Method and device for estimating maximum available capacity in cycle process of lithium ion battery

Abstract

The invention belongs to the technical field of lithium ion battery health state estimation, and discloses a method and device for estimating the maximum available capacity of a lithium ion battery in the cycle process, and the method comprises the steps: obtaining state parameters of a to-be-estimated lithium ion battery in a constant current charging stage; sampling the state parameters for multiple times, calculating a first capacity increment of each sampling point according to the voltage, the sampling time and the current of each sampling point, and generating a first capacity increment curve of the lithium ion battery to be estimated; performing filtering processing on the first capacity increment curve through a filtering algorithm to obtain a filtered second capacity increment curve; extracting a second capacity increment representing the health state of the lithium ion battery to be estimated as a health factor; and inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium ion battery to be estimated. According to the method, the aging information of the lithium ion battery can be more comprehensively excavated, and the maximum available capacity of the lithium ion battery can be accurately estimated.

Description

Technical Field

[0001] The present invention belongs to the technical field of lithium-ion battery health status estimation, and in particular relates to a method and a device for estimating the maximum available capacity of a lithium-ion battery during a cycle. Background Art

[0002] Lithium-ion batteries have the advantages of high energy density, low self-discharge rate and long cycle life, and are widely used in portable electronic devices, electric vehicles and energy storage systems. However, during a long service life, the monomers in the battery system will experience inconsistent attenuation, and some monomers may be overcharged or over-discharged, which is not conducive to the safety of the battery system. The macro factors affecting battery attenuation include the charging and discharging mechanism, ambient temperature and the connection method between batteries; the micro factors are mainly the failure of the material system inside the battery. In addition, the aging mechanism inside the battery involves the coupling of multi-dimensional physical fields. The health state of the battery cannot be directly measured by sensors, and the health state of the battery during the aging process needs to be accurately estimated to ensure the safe and stable operation of the battery system during the service life. Generally, the battery health state is defined as the ratio of the maximum available capacity in the current cycle to the maximum available capacity of a new battery. Therefore, the maximum available capacity of the battery can be used to characterize the battery health state.

[0003] Most of the current battery health status estimation technologies estimate the maximum available capacity based on local aging features extracted during the charge and discharge process. However, studies have shown that even for the same model of battery, the potential aging patterns at different charging stages are different. Therefore, extracting only local aging features cannot effectively characterize the battery aging information, has certain limitations, and cannot accurately estimate the maximum available capacity of the battery. Summary of the invention

[0004] The purpose of the present invention is to more comprehensively mine the aging information of lithium-ion batteries, improve the accuracy of the maximum available capacity estimation model, and then accurately estimate the maximum available capacity of lithium-ion batteries.

[0005] In a first aspect, an embodiment of the present invention provides a method for estimating the maximum available capacity of a lithium-ion battery during a cycle, the method comprising:

[0006] Acquire state parameters of the lithium-ion battery to be estimated in a constant current charging stage, wherein the state parameters include current and voltage;

[0007] Sampling the state parameter multiple times, calculating a first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point, and generating a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment, wherein the first capacity increment curve is a curve of the first capacity increment changing with voltage;

[0008] Filtering the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve;

[0009] Extracting a second capacity increment capable of characterizing the health state of the lithium-ion battery to be estimated from the second capacity increment curve, and using the second capacity increment as a health factor;

[0010] Inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated;

[0011] Wherein, when training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are taken as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is taken as output; the health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health state of the sample lithium-ion battery, and the capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery in the constant current charging stage during the aging and attenuation process.

[0012] Optionally, the sampling the state parameter multiple times and calculating the first capacity increment of the sampling point according to the voltage, sampling time and current of each sampling point includes:

[0013] Sampling the voltage included in the state parameter to obtain a plurality of sampling points;

[0014] For each sampling point, determining a first sampling time and a first sampling voltage corresponding to the sampling point, and a second sampling time and a second sampling voltage corresponding to a previous sampling point adjacent to the sampling point;

[0015] The capacity increment of the sampling point is calculated as the first capacity increment through the constant current, the first sampling time, the first sampling voltage, the second sampling time and the second sampling voltage.

[0016] Optionally, filtering the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve includes:

[0017] Filtering the first capacity increment curve by a moving mean filtering algorithm, a local weighted regression filtering algorithm, and a median filtering algorithm, respectively, to obtain a capacity increment curve processed by a moving mean filtering algorithm, a capacity increment curve processed by a local weighted regression filtering algorithm, and a capacity increment curve processed by a median filtering algorithm;

[0018] respectively calculating a first error between the capacity increment curve processed by the moving mean filter algorithm and the first capacity increment curve, a second error between the capacity increment curve processed by the local weighted regression filter algorithm and the first capacity increment curve, and a third error between the capacity increment curve processed by the median filter algorithm and the first capacity increment curve;

[0019] Using the filtering algorithm corresponding to the minimum error among the first error, the second error and the third error as the target filtering algorithm;

[0020] The capacity increment curve processed by the target filtering algorithm is determined as a filtered second capacity increment curve.

[0021] Optionally, extracting, from the second capacity increment curve, a second capacity increment capable of characterizing the health state of the lithium-ion battery to be estimated includes:

[0022] In the second capacity increment curve, determining a second capacity increment within the target voltage interval that can characterize the health state of the lithium-ion battery to be estimated;

[0023] Among them, the lower limit value of the target voltage interval is the voltage at which the second capacity increment begins to change, and the upper limit value of the target voltage interval is the cut-off voltage of the charging process of the lithium-ion battery to be estimated; the second capacity increment that can characterize the health status of the lithium-ion battery to be estimated is: the second capacity increment extracted from the second capacity increment curve at a preset voltage interval.

[0024] Optionally, the training process of the maximum available capacity estimation model includes:

[0025] Collecting sample state parameters of multiple sample lithium-ion batteries during aging and decay in a constant current charging stage through sensors, wherein the sample state parameters include voltage and current;

[0026] For each sample lithium-ion battery, the sample state parameter is sampled multiple times, the capacity increment of the sampling point is calculated by the voltage, sampling time and current of the sampling point, and a capacity increment curve of each sample lithium-ion battery is generated based on the capacity increment, wherein the capacity increment curve of each sample lithium-ion battery is a curve of capacity increment changing with voltage;

[0027] Filtering the capacity increment curve of each sample lithium-ion battery to obtain a capacity increment of the sample lithium-ion battery within a preset voltage range that can characterize the health status of the sample lithium-ion battery as a health factor of the sample lithium-ion battery;

[0028] Inputting the health factor of each sample lithium-ion battery into the maximum available capacity estimation model to be trained, outputting the maximum available capacity of each sample lithium-ion battery, and calculating the loss function value of the maximum available capacity estimation model to be trained;

[0029] When the loss function value is less than a preset function value, a trained maximum available capacity estimation model is determined.

[0030] Optionally, the loss function of the maximum available capacity estimation model includes a mean square error of model parameters and a sum of squares of model parameters, and the loss function value is obtained by weighted summing the mean square error and the sum of squares.

[0031] In a second aspect, an embodiment of the present invention provides a device for estimating the maximum available capacity of a lithium-ion battery during a cycle, the device comprising:

[0032] A state parameter acquisition module, used to acquire the state parameters of the lithium-ion battery to be estimated in the constant current charging stage, wherein the state parameters include current and voltage;

[0033] a first capacity increment curve generating module, configured to perform multiple sampling of the state parameter, calculate a first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point, and generate a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment, wherein the first capacity increment curve is a curve of the first capacity increment changing with voltage;

[0034] A second capacity increment curve generating module, configured to filter the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve;

[0035] a health factor extraction module, configured to extract, from the second capacity increment curve, a second capacity increment capable of characterizing the health state of the lithium-ion battery to be estimated, and use the second capacity increment as a health factor;

[0036] A maximum available capacity estimation module, used for inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated;

[0037] Wherein, when training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are taken as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is taken as output; the health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health state of the sample lithium-ion battery, and the capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery in the constant current charging stage during the aging and attenuation process.

[0038] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0039] at least one processor;

[0040] a memory for storing the at least one processor-executable instruction;

[0041] The at least one processor is configured to execute the instructions to implement the method described in the first aspect.

[0042] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method described in the first aspect.

[0043] In a fifth aspect, an embodiment of the present invention provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0044] The technical solution provided by the embodiment of the present invention can pre-collect the state parameters of multiple sample lithium-ion batteries during the aging and attenuation process and the constant current charging stage, and calculate the capacity increment curve of the sample lithium-ion battery based on the collected state parameters, and then filter the capacity increment curve through a filtering algorithm, extract the health factor from the filtered capacity increment curve, and use the health factors of multiple sample lithium-ion batteries as input, and use the maximum available capacity of multiple sample lithium-ion batteries during the discharge process as output to train the maximum available capacity estimation model, so that the maximum available capacity estimation model pre-learns the mapping relationship between the health factor and the maximum available capacity. When it is necessary to estimate the maximum available capacity of the lithium-ion battery to be estimated online, the health factor of the lithium-ion battery to be estimated can be extracted, and the health factor can be input into the maximum available capacity estimation model, so that the maximum available capacity of the lithium-ion battery can be accurately estimated.

[0045] It can be seen that the technical solution of the present invention realizes the estimation of the maximum available capacity of lithium-ion batteries during the cycle process. This method overcomes the limitations of lithium-ion battery aging feature extraction technology, can more comprehensively mine the aging information of lithium-ion batteries, improve the accuracy of the maximum available capacity estimation model, and then accurately estimate the maximum available capacity of lithium-ion batteries. In addition, the aging information of lithium-ion batteries at different stages can be comprehensively mined in the case of unknown aging patterns, and the generalization ability of the maximum available capacity estimation model can be improved, so that it can better estimate the maximum available capacity of lithium-ion batteries with different aging paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flow chart of the overall technical solution of an embodiment of the present invention;

[0047] Figure 2 A flowchart of a method for estimating the maximum available capacity of a lithium-ion battery during a cycle according to an embodiment of the present invention;

[0048] Figure 3 for Figure 2 A flowchart of an implementation of S230;

[0049] Figure 4 A flowchart of a training process of a maximum available capacity estimation model provided by an embodiment of the present invention;

[0050] Figure 5 A schematic diagram of the structure of a device for estimating the maximum available capacity during the cycle of a lithium-ion battery according to an embodiment of the present invention;

[0051] Figure 6 The figure is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The present invention will be described in detail below through examples.

[0053] Lithium-ion batteries have the advantages of high energy density, low self-discharge rate and long cycle life, and are widely used in portable electronic devices, electric vehicles and energy storage systems. However, during a long service life, the monomers in the battery system will experience inconsistent attenuation, and some monomers may be overcharged or over-discharged, which is not conducive to the safety of the battery system. The macro factors affecting battery attenuation include the charging and discharging mechanism, ambient temperature and the connection method between batteries; the micro factors are mainly the failure of the material system inside the battery. In addition, the aging mechanism inside the battery involves the coupling of multi-dimensional physical fields. The health state of the battery cannot be directly measured by sensors, and the health state of the battery during the aging process needs to be accurately estimated to ensure the safe and stable operation of the battery system during the service life. Generally, the battery health state is defined as the ratio of the maximum available capacity in the current cycle to the maximum available capacity of a new battery. Therefore, the maximum available capacity of the battery can be used to characterize the battery health state.

[0054] Most of the current battery health status estimation technologies estimate the maximum available capacity based on local aging features extracted during the charge and discharge process. However, studies have shown that even for the same model of battery, the potential aging patterns at different charging stages are different. Therefore, extracting only local aging features cannot effectively characterize the battery aging information, has certain limitations, and cannot accurately estimate the maximum available capacity of the battery.

[0055] To this end, the present invention proposes a maximum available capacity estimation method for a lithium-ion battery during a cycle process, wherein the maximum available capacity estimation method extracts global aging characteristics of the lithium-ion battery during a constant current charging stage.

[0056] like Figure 1 The flowchart of the overall technical solution of the embodiment of the present invention is shown in FIG. The overall technical solution of the embodiment of the present invention may include the following steps, namely:

[0057] S110, obtaining state parameters of the lithium-ion battery in a constant current charging stage during the aging and decay process.

[0058] The state parameters include voltage and current.

[0059] Specifically, the aging and attenuation process of a batch of new lithium-ion batteries can be collected through sensors, and the state parameters such as the charging terminal voltage and current in the constant current charging stage can be used for the training process of the maximum available capacity estimation model. Among them, the aging and attenuation process of new lithium-ion batteries refers to the process in which the maximum available capacity of new lithium-ion batteries decays to 80% of their rated capacity. It can be understood that the new lithium-ion batteries are sample lithium-ion batteries.

[0060] S120, sampling the state parameters to obtain a plurality of discrete sampling points, calculating the capacity increments of the plurality of discrete sampling points respectively according to a capacity increment calculation formula under discrete states, and obtaining a capacity increment curve through the capacity increments of the sampling points.

[0061] The capacity increment curve is a curve showing the change of capacity increment with voltage.

[0062] Specifically, after collecting the state parameters through S110, the state parameters can be sampled. In the actual process, the sampling can be performed every preset time, wherein the preset time can be determined according to the actual situation and will not be repeated here. After obtaining multiple discrete sampling points, the capacity increments of the multiple discrete sampling points can be calculated according to the capacity increment calculation formula under the discrete state.

[0063] The capacity increment calculation formula is as follows:

[0064] IC k =I·(t k _t k-1 ) / (U k _U k-1 )

[0065] IC k is the capacity increment of the kth sampling point, k is the serial number of any sampling point, k is a positive integer, I is the constant current, t k is the sampling time of the kth sampling point, t k-1 is the sampling time of the k-1th sampling point, U k is the voltage at the kth sampling point, U k-1 is the voltage at the k-1th sampling point.

[0066] After the capacity increments of multiple sampling points are obtained, a capacity increment curve may be generated, where the capacity increment curve is a curve showing the change of the capacity increment with the voltage.

[0067] S130, filtering the capacity increment curve using a filtering algorithm to obtain a filtered capacity increment curve, and extracting a capacity increment within a certain terminal voltage range that can characterize the health status of the lithium-ion battery from the filtered capacity increment curve as a health factor.

[0068] Specifically, after the capacity increment curve is obtained, the capacity increment curve is filtered based on a filtering algorithm, wherein there may be multiple filtering algorithms, for example, the filtering algorithm may be a moving mean filtering algorithm, a local weighted regression filtering algorithm, and a median filtering algorithm.

[0069] In this solution, which specific filtering algorithm is used to filter the capacity increment curve can be determined by the average relative error between the capacity increment after processing by different filtering algorithms and the initial capacity increment before processing by the filtering algorithms. In other words, the initial capacity increment curve can be filtered by three different filtering algorithms to obtain the capacity increments after processing by three different filtering algorithms, and the average relative errors between the capacity increments after processing by the three different filtering algorithms and the initial capacity increment are calculated respectively. The filtering algorithm with the smaller average relative error between the capacity increment after processing by the filtering algorithm and the initial capacity increment is selected to filter the capacity increment curve.

[0070] Among them, the specific steps of filtering the capacity increment curve by the moving mean filtering algorithm can be as follows: first, the size of the moving window is selected according to actual needs. Assuming that the size of the moving window is W, which is an odd number, the algorithm uses the arithmetic mean of the capacity increments of all sampling points in the calculation window to replace the capacity increment of the central sampling point of the window. After filtering the capacity increment of the central sampling point in a window, the window will continue to move and filter the capacity increment of the central sampling point in the next window until the end of the capacity increment sequence is reached.

[0071] The core of the local weighted regression filtering algorithm is to consider the weights of the capacity increments of other sampling points near each sampling point, and perform weighted regression on the capacity increments of other sampling points near it to obtain the capacity increment of the sampling point.

[0072] The median filter algorithm is similar to the moving mean filter algorithm, except that the capacity increments of the sampling points within the moving window are sorted, and the capacity increments in the middle of the order are used to replace the capacity increments of the sampling points at the center of the window.

[0073] After filtering the capacity increment curve with a filtering algorithm to obtain the filtered capacity increment curve, the capacity increment within a certain terminal voltage range can be extracted from the filtered capacity increment curve as a health factor. Among them, during the charging cycle, the lower limit of the terminal voltage range is the voltage at which the battery capacity increment begins to change, and the upper limit of the terminal voltage is the cut-off voltage of the charging process. The health factor can be the capacity increment extracted at a preset voltage interval within the target voltage range. Among them, in actual applications, the preset voltage interval can be 0.05V. Of course, the voltage interval can also be set according to actual conditions, and no specific limitation is made here.

[0074] S140, using the extracted health factor as a model input and the maximum available capacity of the lithium-ion battery during discharge as a model output, to train a maximum available capacity estimation model.

[0075] This step establishes a mapping relationship between capacity increment and maximum available capacity through machine learning modeling methods, thereby achieving accurate estimation of the maximum available capacity.

[0076] Specifically, after extracting the health factor of the lithium-ion battery, the health factor can be used as the model input of the maximum available capacity estimation model to be trained, and the maximum available capacity of the lithium-ion battery during discharge can be used as the model output to train the maximum available capacity estimation model. The loss function value of the maximum available capacity estimation model is calculated. If the loss function value is less than the preset loss function value, it means that the accuracy of the maximum available capacity output by the maximum available capacity estimation model is high. At this time, it can be determined that the trained maximum available capacity estimation model is obtained.

[0077] Among them, the structure of the maximum available capacity estimation model may include an input layer, a hidden layer and an output layer. The number of neurons in the input layer can be the same as the number of health factors input into the maximum available capacity estimation model; the number of neurons in the hidden layer can be determined according to actual conditions. For example, the number of neurons in the hidden layer is 5, and the number of neurons in the output layer can be 1, which is used to output the maximum available capacity.

[0078] In an embodiment of the present invention, the loss function of the maximum available capacity estimation model is obtained by weighted summing the mean square error of the model parameters and the sum of the squares of the model parameters, wherein the first weight factor corresponding to the mean square error of the model parameters and the second weight factor corresponding to the sum of the squares of the model parameters can be determined according to actual conditions. By changing the relative size of the first weight factor and the second weight factor, the focus of the maximum available capacity estimation model training can be changed. Since the loss function not only considers the error of the maximum available capacity estimation model, but also considers the complexity of the maximum available capacity estimation model. Therefore, to a certain extent, overfitting can be prevented and the generalization ability of the maximum available capacity estimation model is improved.

[0079] S150, when it is necessary to estimate the maximum available capacity of a lithium-ion battery online, the state parameters of the lithium-ion battery in the constant current charging stage can be obtained, and the health factor of the lithium-ion battery can be extracted according to the method of step S120 and step S130, and the health factor is input into the maximum available capacity estimation model trained in step 4, and the maximum available capacity of the lithium-ion battery is output from the maximum available capacity estimation model.

[0080] Specifically, a maximum available capacity estimation model is obtained through training from S110 to S140. When it is necessary to estimate the maximum available capacity of a lithium-ion battery online, the state parameters of the lithium-ion battery in the constant current charging stage can be obtained. The state parameters may include current and voltage. The state parameters are sampled multiple times to obtain multiple sampling points. For each sampling point, the initial capacity increment of the sampling point is calculated by the voltage, sampling time and current of the sampling point. After obtaining the initial capacity increment of multiple sampling points, an initial capacity increment curve can be generated. The initial capacity increment curve is a curve of the initial capacity increment changing with the voltage. The initial capacity increment curve is filtered by a filtering algorithm to obtain a capacity increment curve after filtering. Then, the capacity increment that can characterize the health status of the lithium-ion battery to be estimated is extracted from the capacity increment curve after filtering, and the extracted capacity increment is used as a health factor.

[0081] After obtaining the health factor of the lithium-ion battery to be estimated, the health factor can be input into the maximum available capacity estimation model obtained by training through S110 to S140, and the maximum available capacity of the lithium-ion battery to be estimated with higher accuracy can be output from the maximum available capacity estimation model.

[0082] Through the technical solution of the present invention, the estimation of the maximum available capacity of lithium-ion batteries during the cycle process is realized. The method overcomes the limitations of lithium-ion battery aging feature extraction technology, can more comprehensively mine the aging information of lithium-ion batteries, improve the accuracy of the maximum available capacity estimation model, and then accurately estimate the maximum available capacity of lithium-ion batteries. In addition, the aging information of lithium-ion batteries at different stages can be comprehensively mined in the case of unknown aging patterns, and the generalization ability of the maximum available capacity estimation model can be improved, so that it can better estimate the maximum available capacity of lithium-ion batteries with different aging paths.

[0083] After the overall technical solution of the embodiment of the present invention is described in detail, a method for estimating the maximum available capacity of a lithium-ion battery during a cycle provided by the embodiment of the present invention will be described in detail below.

[0084] like Figure 2 As shown, an embodiment of the present invention provides a method for estimating the maximum available capacity of a lithium-ion battery during a cycle, and the method may include the following steps:

[0085] S210, obtaining state parameters of the lithium-ion battery to be estimated in a constant current charging stage.

[0086] Among them, the state parameters include current and voltage.

[0087] Specifically, the state parameters such as the charging terminal voltage and current of the lithium-ion battery to be estimated in the constant current charging stage can be collected through sensors.

[0088] S220, sampling the state parameter multiple times, calculating a first capacity increment of each sampling point according to the voltage, sampling time and current of each sampling point, and generating a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment.

[0089] The first capacity increment curve is a curve showing the change of the first capacity increment with the voltage.

[0090] Specifically, after collecting the state parameters through S210, the state parameters can be sampled. In the actual process, the sampling can be performed every preset time period, wherein the preset time period can be determined according to the actual situation and will not be described in detail here. After obtaining multiple discrete sampling points, the capacity increments of multiple discrete sampling points can be calculated respectively according to the capacity increment calculation formula under the discrete state. In order to describe the solution clearly, the capacity increment can be referred to as the first capacity increment. In addition, after calculating the first capacity increment of multiple sampling points, a first capacity increment curve can be generated according to the sampling time sequence of the multiple sampling points. The first capacity increment curve is a curve of the first capacity increment changing with voltage.

[0091] As an implementation of the embodiment of the present invention, S220, sampling the state parameter multiple times, and calculating the first capacity increment of each sampling point through the voltage, sampling time and current of each sampling point, may include the following steps, namely, step a1 to step a3:

[0092] Step a1, sampling the voltage included in the state parameter to obtain a plurality of sampling points.

[0093] Step a2: for each sampling point, determine a first sampling time and a first sampling voltage corresponding to the sampling point, and a second sampling time and a second sampling voltage corresponding to a previous sampling point adjacent to the sampling point.

[0094] Step a3, calculating the capacity increment of the sampling point as the first capacity increment through the constant current, the first sampling time, the first sampling voltage, the second sampling time and the second sampling voltage.

[0095] It should be noted that the formula for calculating the first capacity increment has been explained in detail in step S120 and will not be repeated here.

[0096] S230, filtering the first capacity increment curve by using a filtering algorithm to obtain a filtered second capacity increment curve.

[0097] Specifically, after the first capacity increment curve is obtained, the first capacity increment curve is filtered based on a filtering algorithm, wherein there may be multiple filtering algorithms, for example, the filtering algorithm may be a moving mean filtering algorithm, a local weighted regression filtering algorithm, and a median filtering algorithm.

[0098] In this solution, which filtering algorithm is used to filter the first capacity increment curve can be determined by the average relative error between the capacity increment processed by different filtering algorithms and the first capacity increment. The filtering algorithm with the smaller average relative error is used as the target filtering algorithm, and the first capacity increment curve is filtered by the filtering algorithm to obtain the filtered second capacity increment curve.

[0099] In order to describe the solution clearly, the specific implementation of S230 will be described in detail in the following embodiments.

[0100] S240: extracting a second capacity increment that can characterize the health state of the lithium-ion battery to be estimated from the second capacity increment curve, and using the second capacity increment as a health factor.

[0101] As an implementation of the embodiment of the present invention, S240, extracting a second capacity increment capable of characterizing the health status of the lithium-ion battery to be estimated from the second capacity increment curve, may include the following step b1:

[0102] Step b1, in the second capacity curve, determining a second capacity increment within the target voltage interval that can characterize the health state of the lithium-ion battery to be estimated.

[0103] Among them, the lower limit value of the target voltage interval is the voltage at which the second capacity increment begins to change, and the upper limit value of the target voltage interval is the cut-off voltage of the charging process of the lithium-ion battery to be estimated; the second capacity increment that can characterize the health status of the lithium-ion battery to be estimated is: the second capacity increment extracted from the second capacity increment curve at a preset voltage interval.

[0104] In practical applications, the preset voltage interval may be 0.05V. Of course, the voltage interval may also be set according to actual conditions, and no specific limitation is made here.

[0105] Specifically, after obtaining the second capacity increment curve, the second capacity increment that can characterize the health status of the lithium-ion battery to be estimated within the target voltage range can be extracted from the second capacity increment curve as a health factor. In the charging cycle, the lower limit of the terminal voltage range is the voltage at which the second capacity increment begins to change, and the upper limit of the terminal voltage is the cut-off voltage of the charging process of the lithium-ion battery to be estimated. The health factor can be the second capacity increment extracted at a preset voltage interval within the target voltage range.

[0106] S250, inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated.

[0107] When training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are used as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is used as output. The health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health status of the sample lithium-ion battery. The capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery during the aging and attenuation process in the constant current charging stage.

[0108] Specifically, when it is necessary to estimate the maximum available capacity of the lithium-ion battery to be estimated online, the health factor obtained in S240 is input into the pre-trained maximum available capacity estimation model. Since the maximum available capacity estimation model has pre-learned the mapping relationship between the health factor and the maximum available capacity, the maximum available capacity estimation model can accurately output the maximum available capacity of the lithium-ion battery to be estimated.

[0109] The technical solution provided by the embodiment of the present invention can pre-collect the state parameters of multiple sample lithium-ion batteries during the aging and attenuation process and the constant current charging stage, and calculate the capacity increment curve of the sample lithium-ion battery based on the collected state parameters, and then filter the capacity increment curve through a filtering algorithm, extract the health factor from the filtered capacity increment curve, and use the health factors of multiple sample lithium-ion batteries as input, and use the maximum available capacity of multiple sample lithium-ion batteries during the discharge process as output to train the maximum available capacity estimation model, so that the maximum available capacity estimation model pre-learns the mapping relationship between the health factor and the maximum available capacity. When it is necessary to estimate the maximum available capacity of the lithium-ion battery to be estimated online, the health factor of the lithium-ion battery to be estimated can be extracted, and the health factor can be input into the maximum available capacity estimation model, so that the maximum available capacity of the lithium-ion battery can be accurately estimated.

[0110] It can be seen that the technical solution of the present invention realizes the estimation of the maximum available capacity of lithium-ion batteries during the cycle process. This method overcomes the limitations of lithium-ion battery aging feature extraction technology, can more comprehensively mine the aging information of lithium-ion batteries, improve the accuracy of the maximum available capacity estimation model, and then accurately estimate the maximum available capacity of lithium-ion batteries. In addition, the aging information of lithium-ion batteries at different stages can be comprehensively mined in the case of unknown aging patterns, and the generalization ability of the maximum available capacity estimation model can be improved, so that it can better estimate the maximum available capacity of lithium-ion batteries with different aging paths.

[0111] As an implementation method of the embodiment of the present invention, S230, the first capacity increment curve is filtered by a filtering algorithm to obtain a filtered second capacity increment curve, such as Figure 3 As shown, the following steps may be included:

[0112] S231, filtering the first capacity increment curve by a moving mean filtering algorithm, a local weighted regression filtering algorithm and a median filtering algorithm respectively, to obtain a capacity increment curve processed by a moving mean filtering algorithm, a capacity increment curve processed by a local weighted regression filtering algorithm, and a capacity increment curve processed by a median filtering algorithm.

[0113] Specifically, in Figure 1 In step S130 of the illustrated embodiment, how the moving mean filtering algorithm, the local weighted regression filtering algorithm and the median filtering algorithm perform filtering processing has been described in detail, and will not be repeated here again.

[0114] S232, respectively calculate the first error between the capacity increment curve processed by the moving mean filter algorithm and the first capacity increment curve, the second error between the capacity increment curve processed by the local weighted regression filter algorithm and the first capacity increment curve, and the third error between the capacity increment curve processed by the median filter algorithm and the first capacity increment curve.

[0115] S233: Use the filtering algorithm corresponding to the minimum error among the first error, the second error and the third error as the target filtering algorithm.

[0116] S234, determining the capacity increment curve processed by the target filtering algorithm as a filtered second capacity increment curve.

[0117] Specifically, in order to ensure that the capacity increment after filtering retains the information in the first capacity increment to the greatest extent, the first error between the capacity increment curve after the moving mean filtering algorithm and the first capacity increment curve, the second error between the capacity increment curve after the local weighted regression filtering algorithm and the first capacity increment curve, and the third error between the capacity increment curve after the median filtering algorithm and the first capacity increment curve can be calculated respectively. If the error between the capacity increment curve after filtering and the first capacity increment curve is small, it means that the capacity increment in the capacity increment curve after filtering can retain the information in the first capacity increment to the greatest extent. Therefore, the filtering algorithm corresponding to the minimum error can be used as the target filtering algorithm, and the capacity increment curve processed by the target filtering algorithm is determined as the second capacity increment curve after filtering.

[0118] Among them, the first error, the second error and the third error can all be Figure 1 The average relative error mentioned in the examples.

[0119] It can be seen that the second capacity increment in the filtered second capacity increment curve obtained by the embodiment can retain the information in the first capacity increment to the greatest extent, which is helpful to extract a more accurate health factor and further accurately estimate the maximum available capacity.

[0120] As an implementation method of the embodiment of the present invention, Figure 4 As shown, the training process of the maximum available capacity estimation model may include the following steps:

[0121] S410, collecting, through sensors, sample state parameters of multiple sample lithium-ion batteries during an aging and decaying process at a constant current charging stage.

[0122] The sample state parameters include voltage and current.

[0123] S420, for each sample lithium-ion battery, sampling the sample state parameter multiple times, calculating the capacity increment of the sampling point through the voltage, sampling time and current of the sampling point, and generating a capacity increment curve of each sample lithium-ion battery based on the capacity increment.

[0124] The capacity increment curve of each sample lithium-ion battery is a curve showing the change of capacity increment with voltage.

[0125] S430, filtering the capacity increment curve of each sample lithium-ion battery to obtain a capacity increment of the sample lithium-ion battery within a preset voltage range that can characterize the health status of the sample lithium-ion battery as a health factor of the sample lithium-ion battery.

[0126] S440, inputting the health factor of each sample lithium-ion battery into the maximum available capacity estimation model to be trained, outputting the maximum available capacity of each sample lithium-ion battery, and calculating the loss function value of the maximum available capacity estimation model to be trained.

[0127] S450, when the loss function value is less than the preset function value, determine the trained maximum available capacity estimation model.

[0128] As an implementation manner of an embodiment of the present invention, the loss function of the maximum available capacity estimation model may include a mean square error of model parameters and a sum of squares of model parameters, and the loss function value is obtained by weighted summing the mean square error and the sum of squares.

[0129] Specifically, before training the maximum available capacity estimation model, it is necessary to collect a batch of new lithium-ion batteries during the aging and attenuation process, and state parameters such as the charging terminal voltage and current in the constant current charging stage through sensors. The new lithium-ion batteries are sample lithium-ion batteries, and the collected state parameters are training data. By executing steps S420 and S430 on the training data, the health factor of the sample lithium-ion battery can be extracted. By taking the health factor of each sample lithium-ion battery as the model input and the maximum available capacity of each sample lithium-ion battery as the model output, the maximum available capacity estimation model is trained. When the loss function value is less than the preset loss function value, the trained maximum available capacity estimation model is obtained.

[0130] The lower limit of the preset voltage range is the voltage at which the capacity increment of the sample lithium-ion battery begins to change, and the upper limit of the preset voltage range is the cut-off voltage of the charging process.

[0131] The technical solution of the present invention can fully mine the aging information of lithium-ion batteries at different stages in the case of unknown aging modes, thereby improving the generalization ability of the maximum available capacity estimation model.

[0132] In a second aspect, an embodiment of the present invention provides a device 50 for estimating the maximum available capacity of a lithium-ion battery during a cycle, such as Figure 5 As shown, the device comprises:

[0133] A state parameter acquisition module 510 is used to acquire the state parameters of the lithium-ion battery to be estimated in the constant current charging stage, wherein the state parameters include current and voltage;

[0134] A first capacity increment curve generating module 520 is used to perform multiple sampling of the state parameter, calculate the first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point, and generate a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment, wherein the first capacity increment curve is a curve of the first capacity increment changing with the voltage;

[0135] A second capacity increment curve generating module 530 is used to filter the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve;

[0136] A health factor extraction module 540 is used to extract a second capacity increment that can characterize the health state of the lithium-ion battery to be estimated from the second capacity increment curve, and use the second capacity increment as a health factor;

[0137] A maximum available capacity estimation module 550 is used to input the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated;

[0138] Wherein, when training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are taken as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is taken as output; the health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health state of the sample lithium-ion battery, and the capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery in the constant current charging stage during the aging and attenuation process.

[0139] Optionally, the first capacity increment curve generating module is specifically configured to:

[0140] Sampling the voltage included in the state parameter to obtain a plurality of sampling points;

[0141] For each sampling point, determining a first sampling time and a first sampling voltage corresponding to the sampling point, and a second sampling time and a second sampling voltage corresponding to a previous sampling point adjacent to the sampling point;

[0142] The capacity increment of the sampling point is calculated as the first capacity increment through the constant current, the first sampling time, the first sampling voltage, the second sampling time and the second sampling voltage.

[0143] Optionally, the second capacity increment curve generating module is specifically configured to:

[0144] Filtering the first capacity increment curve by a moving mean filtering algorithm, a local weighted regression filtering algorithm, and a median filtering algorithm, respectively, to obtain a capacity increment curve processed by a moving mean filtering algorithm, a capacity increment curve processed by a local weighted regression filtering algorithm, and a capacity increment curve processed by a median filtering algorithm;

[0145] respectively calculating a first error between the capacity increment curve processed by the moving mean filter algorithm and the first capacity increment curve, a second error between the capacity increment curve processed by the local weighted regression filter algorithm and the first capacity increment curve, and a third error between the capacity increment curve processed by the median filter algorithm and the first capacity increment curve;

[0146] Using the filtering algorithm corresponding to the minimum error among the first error, the second error and the third error as the target filtering algorithm;

[0147] The capacity increment curve processed by the target filtering algorithm is determined as a filtered second capacity increment curve.

[0148] Optionally, the health factor extraction module is specifically used to:

[0149] In the second capacity curve, determining a second capacity increment within the target voltage interval that can characterize the health state of the lithium ions;

[0150] Among them, the lower limit value of the target voltage interval is the voltage at which the second capacity increment begins to change, and the upper limit value of the target voltage interval is the cut-off voltage of the charging process of the lithium-ion battery to be estimated; the second capacity increment that can characterize the health status of the lithium-ion battery to be estimated is: the second capacity increment extracted from the second capacity increment curve at a preset voltage interval.

[0151] Optionally, the device further includes a maximum available capacity estimation model training module, which is specifically used to:

[0152] Collecting sample state parameters of multiple sample lithium-ion batteries during aging and decay in a constant current charging stage through sensors, wherein the sample state parameters include voltage and current;

[0153] For each sample lithium-ion battery, the sample state parameter is sampled multiple times, the capacity increment of the sampling point is calculated by the voltage, sampling time and current of the sampling point, and a capacity increment curve of each sample lithium-ion battery is generated based on the capacity increment, wherein the capacity increment curve of each sample lithium-ion battery is a curve of capacity increment changing with voltage;

[0154] Performing filtering processing on the capacity increment curve of each sample lithium-ion battery to obtain the capacity increment of the sample lithium-ion battery within a preset voltage range as the health factor of the sample lithium-ion battery;

[0155] Inputting the health factor of each sample lithium-ion battery into the maximum available capacity estimation model to be trained, outputting the maximum available capacity of each sample lithium-ion battery, and calculating the loss function value of the maximum available capacity estimation model to be trained;

[0156] When the loss function value is less than a preset function value, a trained maximum available capacity estimation model is determined.

[0157] Optionally, the loss function of the maximum available capacity estimation model includes a mean square error of model parameters and a sum of squares of model parameters, and the loss function value is obtained by weighted summing the mean square error and the sum of squares.

[0158] In a third aspect, an embodiment of the present invention provides an electronic device 600, such as Figure 6 As shown, including:

[0159] at least one processor 601;

[0160] a memory 602 for storing the at least one processor executable instruction;

[0161] The at least one processor is configured to execute the instructions to implement the method described in the first aspect.

[0162] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method described in the first aspect.

[0163] In a fifth aspect, an embodiment of the present invention provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0164] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention without departing from the principles and intent of the present invention.

Claims

1. A method for estimating the maximum available capacity of a lithium-ion battery during cycling, characterized in that: The method comprises: Acquire state parameters of the lithium-ion battery to be estimated in a constant current charging stage, wherein the state parameters include current and voltage; Sampling the state parameter multiple times, calculating a first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point, and generating a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment, wherein the first capacity increment curve is a curve of the first capacity increment changing with voltage; Filtering the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve; Extracting a second capacity increment capable of characterizing the health state of the lithium-ion battery to be estimated from the second capacity increment curve, and using the second capacity increment as a health factor; Inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated; Wherein, when training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are taken as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is taken as output; the health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health state of the sample lithium-ion battery, and the capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery in the constant current charging stage during the aging and attenuation process.

2. The method according to claim 1, characterized in that The sampling of the state parameter for multiple times and calculating the first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point comprises: Sampling the voltage included in the state parameter to obtain a plurality of sampling points; For each sampling point, determining a first sampling time and a first sampling voltage corresponding to the sampling point, and a second sampling time and a second sampling voltage corresponding to a previous sampling point adjacent to the sampling point; The capacity increment of the sampling point is calculated as the first capacity increment through the constant current, the first sampling time, the first sampling voltage, the second sampling time and the second sampling voltage.

3. The method according to claim 1, characterized in that The filtering the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve includes: Filtering the first capacity increment curve by a moving mean filtering algorithm, a local weighted regression filtering algorithm, and a median filtering algorithm, respectively, to obtain a capacity increment curve processed by a moving mean filtering algorithm, a capacity increment curve processed by a local weighted regression filtering algorithm, and a capacity increment curve processed by a median filtering algorithm; respectively calculating a first error between the capacity increment curve processed by the moving mean filter algorithm and the first capacity increment curve, a second error between the capacity increment curve processed by the local weighted regression filter algorithm and the first capacity increment curve, and a third error between the capacity increment curve processed by the median filter algorithm and the first capacity increment curve; Using the filtering algorithm corresponding to the minimum error among the first error, the second error and the third error as the target filtering algorithm; The capacity increment curve processed by the target filtering algorithm is determined as a filtered second capacity increment curve.

4. The method according to any one of claims 1 to 3, characterized in that: The step of extracting, from the second capacity increment curve, a second capacity increment capable of characterizing the health status of the lithium-ion battery to be estimated comprises: In the second capacity increment curve, determining a second capacity increment within the target voltage interval that can characterize the health state of the lithium-ion battery to be estimated; Among them, the lower limit value of the target voltage interval is the voltage at which the second capacity increment begins to change, and the upper limit value of the target voltage interval is the cut-off voltage of the charging process of the lithium-ion battery to be estimated; the second capacity increment that can characterize the health status of the lithium-ion battery to be estimated is: the second capacity increment extracted from the second capacity increment curve at a preset voltage interval.

5. The method according to any one of claims 1 to 3, characterized in that: The training process of the maximum available capacity estimation model includes: Collecting sample state parameters of multiple sample lithium-ion batteries during aging and decay in a constant current charging stage through sensors, wherein the sample state parameters include voltage and current; For each sample lithium-ion battery, the sample state parameter is sampled multiple times, the capacity increment of the sampling point is calculated by the voltage, sampling time and current of the sampling point, and a capacity increment curve of each sample lithium-ion battery is generated based on the capacity increment, wherein the capacity increment curve of each sample lithium-ion battery is a curve of capacity increment changing with voltage; Filtering the capacity increment curve of each sample lithium-ion battery to obtain a capacity increment of the sample lithium-ion battery within a preset voltage range that can characterize the health status of the sample lithium-ion battery as a health factor of the sample lithium-ion battery; Inputting the health factor of each sample lithium-ion battery into the maximum available capacity estimation model to be trained, outputting the maximum available capacity of each sample lithium-ion battery, and calculating the loss function value of the maximum available capacity estimation model to be trained; When the loss function value is less than a preset function value, a trained maximum available capacity estimation model is determined.

6. The method according to claim 5, characterized in that The loss function of the maximum available capacity estimation model includes a mean square error of model parameters and a sum of squares of model parameters, and the loss function value is obtained by weighted summing the mean square error and the sum of squares.

7. A device for estimating the maximum available capacity of a lithium-ion battery during cycling, characterized in that: The device comprises: A state parameter acquisition module, used to acquire the state parameters of the lithium-ion battery to be estimated in the constant current charging stage, wherein the state parameters include current and voltage; a first capacity increment curve generating module, configured to perform multiple sampling of the state parameter, calculate a first capacity increment of the sampling point through the voltage, sampling time and current of each sampling point, and generate a first capacity increment curve of the lithium-ion battery to be estimated based on the first capacity increment, wherein the first capacity increment curve is a curve of the first capacity increment changing with voltage; A second capacity increment curve generating module, configured to filter the first capacity increment curve by a filtering algorithm to obtain a filtered second capacity increment curve; a health factor extraction module, configured to extract, from the second capacity increment curve, a second capacity increment capable of characterizing the health state of the lithium-ion battery to be estimated, and use the second capacity increment as a health factor; A maximum available capacity estimation module, used for inputting the health factor into a pre-trained maximum available capacity estimation model to obtain the maximum available capacity of the lithium-ion battery to be estimated; Wherein, when training the maximum available capacity estimation model, the health factors of multiple sample lithium-ion batteries are taken as input, and the maximum available capacity of multiple sample lithium-ion batteries during discharge is taken as output; the health factor of each sample lithium-ion battery is a capacity increment extracted from the capacity increment curve after filtering, which can characterize the health state of the sample lithium-ion battery, and the capacity increment curve of each sample lithium-ion battery is calculated based on the state parameters of the sample lithium-ion battery in the constant current charging stage during the aging and attenuation process.

8. An electronic device, characterized in that: include: at least one processor; a memory for storing the at least one processor-executable instruction; The at least one processor is configured to execute the instructions to implement the method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises a computer program, which implements the method according to any one of claims 1 to 6 when being executed by a processor.

Images