A lithium-ion battery capacity detection method and system
By segmenting the charging time window during the lithium-ion battery charging process and combining it with temperature suitability to filter data, a neural network is trained, solving the problems of long time consumption and information repetition in traditional detection methods, and achieving efficient and accurate lithium-ion battery capacity detection.
Patent Information
- Application Number
- CN202411207001.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-30
AI Technical Summary
Traditional lithium-ion battery capacity testing methods are time-consuming, which is not conducive to large-batch testing, and the information repetition problem during deep neural network training affects the accuracy of testing.
By acquiring battery parameters during the lithium-ion battery charging process, the charging time window is divided using the change in internal resistance. The charging time window is then selected by combining temperature suitability and overall suitability, and a neural network is trained to improve detection accuracy.
It improves the accuracy and efficiency of lithium-ion battery capacity testing, avoids data duplication and bias, and ensures data representativeness and diversity.
Smart Images

Figure CN118707374B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of battery detection, and in particular to a lithium-ion battery capacity detection method and system. Background Art
[0002] Lithium-ion battery is a highly efficient rechargeable battery. This battery uses lithium ions as the main charge carrier and stores and releases electrical energy by moving lithium ions between the positive and negative electrodes of the battery. Lithium-ion batteries have a long cycle life and can withstand hundreds to thousands of charge and discharge cycles. Therefore, they can be used in cascades. After the first use cycle of lithium-ion batteries, the remaining capacity of the battery needs to be tested to determine whether it meets the conditions for cascade utilization.
[0003] Traditional lithium-ion battery capacity detection methods usually rely on battery charge and discharge cycle testing, which is time-consuming and not conducive to large-scale batch detection. In order to reduce the lithium-ion battery capacity detection time and achieve rapid detection, a deep neural network can be used to predict the complete charging curve to obtain the capacity of the lithium battery. When training a deep neural network, if a charging curve segment is randomly selected, a lot of repeated information may be encountered, which will reduce the accuracy of the prediction of the charging curve. This information duplication problem will affect the accurate detection of the lithium-ion battery capacity. Summary of the invention
[0004] In view of the above, it is necessary to provide a lithium-ion battery capacity detection method and system to solve the above problems.
[0005] The first aspect of the present application provides a lithium ion battery capacity detection method, the method comprising:
[0006] Obtaining battery parameters of a set number of lithium-ion batteries at each moment during the charging process; the battery parameters at least include battery internal resistance and battery temperature;
[0007] The total charging time of the lithium-ion battery is divided according to the change of the internal resistance of the lithium-ion battery during the charging process to obtain several charging time windows;
[0008] According to the distribution of the local range of the battery temperature in each charging time window and the change characteristics, the temperature suitability of each charging time window is obtained; according to the similarity and disorder degree of the battery temperature distribution in each charging time window and other charging time windows, combined with the temperature suitability, the comprehensive suitability of each charging time window is obtained;
[0009] Based on the comprehensive suitability, the charging time window of each lithium-ion battery is screened; according to the battery parameters of the screened charging time window, a neural network is trained to obtain the capacity of the lithium-ion battery to be tested.
[0010] The battery parameters also include: charging current and battery voltage.
[0011] The obtaining of a plurality of charging time windows is specifically as follows:
[0012] The length of the previous charging time window of each charging time window is used as the original window length of each charging time window; the adjustment factor of each charging time window is obtained according to the internal resistance change rate in the original window corresponding to each charging time window and the discrete degree of the internal resistance; wherein the preset length is used as the original window length of the first charging time window;
[0013] The length of the i-th charging time window is recorded as , its formula form is: ;in, represents the length of the i-1th charging time window; represents the adjustment factor of the i-th charging time window; Indicates rounding up.
[0014] The steps for obtaining the internal resistance change rate are as follows:
[0015] For each charging time window corresponding to the original window, the internal resistance difference between the start time and the end time is calculated; the internal resistance change rate within the original window corresponding to each charging time window is obtained by taking the ratio of the internal resistance difference to the internal resistance at the start time.
[0016] The step of obtaining the temperature suitability of each charging time window is as follows:
[0017] For each charging time window, a temperature neighborhood at each moment is preset with the temperature data at each moment as the center; the degree of dispersion of all temperature data in the temperature neighborhood at each moment is calculated; and the temperature suitability of each charging time window is obtained based on the distribution of the degree of dispersion at all moments in each charging time window and the average level of temperature data in each charging time window.
[0018] The process of obtaining the comprehensive suitability of each charging time window includes:
[0019] Calculate the distance measure between each charging time window and the temperature data in other charging time windows, and calculate the mean of all the distance measures in each charging time window; merge the degree of confusion of the temperature data in each charging time window, the mean and the temperature suitability to obtain the comprehensive suitability of each charging time window.
[0020] The specific steps of training the neural network are as follows:
[0021] For each lithium-ion battery, the battery internal resistance, battery temperature, charging current, and battery voltage at all moments in each charging time window after screening are respectively formed into column vectors of the feature matrix corresponding to the charging time window;
[0022] The column vector composed of the battery capacity of each lithium-ion battery at all times is used as the label of all feature matrices of each lithium-ion battery;
[0023] All feature matrices and corresponding labels are used as input of the deep neural network. The deep neural network is pre-trained. The network architecture is three convolutional layers, two maximum pooling layers, one Dropout layer and two dense layers. The optimizer is the Adam optimizer, and the loss function is the mean square error loss function.
[0024] The battery capacity of each lithium-ion battery at all times is obtained as follows:
[0025] The charging current fitting curve of each lithium-ion battery at all times is obtained, and the integral value of the ordinate of the fitting curve corresponding to each time and all previous times is used as the battery capacity at each time.
[0026] Wherein, obtaining the capacity of the lithium-ion battery to be tested includes:
[0027] The feature matrix of a lithium-ion battery to be tested within a preset time period during the charging process is input into the trained deep neural network to obtain the capacity of the lithium-ion battery to be tested at each moment during the charging process, and the maximum value of the capacity is taken as the capacity of the lithium-ion battery to be tested.
[0028] In a second aspect, an embodiment of the present application further provides a lithium-ion battery capacity detection system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.
[0029] In the above scheme, the battery parameters of the lithium-ion battery during the charging process are obtained, and the total charging time of the lithium-ion battery is divided according to the change of the internal resistance during the charging process of the lithium-ion battery. The beneficial effect is that the time period with more information in the charging process is screened based on the change of resistance, so as to better reflect the change of the battery state during the charging process, so that the data characteristics in the acquired charging time window are obvious; since the changes of the two data of battery temperature and battery internal resistance interact with each other, the local distribution and change characteristics of the battery temperature in the charging time window are analyzed to construct the temperature suitability, which has the beneficial effect of evaluating whether the charging time window is suitable as the input for training the neural network based on the distribution characteristics of the temperature data in the charging time window; combined with the similarity characteristics of the battery temperature between the charging time windows, the comprehensive suitability is constructed, which has the beneficial effect of measuring the amount of repeated information contained in different charging time windows, which helps the data in the screened charging time windows to reflect the state of the lithium-ion battery during the charging process to the greatest extent; according to the battery parameters of the screened charging time window, the neural network is trained to better capture the information in the charging process, and ensure that the data is representative and diverse, avoid duplication and deviation, thereby improving the accuracy of lithium-ion battery capacity detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 A flowchart of a lithium-ion battery capacity detection method provided in one embodiment of the present application;
[0031] Figure 2 A schematic diagram of a charging curve of a lithium-ion battery provided in one embodiment of the present application;
[0032] Figure 3 A schematic diagram of a “time-capacity” curve of a battery to be tested provided in one embodiment of the present application. DETAILED DESCRIPTION
[0033] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" and the like are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "or", "for example" and the like is intended to present related concepts in a concrete manner.
[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application.
[0035] It should also be noted that the terms "first" and "second" in this application and the accompanying drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The method disclosed in the embodiments of the present application or the method shown in the flow chart includes one or more steps for implementing the method. Without departing from the scope of protection of the present application, the execution order of multiple steps can be interchanged with each other, and some steps can also be deleted.
[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
[0037] The specific scheme of a lithium-ion battery capacity detection method and system provided by the present application is described in detail below with reference to the accompanying drawings.
[0038] See also Figure 1 , which shows a flow chart of a method for detecting capacity of a lithium-ion battery provided by an embodiment of the present application, the method comprising the following steps:
[0039] The first step is to obtain the battery internal resistance, charging current, battery voltage, and battery temperature at each moment during the charging process of a set number of lithium-ion batteries.
[0040] Prepare N lithium-ion batteries. In this implementation, N is 500 and can be adjusted by the implementer. The rated capacity of the lithium-ion battery is 2.2Ah, the discharge cut-off voltage is 3V, the charge cut-off voltage is 4.2V, the charge current is 1C (i.e., 2.2A), and the charge cut-off current is 0.1C. Since the analysis method for each lithium-ion battery is the same, one of the lithium-ion battery samples is taken as an example for explanation.
[0041] First, the lithium-ion battery is discharged to a discharge cut-off voltage of 3V, and then the lithium-ion battery is charged. During the charging process, a battery tester is used to measure the internal resistance change of the lithium-ion battery in real time to obtain the battery internal resistance of the lithium-ion battery at each moment; a voltage sensor is used to collect the battery voltage of the lithium-ion battery at each moment during the charging process; a current sensor is used to collect the charging current of the lithium-ion battery at each moment during the charging process; and a temperature sensor is used to collect the battery temperature of the lithium-ion battery at each moment; wherein, the present embodiment samples various data synchronously, and the sampling frequency is set to 20Hz; the implementer can adjust the sampling frequency by himself, and the present application does not impose any restriction on this.
[0042] The charging curve of the lithium-ion battery is drawn according to the changes in the charging current and battery voltage during the charging process of the lithium-ion battery. The charging process of the lithium-ion battery can be divided into three stages. The first stage: pre-charging, charging with a relatively small current (0.1C), gradually increasing the voltage; the second stage: constant current charging, charging with the rated current (1C), gradually increasing the voltage to the charging cut-off voltage. It should be noted that since the lithium-ion battery has gone through a use cycle, it cannot reach the rated charging cut-off voltage after charging; the third stage: charging with the rated charging cut-off voltage, gradually reducing the current. When the charging current is reduced to the charging cut-off current, the charging is completed.
[0043] Among them, the charging curve diagram of lithium-ion battery is as follows: Figure 2 As shown, in the figure, a represents voltage and b represents current.
[0044] The second step is to divide the total charging time of the lithium-ion battery according to the change of the internal resistance of the lithium-ion battery during the charging process to obtain several charging time windows.
[0045] The internal resistance of a lithium-ion battery refers to the resistance encountered by the current flowing through the battery when the battery is working. It is mainly composed of ohmic internal resistance, polarization internal resistance and interface impedance. The internal resistance of a lithium-ion battery changes during charging, mainly because the migration of lithium ions between the positive and negative electrodes during charging causes the concentration of lithium salts in the electrolyte to change, thereby affecting the conductivity of the electrolyte. As the use time increases, the lithium-ion battery will age, resulting in an increase in the internal resistance of the lithium-ion battery. The internal resistance of the aged lithium-ion battery changes more significantly during the charging process. Since the internal resistance of lithium-ion batteries with different aging degrees changes differently, the time when the internal resistance changes significantly contains more information. Therefore, the total charging time of the lithium-ion battery can be divided according to the change in the internal resistance of the lithium-ion battery during charging to obtain several charging time windows:
[0046] Specifically, the length of the previous charging time window of each charging time window is used as the original window length of each charging time window, and the adjustment factor of each charging time window is obtained according to the internal resistance change rate within the original window corresponding to each charging time window and the discrete degree of the internal resistance; wherein the preset length is used as the original window length of the first charging time window.
[0047] It should be noted that the method for obtaining the internal resistance change rate is: for each charging time window corresponding to the original window, calculate the internal resistance difference between the start time and the end time; take the ratio of the internal resistance difference to the internal resistance at the start time to obtain the internal resistance change rate of each charging time window corresponding to the original window.
[0048] In this embodiment, the discrete degree of the internal resistance refers to the standard deviation of the internal resistance in the corresponding time period; the adjustment factor of the i-th charging time window is recorded as , its specific formula is: ;in, It represents the internal resistance difference between the starting time and the ending time of the i-th charging time window corresponding to the original window; It represents the internal resistance at the starting time of the i-th charging time window corresponding to the original window; It represents the internal resistance change rate of the i-th charging time window corresponding to the original window; represents the standard deviation of the internal resistance in the original window corresponding to the i-th charging time window; The original window length of the first charging time window is 600, that is, the length of the time period corresponding to the original window is 30 seconds, which can be adjusted by the implementer according to the actual situation; in this embodiment, the normalization function adopts the sigmoid function.
[0049] In other embodiments, the standard deviation may be replaced by the variance; in still other embodiments, the standard deviation may be replaced by the interquartile range.
[0050] Furthermore, the length of each charging time window is obtained according to the length of the previous charging time window of each charging time window and the adjustment factor:
[0051] In this embodiment, the length of the i-th charging time window is recorded as , its formula form is: ;in, represents the length of the i-1th charging time window, that is, the length of the i-th charging time window corresponding to the original window; represents the adjustment factor of the i-th charging time window; Indicates rounding up.
[0052] It should be noted that the minimum length of each charging time window is 200 seconds, that is, the minimum time period corresponding to each charging time window is 10 seconds.
[0053] It should be understood that the change in the internal resistance of a lithium-ion battery can reflect the change in state during the charging process, including aging and immediate performance changes. The length of the charging curve segment is adjusted according to the change rate and stability of the internal resistance of the lithium-ion battery. If the internal resistance change rate is large or the internal resistance stability is poor, the segment length needs to be shortened to capture rapid changes, otherwise the segment length needs to be increased.
[0054] The third step: according to the distribution and change characteristics of the local range of battery temperature in each charging time window, the temperature suitability of each charging time window is obtained; according to the similarity and disorder degree of the battery temperature distribution in each charging time window and other charging time windows, combined with the temperature suitability, the comprehensive suitability of each charging time window is obtained.
[0055] There is an interaction between the change in internal resistance and the change in battery temperature during the charging process of lithium-ion batteries. When current passes through the battery, due to the existence of internal resistance, part of the electrical energy will be converted into heat energy, which will cause the battery temperature to rise during the charging process. When the current is the same, the greater the internal resistance, the more heat is generated and the higher the temperature. The temperature increase will have a counter-effect on the internal resistance of the lithium-ion battery, because the temperature increase will increase the ionic conductivity of the electrolyte, and the electrochemical reaction rate of the electrode material may also increase. At this time, the internal resistance of the lithium-ion battery decreases and the temperature increase rate decreases.
[0056] During the charging process, excessively high temperatures will accelerate the aging of lithium-ion batteries. Lithium-ion batteries with high aging have relatively larger internal resistance and contain more information. Therefore, charging curve segments with small temperature changes and relatively high temperatures are more suitable as inputs for deep neural networks. Based on this, the temperature suitability of each charging time window is calculated according to the temperature information of the lithium-ion battery in each charging time window to characterize their suitability as inputs:
[0057] Specifically, for each charging time window, the temperature neighborhood of each moment is preset with the temperature data at each moment as the center; the discreteness of all temperature data in the temperature neighborhood of each moment is calculated; according to the distribution of the discreteness at all moments of each charging time window, combined with the average level of temperature data in each charging time window, the temperature suitability of each charging time window is obtained.
[0058] In this embodiment, the length of the temperature neighborhood at each moment is 9, which can be adjusted by the implementer according to the actual situation; if the data in the temperature neighborhood at each moment is less than 9, the mean filling method is used to fill it. The degree of dispersion of temperature data is measured by the coefficient of variation. The coefficient of variation at the jth moment in the i-th charging time window is recorded as The average level of temperature data is measured by the average value, and the average value of temperature data at all moments in the i-th charging time window is recorded as ; The temperature suitability of the i-th charging time window is recorded as , its specific formula is: ;in, represents the length of the i-th charging time window, Represents the normalization function, where the normalization function uses the sigmoid function.
[0059] In other embodiments, the standard deviation may be used to measure the dispersion of the temperature data; in some other embodiments, the median may be used to measure the average level of the temperature data.
[0060] It should be understood that the greater the temperature suitability, the higher the average level of the temperature data in the charging time window, and the smaller the discreteness of all temperature data in the charging time window, the more suitable the data in the charging time window is as the input of the deep neural network training model.
[0061] When filtering the data in the charging time window, in addition to considering the impact of the internal resistance change, it is also necessary to consider the amount of information it contains and the relationship between each charging curve segment.
[0062] Furthermore, according to the similarity and disorder degree of the battery temperature distribution between each charging time window and other charging time windows, combined with the temperature suitability, the comprehensive suitability of each charging time window is obtained:
[0063] Calculate the distance measure between each charging time window and the temperature data in other charging time windows, and calculate the mean of all the distance measures in each charging time window; merge the degree of confusion of the temperature data in each charging time window, the mean and the temperature suitability to obtain the comprehensive suitability of each charging time window.
[0064] In this embodiment, the distance measurement of the temperature data between the charging time windows adopts the DTW distance, and the degree of confusion of the temperature data is measured by the information entropy; the multiplication method is adopted to fuse multiple variables, that is, the comprehensive suitability of each charging time window is specifically the product of the information entropy of the temperature data in each charging time window, the mean value and the temperature suitability.
[0065] It should be understood that the greater the comprehensive suitability, the more information the temperature data in the charging time window reflects, and the less similar it is to the information reflected in other charging time windows, that is, the less repeated information it contains, the more suitable the charging time window is as the input for deep neural network model training, and vice versa.
[0066] The fourth step: based on the comprehensive suitability, the charging time window of each lithium-ion battery is screened; according to various data information corresponding to the screened charging time window, a neural network is trained to obtain the capacity of the lithium-ion battery to be tested.
[0067] For each lithium-ion battery, M charging time windows with the largest comprehensive suitability are selected. In this embodiment, the value of M is 30, and the implementer can adjust it according to the actual situation.
[0068] Since the lengths of various charging time windows are inconsistent, the longest charging time window among all lithium-ion batteries including the lithium-ion battery to be tested is taken as a benchmark, the time length of the charging time window is recorded as the standard time length, other charging time windows are filled with 0, and each type of data is normalized.
[0069] A feature matrix is constructed using various data in a filled charging time window. The number of rows in the feature matrix is equal to the length of the filled charging time window, and each row represents data at a time. The number of columns in the feature matrix is equal to the number of types of data in the charging time window. In this embodiment, the value is 4, and each column represents corresponding data. The first column is the battery internal resistance, the second column is the charging current, the third column is the battery voltage, and the fourth column is the battery temperature. A total of M N feature matrices; the column vector composed of the battery capacity of each lithium-ion battery at all times is used as the label of all feature matrices of each lithium-ion battery. It should be noted that the battery capacity at each time can be obtained by integrating the charging current at each time and all previous times, specifically: obtaining the fitting curve of the charging current at all times, and taking the integral value of the ordinate of the fitting curve corresponding to each time and all previous times as the battery capacity at each time.
[0070] These feature matrices and corresponding labels are used as inputs of the deep neural network, and the deep neural network is trained. During training, the neural network can predict the target value based on the input feature matrix, and then adjust the network weights through the loss function to improve the prediction accuracy. Among them, the network architecture is three convolutional layers, two maximum pooling layers, one Dropout layer and two dense layers. The optimizer is the Adam optimizer, and the loss function is the mean square error loss function.
[0071] The feature matrix within a preset time period during the charging process of the lithium-ion battery to be tested is input into the trained deep neural network to obtain the capacity of the lithium-ion battery to be tested at each moment during the charging process, and the maximum value of the capacity is taken as the capacity of the lithium-ion battery to be tested, wherein the length of the preset time period is equal to the standard time length.
[0072] Among them, the schematic diagram of the "time-capacity" curve of the battery to be tested is as follows: Figure 3 shown.
[0073] Based on the same inventive concept as the above method, an embodiment of the present application also provides a lithium-ion battery capacity detection system, including a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the steps of any one of the above-mentioned lithium-ion battery capacity detection methods are implemented.
[0074] In summary, the embodiment of the present application obtains the battery parameters during the charging process of the lithium-ion battery, and divides the total charging time of the lithium-ion battery by the change of the internal resistance during the charging process of the lithium-ion battery. The beneficial effect is that the time period with more information in the charging process is screened based on the change of resistance, so as to better reflect the change of the battery state during the charging process, so that the data characteristics in the acquired charging time window are obvious; since the changes of the two data, battery temperature and battery internal resistance, interact with each other, the local distribution and change characteristics of the battery temperature in the charging time window are analyzed, and the temperature suitability is constructed. The beneficial effect is that the charging time window is evaluated based on the distribution characteristics of the temperature data in the charging time window. Whether it is suitable as an input for training a neural network; combined with the similarity characteristics of the battery temperature between the charging time windows, a comprehensive suitability is constructed, which has the beneficial effect of measuring the amount of repeated information contained in different charging time windows, which helps the data in the screened charging time window to reflect the state of the lithium-ion battery during the charging process to the greatest extent; according to the battery parameters of the screened charging time window, the neural network is trained to better capture the information in the charging process, and ensure that the data is representative and diverse, avoid duplication and deviation, thereby improving the accuracy of lithium-ion battery capacity detection.
[0075] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two continuous boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two continuous operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.
[0076] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the basic features of the present application. Therefore, no matter from which point of view, the above embodiments of the present application should be regarded as exemplary and non-restrictive; the technical solutions recorded in the above embodiments are modified, or some of the technical features are replaced by equivalents, which does not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.
Claims
1. A lithium-ion battery capacity detection method, It is characterized in that The method comprises the following steps: Obtaining battery parameters of a set number of lithium-ion batteries at each moment during the charging process; the battery parameters at least include battery internal resistance and battery temperature; The total charging time of the lithium-ion battery is divided according to the change of the internal resistance of the lithium-ion battery during the charging process to obtain several charging time windows; According to the distribution of the local range of the battery temperature in each charging time window and the change characteristics, the temperature suitability of each charging time window is obtained; according to the similarity and disorder degree of the battery temperature distribution in each charging time window and other charging time windows, combined with the temperature suitability, the comprehensive suitability of each charging time window is obtained; Based on the comprehensive suitability, the charging time window of each lithium-ion battery is screened; according to the battery parameters of the screened charging time window, a neural network is trained to obtain the capacity of the lithium-ion battery to be tested; The temperature suitability of each charging time window is obtained as follows: For each charging time window, taking the temperature data at each moment as the center, a temperature neighborhood at each moment is preset; the degree of dispersion of all temperature data in the temperature neighborhood at each moment is calculated; according to the distribution of the degree of dispersion at all moments of each charging time window, combined with the average level of temperature data in each charging time window, the temperature suitability of each charging time window is obtained; The process of obtaining the comprehensive suitability of each charging time window includes: Calculate the distance measure between each charging time window and the temperature data in other charging time windows, and calculate the mean of all the distance measures in each charging time window; merge the degree of confusion of the temperature data in each charging time window, the mean and the temperature suitability to obtain the comprehensive suitability of each charging time window.
2. A lithium-ion battery capacity detection method as claimed in claim 1, It is characterized in that The battery parameters also include: charging current and battery voltage.
3. A lithium-ion battery capacity detection method as claimed in claim 1, It is characterized in that The obtaining of a plurality of charging time windows is specifically as follows: The length of the previous charging time window of each charging time window is used as the original window length of each charging time window; the adjustment factor of each charging time window is obtained according to the internal resistance change rate in the original window corresponding to each charging time window and the discrete degree of the internal resistance; wherein the preset length is used as the original window length of the first charging time window; The length of the i-th charging time window is recorded as , its formula form is: ;in, represents the length of the i-1th charging time window; represents the adjustment factor of the i-th charging time window; Indicates rounding up.
4. A lithium ion battery capacity detection method as claimed in claim 3, It is characterized in that The steps for obtaining the internal resistance change rate are: For each charging time window corresponding to the original window, the internal resistance difference between the start time and the end time is calculated; the internal resistance change rate within the original window corresponding to each charging time window is obtained by taking the ratio of the internal resistance difference to the internal resistance at the start time.
5. A lithium ion battery capacity detection method as claimed in claim 2, It is characterized in that The specific steps of training the neural network are as follows: For each lithium-ion battery, the battery internal resistance, battery temperature, charging current, and battery voltage at all times in each charging time window after screening are respectively formed into column vectors of the feature matrix corresponding to the charging time window; The column vector composed of the battery capacity of each lithium-ion battery at all times is used as the label of all feature matrices of each lithium-ion battery; All feature matrices and corresponding labels are used as input of the deep neural network. The deep neural network is pre-trained. The network architecture is three convolutional layers, two maximum pooling layers, one Dropout layer and two dense layers. The optimizer is the Adam optimizer, and the loss function is the mean square error loss function.
6. A lithium-ion battery capacity detection method as claimed in claim 5, It is characterized in that The battery capacity of each lithium-ion battery at all times is obtained as follows: The charging current fitting curve of each lithium-ion battery at all times is obtained, and the integral value of the ordinate of the fitting curve corresponding to each time and all previous times is used as the battery capacity at each time.
7. A lithium-ion battery capacity detection method as claimed in claim 1, It is characterized in that The method of obtaining the capacity of the lithium-ion battery to be tested comprises: The feature matrix of a lithium-ion battery to be tested within a preset time period during the charging process is input into the trained deep neural network to obtain the capacity of the lithium-ion battery to be tested at each moment during the charging process, and the maximum value of the capacity is taken as the capacity of the lithium-ion battery to be tested.
8. A lithium-ion battery capacity detection system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
