), is the maximum value of SOC. is the maximum value of the input voltage, satisfying .

[0128] In addition, the present invention uses a single lithium-ion battery for simulation verification of the model and algorithm. The parameters of the lithium-ion battery include a rated voltage of 3.7V, a rated capacity of 30Ah, a discharge cut-off voltage of 3.4V, a charge cut-off voltage of 4.19V, and a rated power of 111W. The lithium-ion battery model, Coulomb counting algorithm, and unscented UKF algorithm are simulated and verified.

[0129] The present invention sets the degradation cycles to 1, 3, 5, and 7, and conducts simulation verification on the new algorithm model. The capacity changes of the lithium-ion battery with different degradation cycles are as Figures 2a - 2d shown, and the corresponding power attenuation is shown in Table 1.

[0130] Table 1 Power attenuation values of lithium-ion batteries under different degradation cycles

[0131]

[0132] Figures 2a through 2d are the changes in the rated capacity values corresponding to the four degradation cycles respectively. The rated capacity of the lithium-ion battery is 29.5Ah after 1 degradation cycle, drops to 27Ah after 3 degradation cycles, drops to 25.2Ah after 5 degradation cycles, and drops to 23Ah after 7 degradation cycles. It is obtained from the experiment that the duration of 1 degradation cycle is 6296s. Therefore, with the increase of the degradation cycle, the degradation of the lithium-ion battery capacity leads to inaccurate SOC estimation values. Table 1 shows that the greater the degradation cycle, the more obvious the power attenuation.

[0133] Further, assume that the total charging or discharging duration of a single lithium-ion battery is 1 hour (h), the sampling time interval is 1 second (s), the maximum initial SOC value of the single battery is 100%, and the maximum input voltage is 4.2V. Assume the parameters , , . Calculate the noise covariance values as and . Using formulas (27)-(30), the SOC change value and SOC estimated value can be calculated. Figure 3 represents the comparison result between the SOC estimated value based on CC-UKF and the actual measured SOC value. The result shows that the SOC estimated value is very close to the true value, and the estimation error is less than 1%.

[0134] It can be seen that the present invention proposes a method for estimating the SOC of a single lithium-ion battery based on the CC-UKF joint algorithm, which improves the performance of the traditional SOC estimation algorithm for lithium-ion batteries. Moreover, the present invention introduces a first-order equivalent circuit model. When the degradation cycle increases, the effectiveness of the joint SOC estimation method is verified by simulation. When the degradation cycles are 1, 3, 5, and 7, the corresponding rated capacities of the single lithium-ion battery are 29.5Ah, 27Ah, 25.2Ah, and 23Ah respectively. In addition, the present invention introduces Gaussian white noise of the voltage parameter value, and the result shows that the SOC estimation error based on the CC-UKF joint algorithm is less than 1%.

[0135] It should be noted that the same or similar parts in this embodiment and Embodiment 1 can be referred to each other and will not be elaborated in this application.

[0136] Embodiment 3:

[0137] Based on Embodiments 1 and 2, Embodiment 3 of the present application provides a lithium-ion battery SOC estimation system based on the coulomb counting and UKF joint algorithm, including:

[0138] A building module for building a first-order equivalent circuit model of the lithium-ion battery;

[0139] A constructing module for constructing an SOC estimation equation based on the coulomb counting and UKF joint algorithm;

[0140] An updating module for updating the SOC estimated value;

[0141] A verifying module for introducing Gaussian noise into the input voltage parameter, establishing a new SOC estimation equation and estimation process, and performing simulation verification at different battery degradation cycles.

[0142] It should be noted that the system provided in this embodiment is the system corresponding to the methods provided in Embodiments 1 and 2. Therefore, for the parts that are the same or similar to those in Embodiment 1 in this embodiment, reference can be made to each other and will not be elaborated in this application.

[0143] In summary, the Coulomb counting method is relatively simple, but there is an accumulation of SOC estimation errors; the UKF algorithm is only applicable to nonlinear systems, and in the present invention, Gaussian noise is added to the battery model, increasing the risk of linear systems; the CC-UKF combined algorithm not only solves the estimation problems of linear and nonlinear systems, but also improves the estimation accuracy.

Claims

1. A lithium-ion battery SOC estimation method based on coulomb counting and UKF combined algorithm, characterized in that: include: Step 1, establishing a first-order equivalent circuit model of a lithium-ion battery; Step 2: Based on the coulomb counting and UKF joint algorithm, construct the SOC estimation equation; Step 2 includes: Step 2.1, based on the coulomb counting method, establish the SOC estimation equation; Step 2.2, based on the parameters of the first-order equivalent circuit model, establish the state equation and observation equation of the UKF algorithm; Step 2.3, using the parameters of the first-order equivalent circuit model and the parameters of the Coulomb counting method, calculate the coefficient parameters of the state equation and the observation equation; Step 3: Update the estimated SOC value; Step 3 includes: Step 3.1, parameter initialization, including: initializing state variables, covariance matrix, Sigma point, average weight parameter and covariance weight parameter; Step 3.2, performing a time update including state estimation, error covariance and output estimation; Step 3.3, update the covariance matrix; Step 3.4, update the state estimation and error covariance; Step 4: Introduce Gaussian noise into the input voltage parameter, establish a new SOC estimation equation and estimation process, and perform simulation verification at different battery degradation cycles.

2. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 1, characterized in that: In step 1, the first-order equivalent circuit model includes a parallel RC network, an ohmic resistor connected in series with the RC network, and an ideal voltage source, wherein the RC network consists of a polarization resistor and a polarization capacitor.

3. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 2, characterized in that: Step 1 also includes: combining Kirchhoff's law and the first-order equivalent circuit model to establish a voltage-current relationship equation.

4. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 1, characterized in that: In step 2.2, the state equation and observation equation are: ； ； in, is the state variable, is the observed variable, K is the sampling time, then the state parameter is and ; The observation parameters are ; , , and is the coefficient variable; is the observation parameter, is the model noise, is Gaussian noise, and and Not related to each other.

5. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 4, characterized in that: In step 2.3, the coefficient parameter expression is: ； ； ； ； Where SOC is the SOC value of lithium-ion battery, represents the derivative, is the polarization resistance, is the polarized capacitance; is an ideal voltage source, is the ohmic resistance, is the output current, It is the battery charging and discharging efficiency.

6. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 5, characterized in that: In step 3.1, initialize the state variables and covariance matrix, expressed as: ； ； In the formula, is the initial value of the state variable, is the initial value of the state covariance, represents the expected value formula; Initialize the Sigma point, expressed as: ； In the formula, and for The mean and covariance values ​​of is the correction value, is the number of sampling points, subscript Represents the i-th Sigma point.

7. The lithium-ion battery SOC estimation method based on the coulomb counting and UKF combined algorithm according to claim 6, characterized in that: In step 3.1, the average weight parameter is initialized, expressed as: ； ； In the formula, is the initial value of the average weight factor, is the i-th average weight factor, As the correction value is defined as the distance between the average point and the sampling point; Initialize the covariance weight parameters, expressed as: ； ； In the formula, is the initial value of the covariance weight factor, is the i-th covariance weight factor, represents the prior parameters.

8. A lithium-ion battery SOC estimation system based on coulomb counting and UKF combined algorithm, characterized in that: Used to perform the method according to any one of claims 1 to 7, comprising: Establish a module for establishing a first-order equivalent circuit model of a lithium-ion battery; A building module for constructing the SOC estimation equation based on the coulomb counting and UKF joint algorithm; An updating module, used for updating the SOC estimation value; The verification module is used to introduce Gaussian noise into the input voltage parameters, establish a new SOC estimation equation and estimation process, and perform simulation verification at different battery degradation cycles.

9. A computer storage medium, characterized in that The computer storage medium stores a computer program; when the computer program is executed on a computer, the computer executes any one of the methods described in claims 1 to 7.

10. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the method according to any one of claims 1 to 7.

Images