Lithium plating detection method, electronic device, battery pack, and energy storage product

Abstract

A lithium plating detection method, an electronic device, a battery pack, and an energy storage product. The method comprises: acquiring data to be processed of a target battery (S101), wherein the data to be processed comprises actual state data of the target battery at a plurality of historical moments within a historical time period prior to the current moment during a current charging process, and the actual state data comprises a full-battery potential value and a charging current value; using a pre-trained potential prediction model to process the data to be processed, so as to obtain predicted negative-electrode potential values of the target battery at the plurality of historical moments (S102), wherein the potential prediction model is obtained by means of performing training on the basis of actual state data of a sample battery at a plurality of sample moments during a historical charging process and a specified actual potential value, the specified actual potential value comprises an actual positive-electrode potential value or an actual negative-electrode potential value; and on the basis of a plurality of obtained predicted negative-electrode potential values, determining whether lithium plating occurs in the target battery at the plurality of historical moments (S103). Therefore, the accuracy of lithium plating detection is improved.

Description

A lithium plating detection method, electronic device, battery pack and energy storage product

[0001] This application claims priority to Chinese Patent Application No. 202510344316.0, filed on March 21, 2025, entitled "A Lithium Plating Detection Method, Electronic Device, Battery Pack and Energy Storage Product", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of battery management technology, and in particular to a lithium plating detection method, electronic equipment, battery pack, and energy storage product. Background Technology

[0003] Due to their advantages such as high energy density, long cycle life, and no memory effect, lithium-ion batteries are widely used in various fields, including new energy vehicles, light electric vehicles / tricycles, drones, and energy storage products. However, during battery use, lithium plating side reactions may occur, leading to a reduction in capacity. In severe cases, this can cause internal short circuits, thermal runaway, and other malfunctions, thereby affecting the personal safety of users.

[0004] Therefore, how to effectively detect lithium plating in batteries and take timely countermeasures has become a key issue in the industry. Summary of the Invention

[0005] The purpose of this application is to provide a lithium plating detection method, an electronic device, a battery pack, and an energy storage product. The specific technical solution is as follows:

[0006] This application provides a lithium plating detection method, the method comprising: acquiring data to be processed from a target battery; wherein the data to be processed includes: real state data of the target battery at multiple historical moments within a historical time period prior to the current moment during the current charging process, the real state data including: full battery potential value and charging current value; processing the data to be processed using a pre-trained potential prediction model to obtain predicted negative electrode potential values ​​of the target battery at the multiple historical moments; wherein the potential prediction model is constructed as follows: trained based on real state data and real specified potential values ​​of sample batteries at multiple sample moments during historical charging processes, the real specified potential values ​​including real positive electrode potential value or real negative electrode potential value; and determining whether lithium plating has occurred in the target battery at the multiple historical moments based on the obtained multiple predicted negative electrode potential values.

[0007] Since there is a mapping relationship between the negative electrode potential value of the target battery at historical time and the actual state data at historical time, the negative electrode potential value at historical time is predicted using the actual state data. Furthermore, during the use of the target battery, the negative electrode potential value characterizes whether lithium plating occurs. Therefore, predicted negative electrode potential values ​​from multiple historical time points are used to determine whether lithium plating has occurred at multiple historical time points, thereby improving the accuracy of lithium plating detection for the target battery.

[0008] In one or more embodiments of this application, processing the data to be processed using a pre-trained potential prediction model to obtain the predicted negative electrode potential value of the target battery at multiple historical times includes: inputting the data to be processed into the pre-trained potential prediction model to obtain the predicted negative electrode potential value of the target battery at multiple historical times; wherein the actual specified potential value includes the actual negative electrode potential value; or, inputting the data to be processed into the pre-trained potential prediction model to obtain the predicted positive electrode potential value of the target battery at multiple historical times; wherein the actual specified potential value includes the actual positive electrode potential value; and combining the actual full-cell potential value of the target battery at multiple historical times and the predicted positive electrode potential value to calculate the predicted negative electrode potential value of the target battery at multiple historical times.

[0009] The negative electrode potential of the target battery can be directly predicted using a potential prediction model, or the positive electrode potential of the target battery can be predicted and combined with the total battery potential to calculate the negative electrode potential of the target battery, thereby improving the accuracy of the obtained negative electrode potential value and further improving the accuracy of lithium plating detection of the target battery.

[0010] In one or more embodiments of this application, determining whether the target battery has undergone lithium plating at multiple historical times based on the obtained multiple predicted negative electrode potential values ​​includes: calculating the average value of the multiple predicted negative electrode potential values; determining that the target battery has not undergone lithium plating at multiple historical times in response to the average value being greater than a preset threshold; and determining that the target battery has undergone lithium plating at multiple historical times in response to the average value not being greater than the preset threshold.

[0011] The average of multiple predicted negative electrode potential values ​​is obtained, which accurately characterizes the negative electrode potential value within a historical time period. Correspondingly, this average value is used to detect whether lithium plating will occur in the target battery at multiple historical moments, thereby improving the accuracy of lithium plating detection results.

[0012] In one or more embodiments of this application, the method further includes: responding to lithium plating occurring in the target battery at multiple historical moments, calculating the lithium plating side reaction current of the target battery at the multiple historical moments using the predicted negative electrode potential values ​​of the target battery at the multiple historical moments; calculating the total amount of lithium plating in the target battery during the historical time period by combining the lithium plating side reaction current of the target battery at the multiple historical moments; and calculating the product of the total amount of lithium plating and a preset ratio to obtain the amount of dead lithium plating in the target battery during the historical time period.

[0013] By predicting the negative electrode potential value, the accuracy of calculating the amount of dead lithium deposited in the target battery over a historical period can be improved, providing a precise basis for the management of the target battery.

[0014] In one or more embodiments of this application, the method further includes: in response to the target battery undergoing lithium plating at the plurality of historical moments, changing the charging conditions during the current charging process and continuing to charge the target battery according to the changed charging conditions; wherein the change includes: reducing the charging rate and / or increasing the battery temperature; and / or, in response to the target battery not undergoing lithium plating at the plurality of historical moments, maintaining the current charging conditions and continuing to charge the target battery; and / or, in response to the target battery undergoing lithium plating at the plurality of historical moments, reducing the output power of the target battery, or reducing the charging rate and / or increasing the battery temperature during the next power supply or charging process.

[0015] If lithium plating is determined to have occurred in the target battery within a historical timeframe, the current charging conditions are adjusted to control the target battery's charging according to these adjusted conditions, further reducing the risk of lithium plating during the entire charging process. Conversely, if lithium plating is not determined to have occurred in the target battery within a historical timeframe, the current charging conditions are maintained, and the target battery continues charging while completing the charging process as quickly as possible. If lithium plating occurs in the target battery at multiple historical points, the output power of the target battery is reduced during the next power supply cycle to mitigate the impact of lithium plating and improve the safety of the target battery's use.

[0016] In one or more embodiments of this application, the potential prediction model includes at least one of the following: convolutional neural network, dense connection network, recurrent neural network, long short-term memory neural network, attention mechanism model, and Transformer network.

[0017] Based on the above network, the accuracy of predicting the negative electrode potential value can be further improved, thereby improving the accuracy of lithium plating detection results.

[0018] In one or more embodiments of this application, the real-state data further includes: battery temperature; and / or, the data to be processed further includes: simulated potential values ​​of the target battery at multiple historical moments, the simulated potential values ​​at multiple historical moments being calculated using a pre-built first electrochemical mechanism model of the target battery, the simulated potential values ​​at multiple historical moments including at least one of the following: simulated full-cell potential value, simulated positive electrode potential value, and simulated negative electrode potential value; the potential prediction model is constructed as follows: trained based on the real-state data, simulated potential values, and real specified potential values ​​of the sample battery at multiple sample moments during historical charging; the simulated potential values ​​at multiple sample moments being calculated using a pre-built second electrochemical mechanism model of the sample battery.

[0019] Since the negative electrode potential of the target battery is affected by temperature, combining the battery temperature with historical data to predict the negative electrode potential further improves the accuracy of the predicted negative electrode potential. Furthermore, since the first electrochemical mechanism model predicts the actual potential value of the target battery at historical moments, combining the simulated potential value obtained from the first electrochemical mechanism model further improves the accuracy of the predicted negative electrode potential, thereby enhancing the accuracy of lithium plating detection results.

[0020] In one or more embodiments of this application, the first electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model; and / or, the second electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model.

[0021] Based on the above electrochemical model, the accuracy of the obtained simulated potential value is improved, thereby further improving the accuracy of the predicted negative electrode potential value and thus improving the accuracy of the lithium plating detection results.

[0022] In one or more embodiments of this application, the step of acquiring the target battery's data to be processed includes: acquiring the target battery's data to be processed when any one of a preset plurality of detection times is reached; wherein the detection interval between any two adjacent detection times is the same as the duration of the historical time period.

[0023] At each detection time, the system checks whether lithium plating has occurred in the target battery within the historical time period preceding that detection time; that is, it checks whether lithium plating occurred in the target battery between the previous detection time and the current detection time. In this way, the system determines whether lithium plating has occurred in a recent period based on the latest state of the target battery in real time, improving the real-time performance of the detection.

[0024] A second aspect of this application provides a lithium plating detection device, the device comprising: a data acquisition module configured to acquire data to be processed from a target battery; wherein the data to be processed includes: real state data of the target battery at multiple historical moments within a historical time period prior to the current moment during the current charging process, the real state data including: full battery potential value and charging current value; a prediction module configured to process the data to be processed using a pre-trained potential prediction model to obtain predicted negative electrode potential values ​​of the target battery at the multiple historical moments; wherein the potential prediction model is constructed by training based on real state data and real specified potential values ​​of sample batteries at multiple sample moments during historical charging processes, the real specified potential values ​​including real positive electrode potential value or real negative electrode potential value; and a judgment module configured to determine whether lithium plating has occurred in the target battery at the multiple historical moments based on the obtained multiple predicted negative electrode potential values.

[0025] Since there is a mapping relationship between the negative electrode potential value of the target battery at historical time and the actual state data at historical time, the negative electrode potential value at historical time is predicted using the actual state data. Furthermore, during the use of the target battery, the negative electrode potential value characterizes whether lithium plating occurs. Therefore, predicted negative electrode potential values ​​from multiple historical time points are used to determine whether lithium plating has occurred at multiple historical time points, thereby improving the accuracy of lithium plating detection for the target battery.

[0026] In one or more embodiments of this application, the prediction module is configured to input the data to be processed into a pre-trained potential prediction model to obtain the predicted negative electrode potential value of the target battery at multiple historical times; wherein the actual specified potential value includes the actual negative electrode potential value; or, input the data to be processed into a pre-trained potential prediction model to obtain the predicted positive electrode potential value of the target battery at multiple historical times; wherein the actual specified potential value includes the actual positive electrode potential value; and combine the actual full-cell potential value of the target battery at multiple historical times with the predicted positive electrode potential value to calculate the predicted negative electrode potential value of the target battery at multiple historical times.

[0027] The negative electrode potential of the target battery can be directly predicted using a potential prediction model, or the positive electrode potential of the target battery can be predicted and combined with the total battery potential to calculate the negative electrode potential of the target battery, thereby improving the accuracy of the obtained negative electrode potential value and further improving the accuracy of lithium plating detection of the target battery.

[0028] In one or more embodiments of this application, the determination module is configured to calculate the average of a plurality of predicted negative electrode potential values; in response to the average value being greater than a preset threshold, determine that the target battery has not undergone lithium plating at the plurality of historical times; in response to the average value not being greater than the preset threshold, determine that the target battery has undergone lithium plating at the plurality of historical times.

[0029] The average of multiple predicted negative electrode potential values ​​is obtained, which accurately characterizes the negative electrode potential value within a historical time period. Correspondingly, this average value is used to detect whether lithium plating will occur in the target battery at multiple historical moments, thereby improving the accuracy of lithium plating detection results.

[0030] In one or more embodiments of this application, the apparatus further includes: a dead lithium deposition calculation module, configured to, in response to lithium deposition occurring in the target battery at a plurality of historical times, calculate the lithium deposition side reaction current of the target battery at the plurality of historical times using the predicted negative electrode potential values ​​of the target battery at the plurality of historical times; calculate the total lithium deposition amount of the target battery within the historical time period by combining the lithium deposition side reaction current of the target battery at the plurality of historical times; and calculate the product of the total lithium deposition amount and a preset ratio to obtain the dead lithium deposition amount of the target battery within the historical time period.

[0031] By predicting the negative electrode potential value, the accuracy of calculating the amount of dead lithium deposited in the target battery over a historical period can be improved, providing a precise basis for the management of the target battery.

[0032] In one or more embodiments of this application, the device further includes: a control module configured to: in response to lithium plating occurring in the target battery at the plurality of historical moments, change the charging conditions during the current charging process and continue charging the target battery according to the changed charging conditions; wherein the change includes: reducing the charging rate and / or increasing the battery temperature; and / or, in response to the target battery not undergoing lithium plating at the plurality of historical moments, maintain the current charging conditions and continue charging the target battery; and / or, in response to the target battery undergoing lithium plating at the plurality of historical moments, reduce the output power of the target battery, or reduce the charging rate and / or increase the battery temperature during the next power supply or charging process.

[0033] If lithium plating is determined to have occurred in the target battery within a historical timeframe, the current charging conditions are adjusted to control the target battery's charging according to these adjusted conditions, further reducing the risk of lithium plating during the entire charging process. Conversely, if lithium plating is not determined to have occurred in the target battery within a historical timeframe, the current charging conditions are maintained, and the target battery continues charging while completing the charging process as quickly as possible. If lithium plating occurs in the target battery at multiple historical points, the output power of the target battery is reduced in the next power supply cycle; for example, by reducing the discharge current, to mitigate the impact of lithium plating and improve the battery's safety.

[0034] In one or more embodiments of this application, the potential prediction model includes at least one of the following: convolutional neural network, dense connection network, recurrent neural network, long short-term memory neural network, attention mechanism model, and Transformer network.

[0035] Based on the above network, the accuracy of predicting the negative electrode potential value can be further improved, thereby improving the accuracy of lithium plating detection results.

[0036] In one or more embodiments of this application, the real-state data further includes: battery temperature; and / or, the data to be processed further includes: simulated potential values ​​of the target battery at multiple historical moments, the simulated potential values ​​at multiple historical moments being calculated using a pre-built first electrochemical mechanism model of the target battery, the simulated potential values ​​at multiple historical moments including at least one of the following: simulated full-cell potential value, simulated positive electrode potential value, and simulated negative electrode potential value; the potential prediction model is constructed as follows: trained based on the real-state data, simulated potential values, and real specified potential values ​​of the sample battery at multiple sample moments during historical charging; the simulated potential values ​​at multiple sample moments being calculated using a pre-built second electrochemical mechanism model of the sample battery.

[0037] Since the negative electrode potential of the target battery is affected by temperature, combining the battery temperature with historical data to predict the negative electrode potential further improves the accuracy of the predicted negative electrode potential. Furthermore, since the first electrochemical mechanism model predicts the actual potential value of the target battery at historical moments, combining the simulated potential value obtained from the first electrochemical mechanism model further improves the accuracy of the predicted negative electrode potential, thereby enhancing the accuracy of lithium plating detection results.

[0038] In one or more embodiments of this application, the first electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model; and / or, the second electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model.

[0039] Based on the above electrochemical model, the accuracy of the obtained simulated potential value is improved, thereby further improving the accuracy of the predicted negative electrode potential value and thus improving the accuracy of the lithium plating detection results.

[0040] In one or more embodiments of this application, the data to be processed acquisition module is configured to acquire the data to be processed of the target battery when any one of a plurality of preset detection times is reached; wherein, the detection interval between any two adjacent detection times is the same as the duration of the historical time period.

[0041] At each detection time, the system checks whether lithium plating has occurred in the target battery within the historical time period preceding that detection time. In other words, it checks whether lithium plating has occurred between the previous and current detection times. This allows for real-time determination of whether lithium plating has occurred in the most recent period based on the latest state of the target battery, improving the real-time performance of the detection.

[0042] A third aspect of this application provides an electronic device configured to perform any of the methods described in the first aspect above.

[0043] A fourth aspect of this application provides a battery pack, the battery pack including a battery module and the electronic device described in the third aspect above.

[0044] The fifth aspect of this application provides an energy storage product, which includes a battery module and the electronic device described in the third aspect above.

[0045] A sixth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements any of the methods described in the first aspect above.

[0046] The seventh aspect of this application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform any of the methods described in the first aspect above.

[0047] Beneficial effects of the embodiments in this application:

[0048] The lithium plating detection method provided in this application embodiment acquires the target battery's data to be processed. The data to be processed includes: the target battery's real state data at multiple historical moments before the current moment during the current charging process, the real state data including: the full battery potential value and the charging current value; then, a pre-trained potential prediction model is used to process the data to be processed to obtain the predicted negative electrode potential value of the target battery at multiple historical moments; the potential prediction model is constructed as follows: it is trained based on the real state data and real specified potential value of the sample battery at multiple sample moments during the historical charging process, the real specified potential value being the real positive electrode potential value or the real negative electrode potential value; furthermore, based on the obtained multiple predicted negative electrode potential values, it is determined whether lithium plating has occurred in the target battery at multiple historical moments.

[0049] Since there is a mapping relationship between the negative electrode potential value of the target battery at historical time and the actual state data at historical time, the negative electrode potential value at historical time is predicted using the actual state data. Furthermore, during the use of the target battery, the negative electrode potential value characterizes whether lithium plating occurs. Therefore, predicted negative electrode potential values ​​from multiple historical time points are used to determine whether lithium plating has occurred at multiple historical time points, thereby improving the accuracy of lithium plating detection for the target battery. Attached Figure Description

[0050] The accompanying drawings, which are provided to further illustrate this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application.

[0051] Figure 1 is a flowchart of a lithium plating detection method provided in an embodiment of this application;

[0052] Figure 2A is a schematic diagram of the actual negative electrode potential value of a battery (referred to as the battery to be used) provided in an embodiment of this application;

[0053] Figure 2B is a schematic diagram of the predicted negative electrode potential value of a battery to be used, obtained by using a trained potential prediction model, according to an embodiment of this application.

[0054] Figure 3 is a flowchart of another lithium plating detection method provided in an embodiment of this application;

[0055] Figure 4 is a flowchart of another lithium plating detection method provided in an embodiment of this application;

[0056] Figure 5 is a structural diagram of a lithium plating detection device provided in an embodiment of this application. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments in this application are within the scope of protection of this application.

[0058] Lithium plating in lithium batteries refers to the deposition of lithium ions on the negative electrode surface during charging, forming metallic lithium. Lithium plating occurs when the battery temperature is low or under other improper charging conditions. Lithium plating can lead to the formation of lithium dendrites inside the battery, potentially puncturing the separator and causing internal short circuits, or even thermal runaway and safety accidents. Furthermore, lithium plating reduces the number of lithium ions, resulting in battery capacity degradation and shortened battery life.

[0059] To improve the accuracy of lithium plating detection in batteries, embodiments of this application provide a lithium plating detection method, electronic device, battery pack, energy storage product, apparatus, equipment, and medium.

[0060] The lithium plating detection method provided in the embodiments of this application will be described below. In some embodiments of this application, the method is applied to electronic devices that provide computing services, such as a battery management system (BMS) for a battery pack, an energy management system (EMS) for an energy storage product, or a server, including local servers and cloud servers. It is understood that a battery pack is a product that includes battery modules and a battery management system electrically connected to the battery modules, wherein a battery module includes at least one battery. When a battery module includes multiple batteries, the electrical connection between the batteries includes series, parallel, or hybrid connection. In this application, hybrid connection indicates that the electrical connection between batteries includes series and parallel connections.

[0061] As shown in Figure 1, this application provides a flowchart of a lithium plating detection method, which includes the following steps:

[0062] S101: Obtain the data to be processed from the target battery.

[0063] The data to be processed includes: real-time state data of the target battery at multiple historical moments prior to the current moment during this charging process. This real-time state data includes: full battery potential value and charging current value.

[0064] In some embodiments of this application, the full-cell potential value of the battery represents the potential value of the positive electrode of the battery relative to the negative electrode, the positive electrode potential value represents the potential value of the positive electrode of the battery relative to the reference electrode, and the negative electrode potential value represents the potential value of the negative electrode of the battery relative to the reference electrode.

[0065] Understandably, the target battery is a battery that is currently being charged. The electronic device detects the target battery based on the lithium plating detection method provided in the embodiments of this application to determine whether lithium plating occurs during the charging process. At this time, the electronic device is a battery management system in a battery pack containing the target battery, or an energy management system in an energy storage product containing the target battery.

[0066] During a complete charging process, the electronic device executes steps S101-S103 once or multiple times. After acquiring the real-state data, the electronic device determines whether lithium plating has occurred in the target battery based on the lithium plating detection method provided in this application embodiment. In this application, the duration between two acquisitions of real-state data is defined as the detection interval. As a specific example of this application, in some practical application scenarios, the electronic device periodically performs detection according to a fixed detection interval.

[0067] In some implementations, the multiple historical moments in step S101 are obtained by uniformly sampling the moments within the historical time period before the current moment, and the duration between each two adjacent historical moments is the same, which is 1 second, 10 seconds, 100 seconds or 1000 seconds, etc. This application does not impose a specific limitation on the duration.

[0068] The number of these historical moments is consistent with the number of sample moments used in step S102 to train the potential prediction model. Correspondingly, the duration between any two adjacent sample moments is the same, and this duration is consistent with the duration between two adjacent historical moments. Thus, when predicting the negative electrode potential value of the target battery, the data input to the potential prediction model is consistent with the method of obtaining the input data during the training of the potential prediction model, improving the accuracy of predicting the negative electrode potential value of the target battery using the potential prediction model.

[0069] S102: The pre-trained potential prediction model is used to process the data to be processed to obtain the predicted negative electrode potential values ​​of the target battery at multiple historical moments.

[0070] The potential prediction model is constructed by training on the actual state data of the sample battery at multiple sample times during historical charging, and the actual specified potential values ​​at those multiple sample times. The actual specified potential values ​​include either the actual positive electrode potential value or the actual negative electrode potential value. The potential prediction model reflects the mapping relationship between the battery's state data at historical times and the actual specified potential values, and is used to predict the negative electrode potential value of the target battery at historical times.

[0071] To further improve the robustness of the potential prediction model and adapt to battery states and operating environments, in one or more embodiments of this application, the actual state data of sample batteries at sample times under various application scenarios, as well as the actual specified potential values ​​at sample times, are obtained. Battery states and operating environments include: battery aging state, ambient temperature, ambient humidity, charging scheme, irradiation, etc., resulting in various application scenarios. Charging schemes include constant current charging, constant current-constant voltage charging, multi-stage constant current charging, and pulse charging, etc.

[0072] To continuously detect whether lithium plating has occurred in the target battery throughout a complete charging process, the duration of the aforementioned historical time period is the same as the duration of the aforementioned detection interval. In one or more embodiments of this application, when any of the preset multiple detection times is reached, the electronic device acquires the target battery's data to be processed and executes steps S101-S103. Each time a detection time is reached, it detects whether lithium plating has occurred in the target battery during the historical time period prior to that time; that is, it detects whether lithium plating has occurred in the target battery between the previous detection time and the current detection time. Thus, it determines whether lithium plating has occurred in the most recent period based on the latest state of the target battery in real time, improving the real-time performance of the detection.

[0073] For example, during the charging process of the target battery, the duration of the historical time period is 1 minute, and the real state data of the target battery is collected once every second. That is, multiple historical moments are the moments corresponding to each second within one minute before the current moment. Accordingly, the number of historical moments is 60, and each time it is detected whether lithium plating has occurred in the target battery within one minute before the current moment.

[0074] That is, according to the above example, if the detection interval is 1 minute, when the charging time reaches one minute during the charging process of the target battery, the potential prediction model is used to process the real state data of the target battery every second in the first minute to obtain the predicted negative electrode potential value in the first minute. Then, according to step S103, it is determined whether lithium plating occurs in the target battery in the first minute.

[0075] Accordingly, when the charging time reaches two minutes, the potential prediction model is used to process the real state data of the target battery for each second in the second minute to obtain the predicted negative electrode potential value in the second minute. Then, according to step S103, it is determined whether lithium plating occurs in the target battery in the second minute.

[0076] This process continues until the target battery's charging process is complete. For example, the target battery is fully charged, or the charging process is interrupted.

[0077] Understandably, since the above detection method uses real-time data of the battery charging process, which is obtained through the battery management system or energy management system, it enables real-time lithium plating detection without disassembling the battery pack, achieving non-destructive testing. The lithium plating detection method of this application embodiment achieves online, real-time, and non-destructive diagnosis of lithium plating in two-electrode batteries. Furthermore, due to real-time detection, manufacturers can select different time periods for detection in different scenarios according to actual needs, adapting to both rapid lithium plating detection and periodic inspections.

[0078] S103: Based on the obtained multiple predicted negative electrode potential values, determine whether lithium plating occurred in the target battery at multiple historical moments.

[0079] In some embodiments of this application, the negative electrode potential value is used to characterize whether lithium plating occurs in the target battery. The negative electrode potential value is predicted using real state data from multiple historical moments to determine whether lithium plating occurs in the target battery at multiple historical moments.

[0080] Because there are predicted negative electrode potential values ​​at multiple historical moments, in some embodiments of this application, the average value of multiple predicted negative electrode potential values ​​is calculated. Based on the relationship between this average value and a preset threshold, it is determined whether lithium plating has occurred in the target battery at multiple historical moments. For example, if the average value is greater than the preset threshold, it is determined that lithium plating has not occurred in the target battery at multiple historical moments; if the average value is not greater than the preset threshold, it is determined that lithium plating has occurred in the target battery at multiple historical moments. As some examples of this application, the preset threshold is 0V, or, in order to improve the fault tolerance of detection, the preset threshold is set to other values ​​near 0V, such as 0.01V, 0.03V, 0.05V, or 0.1V. In this way, by setting a preset threshold and comparing the predicted negative electrode potential value with the preset threshold, real-time, online, and non-destructive lithium plating detection is achieved, improving the safety and reliability of lithium-ion batteries and helping to reduce the risk of internal short circuits and thermal runaway in the early stages of battery development.

[0081] The average of multiple predicted negative electrode potential values ​​obtained more accurately characterizes the negative electrode potential value within a historical time period. Correspondingly, this average value is used to detect whether lithium plating will occur in the target battery at multiple historical moments, thereby improving the accuracy of lithium plating detection results.

[0082] In some embodiments of this application, the maximum and minimum values ​​among multiple predicted negative electrode potential values ​​are removed, and the average value of the remaining predicted negative electrode potential values ​​is calculated to obtain the above average value.

[0083] Because lithium plating lowers the thermal stability boundary of a battery, which represents the safe temperature limit during battery use, in one or more embodiments of this application, after determining that lithium plating has occurred in the target battery at multiple historical moments, the output power of the target battery is reduced, for example, by reducing the discharge current of the target battery, so that the operating temperature of the target battery is lower than the aforementioned safe temperature limit, thereby mitigating the impact of lithium plating and improving the safety of the target battery in use.

[0084] In one or more embodiments of this application, after determining that lithium plating has occurred in the target battery at multiple historical moments, the charging conditions are changed during the current charging process, and the target battery continues to be charged according to the changed charging conditions. The adjustments include reducing the charging rate and / or increasing the battery temperature.

[0085] In some embodiments of this application, the battery temperature is characterized by ambient temperature. Increasing the ambient temperature raises the battery temperature through heat transfer, thereby reducing the subsequent lithium plating degree of the target battery. It is understood that in other embodiments of this application, the battery temperature is characterized by the battery's own temperature.

[0086] In some embodiments of this application, electrical devices (e.g., energy storage products, electric vehicles) include heating components for heating batteries. In these electrical devices, the heating components heat the battery to raise its temperature, thereby reducing the degree of subsequent lithium plating in the target battery.

[0087] Furthermore, reducing the charging rate can decrease the subsequent degree of lithium plating in the target battery. Therefore, once it is determined that lithium plating occurred in the target battery at multiple historical moments, the charging conditions are adjusted by reducing the charging rate and / or increasing the battery temperature, thereby reducing the subsequent degree of lithium plating in the target battery.

[0088] In one or more embodiments of this application, if it is determined that the target battery has not undergone lithium plating at multiple historical moments, the current charging conditions are kept unchanged and the target battery is continued to be charged.

[0089] It is understandable that the initial charging rate of the target battery is set according to actual needs. In one or more embodiments of this application, the charging rate reduction operation may be performed multiple times during a single charging process, following the above-described process. This is because, in order to shorten the charging time, in some embodiments, the target battery starts charging at the maximum charging rate, shortening the charging time of the target battery. Simultaneously, combined with the lithium plating detection method of this application, the charging conditions are adjusted to reduce the risk of lithium plating in the target battery.

[0090] Understandably, for electrical devices without battery temperature control components, such as headphones, mobile phones, and tablets, charging conditions are adjusted by reducing the charging rate. For electrical devices with battery temperature control components, such as energy storage products and electric vehicles, charging conditions are adjusted by raising the battery temperature through the battery temperature control component, or by reducing the charging rate, or by raising the battery temperature and reducing the charging rate.

[0091] In some implementations, the charging rate is reduced according to a preset gradient, for example, reducing the charging rate to 95%, 90%, 85%, or 80% of the current charging rate. In other implementations, the battery temperature is increased according to a preset gradient, for example, increasing the battery temperature to 150%, 200%, or 250% of the current battery temperature.

[0092] In other implementations, a preset charging rate is reduced from the current charging rate, for example, by 0.2C, 0.5C, or 1C. In still other implementations, a preset temperature is increased from the current battery temperature, for example, by 5°C, 8°C, 10°C, 15°C, or 20°C.

[0093] Regarding step S102 above, in some embodiments of this application, it is implemented in one of the following ways:

[0094] Method 1: Predict the negative electrode potential value of the target battery at multiple historical moments to obtain the predicted negative electrode potential value of the target battery at multiple historical moments. That is, the output data of the potential prediction model represents the negative electrode potential value.

[0095] In this approach, the potential prediction model is constructed by training based on the real state data of the sample battery at multiple sample times during historical charging processes, as well as the real negative electrode potential value at multiple sample times.

[0096] The actual state data of the sample battery at multiple sample times are input into the potential prediction model of the initial structure to obtain the predicted negative electrode potential value of the sample battery at multiple sample times. Then, based on the difference between the obtained predicted negative electrode potential value and the actual negative electrode potential value of the sample battery at multiple sample times, the loss value is calculated, and the model parameters of the potential prediction model of the initial structure are adjusted based on the loss value until the convergence condition is reached, and the trained potential prediction model is obtained.

[0097] Referring to Figures 2A and 2B, Figure 2A is a schematic diagram of the actual negative electrode potential value of a battery (referred to as the battery to be utilized) provided in an embodiment of this application, and Figure 2B is a schematic diagram of the predicted negative electrode potential value of the battery to be utilized obtained using a trained potential prediction model. In Figures 2A and 2B, the horizontal axis represents the charging time in seconds (S), and the vertical axis represents the negative electrode potential value in volts (V). As can be seen from Figures 2A and 2B, the difference between the predicted negative electrode potential value and the actual negative electrode potential value provided by the potential prediction model in this embodiment of the application is small, and the accuracy of the negative electrode potential value predicted by the potential prediction model is high.

[0098] Method 2: Predict the positive electrode potential value of the target battery at multiple historical moments to obtain the predicted positive electrode potential value of the target battery at multiple historical moments. Then, combine the actual full cell potential value of the target battery at multiple historical moments to calculate the predicted negative electrode potential value of the target battery at multiple historical moments. That is, the output data of the potential prediction model represents the positive electrode potential value.

[0099] For example, the difference between the predicted positive electrode potential value at each historical moment and the actual full cell potential value at that historical moment is calculated to obtain the predicted negative electrode potential value at that historical moment.

[0100] In this approach, the potential prediction model is constructed by training based on the real state data of the sample battery at multiple sample times during the historical charging process, as well as the real positive electrode potential value at multiple sample times.

[0101] The actual state data of the sample battery at multiple sample times are input into the potential prediction model of the initial structure to obtain the predicted positive electrode potential value of the sample battery at multiple sample times. Then, based on the difference between the obtained predicted positive electrode potential value and the actual positive electrode potential value of the sample battery at multiple sample times, the loss value is calculated, and the model parameters of the potential prediction model of the initial structure are adjusted based on the loss value until the convergence condition is reached, and the trained potential prediction model is obtained.

[0102] Regarding step S102 above, in one or more embodiments of this application, the full battery potential value and charging current value of the target battery at multiple historical moments are input into a pre-trained potential prediction model to obtain the predicted negative electrode potential value of the target battery at multiple historical moments.

[0103] In one or more embodiments of this application, in addition to using the full-cell potential value and charging current value of the target battery at multiple historical moments, other information is combined to predict the negative electrode potential value of the target battery at multiple historical moments, so as to further improve the accuracy of predicting the negative electrode potential value.

[0104] In some embodiments of this application, the true state data of the target battery at each historical moment further includes the temperature of the target battery at that historical moment. Since the negative electrode potential value of the target battery is affected by temperature, combining the temperature of the target battery to predict the negative electrode potential value at historical moments further improves the accuracy of the predicted negative electrode potential value. It is understood that the temperatures of sample batteries at multiple sample moments are also used when training the potential prediction model.

[0105] And / or,

[0106] Using a pre-constructed electrochemical mechanism model of the target battery (referred to as the first electrochemical mechanism model), simulated potential values ​​of the target battery at multiple historical time points are calculated. These simulated potential values ​​include at least one of the following: simulated full-cell potential, simulated positive electrode potential, and simulated negative electrode potential. The actual state data of the target battery at multiple historical time points, along with the calculated simulated potential values ​​at these points, are input into a pre-trained potential prediction model to calculate the predicted negative electrode potential values ​​of the target battery at these multiple historical time points. In other words, the data to be processed for the target battery also includes: simulated potential values ​​of the target battery at multiple historical time points.

[0107] An electrochemical mechanism model is understood as a simulation model of a battery. Technicians construct a corresponding electrochemical mechanism model (i.e., the first electrochemical mechanism model) based on the target battery's dimensions, materials, and other specifications. In some embodiments of this application, after constructing the first electrochemical mechanism model, the target battery's current voltage, remaining charge, and current charging conditions are input into the first electrochemical mechanism model for simulation calculations to obtain the target battery's potential values ​​at historical moments (i.e., simulated potential values).

[0108] Since the first electrochemical mechanism model predicts the potential value of the target battery at historical time points, combining the simulated potential value obtained from the first electrochemical mechanism model further improves the accuracy of the predicted negative electrode potential value. Understandably, the simulated potential values ​​of the sample battery at multiple sample time points were also used when training the potential prediction model.

[0109] In some embodiments of this application, a potential prediction model is trained based on the real state data, real potential value, and simulated potential value of the sample battery at multiple sample moments during historical charging. The real state data and simulated potential value of the sample battery at multiple sample moments are input into the potential prediction model of the initial structure to obtain the predicted potential value of the sample battery at multiple sample moments. Then, based on the difference between the obtained predicted potential value and the real specified potential value of the sample battery at multiple sample moments, a loss value is calculated, and the model parameters of the potential prediction model of the initial structure are adjusted based on the loss value until the convergence condition is met, resulting in a trained potential prediction model.

[0110] The simulated potential values ​​of the sample battery are calculated using the electrochemical mechanism model of the sample battery (referred to as the second electrochemical mechanism model). Technicians construct the corresponding electrochemical mechanism model (i.e., the second electrochemical mechanism model) according to the specifications of the sample battery, such as its size and materials. The calculation method for the simulated potential values ​​of the sample battery at multiple sample times is the same as the description of calculating the simulated potential values ​​of the target battery at historical times in the above embodiments. The simulated potential values ​​of the sample battery at multiple sample times are of the same type as the simulated potential values ​​of the target battery at multiple historical times. That is, if the simulated potential values ​​of the sample battery at multiple sample times include the simulated full-cell potential value, then the simulated potential values ​​of the target battery at multiple historical times also include the simulated full-cell potential value; if the simulated potential values ​​of the sample battery at multiple sample times include the simulated negative electrode potential value, then the simulated potential values ​​of the target battery at multiple historical times also include the simulated negative electrode potential value; if the simulated potential values ​​of the sample battery at multiple sample times include the simulated positive electrode potential value, then the simulated potential values ​​of the target battery at multiple historical times also include the simulated positive electrode potential value.

[0111] In real-world scenarios, the target battery and the sample battery have the same specifications. Consequently, the second electrochemical mechanism model is the same as the first electrochemical mechanism model. Technicians match the target battery, selecting the same positive electrode material, negative electrode material, electrolyte, separator, and other materials, and using the same manufacturing process to assemble a three-electrode battery, which is the sample battery.

[0112] Similarly, to further improve the robustness of the potential prediction model and adapt it to battery states and operating environments, the actual negative electrode potential values ​​of sample batteries at sample times under various application scenarios were obtained. The simulated potential values ​​of these sample batteries under these different application scenarios were then calculated using a second electrochemical mechanism model to train the potential prediction model of the initial structure. Battery states and operating environments included: battery aging state, ambient temperature, ambient humidity, charging scheme, irradiation, etc., resulting in various application scenarios. Charging schemes included constant current charging, constant current-constant voltage charging, multi-stage constant current charging, and pulse charging.

[0113] In one or more embodiments of this application, the potential prediction model mentioned above is a model based on a deep learning algorithm. For example, the potential prediction model includes one or more of the following: convolutional neural network, dense connection network, recurrent neural network, long short-term memory neural network, attention mechanism model, or Transformer.

[0114] By establishing a deep learning algorithm model, the simulation potential values ​​of the two-electrode full cell potential, current, temperature, and electrochemical mechanism model are established, and the mapping relationship between them and the actual positive and negative electrode potential values ​​is established, thereby detecting the lithium plating status of the battery.

[0115] In one or more embodiments of this application, the first electrochemical mechanism model and the second electrochemical mechanism model mentioned above are electrochemical mechanism models that include lithium plating side reactions and thermal field models, such as single-particle models, quasi-two-dimensional models, or multi-dimensional multi-field electrochemical models.

[0116] In one or more embodiments of this application, a target battery already in use in a real-world scenario is used as a sample battery after a period of use. Relevant data is collected according to the methods described in the above embodiments, and the potential prediction model mentioned above is incrementally trained based on the collected data. In this way, the model in this application is updated using battery data from real-world operating conditions, continuously improving the accuracy of model predictions.

[0117] In one or more embodiments of this application, after determining that lithium plating occurred in the target battery at multiple historical moments, the amount of dead lithium plating at the multiple historical moments is further determined.

[0118] After obtaining the predicted negative electrode potential value at each historical moment, the lithium plating side reaction current density is calculated using the Bavo's equation or the Tafel equation. The Bavo's equation is shown in the following formula:

[0119] j is the lithium plating side reaction current density, k0 is the reaction rate constant, and c e Where is the electrolyte concentration, F is the Faraday constant, α is the negative electrode transfer coefficient, R is the gas constant, T is the battery temperature, and η is the negative electrode potential.

[0120] After calculating the lithium plating side reaction current density at each historical moment, the lithium plating side reaction current at each historical moment is calculated using the following formula.

[0121] i is the lithium plating side reaction current, r is the radius of the negative electrode particles, L is twice the length of the negative electrode sheet, W is the width of the negative electrode sheet, D is the thickness of the negative electrode sheet, and epss is the solid phase porosity of the negative electrode.

[0122] By combining the lithium plating side reaction currents at each historical moment, the total amount of lithium plating within the historical time period is calculated using the ampere-hour integration method.

[0123] A and B are the start and end times of the historical time period, respectively. Q loss1 This represents the total amount of lithium deposited due to lithium deposition side reactions.

[0124] Total lithium deposition includes reversible lithium deposition and dead lithium deposition. Reversible lithium can be reintegrated into the battery during charge and discharge without causing battery capacity loss. Dead lithium, however, is difficult to re-participate in the reaction, causing battery capacity loss and affecting battery life. In addition, dendritic dead lithium formed under certain operating conditions can puncture the separator, causing internal short circuit faults and even thermal runaway faults.

[0125] After calculating the total lithium deposition, the predetermined ratio of dead lithium deposition to the total lithium deposition is obtained. Then, the product of the total lithium deposition and this ratio is calculated to obtain the dead lithium deposition.

[0126] Batteries under various operating conditions, including different ambient temperatures, charge / discharge rates, aging conditions, and humidity levels, were pre-acquired and disassembled. The amount of dead lithium deposited within a time period under each operating condition was determined using mass spectrometry titration and nuclear magnetic resonance (NMR) techniques. Furthermore, for each operating condition, the total lithium deposited within the same time period under that condition was calculated using the same method as described above. Finally, the ratio of dead lithium deposited to total lithium deposited under the same operating condition was calculated.

[0127] Understandably, after obtaining the total lithium deposition amount of the target battery, the ratio of the target battery's operating conditions is obtained, and the product of this ratio and the total lithium deposition amount of the target battery is calculated to obtain the dead lithium deposition amount of the target battery, as shown in the following formula:

[0128] Q loss2 This represents the amount of dead lithium deposited, and β is the ratio of dead lithium deposited to the total amount of lithium deposited.

[0129] Based on the above processing, by calculating the amount of dead lithium deposited and predicting the negative electrode potential, online, in-situ, non-destructive lithium deposition detection and quantification can be achieved.

[0130] Referring to Figure 3, Figure 3 is a flowchart of a lithium plating detection method provided in an embodiment of this application.

[0131] The flowchart is divided into two main parts: offline training and online application.

[0132] The offline training process includes the following steps:

[0133] Step 1: Obtain three-electrode battery data. Corresponding to the above embodiment, this involves obtaining the actual state data of the sample battery at multiple sample times, as well as the specified actual potential values ​​at those multiple sample times.

[0134] Step 2: Construct an electrochemical mechanism model that includes lithium plating side reactions and a thermal field model, and generate simulation data. Corresponding to the above embodiments, a second electrochemical mechanism model is constructed, and the simulated potential values ​​of the sample battery at multiple sample times are calculated.

[0135] Step 3: Segment the three-electrode battery data and simulation data into data segments. That is, the acquired data is segmented according to the detection interval in the above embodiment to obtain input data and label data for training the potential prediction model. The data includes: real-state data of the sample battery at multiple sample times, specified real potential values, and simulated potential values. The real-state data and simulated potential values ​​serve as input data, and the specified real potential values ​​serve as the corresponding label data.

[0136] Step 4: Construct a deep learning algorithm to establish a mapping relationship between two-electrode battery data, simulation data, and negative electrode potential values. That is, train a potential prediction model using the segmented data fragments.

[0137] The online application includes the following steps:

[0138] Step 5: Collect actual battery operation data. That is, in the above embodiment, obtain the data to be processed from the target battery.

[0139] Step Six: Use the trained deep learning algorithm. That is, use the pre-trained potential prediction model to process the data to be processed of the target battery to obtain the predicted negative electrode potential values ​​of the target battery at multiple historical moments.

[0140] Step 7: Determine the lithium plating state and calculate the amount of dead lithium deposited. That is, based on the obtained predicted negative electrode potential value, determine whether lithium plating has occurred in the target battery at multiple historical moments, and calculate the amount of dead lithium deposited.

[0141] Step 8: Accumulate actual operating data of the target battery in one or more scenarios.

[0142] Step 9: Update the deep learning algorithm online.

[0143] Steps eight and nine involve incremental training of the model in the above embodiments.

[0144] Referring to Figure 4, which is a flowchart of the lithium plating detection method provided in an embodiment of this application, the method includes the following steps:

[0145] S401: Acquire the voltage, current, and temperature data of the target battery.

[0146] That is, in the above embodiments, the battery management system or energy management system is used to obtain the real state data of the target battery at multiple historical moments.

[0147] S402: Obtain the simulated full cell potential, simulated negative electrode potential, and simulated positive electrode potential obtained from the electrochemical mechanism model calculation.

[0148] That is, to obtain the simulated potential values ​​of the target battery at multiple historical moments calculated by the first electrochemical mechanism model.

[0149] S403: Predicted negative electrode potential value.

[0150] That is, by using a pre-trained potential prediction model to process the real state data and simulated potential values ​​of the target battery, the predicted negative electrode potential values ​​of the target battery at multiple historical moments are obtained.

[0151] S404: Determine whether the predicted negative electrode potential value is greater than the system preset value. If it is greater, execute S401 and S402; if it is not greater, execute S405.

[0152] S405: Determines that lithium plating has occurred in the target battery, outputs the lithium plating status, and calculates the amount of dead lithium deposited.

[0153] That is, based on the obtained predicted negative electrode potential value, it is determined whether lithium plating has occurred in the target battery at multiple historical moments. If lithium plating has not occurred, the target battery continues to be charged under the current charging conditions, and lithium plating detection of the target battery continues according to the detection interval. If lithium plating has occurred, the lithium plating status is output, and the amount of dead lithium plating is calculated.

[0154] In addition, if lithium plating occurs, the charging conditions of the target battery during this charging process shall be adjusted in accordance with the method described in the above embodiments.

[0155] Based on the same inventive concept, this application also provides a lithium plating detection device. Referring to Figure 5, which is a structural diagram of a lithium plating detection device provided in this application, the device includes: a data acquisition module 501, which acquires data to be processed from the target battery; the data to be processed includes: real state data of the target battery at multiple historical moments before the current moment during the current charging process, the real state data including: full battery potential value and charging current value; a prediction module 502, which processes the data to be processed using a pre-trained potential prediction model to obtain predicted negative electrode potential values ​​of the target battery at multiple historical moments. The potential prediction model is constructed as follows: it is trained based on the real state data and real specified potential values ​​of the sample battery at multiple sample moments during historical charging; the real specified potential values ​​include the real positive electrode potential value, or, include the real negative electrode potential value; and a judgment module 503, which judges whether lithium plating has occurred in the target battery at multiple historical moments based on the obtained multiple predicted negative electrode potential values.

[0156] Since there is a mapping relationship between the negative electrode potential value of the target battery at historical time and the actual state data at historical time, the negative electrode potential value at historical time is predicted using the actual state data. Furthermore, during the use of the target battery, the negative electrode potential value characterizes whether lithium plating occurs. Therefore, predicted negative electrode potential values ​​from multiple historical time points are used to determine whether lithium plating has occurred at multiple historical time points, thereby improving the accuracy of lithium plating detection for the target battery.

[0157] In one or more embodiments of this application, the prediction module 502 inputs the data to be processed into a pre-trained potential prediction model to obtain the predicted negative electrode potential value of the target battery at multiple historical moments; the actual specified potential value is the actual negative electrode potential value; or, the prediction module 502 inputs the data to be processed into a pre-trained potential prediction model to obtain the predicted positive electrode potential value of the target battery at multiple historical moments; the actual specified potential value is the actual positive electrode potential value; then, combining the actual full-cell potential value of the target battery at multiple historical moments and the predicted positive electrode potential value, the predicted negative electrode potential value of the target battery at multiple historical moments is calculated.

[0158] The negative electrode potential of the target battery can be directly predicted using a potential prediction model, or the positive electrode potential of the target battery can be predicted and combined with the full cell potential of the target battery to calculate the predicted negative electrode potential value of the target battery, thereby improving the accuracy of the obtained negative electrode potential value and further improving the accuracy of lithium plating detection of the target battery.

[0159] In one or more embodiments of this application, the determination module 503 calculates the average value of multiple predicted negative electrode potential values; in response to the average value being greater than a preset threshold, it determines that the target battery has not undergone lithium plating at multiple historical moments; in response to the average value not being greater than the preset threshold, it determines that the target battery has undergone lithium plating at multiple historical moments.

[0160] The average of multiple predicted negative electrode potential values ​​is obtained, which accurately characterizes the negative electrode potential value within a historical time period. Correspondingly, this average value is used to detect whether lithium plating will occur in the target battery at multiple historical moments, thereby improving the accuracy of lithium plating detection results.

[0161] In one or more embodiments of this application, the above-mentioned apparatus further includes: a dead lithium deposition amount calculation module, which, in response to lithium deposition of the target battery at multiple historical moments, calculates the lithium deposition side reaction current of the target battery at multiple historical moments using the predicted negative electrode potential values ​​of the target battery at multiple historical moments; then, combines the lithium deposition side reaction current of the target battery at multiple historical moments to calculate the total lithium deposition amount of the target battery within the historical time period; and further, calculates the product of the total lithium deposition amount and a preset ratio to obtain the dead lithium deposition amount of the target battery within the historical time period.

[0162] By predicting the negative electrode potential value, the accuracy of calculating the amount of dead lithium deposited in the target battery over a historical period can be improved, providing a more precise basis for the management of the target battery.

[0163] In one or more embodiments of this application, the above-described apparatus further includes: a control module, which, in response to lithium plating occurring in the target battery at multiple historical moments, changes the charging conditions during the current and / or subsequent charging processes, and charges the target battery according to the changed charging conditions. Changing the charging conditions includes: reducing the charging rate and / or increasing the battery temperature. And / or, the control module, in response to no lithium plating occurring in the target battery at multiple historical moments, maintains the current charging conditions and continues charging the target battery. And / or, the control module, in response to lithium plating occurring in the target battery at multiple historical moments, reduces the output power of the target battery during the next power supply process.

[0164] If lithium plating is determined to have occurred in the target battery within a historical timeframe, the current charging conditions are adjusted to control the target battery's charging according to the adjusted conditions, thereby reducing the risk of lithium plating during the charging process. Conversely, if lithium plating is not determined to have occurred in the target battery within a historical timeframe, the current charging conditions are maintained, and the target battery continues charging while completing the charging process as quickly as possible. If lithium plating occurs in the target battery at multiple historical points, the output power of the target battery is reduced during the next charging and / or discharging process, or the adjusted charging conditions are applied to mitigate the impact of lithium plating and improve the safety of the target battery.

[0165] In one or more embodiments of this application, the potential prediction model includes at least one of the following: convolutional neural network, dense connection network, recurrent neural network, long short-term memory neural network, attention mechanism model, and Transformer network.

[0166] Based on the above network, the accuracy of predicting the negative electrode potential value can be further improved, thereby improving the accuracy of lithium plating detection results.

[0167] In one or more embodiments of this application, the real state data of the target battery further includes: battery temperature; and / or, the data to be processed further includes: simulated potential values ​​of the target battery at multiple historical moments, the simulated potential values ​​at multiple historical moments being calculated using a pre-built first electrochemical mechanism model of the target battery, the simulated potential values ​​at multiple historical moments including at least one of the following: simulated full-cell potential value, simulated positive electrode potential value, and simulated negative electrode potential value; the potential prediction model is constructed as follows: trained based on the real state data, simulated potential values, and real specified potential values ​​of the sample battery at multiple sample moments during historical charging; the simulated potential values ​​at multiple sample moments are calculated using a pre-built second electrochemical mechanism model of the sample battery.

[0168] Since the negative electrode potential of the target battery is affected by temperature, combining the battery temperature with historical data to predict the negative electrode potential further improves the accuracy of the predicted negative electrode potential. Furthermore, since the first electrochemical mechanism model predicts the actual potential value of the target battery at historical moments, combining the simulated potential value obtained from the first electrochemical mechanism model further improves the accuracy of the predicted negative electrode potential, thereby enhancing the accuracy of lithium plating detection results.

[0169] In one or more embodiments of this application, the first electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model; and / or, the second electrochemical mechanism model is a single-particle model, a quasi-two-dimensional model, or a multidimensional multi-field electrochemical model.

[0170] Based on the above electrochemical model, the accuracy of the obtained simulated potential value is improved, and the accuracy of the predicted negative electrode potential value is further improved, so as to improve the accuracy of lithium plating detection results.

[0171] In one or more embodiments of this application, the data acquisition module 501 acquires the data to be processed of the target battery when any one of a plurality of preset detection times is reached. In some embodiments of this application, the detection interval between any two adjacent detection times is the same as the duration of a historical time period.

[0172] At each detection time, the system checks whether lithium plating has occurred in the target battery within the historical time period preceding that detection time. In other words, it checks whether lithium plating has occurred between the previous and current detection times. This allows for real-time determination of whether lithium plating has occurred in the most recent period based on the latest state of the target battery, improving the real-time performance of the detection.

[0173] Based on the same inventive concept, this application also provides an electronic device for performing any of the lithium plating detection methods described in the above embodiments.

[0174] Based on the same inventive concept, this application also provides a battery pack, including a battery module and the electronic device described in the above embodiments.

[0175] Based on the same inventive concept, this application also provides an energy storage product, which includes a battery module and an electronic device, wherein the electronic device is used to perform any of the lithium plating detection methods described in the above embodiments.

[0176] Based on the same inventive concept, in another embodiment provided in this application, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the steps of any of the above-described lithium plating detection methods.

[0177] Based on the same inventive concept, in another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the lithium plating detection methods described above.

[0178] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0179] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0180] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for electronic devices, battery packs, charging devices, energy storage products, computer-readable storage media, and computer program products, since they are basically similar to the method embodiments, the descriptions are relatively simple, and relevant parts can be referred to the descriptions of the method embodiments.

[0181] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application are included within the scope of protection of this application.

Claims

1. A method for detecting lithium plating, the method comprising: Acquire the data to be processed for the target battery; wherein, the data to be processed includes: the real state data of the target battery at multiple historical moments in the historical time period before the current moment during this charging process, and the real state data includes: the full battery potential value and the charging current value; The data to be processed is processed using a pre-trained potential prediction model to obtain the predicted negative electrode potential values ​​of the target battery at multiple historical moments. The potential prediction model is constructed as follows: it is trained based on the real state data of the sample battery at multiple sample times during the historical charging process and the real specified potential value, wherein the real specified potential value includes the real positive electrode potential value or the real negative electrode potential value. Based on the obtained multiple predicted negative electrode potential values, it is determined whether the target battery has undergone lithium plating at the multiple historical times.

2. The method of claim 1, wherein, The step of processing the data to be processed using a pre-trained potential prediction model to obtain the predicted negative electrode potential values ​​of the target battery at multiple historical moments includes: The data to be processed is input into a pre-trained potential prediction model to obtain the predicted negative electrode potential values ​​of the target battery at multiple historical moments. The actual specified potential value includes the actual negative electrode potential value; or, The data to be processed is input into a pre-trained potential prediction model to obtain the predicted positive electrode potential values ​​of the target battery at multiple historical moments. The actual specified potential value includes the actual positive electrode potential value; By combining the actual full-cell potential values ​​of the target battery at multiple historical moments with the predicted positive electrode potential values, the predicted negative electrode potential values ​​of the target battery at multiple historical moments are calculated.

3. The method of claim 1, wherein, The step of determining whether lithium plating has occurred in the target battery at the multiple historical moments based on the obtained multiple predicted negative electrode potential values ​​includes: The average of the multiple predicted negative electrode potential values ​​obtained by calculation; In response to the average value being greater than a preset threshold, it is determined that the target battery has not undergone lithium plating at the multiple historical moments; In response to the average value not being greater than the preset threshold, it is determined that the target battery has undergone lithium plating at the multiple historical moments.

4. The method of claim 1, wherein, The method further includes: In response to lithium plating occurring in the target battery at multiple historical moments, the lithium plating side reaction current of the target battery at multiple historical moments is calculated using the predicted negative electrode potential values ​​of the target battery at multiple historical moments. By combining the lithium plating side reaction current of the target battery at multiple historical moments, the total amount of lithium plating of the target battery during the historical time period is calculated. The product of the total lithium deposition amount and the preset ratio is calculated to obtain the dead lithium deposition amount of the target battery during the historical time period.

5. The method of claim 1, wherein, The method further includes: In response to the lithium plating of the target battery at the multiple historical moments, the charging conditions are changed during the current charging process, and the target battery is continued to be charged according to the changed charging conditions. The methods of change include: reducing the charging rate and / or increasing the battery temperature; And / or, In response to the fact that the target battery has not undergone lithium plating at the multiple historical moments, the current charging conditions are kept unchanged, and the target battery continues to be charged; And / or, In response to lithium plating occurring in the target battery at the aforementioned historical moments, the output power of the target battery is reduced, or the charging rate is reduced and / or the battery temperature is increased during the next power supply or charging process.

6. The method of claim 1, wherein, The potential prediction model includes at least one of the following: convolutional neural network, dense connection network, recurrent neural network, long short-term memory neural network, attention mechanism model, and Transformer network.

7. The method according to any one of claims 1 to 6, wherein, The actual status data also includes: battery temperature; And / or, The data to be processed also includes: the simulated potential values ​​of the target battery at multiple historical moments, wherein the simulated potential values ​​at multiple historical moments are calculated using a pre-built first electrochemical mechanism model of the target battery, and the simulated potential values ​​at multiple historical moments include at least one of the following: simulated full cell potential value, simulated positive electrode potential value, and simulated negative electrode potential value; The potential prediction model is constructed by training based on the real state data, simulated potential value and real specified potential value of the sample battery at multiple sample times during the historical charging process. The simulated potential values ​​at the multiple sample times were calculated using a pre-built second electrochemical mechanism model of the sample cells.

8. The method of claim 7, wherein, The first electrochemical mechanism model is: a single-particle model, a quasi-two-dimensional model, or a multi-dimensional multi-field electrochemical model; And / or, The second electrochemical mechanism model is: a single-particle model, a quasi-two-dimensional model, or a multi-dimensional multi-field electrochemical model.

9. The method according to any one of claims 1 to 6, wherein, The process of acquiring the target battery's unprocessed data includes: When any of the preset multiple detection times is reached, the data to be processed of the target battery is acquired; The detection interval between any two adjacent detection times is the same as the duration of the historical time period.

10. An electronic device for performing the method of any one of claims 1 to 9.

11. A battery pack, the battery pack comprising a battery module and the electronic device as claimed in claim 10.

12. An energy storage product, the energy storage product comprising a battery module and the electronic device as described in claim 10.

13. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1 to 9.

Images