A method for estimating the state of charge of lithium-ion batteries based on ultrasonic reflection characteristics and its dielectric material

Abstract

This invention discloses a method and medium for estimating the state of charge (SOC) of a lithium-ion battery based on ultrasonic reflection characteristics, belonging to the field of lithium-ion battery management technology. The method includes: obtaining a SoC-E characterizing the relationship between the SOC of the lithium-ion battery and the energy of the oscillation wave through testing. OC Curve; based on SoC-E OC The acoustic response model of a differential battery, incorporating coupled current, state of charge (SOC), and oscillation wave energy, is constructed and linearized. The parameters to be identified in the differential battery acoustic response model are obtained using the least squares method, yielding a defined differential battery acoustic response model. The acoustic response state-space equation is obtained by combining the SoC expression under the Coulomb counting method with the differential battery acoustic response model. The real-time current and real-time oscillation wave energy of the lithium-ion battery under test are substituted into the acoustic response state-space equation to solve for its real-time SOC. This improves the practicality, accuracy, and robustness of the ultrasonic-based method for detecting the SOC of lithium-ion batteries.

Description

Technical Field

[0001] This invention belongs to the field of lithium-ion battery management technology, and more specifically, relates to a method and medium for estimating the state of charge of lithium-ion batteries based on ultrasonic reflection characteristics. Background Technology

[0002] Lithium-ion batteries, with their advantages of high energy density, long lifespan, low self-discharge rate, and high rated voltage, have become the mainstream batteries for electrochemical energy storage systems and power units in new energy electric vehicles. To promptly prevent battery abuse and ensure safe and efficient battery operation, battery management systems (BMS) need to continuously estimate the battery's state of charge (SoC). Currently, the SoC of lithium-ion batteries cannot be directly measured; it can only be estimated indirectly through voltage, current, and temperature. These external characteristics have low correlation with the SoC, greatly hindering reliable SoC estimation and resulting in low accuracy and poor robustness in existing SoC estimation methods.

[0003] Ultrasonic waves can sense the comprehensive feedback of changes in various physical quantities directly related to the state of charge (SOC) within a battery during charging and discharging, thus enabling the characterization of the SOC using ultrasonic detection technology. However, current methods for estimating SOC using acoustic indicators are limited to constant current conditions. The changes in the acoustic response characteristics of the battery under dynamic operating conditions with real-time current fluctuations have not yet been addressed, significantly hindering the practical application of ultrasound in lithium-ion battery SOC estimation. Therefore, improving the practicality, accuracy, and robustness of ultrasonic-based methods for detecting battery SOC is of significant research importance. Summary of the Invention

[0004] To address the shortcomings and improvement needs of existing technologies, this invention provides a method and medium for estimating the state of charge of lithium-ion batteries based on ultrasonic reflection characteristics. The purpose is to solve the problem that existing ultrasonic characterization technologies for lithium-ion batteries (SoCs) cannot meet the accuracy and robustness requirements under dynamic current conditions.

[0005] To achieve the above objectives, according to one aspect of the present invention, a method for estimating the state of charge (SOC) of a lithium-ion battery based on ultrasonic reflection characteristics is provided, comprising: S1, performing charge-discharge incremental capacity tests on the lithium-ion battery under different SOC states, acquiring the ultrasonic waves reflected by the lithium-ion battery during the test, calculating the oscillation wave energy in each ultrasonic wave, and fitting the SOC and the oscillation wave energy to obtain the SoC-E. OC Curve; S2, according to the SoC-E OCS3. Based on the measured values ​​of the current, state of charge, and oscillation wave energy of the lithium-ion battery under dynamic operating conditions, the parameters to be identified in the differential battery acoustic response model are obtained using the least squares method, thus obtaining the determined differential battery acoustic response model; S4. By combining the SoC expression under the Coulomb counting method with the differential battery acoustic response model, the acoustic response state space equation is obtained; S5. The real-time current and real-time oscillation wave energy of the lithium-ion battery under test are substituted into the acoustic response state space equation to solve for the real-time state of charge of the lithium-ion battery under test.

[0006] Furthermore, the differential battery acoustic response model constructed in S2 is as follows:

[0007] E Ot =E OC (SoC)-γI-E p

[0008] Among them, E p (k+1)=αE p (k)+βI(k),E Ot The oscillation wave energy is measured in real time, and the SoC is in a charged state, E OC (SoC) refers to the SoC-E OC The curve represents the oscillation wave energy corresponding to the SoC, where I is the current and E is the oscillation wave energy corresponding to the SoC. p E represents the dynamic process parameters of the acoustic response of a lithium-ion battery. p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time, I(k) represents the current at the k-th sampling time, and α, β, and γ are three parameters to be identified.

[0009] Furthermore, the linearization process of the differential battery acoustic response model includes: sequentially performing Z-transform and inverse Z-transform on the differential battery acoustic response model to transform the parameters to be identified in the differential battery acoustic response model into linearized parameters.

[0010] Furthermore, the parameters to be identified obtained in S3 are:

[0011]

[0012] Y = [E Oin (2) E Oin (3)…E Oin (m)] T

[0013] H = [h2 h3…h] m ] T

[0014] h k =[E Oin (k-1) -I(k) -I(k-1)] T k = 2, 3, ..., m

[0015] E Oin (k)=E OC [SoC(k)]-E Ot (k)

[0016] in, The three parameters after linearization The vector consists of three parameters to be identified: α, β, and γ. H is the regression matrix, Y is the output vector, and E is the regression matrix. Oin (k) represents E OC [SoC(k)] and E Ot The intermediate variable h is the difference between (k) and (k). k Let I(k) be the intermediate input vector at the k-th sampling time, I(k) be the current at the k-th sampling time, m be the total sampling time, SoC(k) be the state of charge at the k-th sampling time, and E be the intermediate input vector at the k-th sampling time. OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve Ot (k) represents the energy of the oscillation wave measured at the kth sampling time.

[0017] Furthermore, the acoustic response state-space equation includes:

[0018] State equation: x(k+1)=Ax(k)+BI(k)+w(k)

[0019] Observation equation: y(k)=g[x(k),I(k)]+v(k)

[0020] in:

[0021] x(k) = [SoC(k), E p (k)] T

[0022] y(k)=E Ot (k)

[0023] A = diag(1, α)

[0024] B = [Δt / Capacity, β] T

[0025] g[x(k),I(k)]=E OC [SoC(k)]-γI(k)-Ep (k)

[0026] Where x(k) is the state vector at the k-th sampling time, I(k) is the current at the k-th sampling time, w(k) is the process noise at the k-th sampling time, v(k) is the measurement noise at the k-th sampling time, g[x(k),I(k)] is the measurement equation at the k-th sampling time, y(k) is the observation value at the k-th sampling time, A and B are the two coefficient matrices of the state equation, α, β, and γ are the three parameters to be identified, diag() is the function to extract diagonal elements, Δt is the sampling interval, Capacity is the lithium-ion battery capacity, SoC(k) is the state of charge at the k-th sampling time, and E OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time.

[0027] Furthermore, S5 includes: based on the real-time current and real-time oscillation wave energy of the lithium-ion battery under test, using an adaptive extended Kalman filter to perform parameter estimation and noise adaptive update of the acoustic response state space equation, and solving to obtain the real-time state of charge of the lithium-ion battery under test.

[0028] Furthermore, the adaptive extended Kalman filter updates the parameters in the state equation x(k+1)=Ax(k)+BI(k)+w(k) of the acoustic response state space equation in real time as follows:

[0029] α(k+1)=α(k)+r1(k)

[0030] β(k+1)=β(k)+r2(k)

[0031] γ(k+1)=γ(k)+r3(k)

[0032] Where α(k), β(k), and γ(k) are three parameters in the state equation at the k-th sampling time, r1(k), r2(k), and r3(k) are three errors at the k-th sampling time, x(k) is the state vector at the k-th sampling time, I(k) is the current at the k-th sampling time, w(k) is the process noise at the k-th sampling time, and A and B are two coefficient matrices of the state equation;

[0033] The real-time updated acoustic response state-space equations include:

[0034] State equation: x(k+1)=Ax(k)+BI(k)+w(k)

[0035] Observation equation: y(k)=g[x(k),I(k)]+v(k)

[0036] in:

[0037] x(k) = [SoC(k), E p (k),α(k),β(k),γ(k)] T

[0038] y(k)=E Ot (k)

[0039] A = diag(1, α(k), 1, 1, 1)

[0040] B=[Δt / Capacity,β(k),0,0,0] T

[0041] g[x(k),I(k)]=E OC [SoC(k)]-γI(k)-E p (k)

[0042] Where v(k) is the measurement noise at the k-th sampling time, g[x(k),I(k)] is the measurement equation at the k-th sampling time, y(k) is the observation value at the k-th sampling time, diag() is the function to extract diagonal elements, Δt is the sampling interval, Capacity is the lithium-ion battery capacity, SoC(k) is the state of charge at the k-th sampling time, and E OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time.

[0043] Furthermore, the acoustic response state-space equation includes a state equation and an observation equation. The parameter estimation and adaptive noise update specifically include: an initialization phase: setting initial values ​​for the state estimate and the error covariance matrix; a time update phase: updating the prior state estimate and the prior error covariance matrix based on the first-order partial derivative of the state equation with respect to the state vector; a measurement update phase: updating the posterior state estimate and the posterior error covariance matrix based on the first-order partial derivative of the measurement equation with respect to the state vector, as well as the prior state estimate and the prior error covariance matrix; an adaptive error update phase: updating the residual covariance matrix based on the moving window average residual sequence, and updating the process covariance matrix and the measurement covariance matrix based on the residual covariance matrix; repeating the time update phase, the measurement update phase, and the adaptive error update phase to solve the real-time state of charge of the lithium-ion battery under test in real time.

[0044] According to another aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described above.

[0045] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0046] (1) A method for estimating the state of charge (SOC) of a lithium-ion battery based on ultrasonic reflection characteristics is provided. On the one hand, ultrasonic waves can obtain comprehensive feedback on the changes of various physical parameters related to the SOC in the battery in situ and online without damaging the lithium-ion battery. Therefore, compared with the existing relationship between terminal voltage and SOC, the relationship between the oscillating wave energy acoustic index extracted in this embodiment of the invention and the SOC is closer and the linear correlation is better. Therefore, compared with the traditional method of estimating the SOC by using terminal voltage to model the battery equivalent circuit, the method of estimating the SOC of a lithium-ion battery by using the differential battery acoustic response model in this embodiment of the invention has stronger robustness and higher accuracy, especially when the battery is operating in the voltage plateau range, the robustness and accuracy are greatly improved.

[0047] (2) On the other hand, the established differential battery acoustic response model broadens the method of estimating the state of charge of lithium-ion batteries by ultrasound to the application level under dynamic current conditions. The model is simple and the algorithm has low computational complexity, which improves the practicality of estimating the state of charge of lithium-ion batteries by ultrasound technology.

[0048] (3) In the process of solving the acoustic response state space equation using the adaptive extended Kalman filter, the parameters α, β, and γ in the state equation are regarded as slowly changing parameters and updated in real time based on small errors, thereby further improving the accuracy of state of charge estimation. Attached Figure Description

[0049] Figure 1 A flowchart of a lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics provided in an embodiment of the present invention;

[0050] Figure 2 A connection diagram of the experimental apparatus for measuring reflected ultrasound from a lithium-ion battery, provided in an embodiment of the present invention;

[0051] Figure 3 Typical reflected ultrasonic waveform diagram provided for embodiments of the present invention;

[0052] Figure 4 Verification diagram of model estimation results provided in embodiments of the present invention;

[0053] Figure 5 The SoC estimation results based on the acoustic response state-space equation are provided for embodiments of the present invention.

[0054] In all the accompanying drawings, the same reference numerals are used to denote the same elements or structures, wherein:

[0055] 1 is the probe clamp, 2 is the single-crystal ultrasonic probe, 3 is the ultrasonic transmitting and receiving system, 4 is the lithium-ion battery cell, 5 is the battery testing system, 6 is the computer, and 7 is the temperature control box. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0057] In this invention, the terms "first," "second," etc. (if present) in the invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0058] Figure 1 A flowchart illustrating the lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics, provided in an embodiment of the present invention. (See also...) Figure 1 , combined Figures 2-5 The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics in this embodiment is described in detail. The method includes operations S1-S5.

[0059] In this embodiment, the following is provided: Figure 2 The experimental setup shown is used to test the incremental capacity of lithium-ion batteries during charge and discharge. The setup includes: a 3D-printed probe fixture 1, a 2.5MHz single-crystal ultrasonic probe 2, an ultrasonic transmitting and receiving system 3, a lithium-ion battery cell 4, a battery testing system 5, a computer 6, and a temperature control chamber 7.

[0060] A layer of ultrasonic coupling agent is applied to the contact surface between the lithium-ion battery unit 4 and the ultrasonic probe 2. The ultrasonic probe 2 is fixed to the surface of the lithium-ion battery unit 4 by the probe clamp 1. The ultrasonic probe 2 is connected to the ultrasonic transmitting and receiving system 3 and connected to the computer 6 to perform ultrasonic signal acquisition, acoustic index extraction and storage.

[0061] The lithium-ion battery electrodes are connected to the battery testing system 5, and the charging and discharging process is set up. The system is then connected to the computer 6 to store the voltage and current data of the lithium-ion battery. Finally, the probe clamp 1, ultrasonic probe 2, and lithium-ion battery cell 4 are placed together in the temperature control chamber 7 for temperature regulation and control. Typically, the temperature of the temperature control chamber is set to room temperature (25°C).

[0062] Operation S1 performs charge-discharge incremental capacity tests on lithium-ion batteries under different states of charge. The ultrasonic waves reflected by the lithium-ion batteries during the test are acquired, and the oscillation wave energy in each ultrasonic wave is calculated. The state of charge and oscillation wave energy are then fitted to obtain the SoC-E. OC curve.

[0063] In this embodiment, the lithium-ion battery is, for example, a lithium iron phosphate pouch battery. Its voltage characteristics and state of charge exhibit a strongly nonlinear relationship, and the change is flat in most areas, making it difficult to model using terminal voltage and accurately estimate the state of charge. In operation S1, for example, a 2Ah pouch battery with a lithium iron phosphate positive electrode and a graphite negative electrode is selected, placed in a temperature-controlled chamber, and connected to a battery testing system. Capacity calibration is performed using a 1C current cycle to ensure battery cycle stability, with the battery capacity remaining almost unchanged within 3 cycles, and the battery capacity (Capacity) is obtained.

[0064] The charge / discharge incremental capacity test process is as follows: During the discharge test, the lithium-ion battery is charged to 3.65V at a constant current of 1C, then charged to 0.05C at a constant voltage, and left to stand for 2 hours. The current oscillation energy is recorded and denoted as E. OC Then, after discharging 10% of the SoC using a 0.5C current pulse, it was left to stand for 2 hours, and E was recorded. OC Repeat the pulse discharge and resting process for 2 hours until the discharge cutoff voltage of 2V reaches 0SoC, obtaining discharge E at 11 SoC points. OC During the charging test phase, starting from 0 SoC, the process of 0.5C pulse charging and resting for 2 hours was repeated until 90% SoC was reached. Finally, the system was charged to 100% SoC by first using a 0.5C constant current to 3.65V, and then a constant voltage to 0.05C, followed by resting for 2 hours. This yielded charging E values ​​for 11 SoC points. OC .

[0065] Shockwave energy E OC The extraction method is as follows:

[0066]

[0067] Where, x i Let i be the amplitude of the ultrasonic wave at sampling point coordinate i, where i1 and i2 are the coordinates of the starting and ending sampling points of the oscillation wave, respectively. Figure 3As shown, i1 and i2 are, for example, 1 and 171 respectively, indicating that the oscillation wave energy is the sum of the squares of the amplitudes from the first sampling point to the 171st sampling point of the obtained reflected ultrasonic signal.

[0068] Preferably, for example, a fourth-order polynomial is used to measure the measured state of charge (SoC) and oscillation wave energy E. OC By performing fitting, SoC-E is obtained. OC Curve. This testing method eliminates battery polarization and temperature fluctuations, obtaining SoC-E. OC The curve is independent of the current.

[0069] Operation S2, according to SoC-E OC Curves were used to construct a differential battery acoustic response model with coupling current, state of charge, and oscillation wave energy, and the differential battery acoustic response model was linearized.

[0070] According to an embodiment of the present invention, the first-order differential battery acoustic response model constructed in operation S2 is as follows:

[0071] E Ot =E OC (SoC)-γI-E p (2)

[0072] Among them, E p According to the formula E p (k+1)=αE p E is calculated recursively from (k)+βI(k). Ot The oscillation wave energy is measured in real time, and the SoC is in a charged state, E OC (SoC) refers to SoC-E OC The curve represents the oscillation wave energy corresponding to the SoC; I represents current, with discharge current as positive; E p E represents the dynamic process parameters of the acoustic response of a lithium-ion battery. p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time, I(k) represents the current at the k-th sampling time, and α, β, and γ are three parameters to be identified.

[0073] To facilitate the identification of α, β, and γ in the differential battery acoustic response model using the least squares method, the differential battery acoustic response model needs to be linearized.

[0074] According to an embodiment of the present invention, linearization of the differential battery acoustic response model includes: sequentially performing Z-transform and inverse Z-transform on the differential battery acoustic response model to transform the parameters to be identified in the differential battery acoustic response model into linearized parameters.

[0075] Specifically, performing a Z-transform on formula (2) yields:

[0076]

[0077] Then, performing an inverse Z-transform on formula (3), we obtain:

[0078]

[0079] In formula (4), All are linear, and are parameters to be identified by least squares.

[0080] Operation S3: Based on the measured values ​​of current, state of charge and oscillation wave energy of lithium-ion battery under dynamic operating conditions, the parameters to be identified in the differential battery acoustic response model are obtained using the least squares method, thus obtaining the determined differential battery acoustic response model.

[0081] In this embodiment, a dynamic operating condition test is conducted on the lithium-ion battery under urban road cyclic conditions to obtain the measured values ​​of current, state of charge, and oscillation wave energy required for parameter identification. Based on these measured values, the least squares method, as shown in formulas (5)-(9), is used to analyze the parameters in formula (4). To identify.

[0082] According to an embodiment of the present invention, the parameters to be identified obtained in operation S3 are:

[0083]

[0084] Y = [E Oin (2) E Oin (3)…E Oin (m)] T (6)

[0085] H = [h2 h3…h] m ] T (7)

[0086] h k =[E Oin (k-1) -I(k) -I(k-1)] T k = 2, 3, ..., m (8)

[0087] E Oin (k)=E OC [SoC(k)]-E Ot (k) (9)

[0088] in, The three parameters after linearization The vectors formed, where H is the regression matrix, Y is the output vector, and E is the output vector. Oin (k) represents E OC [SoC(k)] and EOt The intermediate variable h is the difference between (k) and (k). k Let be the intermediate input vector at the k-th sampling time, m be the total sampling time, SoC(k) be the state of charge at the k-th sampling time, and E be the intermediate input vector at the k-th sampling time. OC [SoC(k)] refers to SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve Ot (k) represents the energy of the oscillation wave measured at the kth sampling time. Thus, α, β, and γ can be solved, completing the establishment of the differential battery acoustic response model.

[0089] By operating S4, combining the SoC expression under the Coulomb counting method with the differential battery acoustic response model, the acoustic response state-space equation is obtained.

[0090] The SoC expression under the joint Coulomb counting method is:

[0091] SoC(k)=SoC(k-1)+I(k)×Δt / Capacity (10)

[0092] The obtained acoustic response state-space equations include:

[0093] State equation: x(k+1)=Ax(k)+BI(k)+w(k) (11)

[0094] Observation equation: y(k)=g[x(k),I(k)]+v(k) (12)

[0095] in:

[0096] x(k) = [SoC(k), E p (k)] T (13)

[0097] y(k)=E Ot (k) (14)

[0098] A = diag(1, α) (15)

[0099] B = [Δt / Capacity, β] T (16)

[0100] g[x(k),I(k)]=E OC [SoC(k)]-γI(k)-E p (k) (17)

[0101] Where x(k) is the state vector at the k-th sampling time, w(k) is the process noise at the k-th sampling time, v(k) is the measurement noise at the k-th sampling time, g[x(k),I(k)] is the measurement equation at the k-th sampling time, y(k) is the observation value at the k-th sampling time, A and B are the two coefficient matrices of the state equation, diag() is the function to extract diagonal elements, Δt is the sampling interval, Capacity is the lithium-ion battery capacity, SoC(k) is the state of charge at the k-th sampling time, and E OC [SoC(k)] refers to SoC-E OC The energy of the oscillation wave corresponding to SoC(k) in the curve.

[0102] Operation S5 substitutes the real-time current and real-time oscillation wave energy of the lithium-ion battery under test into the acoustic response state-space equation to obtain the real-time state of charge of the lithium-ion battery under test.

[0103] Specifically, in operation S5, based on the real-time current and real-time oscillation wave energy of the lithium-ion battery under test, the acoustic response state space equation is estimated and updated adaptively for noise using an adaptive extended Kalman filter, and the real-time state of charge of the lithium-ion battery under test is obtained by solving the problem.

[0104] For the filter used for state of charge estimation, in this embodiment, an Adaptive Extended Kalman Filter (AEKF) is established to perform parameter estimation, noise adaptive updating, and SoC closed-loop estimation of the model.

[0105] The process of establishing AEKF is as follows: First, the changes in model parameters are regarded as slow changes. It can be assumed that there is a small error between the new parameters and the parameters at the previous moment. Then, the three parameters α, β, and γ of the model can be used as the state to be estimated in the system state equation (11):

[0106] α(k+1)=α(k)+r1(k) (18)

[0107] β(k+1)=β(k)+r2(k) (19)

[0108] γ(k+1)=γ(k)+r3(k) (20)

[0109] Where α(k), β(k), and γ(k) are three parameters in the state equation at the k-th sampling time, and r1(k), r2(k), and r3(k) are three errors at the k-th sampling time.

[0110] After adding equations (18)-(20) to state equation (11), the state vector x becomes a five-dimensional vector. The acoustic response state-space equation after real-time update includes:

[0111] State equation: x(k+1)=Ax(k)+BI(k)+w(k) (21)

[0112] Observation equation: y(k)=g[x(k),I(k)]+v(k) (22)

[0113] in:

[0114] x(k) = [SoC(k), E p (k),α(k),β(k),γ(k)] T (twenty three)

[0115] y(k)=E Ot (k) (24)

[0116] A=diag(1,α(k),1,1,1) (25)

[0117] B=[Δt / Capacity,β(k),0,0,0] T (26)

[0118] g[x(k),I(k)]=E OC [SoC(k)]-γI(k)-E p (k) (27)

[0119] Then, parameter estimation and noise adaptive update are performed, specifically including the initialization stage, time update stage, measurement update stage, and adaptive error update stage.

[0120] Initialization phase: Set initial values ​​for the state estimate and the error covariance matrix. Specifically, initialize the AEKF algorithm and set the initial estimated state vector. Initialize the error covariance matrix P(0), the process covariance matrix Q(0), and the measurement matrix R(0). The acoustic response state-space equation can be expressed as:

[0121] x(k)=[z(k),E p (k),α(k),β(k),γ(k)] T (28)

[0122] y(k)=E Ot (k) (29)

[0123] A=diag(1,α(k),1,1,1) (30)

[0124] B = [Δt / C] bat ,β(k),0,0,0] T (31)

[0125] g[x(k),I(k)]=E OC [z(k)]-γ(k)I(k)-E p (k) (32)

[0126] When discharge begins, the following iterative time update phase, measurement update phase, and adaptive error update phase are initiated.

[0127] Time update phase: Based on the first-order partial derivative of the state equation with respect to the state vector, the prior state estimate and the prior error covariance matrix are updated.

[0128] The first-order partial differentials of the state equation with respect to the state vector are:

[0129]

[0130] The update yields the prior state estimate x. - (k) and the prior error covariance matrix P - (k) is:

[0131] x - (k)=A(k-1)x + (k-1)+B(k-1)I(k-1) (34)

[0132] P - (k)=D(k-1)P + (k-1)D(k-1) T +Q(k-1) (35)

[0133] Measurement update phase: Based on the first-order partial derivative of the measurement equation g[x(k),I(k)] with respect to the state vector, as well as the prior state estimate and prior error covariance matrix, the posterior state estimate and posterior error covariance matrix are updated.

[0134] The first-order partial derivatives of the measurement equation g[x(k),I(k)] with respect to the state vector are:

[0135]

[0136] The Kalman gain matrix K(k) can be calculated using the following formula:

[0137] K(k)=P - (k)C(k) T [C(k)P - (k)C(k) T +R(k)] -1(37)

[0138] The update yields the posterior state estimate x. + (k) and the posterior error covariance matrix P + (k) is:

[0139] x + (k)=x - (k)+K(k){y(k)-g[x - (k),I(k)]} (38)

[0140] P + (k)=(EK(k)C(k))P - (k) (39)

[0141] Here, E is a fifth-order identity matrix. In this stage, we will start from x... + In (k), equation (28) can output the posterior estimated state of charge, complete the accurate estimation of the state of charge at this moment, and update the latest α, β, γ parameters.

[0142] Adaptive error update phase: Update the residual covariance matrix based on the moving window average residual sequence, and update the process covariance matrix and measurement covariance matrix based on the residual covariance matrix.

[0143] Calculate the measurement residual ε(k):

[0144] ε(k)=y(k)-y + (k) (40)

[0145] Using a moving window of length L to average the residual sequence, the residual covariance matrix is:

[0146]

[0147] The updated process covariance matrix and measurement covariance matrix are as follows:

[0148] Q(k)=K(k)C v (k)K(k) T (42)

[0149] R(k)=C v (k)+C(k)P + (k)C(k) T (43)

[0150] During the cycling process of a lithium-ion battery, by repeatedly executing the time update stage, measurement update stage, and adaptive error update stage, the battery acoustic response state-space equation and AEKF can be used to continuously and accurately estimate the real-time battery state of charge closed loop throughout the entire process.

[0151] Through the above operations S1-S5, ultrasonic testing, acoustic feature extraction, battery acoustic dynamic response modeling, offline acoustic model parameter identification, dynamic operating condition battery state of charge prediction based on battery acoustic model, and dynamic acoustic model parameter identification of lithium-ion batteries (such as 2Ah lithium iron phosphate soft pack batteries) can be achieved.

[0152] Taking experimental results as an example, the feasibility of the lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics in this embodiment of the invention is illustrated. The parameter identification results in operation S3 are α = 0.9737, β = 1.5702, and γ = 5.6047. After substituting the above parameters, the model estimates the Et of the urban road cyclic operating condition test results. Ot With real E Ot A control group was formed, and the results were as follows: Figure 4 As shown, the root mean square percentage error is only 0.3%, indicating the accuracy of the model. The initial covariance matrix is ​​P(0) = diag([1, 1×10^2 ... -10 1×10 -6 1×10 -6 1×10 -6 Q(0) = 3.4 × 10 -8 E,R(0)=3.4×10 -8 To verify the model's bias correction capability and estimation accuracy, the initial state of charge was set to 0.8, and the initial E... p The value is set to 0.1, while the actual state of charge is 1, E p If the value is 0, substitute it into the AEKF algorithm for calculation. Experimental results are as follows: Figure 5 As shown, the maximum absolute error is 1.41%, the mean absolute error is 1.15%, and the root mean square error is 1.09%, indicating that the method exhibits extremely fast correction convergence speed and estimation accuracy.

[0153] The lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics provided in this invention utilizes ultrasonic detection technology during battery cycling to obtain the oscillation wave energy E, an acoustic index strongly correlated with the state of charge. O A first-order difference empirical model of the battery acoustic response is used to couple the state of charge, current, and oscillation wave energy E. O The battery state-space equations constructed from the first-order difference model are used to perform closed-loop estimation of the state of charge of lithium-ion batteries through an extended Kalman filter, which improves the practicality of the method of estimating the state of charge of lithium-ion batteries using ultrasonic technology.

[0154] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics as described above.

[0155] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics, characterized in that, include: S1, Perform charge-discharge incremental capacity tests on the lithium-ion battery under different states of charge, acquire the ultrasonic waves reflected by the lithium-ion battery during the test, calculate the oscillation wave energy in each ultrasonic wave, and fit the state of charge and the oscillation wave energy to obtain the SoC-E. OC curve; S2, according to the SoC-E OC Curves were used to construct a differential battery acoustic response model based on coupling current, state of charge, and oscillation wave energy, and the differential battery acoustic response model was linearized. S3. Based on the measured values ​​of the current, state of charge and oscillation wave energy of the lithium-ion battery under dynamic operating conditions, the parameters to be identified in the differential battery acoustic response model are obtained by using the least squares method, and the determined differential battery acoustic response model is obtained. S4. By combining the SoC expression under the Coulomb counting method with the differential battery acoustic response model, the acoustic response state space equation is obtained. S5. Substitute the real-time current and real-time oscillation wave energy of the lithium-ion battery under test into the acoustic response state-space equation to solve for the real-time state of charge of the lithium-ion battery under test.

2. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 1, characterized in that, The differential battery acoustic response model constructed in S2 is as follows: AND Ot =E OC (SoC)-γI-E p Among them, E p (k+1)=αE p (k)+βI(k),E Ot The oscillation wave energy is measured in real time, and the SoC is in a charged state, E OC (SoC) refers to the SoC-E OC The curve represents the oscillation wave energy corresponding to the SoC, where I is the current and E is the oscillation wave energy corresponding to the SoC. p E represents the dynamic process parameters of the acoustic response of a lithium-ion battery. p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time, I(k) represents the current at the k-th sampling time, and α, β, and γ are three parameters to be identified.

3. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 1, characterized in that, The linearization process of the differential battery acoustic response model includes: sequentially performing Z-transform and inverse Z-transform on the differential battery acoustic response model to transform the parameters to be identified in the differential battery acoustic response model into linearized parameters.

4. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 1, characterized in that, The parameters to be identified obtained in S3 are: Y=[E Oin (2) E Oin (3)…E Oin (m)] T H=[h2 h3…h m ] T h k =[E Oin (k-1) -I(k) -I(k-1)] T k=2,3,...,m IN Oin (k)=E OC [SoC(k)]-E Ot (k) in, The three parameters after linearization The vector consists of three parameters to be identified: α, β, and γ. H is the regression matrix, Y is the output vector, and E is the regression matrix. Oin (k) represents E OC [SoC(k)] and E Ot The intermediate variable h is the difference between (k) and (k). k Let I(k) be the intermediate input vector at the k-th sampling time, I(k) be the current at the k-th sampling time, m be the total sampling time, SoC(k) be the state of charge at the k-th sampling time, and E be the intermediate input vector at the k-th sampling time. OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve Ot (k) represents the energy of the oscillation wave measured at the kth sampling time.

5. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 1, characterized in that, The acoustic response state-space equation includes: State equation: x(k+1)=Ax(k)+BI(k)+w(k) Observation equation: y(k)=g[x(k),I(k)]+v(k) in: x(k)=[SoC(k),E p (s)] T y(k)=E Ot (k) A = diag(1, α) B=[Δt / Capacity,β] T g[x(k),I(k)]=E OC [CoC(k)]-γI(k)-E p (k) Where x(k) is the state vector at the k-th sampling time, I(k) is the current at the k-th sampling time, w(k) is the process noise at the k-th sampling time, v(k) is the measurement noise at the k-th sampling time, g[x(k),I(k)] is the measurement equation at the k-th sampling time, y(k) is the observation value at the k-th sampling time, A and B are the two coefficient matrices of the state equation, α, β, and γ are the three parameters to be identified, diag() is the function to extract diagonal elements, Δt is the sampling interval, Capacity is the lithium-ion battery capacity, SoC(k) is the state of charge at the k-th sampling time, and E OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time.

6. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in any one of claims 1-5, characterized in that, S5 includes: Based on the real-time current and real-time oscillation wave energy of the lithium-ion battery under test, the acoustic response state space equation is estimated and updated with noise adaptively using an adaptive extended Kalman filter, and the real-time state of charge of the lithium-ion battery under test is obtained by solving the equation.

7. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 6, characterized in that, The adaptive extended Kalman filter updates the parameters in the state equation x(k+1)=Ax(k)+BI(k)+w(k) of the acoustic response state space equation in real time as follows: α(k+1)=α(k)+r1(k) β(k+1)=β(k)+r2(k) γ(k+1)=γ(k)+r3(k) Where α(k), β(k), and γ(k) are three parameters in the state equation at the k-th sampling time, r1(k), r2(k), and r3(k) are three errors at the k-th sampling time, x(k) is the state vector at the k-th sampling time, I(k) is the current at the k-th sampling time, w(k) is the process noise at the k-th sampling time, and A and B are two coefficient matrices of the state equation; The real-time updated acoustic response state-space equations include: State equation: x(k+1)=Ax(k)+BI(k)+w(k) Observation equation: y(k)=g[x(k),I(k)]+v(k) in: x(k)<[SoC(k),E p (k),α(k),β(k),γ(k)] T y(k)=E Ot (k) A = diag(1, α(k), 1, 1, 1) B=[Δt / Capacity,β(k),0,0,0] T g[x(k),I(k)]=E OC [CoC(k)]-γI(k)-E p (k) Where v(k) is the measurement noise at the k-th sampling time, g[x(k),I(k)] is the measurement equation at the k-th sampling time, y(k) is the observation value at the k-th sampling time, diag() is the function to extract diagonal elements, Δt is the sampling interval, Capacity is the lithium-ion battery capacity, SoC(k) is the state of charge at the k-th sampling time, and E OC [SoC(k)] is the SoC-E OC The oscillation energy E corresponding to SoC(k) in the curve p (k) represents the dynamic process parameters of the acoustic response of the lithium-ion battery at the k-th sampling time.

8. The method for estimating the state of charge of a lithium-ion battery based on ultrasonic reflection characteristics as described in claim 6, characterized in that, The acoustic response state-space equation includes a state equation and an observation equation, and the parameter estimation and noise adaptive update specifically include: Initialization phase: Set initial values ​​for the state estimate and the error covariance matrix; Time update phase: Based on the first-order partial derivative of the state equation with respect to the state vector, the prior state estimate and the prior error covariance matrix are updated. Measurement update phase: Based on the first-order partial derivative of the measurement equation with respect to the state vector, the prior state estimate and the prior error covariance matrix, the posterior state estimate and the posterior error covariance matrix are updated. Adaptive error update phase: Update the residual covariance matrix according to the moving window average residual sequence, and update the process covariance matrix and measurement covariance matrix according to the residual covariance matrix; The time update phase, the measurement update phase, and the adaptive error update phase are repeatedly executed to solve the real-time state of charge of the lithium-ion battery under test in real time.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the lithium-ion battery state-of-charge estimation method based on ultrasonic reflection characteristics as described in any one of claims 1-8.

Images