Battery state of health prediction model training method, apparatus, and computer device

Abstract

The application relates to a battery health state prediction model training method and device and a computer device. The method comprises the following steps: acquiring a theoretical prediction model and reference operation data of a battery; inputting the reference operation data into the theoretical prediction model to obtain a first health state prediction result; determining a differential voltage analysis curve based on the reference operation data; determining a second health state prediction result based on the differential voltage analysis curve; determining training data based on the first health state prediction result and the second health state prediction result; training an initial prediction model by using the training data to obtain a target battery health state prediction model. By using the method, the accuracy of the training data and the prediction model can be ensured when the battery health state is predicted, the prediction result is more accurate and the precision is higher, so that the battery health state can be effectively monitored, and reliable data support is provided for managing and maintaining the battery.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a method, apparatus and computer equipment for training a battery health state prediction model. Background Technology

[0002] With the development of battery technology, battery applications are becoming increasingly widespread, and battery health is receiving widespread attention. Battery Management Systems (BMS) have emerged to address this need. They work together to monitor the status of energy storage batteries, intelligently manage and maintain each battery cell, extend battery life, and monitor battery status. State of Health (SOH) is an indicator of the energy storage battery's capacity, health, and performance status, playing a crucial role in the battery management system.

[0003] In related technologies, SOH prediction usually involves building empirical models of battery aging or using AI algorithms to predict SOH based on existing battery aging data in the laboratory. However, the data sources for this prediction method are not updated in a timely manner, making it difficult to be equivalent to the actual battery conditions. Furthermore, the empirical models are relatively fixed and inflexible, resulting in poor accuracy and low precision in SOH predictions, making it difficult to apply to actual predictions.

[0004] Therefore, there is an urgent need in related technologies for a way to improve the accuracy and precision of battery state of health (SOH) prediction. Summary of the Invention

[0005] Therefore, it is necessary to provide a battery health state prediction model training method, apparatus, and computer equipment that can improve the accuracy and precision of battery health state prediction in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for training a battery health state prediction model. The method includes:

[0007] Obtain theoretical prediction models and reference operating data for the battery;

[0008] The reference operating data is input into the theoretical prediction model to obtain the first health state prediction result;

[0009] The differential voltage analysis curve is determined based on the reference operating data, and the second health state prediction result is determined based on the differential voltage analysis curve.

[0010] Training data is determined based on the first health state prediction result and the second health state prediction result.

[0011] The initial prediction model is trained using the training data to obtain the target battery health status prediction model.

[0012] Optionally, in one embodiment of this application, obtaining the theoretical prediction model includes:

[0013] Initial parameters are determined using an optimization algorithm;

[0014] A theoretical prediction model is obtained based on the initial parameters.

[0015] Optionally, in one embodiment of this application, the step of determining the differential voltage analysis curve based on the reference operating data and determining the second health state prediction result based on the differential voltage analysis curve includes:

[0016] Based on the first health status prediction result, determine the battery data in the reference operating data that meet the preset requirements for battery health status.

[0017] Optionally, in one embodiment of this application, after determining the battery data in the reference operating data that meets the preset requirements based on the first health status prediction result, the method further includes:

[0018] Identify battery data in the battery data where the charging current is less than a preset threshold.

[0019] Optionally, in one embodiment of this application, determining the differential voltage analysis curve based on the reference operating data, and determining the second health state prediction result based on the differential voltage analysis curve, includes:

[0020] Based on the differential voltage analysis curve, a characteristic peak is determined, and the difference between the power value corresponding to the characteristic peak and the power value at the end of a full charge is determined.

[0021] The second health status prediction result is determined based on the power difference and the preset mapping relationship.

[0022] Optionally, in one embodiment of this application, determining the training data based on the first health status prediction result and the second health status prediction result includes:

[0023] The first health status prediction result and the second health status prediction result are matched, and the reference running data corresponding to the successfully matched health status prediction result is used as the training data.

[0024] Optionally, in one embodiment of this application, matching the first health status prediction result and the second health status prediction result includes:

[0025] Calculate the interquartile range based on the second health status prediction results;

[0026] The average value is calculated based on the first health status prediction result and the second health status prediction result;

[0027] If the interquartile range and the average value meet the preset conditions, then the first health state prediction result and the second health state prediction result are successfully matched.

[0028] Secondly, this application also provides a battery health state prediction model training device. The device includes:

[0029] The theoretical prediction model acquisition module is used to acquire the theoretical prediction model and the reference operating data of the battery;

[0030] The first health status prediction module is used to input the reference operating data into the theoretical prediction model to obtain the first health status prediction result;

[0031] The second health status prediction module is used to determine the differential voltage analysis curve based on the reference operating data, and to determine the second health status prediction result based on the differential voltage analysis curve.

[0032] The training data determination module determines training data based on the first health state prediction result and the second health state prediction result.

[0033] The target battery health status prediction model determination module is used to train the initial prediction model using the training data to obtain the target battery health status prediction model.

[0034] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the methods described in the various embodiments above.

[0035] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the methods described in the various embodiments above.

[0036] The aforementioned battery health state prediction model training method, apparatus, computer equipment, and storage medium acquire a theoretical prediction model and reference operating data of the battery. Then, the reference operating data is input into the theoretical prediction model to obtain a first health state prediction result. Next, a differential voltage analysis curve is determined based on the reference operating data, and a second health state prediction result is determined based on the differential voltage analysis curve. Then, training data is determined based on the first and second health state prediction results. Finally, the initial prediction model is trained using the training data to obtain a target battery health state prediction model. In other words, when predicting battery health state, the prediction results of a theoretical prediction model based on theoretical principles are compared with those of a prediction model based on actual data, resulting in highly consistent prediction results. This ensures the accuracy of the training data and the prediction model, making the prediction results more accurate and precise, thereby effectively monitoring battery health state and providing reliable data support for battery management and maintenance. Attached Figure Description

[0037] Figure 1 This is a diagram illustrating the application environment of a battery health state prediction model training method in one embodiment.

[0038] Figure 2 This is a flowchart illustrating a battery health status prediction model training method in one embodiment;

[0039] Figure 3 This is a schematic diagram of the differential voltage analysis curve in one embodiment;

[0040] Figure 4 This is a schematic diagram illustrating the specific process of training a battery health status prediction model in one embodiment.

[0041] Figure 5 This is a structural block diagram of a battery health state prediction model training device in one embodiment;

[0042] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0044] The battery health state prediction model training method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located in the cloud or on other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0045] In one embodiment, such as Figure 2 As shown, a method for training a battery health state prediction model is provided, which can be applied to... Figure 1 Taking the server in the example, the following steps are included:

[0046] S201: Obtain theoretical prediction models and reference operating data for the battery.

[0047] Under the control of the Battery Management System (BMS), and with proper battery use, the health status of most batteries follows a certain battery aging pattern. Based on this battery aging pattern and laboratory test data of battery health status, a theoretical prediction model for battery health status can be obtained. Most battery health statuses can be predicted using this theoretical prediction model.

[0048] In one embodiment of this application, an empirical formula for battery capacity health is selected as the theoretical prediction model for prediction. The empirical formula for battery capacity health is as follows:

[0049]

[0050] Where A1, A2, A3, B1, B2, B3 are parameters to be identified, T represents the ambient temperature, t represents the service time, and Ah represents the cumulative charge throughput.

[0051] At the same time, the actual operating conditions of the car's battery are obtained as reference operating data.

[0052] S203: Input the reference operating data into the theoretical prediction model to obtain the first health state prediction result.

[0053] In this embodiment of the application, the reference operating data is input into the theoretical prediction model, that is, the actual operating conditions of the car's battery are used as the input value, and the prediction result of the car's battery health status is calculated by the battery health status calculation formula, which is the first health status prediction result.

[0054] S205: Determine the differential voltage analysis curve based on the reference operating data, and determine the second health state prediction result based on the differential voltage analysis curve.

[0055] In this embodiment, based on the reference operating data, a differential voltage analysis curve is determined using the charged capacity and voltage data. This differential voltage analysis curve is an indicator of electrode degradation, and the peak shift and peak capacity changes of the curve are useful indicators for understanding the capacity decay of the electrodes within the battery. The second health state prediction result of the battery is determined based on the curve characteristics of the differential voltage analysis curve.

[0056] S207: Determine training data based on the first health state prediction result and the second health state prediction result.

[0057] In this embodiment of the application, the first health status prediction result and the second health status prediction result are compared with each other to determine whether they are consistent, thereby determining whether the second health status prediction result is accurate. If the prediction result is accurate, the corresponding part of the reference running data is determined as the training data for training the prediction model.

[0058] S209: The initial prediction model is trained using the training data to obtain the target battery health status prediction model.

[0059] In this embodiment of the application, after determining the above-mentioned training data, the initial prediction model is trained using the training data. The initial prediction model is the AI ​​prediction model used in most predictions on the market. Finally, a target battery health status prediction model applicable to most actual operating conditions of automobiles is obtained.

[0060] The battery health state prediction model training method provided in this application obtains a theoretical prediction model and reference operating data of the battery. Then, the reference operating data is input into the theoretical prediction model to obtain a first health state prediction result. Next, a differential voltage analysis curve is determined based on the reference operating data, and a second health state prediction result is determined based on the differential voltage analysis curve. Then, training data is determined based on the first and second health state prediction results. Finally, the initial prediction model is trained using the training data to obtain a target battery health state prediction model. In other words, when predicting battery health state, the prediction results of the theoretical prediction model based on theoretical principles are compared with the prediction results of the prediction model based on actual data, obtaining prediction results with high consistency. This ensures the accuracy of the training data and the prediction model, making the prediction results more accurate and precise, thereby effectively monitoring battery health state and providing reliable data support for battery management and maintenance.

[0061] In one embodiment of this application, obtaining the theoretical prediction model includes:

[0062] S301: Initial parameters are determined using an optimization algorithm.

[0063] In one embodiment of this application, an optimization algorithm is used to update the aforementioned parameters A1, A2, A3, B1, B2, B3 to be identified, to obtain initial parameters. In practical applications, the optimization algorithm can be genetic algorithm parameter optimization, cross-validation parameter optimization, quantum particle swarm optimization parameter optimization, or particle swarm optimization parameter optimization, etc., and is not limited thereto.

[0064] S303: Obtain the theoretical prediction model based on the initial parameters.

[0065] In one embodiment of this application, after the initial parameters are determined, a theoretical prediction model can be obtained. Based on the theoretical prediction model and reference operating data, a first health state prediction result can be obtained.

[0066] In this embodiment, the initial parameters of the theoretical prediction model are determined by an optimization algorithm, which can improve the prediction effect of the theoretical prediction model and make the prediction results more accurate.

[0067] In one embodiment of this application, the step of determining the differential voltage analysis curve based on the reference operating data, and determining the second health state prediction result based on the differential voltage analysis curve, includes:

[0068] Based on the first health status prediction result, determine the battery data in the reference operating data that meet the preset requirements for battery health status.

[0069] In one embodiment of this application, after obtaining the first health state prediction result, the reference operating data is classified according to the first health state prediction result. Reference operating data with consistent battery health states are grouped into the same set, which is the battery data that meets preset requirements. For example, reference operating data with a first health state in the range of [X-1%, X+1%] are considered as a set with consistent health states, where X takes values ​​of 100%, 95%, 90%, 85%, 80%, 75%, and 70%, respectively.

[0070] In this embodiment, by classifying and statistically analyzing battery data with consistent battery health status, the classification of reference operating data can be made more detailed. Using the detailed reference operating data as battery data makes the battery health status prediction results more accurate.

[0071] In one embodiment of this application, after determining the battery data in the reference operating data that meets the preset requirements based on the first health status prediction result, the method further includes:

[0072] Identify battery data in the battery data where the charging current is less than a preset threshold.

[0073] In one embodiment of this application, battery data in which the charging current is less than a preset threshold within a fixed time period is selected from the battery data, wherein the fixed time period is one week and the preset threshold is 0.1C.

[0074] In this embodiment, by selecting battery data whose battery health status meets the preset requirements from the reference operating data, the change in battery health status of a fully charged battery with a small current in a short period of time is negligible, thereby ensuring the prediction accuracy of the second health status prediction result.

[0075] In one embodiment of this application, determining the differential voltage analysis curve based on the reference operating data, and determining the second health state prediction result based on the differential voltage analysis curve, includes:

[0076] S601: Based on the differential voltage analysis curve, determine the characteristic peak, and determine the difference between the power value corresponding to the characteristic peak and the power value at the end of full charging.

[0077] S603: Determine the second health status prediction result based on the power difference and the preset mapping relationship.

[0078] In one embodiment of this application, such as Figure 3 The figure shows a differential voltage analysis curve obtained based on the data in the reference operating data, determined by the charge and voltage data. Characteristic peaks are determined based on these differential voltage analysis curves. These characteristic peaks reflect the phase transition of the active material during lithium insertion and delithiation processes. Figure 3 After determining the inflection point P1, the difference between the power value corresponding to the characteristic peak and the power value of a full charge is determined, that is, the power charged from the inflection point P1 to the full charge state is determined, which is △Q in the figure.

[0079] In one embodiment of this application, the power difference is used as a feature to determine the second health state prediction result based on a preset mapping relationship. The preset mapping relationship is a mapping relationship obtained from the standard battery cycle aging test. In a specific application, the standard battery cycle aging test is the enterprise battery aging DV test.

[0080] In this embodiment, the second health status prediction result of the reference operating data is obtained by predicting the reference operating data based on the differential voltage analysis curve. This can obtain accurate health status prediction results based on the actual capacity of different batteries, avoiding the impact of directly using uniform estimated battery capacity data on the accuracy of battery health status prediction results.

[0081] In one embodiment of this application, determining the training data based on the first health status prediction result and the second health status prediction result includes:

[0082] S701: Match the first health status prediction result and the second health status prediction result, and use the reference running data corresponding to the successfully matched health status prediction result as the training data.

[0083] In one embodiment of this application, the first health status prediction result and the second health status prediction result are matched with each other to determine whether the first health status prediction result and the second health status prediction result are accurate. If the match is successful, the reference running data corresponding to the health status prediction result is used as the training data to train the AI ​​prediction model.

[0084] In this embodiment, by matching the first health status prediction result and the second health status prediction result and using the reference running data corresponding to the successfully matched health status prediction result as the training data, the health status prediction result corresponding to the preferred battery data can be made more accurate.

[0085] To determine whether the prediction result is accurate, it can be judged through some matching data. Specifically, in one embodiment of this application, matching the first health state prediction result and the second health state prediction result includes:

[0086] S801: Calculate the interquartile range based on the second health status prediction result.

[0087] S803: Calculate the average value based on the first health status prediction result and the second health status prediction result.

[0088] S805: If the interquartile range and the average value meet the preset conditions, then the first health state prediction result and the second health state prediction result are successfully matched.

[0089] In one embodiment of this application, the interquartile range of the second health status prediction results is calculated. Specifically, the interquartile range includes two calculation methods: one for results divisible by 4 and the other for results not divisible by 4. For example, if there are n second health status prediction results that are exactly divisible by 4, the n numbers are first sorted in ascending order. Then, assuming the first quartile Q1 is in the n / 4th position and the third quartile Q3 is in the 3*n / 4th position, the interquartile range is equal to Q3 - Q1. If there are n second health status prediction results that are not divisible by 4, the n numbers are first sorted in ascending order. Then, assuming X1 is in the floor(n / 4th position) and X2 is in the ceil(n / 4th position), the first quartile Q1 = X1*( 1 - n / 4 - floor(n / 4)) + X2 * (n / 4 - floor(n / 4)); X3 is located at floor(3*n / 4), and X4 is located at ceil(3*n / 4). Therefore, the third quartile Q3 = X3 * (1 - 3*n / 4 - floor(3*n / 4)) + X4 * (3*n / 4 - floor(3*n / 4)). Thus, the interquartile range is equal to Q3 - Q1, where floor() represents the floor function and ceil() represents the floor function.

[0090] In one embodiment of this application, the average of the first health status prediction result and the second health status prediction result is calculated. In a specific application, the average of the first health status prediction result and the second health status prediction result is the value obtained by summing the results and dividing by the total number.

[0091] In one embodiment of this application, if the interquartile range and the average value meet a preset condition, the first health state prediction result and the second health state prediction result are successfully matched, indicating that the first health state prediction result and the second health state prediction result are close and accurate. The preset condition is that the interquartile range of the second health state prediction result is less than 2%, and the absolute value of the difference between the average value of the first health state prediction result and the average value of the second health state prediction result is less than 2%. In other embodiments, the preset condition can be set according to actual needs, as long as it reflects the consistency between the first health state prediction result and the second health state prediction result; no specific limitation is made here.

[0092] In this embodiment, by matching the first health status prediction result and the second health status prediction result, a large amount of labeled data with high battery health status prediction accuracy can be obtained, which can be used to train the battery health status prediction AI model.

[0093] The following is a specific embodiment illustrating the detailed process of the battery health state prediction model training method, such as... Figure 4As shown, in one embodiment of this application, firstly, in step 401, an optimization algorithm is used to determine initial parameters, and a theoretical prediction model is obtained based on the initial parameters. Specifically, the theoretical prediction model is a battery capacity health empirical formula model in the battery management system (BMS). Next, in step 403, reference operating data of the battery is obtained, i.e., the actual operating conditions of the vehicle's battery are obtained. Then, in step 405, the reference operating data is input into the theoretical prediction model to obtain a first health state prediction result. Specifically, the service time t and cumulative charge throughput Ah of each battery at different temperatures are statistically analyzed, and the first health state prediction result is calculated according to the aforementioned battery capacity health empirical formula model.

[0094] Next, in step 407, based on the first health status prediction result, determine the battery data in the reference operating data whose battery health status meets the preset requirements. In step 409, determine the battery data in the battery data whose charging current is less than a preset threshold. Specifically, classify the reference operating data according to the first health status prediction result, group the reference operating data with consistent battery health status into the same set, and select the battery data whose charging current is less than the preset threshold within a fixed time period from the set.

[0095] Next, in step 4011, a differential voltage analysis curve is determined based on the reference operating data, and a second health state prediction result is obtained based on the differential voltage analysis curve. Specifically, a characteristic peak is determined based on the differential voltage analysis curve, and the difference between the power value corresponding to the characteristic peak and the power value at the end of full charging is determined. The second health state prediction result is determined based on the power difference and a preset mapping relationship.

[0096] Next, in step 4013, the first health state prediction result and the second health state prediction result are matched to determine if they match. In step 4015, if the match is successful, the reference running data corresponding to the successfully matched health state prediction result is used as training data; if the match fails, the initial parameters are redefined, and prediction is performed again. Specifically, whether the interquartile range of the second health state prediction result and the average value of the first and second health state prediction results meet preset conditions is used as the criterion for determining whether the match is successful. Finally, in step 4017, the initial prediction model is trained using the training data to obtain the target battery health state prediction model.

[0097] In one embodiment of this application, after obtaining the target battery health status prediction model, the actual battery operating condition data of the target vehicle whose state to be determined is input into the target battery health status prediction model. The type of the actual battery operating condition data is consistent with the battery data used as input during training. In specific applications, it includes the ambient temperature of the battery, service time, and cumulative charge throughput. Then, the battery health status prediction result of the target vehicle is obtained.

[0098] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0099] Based on the same inventive concept, this application also provides a battery health state prediction model training device for implementing the battery health state prediction model training method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more battery health state prediction model training device embodiments provided below can be found in the limitations of the battery health state prediction model training method described above, and will not be repeated here.

[0100] In one embodiment, such as Figure 5 As shown, a battery health state prediction model training device 500 is provided, including: a theoretical prediction model acquisition module 501, a first health state prediction module 503, a second health state prediction module 505, a training data determination module 507, and a target battery health state prediction model determination module 509, wherein:

[0101] The theoretical prediction model acquisition module 501 is used to acquire the theoretical prediction model and the reference operating data of the battery.

[0102] The first health status prediction module 503 is used to input the reference operating data into the theoretical prediction model to obtain the first health status prediction result.

[0103] The second health status prediction module 505 is used to determine the differential voltage analysis curve based on the reference operating data, and to determine the second health status prediction result based on the differential voltage analysis curve.

[0104] The training data determination module 507 determines training data based on the first health state prediction result and the second health state prediction result.

[0105] The target battery health status prediction model determination module 509 is used to train the initial prediction model using the training data to obtain the target battery health status prediction model.

[0106] Optionally, in one embodiment of this application, the theoretical prediction model acquisition module is further configured to:

[0107] Initial parameters are determined using an optimization algorithm;

[0108] A theoretical prediction model is obtained based on the initial parameters.

[0109] The battery health status prediction model training device further includes a battery data determination module.

[0110] The battery data determination module is used to determine, based on the first health status prediction result, battery data in the reference operating data that meets preset requirements for battery health status.

[0111] Optionally, in one embodiment of this application, the battery data determination module is further configured to determine battery data in which the charging current is less than a preset threshold.

[0112] Optionally, in one embodiment of this application, the second health status prediction module is further configured to:

[0113] Based on the differential voltage analysis curve, a characteristic peak is determined, and the difference between the power value corresponding to the characteristic peak and the power value at the end of a full charge is determined.

[0114] The second health status prediction result is determined based on the power difference and the preset mapping relationship.

[0115] Optionally, in one embodiment of this application, the training data determination module is further configured to:

[0116] The first health status prediction result and the second health status prediction result are matched, and the reference running data corresponding to the successfully matched health status prediction result is used as the training data.

[0117] Optionally, in one embodiment of this application, the training data determination module is further configured to:

[0118] Calculate the interquartile range based on the second health status prediction results;

[0119] The average value is calculated based on the first health status prediction result and the second health status prediction result;

[0120] If the interquartile range and the average value meet the preset conditions, then the first health state prediction result and the second health state prediction result are successfully matched.

[0121] Each module in the aforementioned battery health status prediction model training device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0122] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores the training data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a battery health state prediction model training method.

[0123] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0124] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0125] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0126] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0127] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for training a battery health status prediction model, characterized in that, The method includes: Obtain theoretical prediction models and reference operating data for the battery; The reference operating data is input into the theoretical prediction model to obtain the first health state prediction result; The differential voltage analysis curve is determined based on the reference operating data, and the second health state prediction result is determined based on the differential voltage analysis curve. Training data is determined based on the first health state prediction result and the second health state prediction result. The initial prediction model is trained using the training data to obtain the target battery health status prediction model; The step of determining the training data based on the first health state prediction result and the second health state prediction result includes: The first health status prediction result and the second health status prediction result are matched, and the reference running data corresponding to the successfully matched health status prediction result is used as the training data.

2. The method according to claim 1, characterized in that, The obtained theoretical prediction model includes: Initial parameters are determined using an optimization algorithm; A theoretical prediction model is obtained based on the initial parameters.

3. The method according to claim 1, characterized in that, Before determining the differential voltage analysis curve based on the reference operating data, and before determining the second health state prediction result based on the differential voltage analysis curve, the following steps are included: Based on the first health status prediction result, determine the battery data in the reference operating data that meet the preset requirements for battery health status.

4. The method according to claim 3, characterized in that, After determining the battery data in the reference operating data that meet the preset requirements based on the first health status prediction result, the method further includes: Identify battery data in the battery data where the charging current is less than a preset threshold.

5. The method according to claim 1, characterized in that, The step of determining the differential voltage analysis curve based on the reference operating data, and determining the second health state prediction result based on the differential voltage analysis curve, includes: Based on the differential voltage analysis curve, a characteristic peak is determined, and the difference between the power value corresponding to the characteristic peak and the power value at the end of a full charge is determined. The second health status prediction result is determined based on the power difference and the preset mapping relationship.

6. The method according to claim 1, characterized in that, The step of matching the first health status prediction result and the second health status prediction result includes: Calculate the interquartile range based on the second health status prediction results; The average value is calculated based on the first health status prediction result and the second health status prediction result; If the interquartile range and the average value meet the preset conditions, then the first health state prediction result and the second health state prediction result are successfully matched.

7. A battery health status prediction model training device, characterized in that, The device includes: The theoretical prediction model acquisition module is used to acquire the theoretical prediction model and the reference operating data of the battery; The first health status prediction module is used to input the reference operating data into the theoretical prediction model to obtain the first health status prediction result; The second health status prediction module is used to determine the differential voltage analysis curve based on the reference operating data, and to determine the second health status prediction result based on the differential voltage analysis curve. The training data determination module determines training data based on the first health state prediction result and the second health state prediction result. The target battery health status prediction model determination module is used to train the initial prediction model using the training data to obtain the target battery health status prediction model. The training data determination module is further configured to match the first health state prediction result and the second health state prediction result, and use the reference running data corresponding to the successfully matched health state prediction result as the training data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images