Method and system for rapidly calculating impedance spectrum data of lithium ion battery in multiple states

Abstract

The invention provides a method and a system for rapidly calculating impedance spectrum data of a lithium ion battery in multiple states. The method comprises the following steps: S1, acquiring impedance spectrum data and constructing a data set; s2, selecting a corresponding machine learning algorithm to build a battery impedance spectrum calculation model; s3, taking the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model; and S4, deploying the trained battery impedance spectrum calculation model, and completing parameter calculation. The method provided by the invention can be applied to an experiment or engineering scene in which the impedance spectrums of the lithium ion battery in a plurality of different states need to be measured in sequence, the total time of multiple measurements is greatly shortened, especially a large amount of standing waiting time is saved, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium ion battery detection, and in particular to a method and system for quickly calculating impedance spectrum data of a lithium ion battery under multiple states. Background Art

[0002] In recent years, my country's advancement and development of new energy technology has reached an unprecedented level. As the current mainstream new energy carrier, lithium-ion batteries are widely used in electric vehicles and energy storage power stations. On the one hand, since 2015, my country's sales and ownership of electric vehicles powered by lithium-ion batteries have reached the first place in the world. On the other hand, with the construction of energy storage power station projects in various places in recent years, the demand for lithium-ion batteries as energy storage carriers has risen again. The widespread use of lithium-ion batteries has driven the prosperity of related industries, but some key issues have not been effectively resolved so far.

[0003] The impedance data of lithium-ion batteries contains important information about the internal electrochemical state of the battery. However, the current measurement of accurate impedance data inside lithium-ion batteries requires high measurement equipment and is time-consuming and labor-intensive, which has caused certain obstacles to in-depth research on the internal electrochemical state and electrochemical reaction mechanism of the battery. These current situations show that it is of great significance to propose some improvement strategies for the current impedance data acquisition methods of lithium-ion batteries.

[0004] At present, the traditional method of measuring the impedance spectrum of lithium-ion batteries is mainly to use a professional electrochemical workstation or impedance meter. However, these measurement methods have some problems, as follows:

[0005] 1) Electrochemical workstations or impedance meters are expensive;

[0006] 2) Each measurement can only measure the battery impedance spectrum data under a single temperature, a single health state (SOH), and a single charge state (SOC, for example: when the battery is fully charged, the charge state is 100%; when the battery is fully discharged, the charge state is 0%), and it is impossible to obtain impedance spectrum data under multiple charge states at the same time;

[0007] 3) It takes a long time to measure the impedance spectrum of a battery using an electrochemical workstation or impedance meter. A single measurement usually takes at least 10 minutes, and the battery needs to be left to stand for at least 6 to 12 hours before measurement to allow the battery to reach a relatively stable electrochemical state. If you want to measure impedance data at multiple different charge states, it will take at least several days.

[0008] In addition, there are some novel impedance spectrum measurement methods and technologies for lithium-ion batteries. One type of method applies a current or voltage excitation signal to the battery to obtain the voltage or current response signal of the battery, uses digital signal conversion to process the signal, and calculates the impedance spectrum data of the battery. Although this type of method has lower requirements for equipment and is easy to implement, due to the presence of noise in the high-frequency signal obtained, its measurement accuracy is worse than that of traditional impedance spectrum measurement methods, and it cannot save time in measuring the impedance spectrum.

[0009] Another method is to build an equivalent circuit model of the battery impedance spectrum, identify the model parameters with the help of the existing current and voltage data during the battery charging and discharging process, and then calculate the battery impedance spectrum data. Although this method can save a lot of measurement time, it cannot make a completely accurate model of the equivalent circuit of the lithium battery, and its measurement error is greater than that of the traditional impedance spectrum measurement method.

[0010] In summary, the currently commonly used battery impedance spectrum data measurement method faces the problem of expensive professional measurement equipment and difficulty in universal application on the one hand, and the problem of time-consuming and labor-intensive measurement process and poor measurement data accuracy on the other hand.

[0011] In the Chinese patent document with publication number CN115792670A, a rapid test method for lithium battery impedance spectrum using an optimized binary sequence is disclosed, which uses an optimized binary sequence as a current input to excite the battery to obtain voltage response data, and uses the Fourier transform method to calculate the value of the battery impedance spectrum. However, due to the limitations of the Fourier transform method and the digital signal sampling rate, the calculated impedance value has average accuracy in the low-frequency range and high-frequency range.

[0012] In the Chinese patent document with publication number CN116008835B, a battery impedance spectrum test method based on single pulse response is disclosed, which obtains the parameter data of the nonlinear analog equivalent circuit by applying a short pulse current to the lithium battery and using the collected current and voltage signals. However, the patent document also uses the time-frequency transformation method to process the voltage and current signals to obtain the impedance spectrum data, so the calculated impedance value has a general accuracy in the low-frequency range and the high-frequency range.

[0013] In a Chinese patent document with publication number CN117872192A, a rapid measurement method for the impedance spectrum of an energy storage battery is disclosed, which applies an excitation current signal to the energy storage battery to be measured, samples the excitation current signal and the response voltage signal, and then obtains the voltage and current signal components at different frequencies through discrete Fourier transform, and obtains the electrochemical impedance spectrum data of the energy storage battery according to the corresponding voltage and current signal components. Similarly, the patent document uses the discrete Fourier transform method to process the voltage and current signal components to obtain the impedance spectrum data, so the calculated impedance value has a general accuracy in the low-frequency range and the high-frequency range.

[0014] In the Chinese patent document with the publication number CN109459465B, a rapid electrochemical impedance spectroscopy measurement method based on current pulse injection is disclosed, which injects a pulse perturbation current in the form of an M sequence into the output current of the electrochemical system under test by controlling a controllable constant current source circuit, and then collects the perturbation current and response voltage at the output end of the electrochemical system under test, and uses an impedance calculation method to obtain the electrochemical impedance spectrum for the collected perturbation current and response voltage data. However, since the pulse current duration is very short, this method can only accurately measure and reflect the transient electrochemical response part inside the battery (also belonging to the electrochemical system), and cannot measure the steady-state electrochemical response well, so the data error in the low-frequency impedance part is large.

[0015] In the Chinese patent document with publication number CN115575714A, a method is disclosed that uses a multi-frequency mixed non-standard sinusoidal signal as an excitation signal to act on a lithium battery, collects the voltage response signal and loop current signal of the lithium battery, performs Fourier transform after digital filtering, and calculates the impedance spectrum data. However, the patent document also uses the discrete Fourier transform method to process the voltage and current signal components to obtain the impedance spectrum data, so the calculated impedance value has average accuracy in the low-frequency range and the high-frequency range.

[0016] In a Chinese patent document with publication number CN106970266A, a method for rapid measurement of impedance spectrum of lithium-ion batteries is disclosed, which applies a square wave current of a selected frequency point and a specific amplitude to the lithium-ion battery, performs Fourier transform on the sampled square wave current and response voltage, obtains the impedance at a specific frequency point, and then obtains the electrochemical impedance spectrum data of the lithium-ion battery. However, the patent document uses the Fourier transform method to process the voltage and current signal components to obtain the impedance spectrum data, so the calculated impedance value has a general accuracy in the low-frequency range and the high-frequency range.

[0017] In the Chinese patent document with publication number CN112881929B, a step-wave-based online measurement method for low-frequency band EIS of lithium-ion batteries is disclosed, which applies a step-wave current of a suitable frequency range to the lithium-ion battery, performs a sine fit on the sampled step-wave current and response voltage, obtains the impedance value of the required low-frequency band, and then obtains the low-frequency band electrochemical impedance spectrum data of the lithium-ion battery. The patent document uses a sine fitting method to calculate the current and voltage signal components on different frequency components, processes the voltage and current signal components obtained by Fourier transform to finally calculate the impedance spectrum value, but the calculated impedance value has average accuracy in the low-frequency range and high-frequency range.

[0018] In a Chinese patent document with publication number CN117368774A, a method is disclosed for quickly testing the battery impedance spectrum by designing a current excitation signal containing multi-frequency components using an inversely repeated M sequence. The battery current excitation and the measured voltage response are subjected to continuous wavelet transform by selecting the Morse complex wavelet as the mother wavelet to obtain the battery impedance spectrum data. The patent document uses the Morse wavelet transform method in the signal time-frequency domain to process the voltage and current signal components to obtain the impedance spectrum data. This method is more accurate than the extraction method using Fourier transform, but there is still a certain error in the impedance value obtained in the low-frequency range, and the impedance value finally calculated has a general accuracy in the low-frequency range. Summary of the invention

[0019] In view of the defects in the prior art, the object of the present invention is to provide a method and system for quickly calculating impedance spectrum data of lithium-ion batteries under multiple states.

[0020] A method for quickly calculating impedance spectrum data of a lithium-ion battery under multiple states provided by the present invention comprises:

[0021] Step S1: collecting impedance spectrum data and constructing a data set;

[0022] Step S2: Select a corresponding machine learning algorithm to build a battery impedance spectrum calculation model;

[0023] Step S3: using the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model;

[0024] Step S4: deploy the trained battery impedance spectrum calculation model to complete parameter calculation.

[0025] Preferably, step S1 includes the following sub-steps:

[0026] Step S1.1: changing the battery to different preset states and leaving it alone for a certain period of time;

[0027] The preset state includes an aging state, a charge state, and a temperature state;

[0028] Step S1.2: using a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under the corresponding state;

[0029] Step S1.3: classify and organize the measured impedance spectrum data to construct an impedance spectrum data set;

[0030] Step S1.4: Record the current battery status data and enter it into the database;

[0031] Step S1.5: Repeat steps S1.1-S1.4 until the impedance spectrum data set contains a preset number of impedance spectrum data.

[0032] Preferably, step S2 includes the following sub-steps:

[0033] Step S2.1: Select the core machine learning algorithm of the battery impedance spectrum calculation model;

[0034] Step S2.2: setting the input variables, output function values ​​and model parameters of the battery impedance spectrum calculation model;

[0035] Step S2.3: Use programming software to build the battery impedance spectrum calculation model.

[0036] Preferably, the input variables of the battery impedance spectrum calculation model are composed of a battery state vector, and the battery state includes a health state, an ambient temperature state, and a charge state; the output function value includes a vector composed of impedance values ​​of the battery in different states.

[0037] Preferably, step S3 includes dividing the data set in the impedance spectrum data into input variable vectors according to the required battery state; substituting the input variable vectors into the battery impedance spectrum calculation model as training samples to calculate the impedance spectrum data of the battery in other states.

[0038] A system for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states provided by the present invention comprises:

[0039] Module M1: Collect impedance spectroscopy data and build a data set;

[0040] Module M2: Select the corresponding machine learning algorithm to build a battery impedance spectrum calculation model;

[0041] Module M3: using the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model;

[0042] Module M4: Deploy the trained battery impedance spectrum calculation model and complete parameter calculation.

[0043] Preferably, the module M1 includes the following submodules:

[0044] Module M1.1: Change the battery to different preset states and leave it for a certain period of time;

[0045] The preset state includes an aging state, a charge state, and a temperature state;

[0046] Module M1.2: Use a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under the corresponding state;

[0047] Module M1.3: Classify and organize the measured impedance spectrum data and construct an impedance spectrum data set;

[0048] Module M1.4: record the current battery status data and enter it into the database;

[0049] Module M1.5: Repeat modules M1.1-M1.4 until the impedance spectrum data set contains a preset number of impedance spectrum data.

[0050] Preferably, the module M2 includes the following submodules:

[0051] Module M2.1: Select the core machine learning algorithm for the battery impedance spectrum calculation model;

[0052] Module M2.2: Setting the input variables, output function values ​​and model parameters of the battery impedance spectrum calculation model;

[0053] Module M2.3: Use programming software to build the battery impedance spectrum calculation model.

[0054] Preferably, the input variables of the battery impedance spectrum calculation model are composed of a battery state vector, and the battery state includes a health state, an ambient temperature state, and a charge state; the output function value includes a vector composed of impedance values ​​of the battery in different states.

[0055] Preferably, the module M3 includes dividing the data set in the impedance spectrum data into input variable vectors according to the required battery state; substituting the input variable vectors into the battery impedance spectrum calculation model as training samples to calculate the impedance spectrum data of the battery in other states.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. The method provided by the present invention can be applied in experiments or engineering scenarios where the impedance spectra of lithium-ion batteries in multiple different states need to be measured sequentially, which greatly shortens the overall time of multiple measurements, especially saves a lot of static waiting time, and improves work efficiency.

[0058] 2. The present invention incorporates the influence of various different states on battery impedance in the calculation process of battery impedance spectrum data, including battery aging state, battery charge state, and the influence of external temperature on impedance. The method of obtaining lithium-ion battery impedance spectrum data under different states is improved, and signal time-frequency transformation (such as Fourier transform, wavelet transform, etc.) is not used, thereby avoiding sampling signal noise and transformation errors introduced in the current and voltage signal processing process, and greatly improving the accuracy of the calculation results.

[0059] 3. The present invention uses a machine learning model to map and model the relationship between the battery impedance spectrum data under the measured battery state and the battery impedance spectrum data under the unmeasured battery state; the measured battery impedance spectrum data and the battery state data are substituted into the model variables, and through the training and calculation of the machine learning model, the battery impedance spectrum data of the battery in other different states can be calculated, which has good flexibility.

[0060] 4. The present invention reduces the requirements on the processor's memory, computing power and other performance, and can calculate the battery impedance spectrum data under multiple different states using a simple machine learning model, which has good practicality.

[0061] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0063] Figure 1 It is a framework diagram of the method flow of the present invention.

[0064] Figure 2 Schematic diagram of the calculation effect of battery impedance spectrum data in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0066] Reference Figure 1 As shown, a method for quickly calculating impedance spectrum data of lithium-ion batteries under multiple states includes:

[0067] Step 1: Construct an impedance spectroscopy dataset based on a small amount of measured impedance spectroscopy data:

[0068] 1. Change the battery to different preset states

[0069] Use common battery charging and discharging instruments and constant temperature incubators in laboratories or projects to change the battery state to the preset aging state, charge state, and temperature state, and leave it to stand for at least 12 hours to allow the battery to basically reach electrochemical stability under the preset state.

[0070] 2. Use a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery in this state.

[0071] The impedance spectrum data of the battery in this state is measured using a battery impedance spectrum measuring instrument according to a conventional electrochemical impedance spectrum measurement method.

[0072] 3. Construct impedance spectrum data set

[0073] The impedance spectrum data obtained in the first two steps are classified and sorted according to the state to construct an impedance spectrum data set. At the same time, the battery status data at this time (including battery aging status, temperature, battery charge state, etc.) is recorded and also recorded in the data file.

[0074] 4. Repeat steps 1 to 3 above.

[0075] Repeat the steps until a certain number of battery impedance spectrum data sets under different states are constructed, and the constructed data sets will be used as training samples for the subsequent machine learning model.

[0076] Step 2: Build a battery impedance spectrum calculation model based on machine learning methods:

[0077] 1. Choose the right machine learning model.

[0078] Since the electrochemical impedance spectrum data of the battery changes with the battery state in a certain pattern, regression models (such as support vector regression, Gaussian process regression, etc.), supervised learning, neural networks and other machine learning methods can be used as the core algorithm of the battery impedance spectrum calculation model.

[0079] 2. Determine the input variables, output function values, and model parameters of the machine learning model.

[0080] Specifically, the input variables of the machine learning model are a vector composed of the state of the battery, which must include information such as the battery's health status, ambient temperature status, and battery charge status; the output function value is a vector composed of the impedance values ​​of the battery under different states.

[0081] Since the input and output vectors of the model have a small amount of data and low dimensions, the level and complexity of the model can be appropriately reduced when adjusting the model parameters and architecture.

[0082] 3. Implement the battery impedance spectrum calculation model based on machine learning in a computer programming environment.

[0083] Step 3: Calculate the impedance spectrum data of the battery under different conditions through the constructed model

[0084] According to the battery state to be calculated (for example, at the same temperature, under the same battery aging state, etc.), the data set in the impedance spectrum data obtained in step one is divided into input variable vectors, and the input variables are substituted as training samples into the battery impedance spectrum calculation model based on the machine learning method described in step two, so that the impedance spectrum data of the battery in other states can be calculated.

[0085] The present invention incorporates the influence of various different states on the battery impedance in the calculation process of the battery impedance spectrum data, including the battery aging state, the battery charge state, the influence of the external temperature on the impedance, etc.; improves the method of obtaining the lithium-ion battery impedance spectrum data in different states, does not use signal time-frequency transformation (such as Fourier transformation, wavelet transformation, etc.), avoids the sampling signal noise and transformation error introduced in the current and voltage signal processing process, and greatly improves the accuracy of the calculation result; at the same time, this method can be applied to experiments or engineering scenarios that need to measure the impedance spectra of lithium-ion batteries in multiple different states in sequence, greatly shortens the overall time of multiple measurements, especially saves a lot of static waiting time, and improves work efficiency.

[0086] The above is a basic embodiment of the present invention. The technical solution of the present invention is further described below through a preferred embodiment.

[0087] Example 1

[0088] Reference Figure 1 As shown, a method for quickly calculating impedance spectrum data of lithium-ion batteries under multiple states includes:

[0089] 1. Data preparation stage

[0090] 1. Change the battery to several different preset states

[0091] Use common battery charging and discharging instruments and constant temperature incubators in laboratories or projects to change the battery state to a preset health state, charge state, and temperature state, and leave it to stand for at least 12 hours so that the battery can basically achieve electrochemical stability in the preset state. In this embodiment, the battery is changed to different charge states SOC1, SOC2, and SOC3 in turn, keeping the ambient temperature the same as T, and the battery health state is SOH1.

[0092] 2. Use a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under these conditions.

[0093] The impedance spectrum data of the battery in this state is measured using a battery impedance spectrum measuring instrument according to a conventional electrochemical impedance spectrum measurement method. In the three different battery states of this embodiment, the impedance spectrum data of the battery are measured respectively, the measurement frequency range is set to 0.01 Hz to 5 kHz, and the amplitude of the sinusoidal excitation current is set to 0.2 A.

[0094] 3. Construct impedance spectrum data set

[0095] The impedance spectrum data obtained through the first two steps are classified and sorted according to the state to construct an impedance spectrum data set. In this embodiment, the battery state information (including health state, temperature state, charge state) and the measured battery impedance spectrum data need to be saved in a data file.

[0096] 4. Repeatedly change battery status, battery static, impedance spectrum data measurement and data collation

[0097] Repeat the contents of steps 2 and 3 above to change the battery to these three battery states in turn; repeatedly let the battery stand, measure the battery impedance spectrum data and organize the data until the impedance spectrum data of the battery in all different states are fully measured.

[0098] 5. Constructing impedance spectrum data set

[0099] The battery impedance spectrum data obtained through the first two steps are classified and sorted according to different state dimensions, such as the same temperature, the same aging state, the same state of charge, etc., to construct an impedance spectrum data set.

[0100] 2. Preparation stage for machine learning model construction

[0101] 1. Build the impedance spectrum calculation model and calculation formula based on machine learning in the computer programming environment, and determine the model's hyperparameters. Define the model's input variable vector and output variable vector, define the input variable as the battery's state vector (list the health state, temperature state, and charge state into vectors) and the impedance spectrum data under known states, and the output variable is the battery impedance spectrum data vector to be calculated and its confidence interval.

[0102] In this embodiment, a Gaussian process regression machine learning model is used. The Gaussian process regression method is a machine learning method of non-parametric regression based on Bayesian inference, which can establish a nonlinear functional relationship between any input variable and output value. In the Gaussian process regression method, multiple function values ​​are multidimensional joint normal distributions of multiple input variables, as shown in the following formula.

[0103]

[0104] in, Indicates that the function value conforms to the multidimensional normal distribution relationship of the input variables, k means there are k groups of corresponding input variables and function values, x i is the i-th input variable (specifically, in this embodiment, the input variables are the battery's state of health SOH1, temperature state T, state of charge SOC1, SOC2, and SOC3), f(x i ) is the output function value corresponding to the input variable vector (specifically, in the implementation process, the output function value is the impedance value of the battery in this state), μ i represents the expected value of the output function value under multidimensional joint normal distribution, ∑ ij Input variable x in multidimensional joint normal distribution i and x j The corresponding covariance matrix, in the formula, the subscripts i and j of the variables represent the order of the variables, which are any integers between 1 and k, that is, i = 1, 2..., k; j = 1, 2..., k;

[0105] If the input variable x is regarded as a continuous variable domain, the Gaussian process between the input variable and the function value can be further expressed as follows:

[0106] f(x)~GP(m(x),K)

[0108] Among them, m(x) is the mean function vector on the input variable domain, and K is the input variable covariance matrix.

[0109] 2. Define the kernel function of the covariance matrix

[0110] Each element in the covariance matrix is ​​expressed by kernel function calculation. The commonly used Gaussian kernel function (or radial basis function) is calculated as follows:

[0111]

[0112] Among them, α 2 is the vertical scale of the kernel function (i.e., the amplitude of the function), β is the length scale (i.e., the rate at which the correlation decreases with increasing distance), and ||x i -xj ||2 is the variable x i and x j In the formula, the subscripts i and j of the variables represent the order of the variables and are any integers between 1 and k, i.e., i = 1, 2, ..., k; j = 1, 2, ..., k.

[0113] 3. Use the built model to derive the formula for calculating the unknown function value using the known function value

[0114] The output function value vector (a vector of battery impedance values ​​under known conditions) f(x) and its noise variance Now there is a new set of unknown output function values ​​to be calculated (a vector of battery impedance values ​​in other states) f(x * ), assuming that the mean function m(x) of the Gaussian process is a 0 vector, the joint Gaussian probability distribution of the training sample and the estimated sample is as follows:

[0115]

[0116] Among them, the covariance matrix K = K(x, x), K * =K(x,x * ), K ** =K(x * , x * ). These covariance matrices can be defined and represented using square matrices.

[0117] Finally, the battery impedance value and variance in other states can be calculated by the following formula:

[0118]

[0119] Among them, the vector That is, the calculated battery impedance value vector under other conditions, the variance matrix Represents the upper and lower 95% confidence interval range of the battery impedance value vector.

[0120] 3. Calculate the battery impedance spectrum data under other conditions

[0121] In this embodiment, the impedance spectrum calculation model program built above is opened, and the data file recording the battery impedance spectrum data and the corresponding battery state is imported and substituted into the input variable vector of the model. After adjusting the hyperparameters of the appropriate Gaussian process regression model, the battery impedance spectrum data under other battery states is finally calculated. Figure 2As shown, by comparing the impedance spectrum data obtained by the calculation method proposed by the present invention with the actual measured data, a calculation error of less than 2% can be achieved (wherein the curve marked with hollow circles is the impedance spectrum data actually measured for the battery, and the curve marked with solid points is the impedance spectrum data in this state calculated using data in other states). The battery state is: ambient temperature 25°C, battery state of charge SOC=20%, and battery health state SOH=0.98.

[0122] It should be noted that other supervised learning models or data-driven methods such as neural networks can be used to take the battery state and the corresponding battery impedance spectrum data as input variables and training samples, and output and calculate the battery impedance spectrum data values ​​under other multiple states. Although other supervised learning methods or data-driven methods may bring greater computational complexity and sacrifice certain simplicity and ease of implementation, they can also solve the problem solved by the present invention.

[0123] The present invention also provides a system for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states. The system for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states can be realized by executing the process steps of the method for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states, that is, those skilled in the art can understand the method for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states as a preferred implementation of the system for rapidly calculating impedance spectrum data of a lithium-ion battery under multiple states.

[0124] Specifically, a system for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states includes:

[0125] Module M1: Collect impedance spectroscopy data and build a data set;

[0126] Module M2: Select the corresponding machine learning algorithm to build a battery impedance spectrum calculation model;

[0127] Module M3: using the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model;

[0128] Module M4: Deploy the trained battery impedance spectrum calculation model and complete parameter calculation.

[0129] The module M1 includes the following submodules:

[0130] Module M1.1: Change the battery to different preset states and leave it for a certain period of time;

[0131] The preset state includes an aging state, a charge state, and a temperature state;

[0132] Module M1.2: Use a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under the corresponding state;

[0133] Module M1.3: Classify and organize the measured impedance spectrum data and construct an impedance spectrum data set;

[0134] Module M1.4: record the current battery status data and enter it into the database;

[0135] Module M1.5: Repeat modules M1.1-M1.4 until the impedance spectrum data set contains a preset number of impedance spectrum data.

[0136] The module M2 includes the following submodules:

[0137] Module M2.1: Select the core machine learning algorithm for the battery impedance spectrum calculation model;

[0138] Module M2.2: Setting the input variables, output function values ​​and model parameters of the battery impedance spectrum calculation model;

[0139] Module M2.3: Use programming software to build the battery impedance spectrum calculation model.

[0140] The input variables of the battery impedance spectrum calculation model are composed of a battery state vector, and the battery state includes a health state, an ambient temperature state, and a charge state; the output function value includes a vector composed of impedance values ​​of the battery in different states.

[0141] The module M3 includes dividing the data set in the impedance spectrum data into input variable vectors according to the required battery state; substituting the input variable vectors into the battery impedance spectrum calculation model as training samples to calculate the impedance spectrum data of the battery in other states.

[0142] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0143] In the description of the present application, it should be understood that the terms "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0144] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states, characterized in that: include: Step S1: collecting impedance spectrum data and constructing a data set; Step S2: Select a corresponding machine learning algorithm to build a battery impedance spectrum calculation model; Step S3: using the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model; Step S4: deploy the trained battery impedance spectrum calculation model to complete parameter calculation.

2. The method for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states according to claim 1, characterized in that: The step S1 comprises the following sub-steps: Step S1.1: changing the battery to different preset states and leaving it alone for a certain period of time; The preset state includes an aging state, a charge state, and a temperature state; Step S1.2: using a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under the corresponding state; Step S1.3: classify and organize the measured impedance spectrum data to construct an impedance spectrum data set; Step S1.4: Record the current battery status data and enter it into the database; Step S1.5: Repeat steps S1.1-S1.4 until the impedance spectrum data set contains a preset number of impedance spectrum data.

3. The method for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states according to claim 1, characterized in that: The step S2 comprises the following sub-steps: Step S2.1: Select the core machine learning algorithm of the battery impedance spectrum calculation model; Step S2.2: setting the input variables, output function values ​​and model parameters of the battery impedance spectrum calculation model; Step S2.3: Use programming software to build the battery impedance spectrum calculation model.

4. The method for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states according to claim 3, characterized in that: The input variables of the battery impedance spectrum calculation model are composed of a battery state vector, and the battery state includes a health state, an ambient temperature state, and a charge state; the output function value includes a vector composed of impedance values ​​of the battery in different states.

5. The method for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states according to claim 1, characterized in that: The step S3 includes dividing the data set in the impedance spectrum data into input variable vectors according to the desired battery state; substituting the input variable vectors into the battery impedance spectrum calculation model as training samples to calculate the impedance spectrum data of the battery in other states.

6. A system for rapidly calculating impedance spectrum data of lithium-ion batteries under multiple states, characterized in that: include: Module M1: Collect impedance spectroscopy data and build a data set; Module M2: Select the corresponding machine learning algorithm to build a battery impedance spectrum calculation model; Module M3: using the impedance spectrum data set as a training sample to train the battery impedance spectrum calculation model; Module M4: Deploy the trained battery impedance spectrum calculation model and complete parameter calculation.

7. The lithium-ion battery multi-state impedance spectrum data rapid calculation system according to claim 6, characterized in that: The module M1 Includes the following submodules: Module M1.1: Change the battery to different preset states and leave it for a certain period of time; The preset state includes an aging state, a charge state, and a temperature state; Module M1.2: Use a battery impedance spectrum measuring instrument to measure the impedance spectrum data of the battery under the corresponding state; Module M1.3: Classify and organize the measured impedance spectrum data and construct an impedance spectrum data set; Module M1.4: record the current battery status data and enter it into the database; Module M1.5: Repeat modules M1.1-M1.4 until the impedance spectrum data set contains a preset number of impedance spectrum data.

8. The lithium-ion battery multi-state impedance spectrum data rapid calculation system according to claim 6, characterized in that: The module M2 Includes the following submodules: Module M2.1: Select the core machine learning algorithm for the battery impedance spectrum calculation model; Module M2.2: Setting the input variables, output function values ​​and model parameters of the battery impedance spectrum calculation model; Module M2.3: Use programming software to build the battery impedance spectrum calculation model.

9. The lithium-ion battery multi-state impedance spectrum data rapid calculation system according to claim 8, characterized in that: The input variables of the battery impedance spectrum calculation model are composed of a battery state vector, and the battery state includes a health state, an ambient temperature state, and a charge state; the output function value includes a vector composed of impedance values ​​of the battery in different states.

10. The lithium-ion battery multi-state impedance spectrum data rapid calculation system according to claim 6, characterized in that: The module M3 includes dividing the data set in the impedance spectrum data into input variable vectors according to the required battery state; substituting the input variable vectors into the battery impedance spectrum calculation model as training samples to calculate the impedance spectrum data of the battery in other states.

Images