Battery state detection method and device based on battery characteristic parameters

Abstract

The application relates to a battery state detection method and device based on battery characteristic parameters, a computer device, a storage medium and a computer program product. The method comprises the following steps: acquiring a charging characteristic parameter measured value and a discharging characteristic parameter measured value of a to-be-detected battery; the charging characteristic parameter measured value and the discharging characteristic parameter measured value are generated by charging and discharging treatment of the to-be-detected battery; inputting the charging characteristic parameter measured value into a trained machine learning model to obtain a discharging characteristic parameter predicted value; determining the difference degree of the discharging characteristic parameter measured value and the discharging characteristic parameter predicted value, and determining the state of the to-be-detected battery according to the difference degree. The method can improve the accuracy of detecting the battery state.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a method, apparatus, computer device, storage medium, and computer program product for detecting battery state based on battery characteristic parameters. Background Technology

[0002] With the development of the new energy industry, the demand for power batteries and energy storage batteries is rising rapidly. Lithium-ion batteries, due to their high specific energy, low self-discharge rate, and long cycle life, are currently the most practical energy source for electrification.

[0003] With the widespread application of lithium-ion batteries, safety incidents, such as thermal runaway, occur frequently. Internal battery anomalies (such as defects introduced during manufacturing or damage caused during use) are one of the potential causes of these accidents. Therefore, to improve battery safety, it is necessary to test the battery's condition before it leaves the factory or during after-sales service. If an abnormal battery condition is detected, its use can be avoided, thereby improving battery safety and reducing the accident rate.

[0004] In related technologies, the charge / discharge test data of the battery under test is typically compared with the average test data of normal batteries of the same model (i.e., the average of the charging test data and the average of the discharging test data of multiple normal batteries of the same model). If the difference between the test data of the battery under test and the average test data exceeds a preset threshold, the battery under test is judged as an abnormal battery. However, due to individual differences among batteries, comparing each battery with the average test data can easily lead to false detections, meaning the accuracy of battery status detection is low. Summary of the Invention

[0005] Therefore, it is necessary to provide a battery state detection method, apparatus, computer equipment, computer-readable storage medium, and computer program product based on battery characteristic parameters that can improve the accuracy of battery state detection, addressing the aforementioned technical problems.

[0006] Firstly, this application provides a battery state detection method based on battery characteristic parameters. The method includes:

[0007] The measured values ​​of the charging characteristic parameters and the discharging characteristic parameters of the battery under test are obtained; the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters are generated by charging and discharging the battery under test, respectively.

[0008] The measured values ​​of the charging characteristic parameters are input into the trained machine learning model to obtain the predicted values ​​of the discharging characteristic parameters.

[0009] The degree of difference between the measured value and the predicted value of the discharge characteristic parameter is determined, and the state of the battery under test is determined based on the degree of difference.

[0010] In one embodiment, obtaining the measured values ​​of the charging characteristic parameters and the measured values ​​of the discharging characteristic parameters of the battery under test includes:

[0011] Acquire charging and discharging test data generated by the battery under test during the formation and / or capacitance process; the charging and discharging test data include one or more of temperature, voltage, and current data.

[0012] The measured values ​​of charging characteristic parameters and discharging characteristic parameters are extracted from the charging test data and the discharging test data, respectively.

[0013] In one embodiment, the measured values ​​of the discharge characteristic parameters include measured values ​​corresponding to multiple discharge characteristic parameters, and the predicted values ​​of the discharge characteristic parameters include predicted values ​​corresponding to the multiple discharge characteristic parameters. Determining the difference between the measured values ​​of the discharge characteristic parameters and the predicted values ​​of the discharge characteristic parameters includes:

[0014] For each discharge characteristic parameter, the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter is determined, and a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the preset standard deviation.

[0015] The difference between the measured value and the predicted value of the discharge characteristic parameter is obtained by summing the individual deviation values ​​corresponding to each discharge characteristic parameter.

[0016] In one embodiment, obtaining the single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to a preset standard deviation includes:

[0017] Obtain reference values ​​for each of the discharge characteristic parameters corresponding to multiple reference batteries;

[0018] Based on the reference values ​​of each discharge characteristic parameter and the number of reference batteries, the standard deviation corresponding to each discharge characteristic parameter is determined;

[0019] For each of the discharge characteristic parameters, a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the standard deviation corresponding to the discharge characteristic parameter.

[0020] In one embodiment, determining the state of the battery under test based on the difference includes:

[0021] If the difference is greater than a preset threshold, the state of the battery under test is determined to be an abnormal state.

[0022] If the difference is less than or equal to the preset threshold, the state of the battery under test is determined to be normal.

[0023] In one embodiment, the training process of the machine learning model includes:

[0024] Measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained for multiple sample batteries; the measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained based on test data generated by the sample batteries during the formation process and / or the capacity determination process;

[0025] The measured sample values ​​of the charging characteristic parameters are input into the initial machine learning model to obtain the predicted sample values ​​of the discharging characteristic parameters;

[0026] The loss value is calculated based on the predicted sample value of the discharge feature parameter and the measured sample value of the discharge feature parameter, and the parameters of the machine learning model are updated based on the loss value to obtain the trained machine learning model.

[0027] Secondly, this application also provides a battery state detection device based on battery characteristic parameters.

[0028] The device includes:

[0029] The first acquisition module is used to acquire the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters of the battery under test; the measured values ​​of the charging characteristic parameters and the measured values ​​of the discharging characteristic parameters are generated by performing charging and discharging processes on the battery under test, respectively.

[0030] The prediction module is used to input the measured values ​​of the charging characteristic parameters into the trained machine learning model to obtain the predicted values ​​of the discharging characteristic parameters;

[0031] The determination module is used to determine the degree of difference between the measured value and the predicted value of the discharge characteristic parameter, and to determine the state of the battery under test based on the degree of difference.

[0032] In one embodiment, the first acquisition module is specifically used for:

[0033] Acquire charging and discharging test data generated by the battery under test during the formation and / or capacitance process; the charging and discharging test data include one or more of temperature, voltage, and current data; extract the measured values ​​of charging characteristic parameters and discharging characteristic parameters based on the charging and discharging test data, respectively.

[0034] In one embodiment, the measured values ​​of the discharge characteristic parameters include measured values ​​corresponding to multiple discharge characteristic parameters, and the predicted values ​​of the discharge characteristic parameters include predicted values ​​corresponding to the multiple discharge characteristic parameters. The determining module is specifically used for:

[0035] For each discharge characteristic parameter, the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter is determined, and a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the preset standard deviation; the degree of difference between the measured value and the predicted value of the discharge characteristic parameter is obtained based on the sum of the single deviation values ​​corresponding to each discharge characteristic parameter.

[0036] In one embodiment, the determining module is specifically used for:

[0037] Obtain reference values ​​for each discharge characteristic parameter corresponding to multiple reference batteries; determine the standard deviation corresponding to each discharge characteristic parameter based on the reference value of each discharge characteristic parameter and the number of reference batteries; for each discharge characteristic parameter, obtain a single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference corresponding to the discharge characteristic parameter to the standard deviation.

[0038] In one embodiment, the determining module is specifically used for:

[0039] If the difference is greater than a preset threshold, the state of the battery under test is determined to be an abnormal state; if the difference is less than or equal to the preset threshold, the state of the battery under test is determined to be a normal state.

[0040] In one embodiment, the device further includes:

[0041] The second acquisition module is used to acquire measured sample values ​​of charging characteristic parameters and discharge characteristic parameters corresponding to multiple sample batteries; the measured sample values ​​of charging characteristic parameters and discharge characteristic parameters are obtained based on test data generated by the sample batteries during the formation process and / or the capacity determination process;

[0042] The input module is used to input the measured sample values ​​of the charging characteristic parameters into the initial machine learning model to obtain the predicted sample values ​​of the discharging characteristic parameters.

[0043] The update module is used to calculate the loss value based on the predicted sample value of the discharge feature parameter and the measured sample value of the discharge feature parameter, and update the parameters of the machine learning model based on the loss value to obtain the trained machine learning model.

[0044] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in the first aspect.

[0045] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0046] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0047] The aforementioned battery state detection method, apparatus, computer equipment, storage medium, and computer program products based on battery characteristic parameters acquire measured values ​​of charging and discharging characteristic parameters of the battery under test. The measured charging characteristic parameters are then input into a trained machine learning model to obtain predicted discharging characteristic parameters. The state of the battery is determined based on the difference between the measured and predicted discharging characteristic parameters. This method compares the battery's own relevant data (measured and predicted discharging characteristic parameters) to determine whether the battery state is abnormal. This approach considers the differences between batteries. Compared to comparing each battery with the same standard (average test data), this method reduces interference from differences or inconsistencies between batteries, improving the accuracy of battery state detection. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating a battery state detection method based on battery characteristic parameters in one embodiment.

[0049] Figure 2 This is a schematic diagram of the process for obtaining the measured values ​​of feature parameters in one embodiment;

[0050] Figure 3 This is a flowchart illustrating the process of determining the degree of difference in one embodiment;

[0051] Figure 4 This is a flowchart illustrating the process of determining a single deviation value in one embodiment;

[0052] Figure 5 This is a flowchart illustrating the training process of a machine learning model in one embodiment.

[0053] Figure 6 This is a comparison chart of the voltage curves of an abnormal battery and a normal battery in an example.

[0054] Figure 7 This is a structural block diagram of a battery state detection device based on battery characteristic parameters in one embodiment;

[0055] Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0057] First, before introducing the technical solutions of the embodiments of this application, the technical background or evolution of the embodiments of this application will be introduced. To improve battery safety, it is necessary to test the battery status before it leaves the factory or during after-sales service to avoid using abnormal batteries. In related technologies, the charge / discharge test data of a battery is generally compared with the average test data of normal batteries of the same model (i.e., the average of the charging test data and the average of the discharging test data of multiple normal batteries of the same model). If the difference between the test data and the average test data is greater than a preset threshold, it is judged as an abnormal battery. However, due to individual differences in batteries, comparing each battery with the average test data can easily lead to false detections and low detection accuracy. Based on this background, the applicant, through long-term research and development and experimental verification, proposes the battery status detection method based on battery characteristic parameters, which can reduce the interference of differences or inconsistencies between batteries and improve the accuracy of battery status detection. Furthermore, it should be noted that the applicant has devoted considerable creative effort to discovering the technical problems of this application and the technical solutions described in the following embodiments.

[0058] The battery state detection method based on battery characteristic parameters provided in this application can be applied to terminals, servers, and systems including both terminals and servers, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, tablets, smartphones, IoT devices, portable wearable devices, and other electronic devices (such as battery state detection devices). The server can be a standalone server or a server cluster consisting of multiple servers.

[0059] In one embodiment, such as Figure 1 As shown, a battery state detection method based on battery characteristic parameters is provided. Taking the application of this method to a terminal as an example, the method includes the following steps:

[0060] Step 101: Obtain the measured values ​​of the charging characteristic parameters and the discharge characteristic parameters of the battery under test.

[0061] The battery under test can be a battery before it leaves the factory (not yet used) or a battery after it has been used. The measured values ​​of charging and discharging characteristic parameters are generated by charging and discharging the battery under test, respectively. For example, during the charging and discharging process, temperature, voltage, or current data of the battery under test (generally sequential data, i.e., temperature, voltage, or current data at each moment during charging and discharging) can be collected as measured values ​​of charging / discharging characteristic parameters. Alternatively, the measured values ​​of charging / discharging characteristic parameters can be extracted from the collected temperature, voltage, or current data. Thus, the terminal can obtain the measured values ​​of the charging and discharging characteristic parameters of the battery under test.

[0062] Step 102: Input the measured values ​​of the charging characteristic parameters into the trained machine learning model to obtain the predicted values ​​of the discharging characteristic parameters.

[0063] The machine learning model is a pre-trained model used to predict the battery's discharge characteristics under normal conditions based on the battery's charging characteristics. An example of the machine learning model's training process will be provided later, and will not be elaborated upon here.

[0064] In practice, the terminal can input the measured values ​​of the charging characteristic parameters of the battery under test into the trained machine learning model. The machine learning model then processes the input data and outputs predicted values ​​of the discharge characteristic parameters. These predicted values ​​represent the values ​​of the discharge characteristic parameters predicted when the battery is in a normal state, based on the measured values ​​of the charging characteristic parameters.

[0065] Step 103: Determine the degree of difference between the measured values ​​and the predicted values ​​of the discharge characteristic parameters, and determine the state of the battery under test based on the degree of difference.

[0066] In implementation, the terminal can calculate the difference between the measured values ​​of the discharge characteristic parameters obtained in step 101 and the predicted values ​​of the discharge characteristic parameters obtained in step 102, and determine the state of the battery under test based on this difference. For example, the difference between the two can be used as the difference, and then the difference can be compared with a preset difference threshold. If the difference is greater than the preset difference threshold, the battery under test can be considered to be in an abnormal state, i.e., an abnormal battery. Conversely, if the difference is less than or equal to the preset difference threshold, the battery under test can be considered to be in a normal state. It is understood that there can be multiple discharge characteristic parameters. A corresponding difference can be calculated for each discharge characteristic parameter, or the sub-differences of each discharge characteristic parameter can be integrated into a total difference. That is, the difference can contain one value or multiple values. If it contains multiple values, they can be compared with their respective preset thresholds to determine the state of the battery under test.

[0067] In the aforementioned battery state detection method based on battery characteristic parameters, the measured values ​​of the charging and discharging characteristic parameters of the battery under test are obtained. These measured values ​​are then input into a trained machine learning model to obtain predicted values ​​of the discharging characteristic parameters during the discharging phase. The state of the battery is then determined based on the difference between the measured and predicted values ​​of the discharging characteristic parameters. The machine learning model is a pre-trained model used to predict the discharging characteristic parameters of the battery under normal conditions based on its charging characteristic parameters. Therefore, the predicted values ​​of the discharging characteristic parameters represent the values ​​predicted for the battery under normal conditions based on its charging characteristic parameters. Thus, the state of the battery can be determined based on the difference between the predicted and measured values ​​of the discharging characteristic parameters. This method compares the battery's own relevant data (measured and predicted values ​​of discharging characteristic parameters) to determine whether the battery state is abnormal. This approach considers the differences between batteries. Compared to comparing each battery with the same standard (average test data), this method reduces the interference of differences or inconsistencies between batteries, improving the accuracy of battery state detection.

[0068] In one embodiment, the battery under test can be a battery before it leaves the factory; that is, this embodiment can be used to perform pre-shipment condition testing on batteries on the production line. Figure 2 As shown, the process of obtaining the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters of the battery under test in step 101 specifically includes the following steps:

[0069] Step 201: Obtain charging test data and discharging test data generated by the battery under test during the formation process and / or the capacitance process.

[0070] Understandably, a battery production line generally includes a formation process and a capacity setting process. The formation process is the first charge and discharge of the manufactured battery to activate it. The capacity setting process is to charge the battery to full capacity and then discharge it to empty capacity to determine the actual capacity of the battery.

[0071] In implementation, during the formation and / or capacity setting stages of the battery under test, charging and discharging test data generated during these processes can be collected using sensors, acquisition circuits, and other relevant instruments. The collected data is then transmitted to the terminal. The charging and discharging test data includes one or more of temperature, voltage, and current data. It is understood that the collected test data is sequential data; for example, for voltage data during the charging process, the voltage values ​​of the battery under test at various moments during charging can be obtained, and a voltage-time curve can be generated based on the voltage data. Optionally, the charging and discharging test data can both be generated from the same process, i.e., both from the formation process or both from the capacity setting process. Optionally, the charging and discharging test data can be generated within the same charge-discharge cycle, where the battery is charged and then discharged, constituting one charge-discharge cycle.

[0072] Step 202: Extract the measured values ​​of charging characteristic parameters and discharging characteristic parameters based on the charging test data and discharging test data, respectively.

[0073] In implementation, the terminal can extract measured values ​​of charging characteristic parameters and discharging characteristic parameters from charging test data and discharging test data, respectively. Charging characteristic parameters can be one or more of the following: charging capacity, average voltage, voltage standard deviation, voltage skewness, voltage kurtosis, average temperature, average polarization voltage, and minimum slope of the charge-voltage curve. Discharging characteristic parameters can be one or more of the following: discharging capacity, average voltage, voltage standard deviation, voltage skewness, voltage kurtosis, average temperature, average polarization voltage, and minimum slope of the charge-voltage curve. For example, if the charging characteristic parameter is charging capacity, the charging capacity of the battery under test can be calculated from the current data during the charging stage (the charging stage in the formation process and / or the calibrating process), and used as the measured value of the charging characteristic parameter. If the charging characteristic parameters are average voltage, voltage standard deviation, and average polarization voltage, they can be calculated from the voltage data in the charging test data. For voltage skewness and voltage kurtosis, a voltage-time curve can be generated from the voltage data to calculate the voltage skewness and voltage kurtosis. The discharging characteristic parameters can be obtained similarly.

[0074] In some specific examples, charging characteristic parameters can be the charging capacity, average voltage, voltage skewness, voltage kurtosis, average temperature, and average polarization voltage during the charging stage. Discharging characteristic parameters can be the discharged capacity and average voltage during the discharge stage (the discharge stage in the formation process and / or the stabilization process), thus achieving high detection accuracy and efficiency.

[0075] In this embodiment, the measured values ​​of charging characteristic parameters and discharging characteristic parameters are extracted from the charging test data and discharging test data generated by the battery under test during the formation process and / or the capacity setting process. This allows for subsequent processing based on the measured values ​​of charging characteristic parameters and discharging characteristic parameters, enabling pre-shipment status detection of the battery under test. Furthermore, this method has high detection accuracy and can also take into account detection efficiency.

[0076] In one embodiment, the measured values ​​of discharge characteristic parameters include the measured values ​​corresponding to multiple discharge characteristic parameters, and the predicted values ​​of discharge characteristic parameters include the predicted values ​​corresponding to multiple discharge characteristic parameters. For example... Figure 3 As shown, step 103, determining the difference between the measured and predicted values ​​of the discharge characteristic parameters, specifically includes the following steps:

[0077] Step 301: For each discharge characteristic parameter, determine the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter, and obtain the single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to the preset standard deviation.

[0078] In practice, there can be multiple discharge characteristic parameters, such as discharge capacity, average voltage, voltage standard deviation, voltage skewness, voltage kurtosis, average temperature, average polarization voltage, and minimum slope of the charge-voltage curve. There can also be multiple charging characteristic parameters, such as charge capacity, average voltage, voltage standard deviation, voltage skewness, voltage kurtosis, average temperature, average polarization voltage, and minimum slope of the charge-voltage curve. If the charging characteristic parameters are denoted as n and the discharge characteristic parameters as m, then the measured values ​​of the charging characteristic parameters of the battery under test (which can be denoted as k) will include the measured values ​​corresponding to the n charging characteristic parameters, which can be denoted as X. k ={X k，1 X k，2 , ..., X k，i , ..., X k，n}, where X k，i This represents the measured value corresponding to the charging characteristic parameter i. The measured values ​​of the discharge characteristic parameters of the battery under test k will include the measured values ​​corresponding to m discharge characteristic parameters, which can be denoted as Y. k ={Y k，1 Y k，2 , ..., Y k，j , ..., Yk，m}, where Y k，j Let X be the measured value corresponding to the discharge characteristic parameter j. Let X be the measured value of the charging characteristic parameter X of the battery under test k. k ={X k，1 X k，2 , ..., X k，i , ..., X k，n After being input into the trained machine learning model, the predicted values ​​corresponding to m discharge feature parameters can be obtained, which can be denoted as: in This is the predicted value corresponding to the discharge characteristic parameter j.

[0079] For each discharge characteristic parameter, the terminal can calculate the difference between the measured value and the predicted value corresponding to that discharge characteristic parameter. For example, for discharge characteristic parameter j, the measured value Y can be calculated. k，j and predicted value The difference (which is understandable; the difference can be taken as its absolute value) Then, the terminal can calculate the difference. and the preset standard deviation (which can be denoted as σ) j The ratio of ) is used to obtain the single deviation value (which can be denoted as e) corresponding to the discharge characteristic parameter. k，j , representing a single deviation value of the discharge characteristic parameter j corresponding to the battery k under test). Wherein, the preset standard deviation σ j It can be set based on experience, or calculated based on discharge test data of multiple batteries from the same production line or model as the battery under test. Detailed examples will be given later, and will not be repeated here.

[0080] Step 302: Based on the sum of the individual deviation values ​​corresponding to each discharge characteristic parameter, obtain the degree of difference between the measured value and the predicted value of the discharge characteristic parameter.

[0081] In implementation, the terminal calculates the single deviation value e corresponding to each discharge characteristic parameter according to step 301. k，j After (j = 1, 2, ..., m), the individual deviation values ​​can be summed to obtain the sum value, which serves as the difference between the measured and predicted values ​​of the discharge characteristic parameters of the battery under test (denoted as E). k Difference E k The formula for calculating can be shown below:

[0082]

[0083] Thus, the terminal can adjust the difference based on E. k Determine the state of the battery k under test. For example, the difference E can be used. k Compared with a preset threshold (which can be denoted as θ), if the difference is greater than the preset threshold (E) kIf the difference is greater than or equal to the preset threshold (E), then the state of the battery under test k can be determined as an abnormal state, that is, the battery under test k is considered to be an abnormal battery; if the difference is less than or equal to the preset threshold (E), then the state of the battery under test k can be determined as an abnormal state. k If the threshold value ≤ θ), then the state of the battery under test k is determined to be normal, that is, the battery under test k is considered to be a normal battery. The preset threshold value θ can be set according to the situation.

[0084] This embodiment provides a method for determining the degree of difference, and the state of the battery under test is determined based on the degree of difference, which can improve the accuracy of detecting the battery state.

[0085] In one embodiment, such as Figure 4 As shown, the process of determining the single deviation value corresponding to the discharge characteristic parameter in step 301 specifically includes the following steps:

[0086] Step 401: Obtain reference values ​​for each discharge characteristic parameter corresponding to multiple reference batteries.

[0087] The reference battery can be a battery of the same model as the battery under test, such as multiple batteries from the same production line, and can include normal and abnormal batteries. The reference values ​​for each discharge characteristic parameter corresponding to the reference battery are extracted from discharge test data collected during the formation and / or capacity setting processes of the reference battery. Thus, the terminal can obtain the reference values ​​for each discharge characteristic parameter corresponding to multiple reference batteries. The reference battery can include the battery under test itself. For example, if it is necessary to perform pre-shipment status testing on N batteries from the same production line, these N batteries can each serve as a battery under test, and can also serve as a reference battery.

[0088] Step 402: Determine the standard deviation of each discharge characteristic parameter based on the reference values ​​of each discharge characteristic parameter and the number of reference batteries.

[0089] In practice, the terminal can determine the standard deviation of each discharge characteristic parameter based on the reference values ​​of each parameter and the number of reference batteries. For example, N reference batteries from the same production line as the battery under test can be selected. Based on the discharge test data collected during the formation and / or capacity setting stages of these N reference batteries, the reference value Y of the discharge characteristic parameter j of reference battery c can be extracted. c，j (c = 1, 2, ..., N; j = 1, 2, ..., m). Then, based on the reference value Y of the discharge characteristic parameter j... c，j The standard deviation σ corresponding to the discharge characteristic parameter j is calculated based on the number of reference batteries N. j In one example, the standard deviation σ j The calculation formula is as follows:

[0090]

[0091] in, Y is the reference value of the discharge characteristic parameter j based on N reference batteries c. c，j The average value corresponding to the calculated discharge characteristic parameter j.

[0092] Step 403: For each discharge characteristic parameter, obtain the single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to the standard deviation of the discharge characteristic parameter.

[0093] In practice, the terminal can determine the differences between the discharge characteristic parameters j (j = 1, 2, ..., m) of the battery under test k. and standard deviation σ j The ratio of these values ​​yields the single deviation value e corresponding to each discharge characteristic parameter j. k，j In one example, the single deviation value e corresponds to the discharge characteristic parameter j of the battery under test k. k，j The calculation formula can be expressed as:

[0094]

[0095] in, Y is the measured value of the discharge characteristic parameter j of the battery k under test. k，j and predicted value The difference, σ j Let j be the standard deviation corresponding to the discharge characteristic parameter j.

[0096] In this embodiment, there can be multiple discharge characteristic parameters. The standard deviation of each discharge characteristic parameter can be calculated based on the reference values ​​of the discharge characteristic parameters of multiple reference batteries, thereby obtaining the single deviation value corresponding to each discharge characteristic parameter. Then, the difference degree can be obtained based on the sum of the single deviation values, and the state of the battery under test can be more accurately determined based on the difference degree.

[0097] In one embodiment, such as Figure 5 The training process for the machine learning model used in step 102 is described above, including the following steps:

[0098] Step 501: Obtain the measured sample values ​​of charging characteristic parameters and discharging characteristic parameters corresponding to multiple sample batteries.

[0099] The measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained based on test data generated during the formation and / or capacitance processes of the sample batteries. The sample batteries can be multiple batteries of the same model as the battery under test, for example, multiple batteries from the same production line. The sample batteries can be all normal batteries, or they can include a small number of abnormal batteries (e.g., multiple batteries from the same production line that require pre-shipment condition testing can be directly used as sample batteries for model training; these sample batteries generally only contain a small number of abnormal batteries).

[0100] In implementation, during the formation and / or capacity stabilization stages of each sample battery, charging and discharging test data generated during the formation and / or capacity stabilization processes can be collected using sensors, acquisition circuits, and other relevant instruments. This collected data is then transmitted to the terminal. The terminal can then extract measured sample values ​​of charging characteristic parameters (such as charged capacity, average voltage, voltage skewness, voltage kurtosis, average temperature, and average polarization voltage) and discharging characteristic parameters (such as discharged capacity and average voltage) based on the acquired charging and discharging test data.

[0101] Step 502: Input the measured sample values ​​of the charging characteristic parameters into the initial machine learning model to obtain the predicted sample values ​​of the discharging characteristic parameters.

[0102] In implementation, the terminal can input the measured sample values ​​of the charging characteristic parameters obtained in step 501 into the initial machine learning model. The machine learning model can be any model with fitting capabilities, and the specific model can be selected according to the situation. After processing the input measured samples of charging characteristic parameters, the machine learning model can output the predicted values ​​(predicted sample values ​​of discharge characteristic parameters) corresponding to each discharge characteristic parameter.

[0103] Step 503: Calculate the loss value based on the predicted sample value of the discharge characteristic parameter and the measured sample value of the discharge characteristic parameter, and update the parameters of the machine learning model based on the loss value to obtain the trained machine learning model.

[0104] In implementation, the terminal can calculate the loss value (denoted as ε) based on the predicted sample values ​​of the discharge characteristic parameters and the measured sample values ​​of the discharge characteristic parameters. In one example, the loss value ε can be calculated using the following loss function:

[0105]

[0106] Where M is the number of sample batteries e, Y e，j This represents the measured sample value of the charging characteristic parameter j (j = 1, 2, ..., m) corresponding to sample battery e. This represents the predicted sample value of the charging characteristic parameter j corresponding to sample battery e.

[0107] Then, the terminal can update the parameters of the machine learning model based on the loss value ε, so that the loss value ε is minimized (e.g., ε is less than a preset value), or the preset number of training iterations is reached, thus obtaining the trained machine learning model.

[0108] In this embodiment, the machine learning model is trained by using the measured sample values ​​of charging characteristic parameters and discharge characteristic parameters extracted from the test data generated by the sample battery during the formation process and / or the capacity setting process. This makes the predicted values ​​of discharge characteristic parameters by the trained machine learning model closer to the values ​​of discharge characteristic parameters when the battery under test is in a normal state. Thus, the difference between the predicted value and the measured value can more accurately reflect the state of the battery under test.

[0109] This application also provides a specific example of a battery state detection method based on battery characteristic parameters. In this example, there are multiple batteries to be tested, which are a batch of batteries from the same production line. Each battery needs to be state-tested before leaving the factory. In this example, the reference battery and the sample battery are both batteries from this batch to be tested, and the number can be 5000. This example illustrates the process of performing state detection on any one of the batteries to be tested, specifically including the following steps:

[0110] Step 1: Obtain charging test data and discharging test data generated by multiple sample batteries during the formation process or the capacity determination process, and obtain the measured sample values ​​of charging characteristic parameters and discharging characteristic parameters based on the charging test data.

[0111] The charging characteristic parameters include charged capacity, average voltage, voltage skewness, voltage kurtosis, average temperature, and average polarization voltage (a total of 6 features), while the discharging characteristic parameters include discharged capacity and average voltage (a total of 2 features). The sample battery includes the battery under test.

[0112] Step 2: The machine learning model is trained using the measured sample values ​​of the charging characteristic parameters and the discharge characteristic parameters of each sample battery to obtain the trained machine learning model.

[0113] The machine learning model employs a feedforward neural network model to process the measured values ​​of the input charging characteristic parameters and obtain the predicted values ​​of the discharging characteristic parameters.

[0114] Step 3: Based on the measured sample values ​​of the charging characteristic parameters of each sample battery and the number of sample batteries, determine the standard deviation σ corresponding to each discharge characteristic parameter. j .

[0115] Step 4: Take the measured sample values ​​of the charging characteristic parameters and the measured sample values ​​of the discharging characteristic parameters corresponding to the battery k under test as the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters Y, respectively. k，j The measured values ​​of charging characteristic parameters are input into the trained machine learning model to obtain the predicted values ​​of discharging characteristic parameters.

[0116] Step 5, based on the measured value of the discharge characteristic parameter Y k，j and predicted values ​​of discharge characteristic parameters Determine the difference And based on this difference and the standard deviation σ j The single deviation value e corresponding to each discharge characteristic parameter of the battery k under test is obtained. k，j The sum of each individual deviation value is calculated to obtain the difference E between the measured and predicted values ​​of the discharge characteristic parameters of the battery under test. k .

[0117] Step 6, calculate the difference E. k The state of the battery k under test is determined by comparing it with a preset deviation threshold. The preset deviation threshold (which can be denoted as θ) can be set to 3. If E k If the value is greater than θ, then the battery under test, k, is determined to be an abnormal battery.

[0118] In this example, 5,000 batteries to be tested were inspected on the same production line, and 4 abnormal batteries were detected.

[0119] Figure 6 A comparison of the charge and discharge voltage curves of a typical abnormal battery with those of a normal battery is shown. Due to the small deviation in the battery voltage curves, there is a high degree of similarity between the voltage curves of different batteries. Therefore, discharge characteristic parameters can be used to predict values. The battery voltage curve was reconstructed for easier visual display. Figure 6 The predicted voltage curve shown is based on This was obtained by performing a translation and scaling transformation on the voltage curve.

[0120] according to Figure 6 It can be seen that the voltage curves of the abnormal battery and the normal battery almost overlap during the charging stage. Let the normal battery be a and the abnormal battery be b. Then the measured values ​​of the charging characteristic parameters extracted from the charging stage data of these two batteries are similar. a ≈X b X a X b After being input into the neural network, they predict respectively And the two are similar For a normal battery a, the relationship between its two data segments conforms to the average pattern learned by the neural network from a large amount of battery data. Therefore (The predicted values ​​of the discharge characteristic parameters are approximately equal to the measured values). However, abnormal battery b exhibits significant differences from normal batteries during the discharge phase; the abnormal battery discharges significantly less charge than the normal battery, while its average voltage is not significantly different. Considering both capacity and voltage deviations, Y... a，j With Y b，j There are significant differences in the measured discharge characteristic parameters between normal battery a and abnormal battery b. Furthermore, because... Therefore, the measured value of the discharge characteristic parameter Y of the abnormal battery b，j Compared with the predicted value Significant differences exist. Looking at the prediction bias (degree of difference) of the characteristic parameters, E... a <θ<E b Therefore, battery a is judged to be a normal battery, and battery b is judged to be an abnormal battery.

[0121] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0122] Based on the same inventive concept, this application also provides a battery state detection device based on battery characteristic parameters for implementing the battery state detection method based on battery characteristic parameters described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the battery state detection device based on battery characteristic parameters provided below can be found in the limitations of the battery state detection method based on battery characteristic parameters described above, and will not be repeated here.

[0123] In one embodiment, such as Figure 7 As shown, a battery state detection device 700 based on battery characteristic parameters is provided, including: a first acquisition module 701, a prediction module 702, and a determination module 703, wherein:

[0124] The first acquisition module 701 is used to acquire the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters of the battery under test; the measured values ​​of the charging characteristic parameters and the discharging characteristic parameters are generated by charging and discharging the battery under test, respectively.

[0125] The prediction module 702 is used to input the measured values ​​of charging characteristic parameters into the trained machine learning model to obtain the predicted values ​​of discharging characteristic parameters.

[0126] The determination module 703 is used to determine the degree of difference between the measured values ​​and the predicted values ​​of the discharge characteristic parameters, and to determine the state of the battery under test based on the degree of difference.

[0127] In one embodiment, the first acquisition module 701 is specifically used to: acquire charging test data and discharging test data generated by the battery under test during the formation process and / or the capacity determination process; the charging test data and discharging test data include one or more of temperature, voltage, and current data; and extract the measured values ​​of charging characteristic parameters and discharging characteristic parameters respectively based on the charging test data and discharging test data.

[0128] In one embodiment, the measured values ​​of discharge characteristic parameters include the measured values ​​corresponding to multiple discharge characteristic parameters, and the predicted values ​​of discharge characteristic parameters include the predicted values ​​corresponding to multiple discharge characteristic parameters. The determining module 703 is specifically used to: for each discharge characteristic parameter, determine the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter, and obtain a single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to the preset standard deviation; and obtain the degree of difference between the measured value and the predicted value of the discharge characteristic parameter based on the sum of the single deviation values ​​corresponding to each discharge characteristic parameter.

[0129] In one embodiment, the determining module 703 is specifically used to: obtain reference values ​​for each discharge characteristic parameter corresponding to multiple reference batteries; determine the standard deviation corresponding to each discharge characteristic parameter based on the reference value of each discharge characteristic parameter and the number of reference batteries; and for each discharge characteristic parameter, obtain a single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to the standard deviation.

[0130] In one embodiment, the determining module 703 is specifically used to: determine the state of the battery under test as an abnormal state when the difference is greater than a preset threshold; and determine the state of the battery under test as a normal state when the difference is less than or equal to the preset threshold.

[0131] In one embodiment, the device further includes a second acquisition module, an input module, and an update module, wherein:

[0132] The second acquisition module is used to acquire the measured sample values ​​of charging characteristic parameters and discharging characteristic parameters corresponding to multiple sample batteries. The measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained based on the test data generated by the sample batteries during the formation process and / or the capacitance process.

[0133] The input module is used to input the measured sample values ​​of charging characteristic parameters into the initial machine learning model to obtain the predicted sample values ​​of discharging characteristic parameters.

[0134] The update module is used to calculate the loss value based on the predicted sample value and the measured sample value of the discharge characteristic parameter, and update the parameters of the machine learning model based on the loss value to obtain the trained machine learning model.

[0135] Each module in the aforementioned battery state detection device based on battery characteristic parameters can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0136] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a battery state detection method based on battery characteristic parameters. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0137] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0138] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0139] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0140] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0141] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A battery state detection method based on battery characteristic parameters, characterized in that, The method includes: Acquire charging and discharging test data generated by the battery under test during the formation and / or capacitance process; the charging and discharging test data include at least two of temperature, voltage, and current data. The measured values ​​of charging characteristic parameters and discharging characteristic parameters are extracted from the charging test data and the discharging test data, respectively. The measured values ​​of charging characteristic parameters and discharging characteristic parameters are generated by charging and discharging the battery under test, respectively. The measured values ​​of discharging characteristic parameters include the measured values ​​corresponding to multiple discharging characteristic parameters. The measured values ​​of the charging feature parameters are input into the trained machine learning model to obtain the predicted values ​​of the discharging feature parameters, which include the predicted values ​​corresponding to the plurality of discharging feature parameters. For each discharge characteristic parameter, the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter is determined, and a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the preset standard deviation; the degree of difference between the measured value and the predicted value of the discharge characteristic parameter is obtained based on the sum of the single deviation values ​​corresponding to each discharge characteristic parameter, and the state of the battery under test is determined based on the degree of difference.

2. The method according to claim 1, characterized in that, The step of obtaining a single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to a preset standard deviation includes: Obtain reference values ​​for each of the discharge characteristic parameters corresponding to multiple reference batteries; Based on the reference values ​​of each discharge characteristic parameter and the number of reference batteries, the standard deviation corresponding to each discharge characteristic parameter is determined; For each of the discharge characteristic parameters, a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the standard deviation corresponding to the discharge characteristic parameter.

3. The method according to claim 1, characterized in that, Determining the state of the battery under test based on the difference includes: If the difference is greater than a preset threshold, the state of the battery under test is determined to be an abnormal state. If the difference is less than or equal to the preset threshold, the state of the battery under test is determined to be normal.

4. The method according to claim 1, characterized in that, The training process of the machine learning model includes: Measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained for multiple sample batteries; the measured sample values ​​of charging characteristic parameters and discharging characteristic parameters are obtained based on test data generated by the sample batteries during the formation process and / or the capacity determination process; The measured sample values ​​of the charging characteristic parameters are input into the initial machine learning model to obtain the predicted sample values ​​of the discharging characteristic parameters; The loss value is calculated based on the predicted sample value of the discharge feature parameter and the measured sample value of the discharge feature parameter, and the parameters of the machine learning model are updated based on the loss value to obtain the trained machine learning model.

5. A battery state detection device based on battery characteristic parameters, characterized in that, The device includes: The first acquisition module is used to acquire charging test data and discharging test data generated by the battery under test during the formation process and / or the capacity determination process; the charging test data and the discharging test data include at least two of temperature, voltage and current data. The first acquisition module is used to extract measured values ​​of charging characteristic parameters and measured values ​​of discharging characteristic parameters based on the charging test data and the discharging test data, respectively; the measured values ​​of charging characteristic parameters and the measured values ​​of discharging characteristic parameters are generated by charging and discharging the battery under test, respectively, and the measured values ​​of discharging characteristic parameters include measured values ​​corresponding to multiple discharging characteristic parameters; The prediction module is used to input the measured values ​​of the charging characteristic parameters into the trained machine learning model to obtain the predicted values ​​of the discharging characteristic parameters, which include the predicted values ​​corresponding to multiple discharging characteristic parameters. The determination module is used to determine, for each discharge characteristic parameter, the difference between the measured value and the predicted value corresponding to the discharge characteristic parameter, and obtain a single deviation value corresponding to the discharge characteristic parameter based on the ratio of the difference to a preset standard deviation; obtain the degree of difference between the measured value and the predicted value of the discharge characteristic parameter based on the sum of the single deviation values ​​corresponding to each discharge characteristic parameter, and determine the state of the battery under test based on the degree of difference.

6. The apparatus according to claim 5, characterized in that, The determining module is specifically used to obtain reference values ​​for each of the discharge characteristic parameters corresponding to multiple reference batteries; Based on the reference values ​​of each discharge characteristic parameter and the number of reference batteries, the standard deviation corresponding to each discharge characteristic parameter is determined; For each of the discharge characteristic parameters, a single deviation value corresponding to the discharge characteristic parameter is obtained based on the ratio of the difference to the standard deviation corresponding to the discharge characteristic parameter.

7. The apparatus according to claim 5, characterized in that, Specifically, when the difference is greater than a preset threshold, the determining module determines the state of the battery under test as an abnormal state. If the difference is less than or equal to the preset threshold, the state of the battery under test is determined to be normal.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

Images