Method and device for estimating state of health of battery

Abstract

The invention provides a method and device for estimating the state of health of a battery, and the method comprises the steps: obtaining battery cell parameters which comprise the current value, the voltage value and the actual impedance value of a to-be-detected battery cell; according to the cell parameters, a target equivalent circuit of the to-be-tested cell is constructed, the target equivalent circuit comprises a resistor and a plurality of sub-circuit elements, the number of the sub-circuit elements is the same as the order number of the target equivalent circuit, and the order number of the target equivalent circuit is determined according to the measurement coefficient of the to-be-tested cell; and based on the target equivalent circuit, predicting a health state estimation value of the to-be-tested cell. According to the invention, the efficiency of predicting the health state of the to-be-detected battery cell and the prediction precision are improved.

Description

Technical Field

[0001] The present application relates to the technical field of battery health status assessment, and in particular, to a battery health status estimation method and device. Background Art

[0002] Responding to the pressure brought by environmental damage and energy crisis has always been an important task. As a widely accepted means of transportation, new energy vehicles can reduce dependence on fossil fuels and play an important role in slowing down global warming and promoting the development of energy management technology. Among all categories of new energy vehicles, lithium-ion power battery vehicles are the most promising development direction. Lithium batteries are widely used in many fields such as mobile phones, electric vehicles, and mobile power supplies due to their high energy density, high cycle life, and high environmental protection.

[0003] During the charging process of lithium batteries, lithium ions are reduced to form lithium dendrites. The growth of lithium dendrites will continue until the internal short circuit of the battery occurs. This is an irreversible process. The formation of lithium dendrites will lead to battery capacity loss and safety hazards. Therefore, it is very important to detect the state of health (SOH) of lithium batteries.

[0004] The health state of a lithium battery can be understood as a measurement that reflects the overall condition of a lithium-ion battery, and is usually defined in terms of remaining available capacity (SOC). Currently, the methods for evaluating lithium-ion SOH include model-based methods and data-driven methods. Data-driven methods are widely used due to their lower computational complexity and higher flexibility. They extract features by analyzing the current and voltage response data in the time domain, and based on this, establish a SOH health state prediction model. However, when the actual operating conditions of the battery are inconsistent with the experimental settings, the accuracy of these models may be affected or even fail. Summary of the invention

[0005] In view of this, the purpose of the present application is to provide a method and device for estimating the health status of a battery to overcome at least one of the above-mentioned defects.

[0006] In a first aspect, an embodiment of the present application provides a method for estimating a battery health state, the method comprising: obtaining battery cell parameters, the battery cell parameters comprising a current value, a voltage value, and an actual impedance value of the battery cell to be tested; constructing a target equivalent circuit of the battery cell to be tested according to the battery cell parameters, wherein the target equivalent circuit comprises a resistor and a plurality of sub-circuit elements, the number of the plurality of sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the measurement coefficient of the battery cell to be tested; and predicting an estimated value of the health state of the battery cell to be tested based on the target equivalent circuit.

[0007] In an optional embodiment of the present application, the battery cell parameters are determined in the following manner: receiving a start test instruction, the start test instruction is used to indicate a preset frequency of the charging current applied to the battery cell to be tested; for each preset frequency, applying the charging current at the preset frequency to the battery cell to be tested, and measuring the voltage value of the battery cell to be tested at the preset frequency; for each preset frequency, determining the actual impedance value of the battery cell to be tested under the charging current of the preset frequency based on the current value of the charging current at the preset frequency and the voltage value.

[0008] In an optional embodiment of the present application, the target equivalent circuit is constructed in the following manner: (A) an initial equivalent circuit is formed according to the current order; (B) a charging current at a preset frequency is applied to the initial equivalent circuit to determine the fitted impedance value of the battery cell to be tested, and a measurement coefficient of the initial equivalent circuit is determined based on the actual impedance value and the fitted impedance value at the preset frequency; (C) a determination is made as to whether the measurement coefficient is less than a design parameter; (D) if the measurement coefficient of the initial equivalent circuit is less than the design parameter, the target equivalent circuit is constructed using the current order; (E) if the measurement coefficient of the initial equivalent circuit is not less than the design parameter, the value of the current order is increased by one to update the current order, and the process returns to step (A).

[0009] In an optional embodiment of the present application, the measurement coefficient is determined in the following manner: for each actual impedance value, the angular frequency corresponding to the actual impedance value is determined; the maximum angular frequency and the minimum angular frequency are selected from multiple angular frequencies to determine the time coefficient; based on the time coefficient and the fitted impedance value, the imaginary impedance values ​​corresponding to multiple sub-circuit elements in the fitted impedance value are calculated; based on the imaginary impedance values ​​corresponding to multiple sub-circuit elements in the fitted impedance value, the measurement coefficient is determined.

[0010] In an optional embodiment of the present application, the measurement coefficient is determined by the following formula:

[0011]

[0012] Among them, μ represents the measurement coefficient, R k1 Represents the imaginary impedance value corresponding to the k1th subcircuit element among all imaginary impedance values ​​less than 0 in the target equivalent circuit, R k2 It represents the imaginary impedance value corresponding to the k2th sub-circuit element among all imaginary impedance values ​​not less than 0 in the target equivalent circuit, 1≤k1≤M, 1≤k2≤M, and M is the order of the target equivalent circuit.

[0013] In an optional embodiment of the present application, the time coefficient is determined by the following formula:

[0014]

[0015] Among them, τ k represents the time coefficient corresponding to the kth subcircuit element in the target equivalent circuit, τ min represents the minimum time coefficient determined by the minimum angular frequency, τ max represents the maximum time coefficient determined according to the maximum angular frequency, and M represents the order of the target equivalent circuit.

[0016] In an optional embodiment of the present application, the actual impedance value and the fitted impedance value are both complex numbers, the actual impedance value includes a first real impedance value and a first imaginary impedance value, and the fitted impedance value includes a second real impedance value and a second imaginary impedance value, wherein the method further includes: obtaining the fitted impedance value of the initial equivalent circuit under the same preset frequency charging current and the actual impedance value of the battery cell to be tested; determining the difference between the first real impedance value and the second real impedance value, and calculating the ratio of the difference to the modulus value of the fitted impedance value to determine the real fitting residual ; Determine the difference between the first imaginary impedance value and the second imaginary impedance value, and determine the imaginary fitting residual by taking the ratio of the difference to the modulus value of the fitting impedance value; determine whether the real fitting residual and the imaginary fitting residual are both smaller than the preset residual; if the real fitting residual and the imaginary fitting residual are both smaller than the preset residual, it means that the current order of the initial equivalent circuit is the order of the target equivalent circuit; if the real fitting residual and / or the imaginary fitting residual are not smaller than the preset residual, adjust the order of the target equivalent circuit according to the metric coefficient.

[0017] In a second aspect, an embodiment of the present application also provides a device for estimating a battery health status, the device comprising: an acquisition module for acquiring battery cell parameters, the battery cell parameters comprising current value, voltage value and actual impedance value of the battery cell to be tested; a construction module for constructing a target equivalent circuit of the battery cell to be tested according to the battery cell parameters, wherein the target equivalent circuit comprises a resistor and a plurality of sub-circuit elements, the number of the plurality of sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the measurement coefficient of the battery cell to be tested; a prediction module for predicting an estimated value of the health status of the battery cell to be tested based on the target equivalent circuit.

[0018] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the method described above are performed.

[0019] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described above are executed.

[0020] The method and device for estimating the battery health state provided in the embodiment of the present application obtain battery cell parameters, which include the current value, voltage value and actual impedance value of the battery cell to be tested; construct a target equivalent circuit of the battery cell to be tested according to the battery cell parameters, wherein the target equivalent circuit includes a resistor and multiple sub-circuit elements, the number of the multiple sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the measurement coefficient of the battery cell to be tested; based on the target equivalent circuit, predict the estimated value of the health state of the battery cell to be tested. Through this application, the efficiency and prediction accuracy of predicting the health state of the battery cell to be tested are improved.

[0021] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 A flowchart of a method for estimating a battery health status provided in an embodiment of the present application;

[0024] Figure 2 A flowchart of constructing a target equivalent circuit provided by an embodiment of the present application;

[0025] Figure 3 A schematic diagram of the structure of the target equivalent circuit provided in the embodiment of the present application;

[0026] Figure 4 A schematic diagram of the structure of a battery health status estimation device provided in an embodiment of the present application;

[0027] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0029] First, the application scenarios to which the present application is applicable are introduced. The present application can be applied to the technical field of battery health status assessment.

[0030] Research has found that during the charging process of lithium batteries, lithium ions are reduced to form lithium dendrites. The growth of lithium dendrites will continue until the internal short circuit of the battery occurs. This is an irreversible process. The formed lithium dendrites will lead to battery capacity loss and safety hazards. At present, the means of evaluating lithium ion SOH include model-based methods and data-driven methods. Data-driven methods are widely used due to their lower computational complexity and higher flexibility. Features are extracted by analyzing the current and voltage response data in the time domain, and a SOH health status prediction model is established based on this. However, when the actual operating conditions of the battery are inconsistent with the experimental settings, the accuracy of these models may be affected or even fail.

[0031] Based on this, the embodiment of the present application provides a method and device for estimating the health status of a battery, obtaining cell parameters, which include the current value, voltage value, and actual impedance value of the cell to be tested; constructing a target equivalent circuit of the cell to be tested based on the cell parameters, wherein the target equivalent circuit includes a resistor and multiple sub-circuit elements, the number of the multiple sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the measurement coefficient of the cell to be tested; based on the target equivalent circuit, predicting the estimated value of the health status of the cell to be tested. Through this application, the efficiency and accuracy of predicting the health status of the cell to be tested are improved.

[0032] See also Figure 1 , Figure 1 This is a flow chart of a method for estimating the battery health status provided in an embodiment of the present application. Figure 1 As shown in , the method for estimating the battery health status provided in the embodiment of the present application includes:

[0033] S101, obtaining battery cell parameters.

[0034] The cell parameters include the current value, voltage value and actual impedance value of the cell to be tested.

[0035] Here, the battery is composed of multiple cells to be tested, and the prediction results of the SOH test on the cells to be tested can be used to estimate the health status of the battery.

[0036] Preferably, the object to be tested is a lithium battery cell, because when the lithium battery is charged, lithium ions are reduced to form lithium dendrites, which will affect the battery and reduce the battery capacity. More seriously, it will cause high temperature, loss of control, fire, combustion and explosion, etc. Therefore, the SOH of the lithium battery needs to be tested.

[0037] Specifically, the cell parameters are determined in the following ways:

[0038] Receive a start test instruction;

[0039] Here, the start test instruction is used to indicate the preset frequency of the charging current applied to the battery cell to be tested;

[0040] For each preset frequency, applying a charging current at the preset frequency to the battery cell to be tested, and measuring the voltage value of the battery cell to be tested at the preset frequency;

[0041] For each preset frequency, based on the current value of the charging current at the preset frequency and the voltage value, the actual impedance value of the battery cell to be tested at the charging current at the preset frequency is determined.

[0042] Frequency domain analysis, specifically EIS analysis, is used to screen known power battery data through prerequisite judgment, and then the EIS data is preprocessed to ensure the validity of the EIS data.

[0043] Here, the prerequisite is to determine that the battery cell to be tested is not damaged and is a qualified sample. The battery to be tested is selected randomly to ensure that the battery cell to be tested is highly consistent with the samples on the market. The known power battery data can be voltage, cycle performance (normal battery cycle performance is 100%, and after 100 cycles it is above 99.7%, which proves that the battery cycle performance can reach 100%) and battery capacity.

[0044] Preprocessing is to screen whether these data are the same as the market average. If not, replace the same battery cell to be tested. In addition, the collected known power battery data is processed to eliminate noise, calibration errors, etc. At the same time, it may be necessary to segment the data to better reflect the contribution of different physical processes (such as charge transfer, diffusion, etc.).

[0045] Here, the EIS measurement principle is that under the action of a sinusoidal current of a given specific frequency, a corresponding voltage drop will be generated at both ends of the battery cell to be tested. The impedance value of the battery cell to be tested can be calculated by applying Ohm's law and Euler's formula. Here, the sinusoidal current is not limited, and alternating current can be used to change the voltage of the battery cell to be tested.

[0046] The actual impedance value of the battery cell to be tested under the preset frequency charging current is calculated by the following formula:

[0047]

[0048] Wherein, V represents the voltage value of the battery cell to be tested under the charging current of the preset frequency, I represents the current value of the preset frequency, It represents the phase difference between the current value and the voltage value, Z' represents the real impedance value of the actual impedance value, Z" represents the imaginary impedance value of the actual impedance value, and ω represents the angular frequency corresponding to the actual impedance value.

[0049] By changing the test frequency, a series of impedance data can be obtained, including real impedance values ​​and imaginary impedance values, etc., and then these impedance data are plotted on the complex plane to obtain the electrochemical impedance spectrum curve.

[0050] The core of EIS is to evaluate the frequency domain response of the electrochemical system. It is necessary to determine whether the EIS measurement results are meaningful through three basic conditions: causality, linearity, and stability. The details are as follows:

[0051] 1) Causality. In order to ensure that the measured AC signal is an independent response to the stimulus, the system response must not appear earlier than the stimulus signal.

[0052] 2) Linearity: The system’s response signal is required to be unaffected by the amplitude and phase of the excitation signal, that is, there is a linear function relationship.

[0053] 3) Stability: It requires that the system’s response to the excitation signal will not continue indefinitely, nor will it cause changes in the internal structure of the electrochemical system.

[0054] Here, the excitation signal is a signal sent by the test terminal to the battery cell to be tested, and the response signal is a response signal generated by the battery cell to be tested after receiving the excitation signal. For example, the excitation signal is a battery cell parameter such as voltage value and current value recorded by the test terminal when the battery cell to be tested does not receive a charging current, and the response signal is a signal for the battery cell to be tested to send a battery cell parameter signal to the test terminal.

[0055] S102: construct a target equivalent circuit of the battery cell to be tested according to the battery cell parameters.

[0056] The target equivalent circuit includes a resistor and multiple sub-circuit elements, the number of the multiple sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the measurement coefficient of the battery cell to be tested.

[0057] See also Figure 2 , Figure 2 This is a flow chart of constructing a target equivalent circuit provided by an embodiment of the present application. Figure 2 As shown in , the flowchart of constructing the target equivalent circuit provided by the embodiment of the present application includes:

[0058] S201, (A) forming an initial equivalent circuit according to the current order;

[0059] For example, the current order of the initial equivalent circuit is set to 1st order by default.

[0060] S202, (B) applying a charging current at a preset frequency to the initial equivalent circuit, determining a fitted impedance value of the battery cell to be tested, and determining a measurement coefficient of the initial equivalent circuit based on the actual impedance value and the fitted impedance value at the preset frequency;

[0061] Specifically, the measurement coefficient is determined in the following way:

[0062] For each actual impedance value, determining the angular frequency corresponding to the actual impedance value;

[0063] Select a maximum angular frequency and a minimum angular frequency from a plurality of angular frequencies, and determine a time coefficient;

[0064] The minimum time factor is determined by the following formula:

[0065]

[0066] Among them, τ min Indicates the minimum time coefficient corresponding to the actual impedance value, ω max It represents the maximum angular frequency corresponding to the actual impedance value, and τ1 represents the time coefficient corresponding to the first actual impedance value.

[0067] The maximum time factor is determined by the following formula:

[0068]

[0069] Among them, τ max represents the maximum time coefficient, ω min represents the minimum angular frequency corresponding to the actual impedance value, M is the order of the target equivalent circuit, τ M Indicates the time coefficient corresponding to the Mth actual impedance value.

[0070] The time factor is determined by the following formula:

[0071]

[0072] Among them, τ k represents the time coefficient corresponding to the kth subcircuit element in the target equivalent circuit, τ min represents the minimum time coefficient determined by the minimum angular frequency, τ max represents the maximum time coefficient determined according to the maximum angular frequency, and M represents the order of the target equivalent circuit.

[0073] Calculate the imaginary impedance values ​​corresponding to the multiple sub-circuit elements in the fitting impedance value according to the time coefficient and the fitting impedance value;

[0074] Based on applying a charging current at a preset frequency to the battery cell to be tested, applying a charging current with the same preset frequency to the initial equivalent circuit, and determining a fitting impedance value according to a voltage value, the fitting impedance value is a complex number.

[0075] The real and imaginary impedance values ​​of the actual impedance value are determined by the following formula:

[0076]

[0077]

[0078] Among them, Z Re (ω) represents the real impedance value of the actual impedance value, Z Re (ω') represents the real impedance value of the fitted impedance value, ω' represents the angular frequency corresponding to the fitted impedance value, Z Im (ω) represents the imaginary impedance value of the actual impedance value, Z Im (ω') represents the imaginary impedance value of the fitted impedance value, and ω represents the angular frequency corresponding to the actual impedance value.

[0079] Calculate formula (5) and (6) together to obtain the real impedance value and the imaginary impedance value of the actual impedance value.

[0080] Here, the key to accurately estimating SOH is to ensure that the measured impedance spectrum characteristics are linear and stable. Reasonable impedance data should satisfy the Cramer-Kronym relationship (ie, KK). The above formula determines the real and imaginary impedance values ​​of the actual impedance value by integration, which is more accurate.

[0081] The real impedance value and imaginary impedance value of the fitted impedance value are determined by the following formula:

[0082]

[0083] Where Z(ω) represents the fitted impedance value, R^ ohm Indicates the fitted resistance value of the resistor, R^ krepresents the fitted resistance value of the kth sub-circuit element, ω represents the angular frequency corresponding to the actual impedance value, τ k It represents the time coefficient corresponding to the kth sub-circuit element in the target equivalent circuit, j represents a constant, 1≤k≤M, and M is the order of the target equivalent circuit.

[0084] The metric coefficient is determined according to the imaginary impedance values ​​corresponding to the plurality of sub-circuit elements in the fitted impedance value.

[0085] Specifically, the measurement coefficient is determined by the following formula:

[0086]

[0087] Among them, μ represents the measurement coefficient, R k1 Represents the imaginary impedance value corresponding to the k1th subcircuit element among all imaginary impedance values ​​less than 0 in the target equivalent circuit, R k2 Represents the imaginary impedance value corresponding to the k2th sub-circuit element among all imaginary impedance values ​​not less than 0 in the target equivalent circuit, 1≤k1≤M, 1≤k2≤M, and M is the order of the target equivalent circuit.

[0088] Here, R k1 and R k2 are all one of the imaginary impedance values ​​corresponding to multiple sub-circuit elements. The difference is that R k1 is the imaginary impedance value corresponding to the k1th subcircuit element among all imaginary impedance values ​​less than 0, R k2 is the imaginary impedance value corresponding to the k2th sub-circuit element among all imaginary impedance values ​​not less than 0.

[0089] S203, (C) determining whether the measurement coefficient is less than the design parameter;

[0090] Here, it is crucial to select the number of sub-circuit elements. If the number of selected sub-circuit elements is too small, the measured cell parameters may not be accurately fitted due to the limited order of the equivalent circuit, thereby increasing the fitting residual. Conversely, if the number of sub-circuit elements is too large, noise information may be introduced, causing the effective information in the data to become blurred. Although the fitting residual may be within a reasonable range in this case, the accuracy of the cell parameters cannot be effectively verified. In order to quantify this impact more accurately, the number of sub-circuit elements, that is, the order of the target equivalent circuit, is determined by judging the metric coefficient μ and the design parameters.

[0091] In an optional embodiment, the μ value can be 0 to 1. When overfitting occurs, the number of imaginary impedance values ​​less than 0 will gradually increase, resulting in a monotonically decreasing μ value. To determine the appropriate number of sub-circuit elements in the equivalent circuit, this can be achieved by judging the size of the μ value during the fitting process.

[0092] In another optional embodiment, the value of the design parameter c has a significant impact on the result, and preferably, c is 0.85.

[0093] S204, (D) if the metric coefficient of the initial equivalent circuit is less than the design parameter, constructing the target equivalent circuit using the current order;

[0094] S205. (E) If the metric coefficient of the initial equivalent circuit is not less than the design parameter, the value of the current order is increased by one to update the current order, and the process returns to step (A).

[0095] According to the above-mentioned loop steps, the order of the target equivalent circuit is determined to construct the target equivalent circuit.

[0096] S103: predicting an estimated health status of the battery cell to be tested based on the target equivalent circuit.

[0097] After constructing the target equivalent circuit according to the cell parameters of the cell to be tested, the preprocessed data is input into the target equivalent circuit, which will calculate the predicted value of SOH based on the parameters of the equivalent circuit model and the current state of the cell. This predicted value can be expressed as the ratio of the current available capacity to the initial available capacity, or other forms such as the internal resistance growth rate, depending on the application scenario and requirements of the cell.

[0098] Here, the target equivalent circuit fits the resistor and the time constant of the equivalent circuit model (ECM) to the measured impedance spectrum. The advantage of this synchronous fitting is that it can reveal the random errors in the measurement process and then evaluate the frequency-dependent confidence interval of the predicted impedance.

[0099] like Figure 3 As shown, Figure 3 Schematic diagram of the structure of the target equivalent circuit provided in the embodiment of the present application, Rohm represents the real resistance, R1 represents the first imaginary resistance, C1 represents the first imaginary capacitance, R M Represents the Mth imaginary resistance, C M Represents the Mth imaginary capacitor. An imaginary resistor and an imaginary capacitor form a sub-circuit element. In the equivalent circuit, the resistor is used to represent the impedance or energy loss in the circuit, and the capacitor is used to represent the energy storage element in the circuit. The capacitor has a phase difference and frequency-dependent impedance to the current.

[0100] The structural diagram of the target equivalent circuit selects the commonly used parallel resistor-capacitor unit, and the size of the fitting residual can be observed to determine whether the data conforms to the KK relationship.

[0101] Specifically, the actual impedance value and the fitted impedance value are both complex numbers. The actual impedance value includes a first real impedance value and a first imaginary impedance value, and the fitted impedance value includes a second real impedance value and a second imaginary impedance value. The order of the target equivalent circuit can also be determined by calculating the real fitting residual and the imaginary fitting residual.

[0102] Obtain the fitted impedance value of the initial equivalent circuit under the same preset frequency charging current and the actual impedance value of the battery cell to be tested;

[0103] Determine a difference between the first real impedance value and the second real impedance value, and determine a real fitting residual by taking the ratio of the difference to the modulus value of the fitting impedance value;

[0104] The prediction results are verified and optimized. By comparing them with the actual battery cell test data, the accuracy and reliability of the target equivalent circuit can be evaluated. If there are deviations in the prediction results, the target equivalent circuit parameters can be further adjusted or the optimization algorithm can be used to improve the prediction accuracy.

[0105] In order to reasonably judge the correlation between the measured EIS data and the ECM (equivalent circuit model), the real part fitting residual and the imaginary part fitting residual are used as the basis for judging the rationality of the impedance data. The real part fitting residual is determined by the following formula:

[0106]

[0107] Among them, Δ Re (ω) represents the real part of the fitting residual, Z Re (ω) represents the first real impedance value of the actual impedance value, Z Re (ω') represents the second real impedance value of the fitting impedance value, and |Z(ω)| represents the modulus of the fitting impedance value.

[0108] Determine the difference between the first imaginary impedance value and the second imaginary impedance value, and determine the imaginary fitting residual by taking the ratio of the difference to the modulus value of the actual impedance value;

[0109] The imaginary fit residual is determined by the following formula:

[0110]

[0111] Among them, Δ Im (ω) represents the imaginary fitting residual, Z Im (ω) represents the first imaginary impedance value of the real impedance value, Z Im (ω') represents the second imaginary impedance value of the fitting impedance value, and |Z(ω)| represents the modulus of the fitting impedance value.

[0112] Determine whether the real part fitting residual and the imaginary part fitting residual are both less than the preset residual;

[0113] In an optional embodiment, if the real fitting residual and the imaginary fitting residual are both smaller than the preset residual, it indicates that the current order of the initial equivalent circuit is the order of the target equivalent circuit; if the real fitting residual and / or the imaginary fitting residual is not smaller than the preset residual, the order of the target equivalent circuit is adjusted according to the metric coefficient.

[0114] In another optional embodiment, in order to further explore the impedance characteristics of lithium-ion batteries under different conditions, the Lin-KK method can be used to perform cyclic impedance tests on 6 lithium-ion batteries under various temperatures and SOC conditions until the end of their life, and the real and imaginary impedance residuals of all cycles of each battery are calculated, and then the average value is taken to more accurately explore the impedance characteristics of the lithium-ion battery based on the average value.

[0115] One of the main problems in SOH estimation is feature extraction, including poor quality of battery feature extraction, complex extraction process, and insufficient adaptability of the algorithm. In the process of SOH estimation, it is often difficult to achieve accurate estimation requirements by simply relying on current, voltage, and temperature data as input, and manual processing of these raw data is extremely challenging. Therefore, using convolutional neural networks to automatically extract feature parameters of raw data has become a strategy worth trying.

[0116] Convolutional neural networks usually include an input layer, several hidden layers, and an output layer. Among them, Relu, as a commonly used activation function, has shown significant effects in solving the gradient vanishing problem. The convolution layer is responsible for processing the input data and maintaining a local connection with the previous layer, thereby effectively extracting the common features of the data. Suppose there is an image of size 4×4 (labeled ap), using a 2×2 convolution kernel (labeled α, β, γ, λ), and a convolution operation with a sliding step of 2. The pooling layer is often called a downsampling layer. Its core function is to compress the data size and reduce the processing complexity while ensuring the retention of key features. Regarding specific pooling techniques, there are many methods to choose from, such as Max Pooling, Mean Pooling, and Stochastic Pooling.

[0117] In the development of neural networks, a new type of recursive neural network, the long short-term memory (LSTM) neural network, is widely used. The core mechanism of this network lies in the three major logical control units: forget gate, input gate, and output gate. These units work together to effectively manage the maintenance and update of cell states, thereby significantly enhancing the long-term memory capacity of the network and overcoming the common gradient vanishing and gradient exploding problems in traditional recurrent neural networks (RNNs). Specifically, they include:

[0118] 1) In the LSTM network, the core function of the forget gate is to determine which information in the cell state at time t-1 should be discarded. The forget gate will generate an output value between 0 and 1, where a value close to 1 means that the corresponding information should be completely retained, and a value close to 0 means that the information should be completely forgotten. The calculation formula of the forget gate is shown as follows (11):

[0119] f (t) =σ(W f ×[h (t-1) , x (t) ]+b f ) (11)

[0120] Among them, f (t) is the state of the forget gate at time t; W f , b f are the weight and bias of the forget gate respectively; σ is the activation function; h (t-1) , x (t) are the hidden layer state at time t-1 and the input sequence at time t respectively.

[0121] 2) The input gate is responsible for reading the input x at time t (t) , and decides the information to be stored in the neuron accordingly. The tanh layer generates the temporary state C of the memory unit at time t (t) , by updating the cell state C (t) , ensuring the effective processing and transmission of information. The update process of the input gate is as follows:

[0122] I (t) =σ(Wi·[h (t-1) ,x (t) ]+b i ) (12)

[0123] C (t)’ =tanh(W c ×[h (t-1) ,x (t) ]+b c ) (13)

[0124]

[0125] Among them, i (t) is the input gate state at time t; W i , b i are the weight and bias of the input gate respectively; W c , b c are the weight matrix and bias term of the cell state respectively; tanh is the activation.

[0126] 3) The output gate ensures that the LSTM network can accurately control the output of information and realize the effective transmission and utilization of information when processing time series data. The calculation formula of the output gate is as follows:

[0127] o (t) =σ(W o ×[h (t-1) ,x (t) ]+b c ) (15)

[0128]

[0129] Among them, (t) is the state of the output gate at time t; Wo and bo are the weight matrix and bias term of the output gate respectively.

[0130] The BiLSTM neural network (Bidirectional Long Short-Term Memory) is an important optimization and improvement of the traditional unidirectional LSTM. BiLSTM combines a forward LSTM layer and a backward LSTM layer so that both layers can have a significant impact on the output. When processing time series data, the unidirectional LSTM can make full use of its historical information, effectively avoid long-distance dependency problems, and ensure the integrity and accuracy of the information. However, the unidirectional LSTM may be limited in processing certain complex time series data because it can only focus on one direction of the sequence. In contrast, BiLSTM can more comprehensively capture and utilize the information in the time series data by considering the two directions of the sequence at the same time, thereby improving the performance of the model. The output process is shown below:

[0131] A i =f1(ω1x i +ω2A i-1 ) (17)

[0132] B j =f2(ω3x i +ω5B i+1 ) (18)

[0133] Y i =f3(ω4A i +ω6B i ) (19)

[0134] Among them, x t represents the input data at time t, A t , B t denote the LSTM hidden states of the forward and backward iterations, respectively, and Y trepresents the output data, ω1, ω2, ω3, …, ω6 represent the corresponding weights of each layer, and f1, f2, f3 are the activation functions between different layers.

[0135] According to the study of human vision, people usually ignore minor information when paying attention to information and focus on specific parts. Based on this observation mode, a bionic mechanism, namely the attention mechanism, is proposed. This mechanism simulates the attention behavior of humans in hearing, language and vision, and improves the efficiency and accuracy of the model in processing data by forgetting low-probability and low-value information. In the attention mechanism, the weight selection of information directly depends on the attention value. Generally speaking, the calculation of the attention value includes two main steps: first, the attention is allocated to all input information, and then based on this allocation result, the weighted average of the input information is calculated to achieve accurate information processing.

[0136] The attention state transition process is as follows:

[0137] S ti =V tanh(Wh t +Uh i +b), i=1,2,3,...,t-1 (20)

[0138]

[0139] h′ t =f(F,h t ,y t ) (twenty three)

[0140] Where: S ti h t and h i The relevance of ti is the attention weight value of the hidden layer output value hi to the current input; y t is the input sequence; h t corresponds to the input sequence y t The hidden layer state value; h t ' is the final feature vector; V, W, U, b are the learning parameters of the model.

[0141] In the field of neural networks, overfitting is particularly common, especially when there is little training data or the learning time is too long. In addition, there is often a lack of clear and explicit causal relationship between the random features generated by the model and the target results, which further exacerbates the problem of overfitting. In order to ensure that the neural network performs well on both new and old data, overfitting needs to be avoided. At this time, regularization algorithms can be introduced into the neural network. As a widely used overfitting control method, Dropout plays a key role in the fully connected layer of the neural network. It can effectively reduce the model's excessive dependence on specific neurons by randomly adjusting the activation state of neurons, thereby improving the generalization ability of the model.

[0142] During model training, randomly inactivating neurons may lead to the loss of key neuron information, thereby weakening the network's ability to predict results. In order to effectively deal with this problem, this paper proposes a new strategy that combines Dropout and attention weight mechanisms. After introducing the attention weight mechanism, the inactivation probability of neurons is adjusted according to their weights. For neurons with higher attention weights, their inactivation probability will be reduced accordingly to ensure that those neurons that are critical to the prediction results can be retained to the greatest extent.

[0143] Then the CBA prediction model is formed, and its parameter settings are shown in Table 1 and Table 2:

[0144] Table 1 One-dimensional CNN structure settings

[0145]

[0146] Table 2 BiLSTM structure settings

[0147]

[0148] The model consists of three parts: one-dimensional CNN, BiLSTM neural network and attention mechanism.

[0149] First, the input layer preprocesses the impedance data;

[0150] Secondly, CNN automatically extracts effective nonlinear local features of EIS data;

[0151] Thirdly, the BiLSTM neural network is used to learn the nonlinear mapping relationship between the high-level features extracted by CNN and SOH;

[0152] Finally, the attention mechanism is used to optimize the BiLSTM network, automatically learning and focusing on the most critical and influential features for SOH estimation, thereby improving the model's efficiency in utilizing EIS data and SOH prediction performance. Since the amount of training data for SOH prediction is small, overfitting is prone to occur. Therefore, a Dropout layer is added to the model to prevent overfitting.

[0153] CNN can automatically process impedance data and extract features, which can effectively solve the problem of missing important features when manually extracting features. The Attention mechanism can optimize the BiLSTM prediction model and improve the prediction accuracy. Obtaining timely, accurate and more effective electric vehicle and battery big data can achieve accurate and efficient prediction of lithium-ion battery SOH with high prediction accuracy and good practicality. It can directly guide the design improvement of electric vehicles and power batteries from a mechanism perspective, help improve the life and safety of electric vehicles and batteries, and further promote their commercial application. The above method can form an empirical analysis model and be applied to the design and optimization of electric vehicles and power batteries. This method is original and extensible.

[0154] The battery health status estimation method and device provided in the embodiments of the present application obtain battery cell parameters, where the battery cell parameters include the current value, voltage value and actual impedance value of the battery cell to be tested; construct a target equivalent circuit of the battery cell to be tested based on the battery cell parameters, wherein the target equivalent circuit includes a resistor and multiple sub-circuit elements, the number of the multiple sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined based on the measurement coefficient of the battery cell to be tested; based on the target equivalent circuit, predict the estimated value of the health status of the battery cell to be tested. Through the present application, the efficiency and prediction accuracy of predicting the health status of the battery cell to be tested are improved.

[0155] This application uses big data analysis to achieve the effect of battery SOH prediction and avoid signal interference and lack of authenticity caused by objective factors; the convolutional neural network (CNN) can automatically process impedance data and extract features, which can effectively solve the problem of omission of important features when manually extracting features. The Attention mechanism can optimize the BiLSTM prediction model and improve the prediction accuracy; timely, accurate and more effective electric vehicle and battery big data can be obtained, which can achieve accurate and efficient prediction of lithium-ion battery SOH, with high prediction accuracy and good practicality.

[0156] This application can also directly guide the design improvement of electric vehicles and power batteries from a mechanism perspective, help improve the life and safety of electric vehicles and batteries, and further promote their commercial application. The above method can form an empirical analysis model and be applied to the design and optimization of electric vehicles and power batteries.

[0157] Based on the same inventive concept, an embodiment of the present application also provides a battery health status estimation device corresponding to the battery health status estimation method. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned battery health status estimation method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0158] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of the battery health status estimation device provided in the embodiment of the present application. Figure 4 As shown in FIG. 4 , the battery health status estimation device 400 includes:

[0159] An acquisition module 401 is used to acquire cell parameters, where the cell parameters include a current value, a voltage value, and an actual impedance value of the cell to be tested;

[0160] A construction module 402 is used to construct a target equivalent circuit of the battery cell to be tested according to the battery cell parameters, wherein the target equivalent circuit includes a resistor and a plurality of sub-circuit elements, the number of the plurality of sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the metric coefficient of the battery cell to be tested;

[0161] The prediction module 403 is used to predict the estimated health state value of the battery cell to be tested based on the target equivalent circuit.

[0162] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present application. Figure 5 As shown in , the electronic device 500 includes a processor 510 , a memory 520 and a bus 530 .

[0163] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 via the bus 530. When the machine-readable instructions are executed by the processor 510, the above-mentioned Figure 1 The steps of the method for estimating the battery health status in the method embodiment shown, the specific implementation method can be found in the method embodiment, and will not be repeated here.

[0164] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the method for estimating the battery health status in the method embodiment shown, the specific implementation method can be found in the method embodiment, and will not be repeated here.

[0165] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0166] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0167] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0169] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0170] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for estimating a battery health state, characterized in that: include: Obtaining battery cell parameters, wherein the battery cell parameters include a current value, a voltage value, and an actual impedance value of the battery cell to be tested; According to the cell parameters, construct a target equivalent circuit of the cell to be tested, wherein the target equivalent circuit includes a resistor and a plurality of sub-circuit elements, the number of the plurality of sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the metric coefficient of the cell to be tested; Based on the target equivalent circuit, an estimated health state value of the battery cell to be tested is predicted.

2. The method according to claim 1, characterized in that The cell parameters are determined in the following manner: receiving a start test instruction, wherein the start test instruction is used to indicate a preset frequency of a charging current applied to the battery cell to be tested; For each preset frequency, applying a charging current at the preset frequency to the battery cell to be tested, and measuring the voltage value of the battery cell to be tested at the preset frequency; For each preset frequency, based on the current value of the charging current at the preset frequency and the voltage value, the actual impedance value of the battery cell to be tested at the charging current at the preset frequency is determined.

3. The method according to claim 2, characterized in that The target equivalent circuit is constructed in the following way: (A) Based on the current order, an initial equivalent circuit is formed; (B) applying a charging current at a preset frequency to the initial equivalent circuit, determining a fitted impedance value of the battery cell to be tested, and determining a measurement coefficient of the initial equivalent circuit based on the actual impedance value and the fitted impedance value at the preset frequency; (C) determining whether the metric coefficient is less than a design parameter; (D) if the metric coefficient of the initial equivalent circuit is less than the design parameter, constructing the target equivalent circuit using the current order; (E) If the metric coefficient of the initial equivalent circuit is not less than the design parameter, the value of the current order is increased by one to update the current order, and the process returns to step (A).

4. The method according to claim 2, characterized in that: The scaling factor is determined by: For each actual impedance value, determining the angular frequency corresponding to the actual impedance value; Select a maximum angular frequency and a minimum angular frequency from a plurality of angular frequencies, and determine a time coefficient; Calculating the imaginary impedance values ​​corresponding to the plurality of sub-circuit elements in the fitting impedance value according to the time coefficient and the fitting impedance value; The metric coefficient is determined according to the imaginary impedance values ​​corresponding to the plurality of sub-circuit elements in the fitting impedance value.

5. The method according to claim 3, characterized in that: The scaling factor is determined by the following formula: Among them, μ represents the measurement coefficient, R k1 Represents the imaginary impedance value corresponding to the k1th subcircuit element among all imaginary impedance values ​​less than 0 in the target equivalent circuit, R k2 It represents the imaginary impedance value corresponding to the k2th sub-circuit element among all imaginary impedance values ​​not less than 0 in the target equivalent circuit, 1≤k1≤M, 1≤k2≤M, and M is the order of the target equivalent circuit.

6. The method according to claim 5, characterized in that The time factor is determined by the following formula: Among them, τ k represents the time coefficient corresponding to the kth subcircuit element in the target equivalent circuit, τ min represents the minimum time coefficient determined by the minimum angular frequency, τ max represents the maximum time coefficient determined according to the maximum angular frequency, and M represents the order of the target equivalent circuit.

7. The method according to claim 3, characterized in that The actual impedance value and the fitted impedance value are both complex numbers, the actual impedance value includes a first real impedance value and a first imaginary impedance value, and the fitted impedance value includes a second real impedance value and a second imaginary impedance value, Wherein, the method further comprises: Obtaining the fitted impedance value of the initial equivalent circuit under the same preset frequency charging current and the actual impedance value of the battery cell to be tested; Determine a difference between the first real impedance value and the second real impedance value, and determine a real fitting residual by taking the ratio of the difference to the modulus value of the fitting impedance value; Determine a difference between the first imaginary impedance value and the second imaginary impedance value, and determine an imaginary fitting residual by taking the ratio of the difference to the modulus of the fitting impedance value; Determine whether the real part fitting residual and the imaginary part fitting residual are both less than a preset residual; If the real part fitting residual and the imaginary part fitting residual are both smaller than the preset residual, it indicates that the current order of the initial equivalent circuit is the order of the target equivalent circuit; If the real part fitting residual and / or the imaginary part fitting residual is not less than a preset residual, the order of the target equivalent circuit is adjusted according to the metric coefficient.

8. A battery health status estimation device, characterized in that: include: An acquisition module is used to acquire cell parameters, wherein the cell parameters include a current value, a voltage value, and an actual impedance value of the cell to be tested; A construction module, configured to construct a target equivalent circuit of the cell to be tested according to the cell parameters, wherein the target equivalent circuit comprises a resistor and a plurality of sub-circuit elements, the number of the plurality of sub-circuit elements is the same as the order of the target equivalent circuit, and the order of the target equivalent circuit is determined according to the metric coefficient of the cell to be tested; A prediction module is used to predict the estimated health state of the battery cell to be tested based on the target equivalent circuit.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and the processor executes the machine-readable instructions to perform the steps of any method as claimed in claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are executed.

Images