Battery cell capacity prediction method and device, computer equipment and storage medium

Abstract

The invention provides a cell capacity prediction method and device, computer equipment and a storage medium. The method comprises the following steps: collecting full-capacity test process data of a sample battery; dividing the full-capacity test process data into multiple parts of simplified capacity data; inputting each part of simplified capacity data into a capacity prediction model for capacity prediction to obtain a plurality of predicted capacities; comparing each predicted capacity with the actual capacity to obtain an error value between each predicted capacity and the actual capacity; and selecting the test process corresponding to the prediction capacity with the minimum error value, and determining the test process as a simplified capacity test process. According to the method, the capacity of the battery can be tested according to the selected simplified capacity testing process, so that a full-capacity testing process is not needed for testing, the battery does not need to be fully charged and discharged, the testing process is greatly saved, and the capacity testing efficiency is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery cell capacity prediction, and in particular to a battery cell capacity prediction method, device, computer equipment and storage medium. Background Art

[0002] With the continuous advancement of energy storage and electric vehicle technologies, lithium-ion batteries, as core components, are becoming increasingly important. Improving the manufacturing process for these batteries has become a key approach to improving battery quality and reducing production costs. However, in actual production, the lithium-ion battery capacity grading process involves full charge and discharge, which is energy-intensive and time-consuming. This results in long production cycles, low battery channel utilization, and limited production efficiency. Therefore, finding an efficient and accurate method to predict battery cell capacity has become a pressing issue for companies.

[0003] In the existing technology, the models used for full-capacity result prediction mainly include linear regression models, nonlinear regression models, convolutional neural network models, etc. The data used for full-capacity result prediction mainly include front-end process data (coating, winding, shelling), mid-stage process data (baking, liquid injection, formation and static), and dynamic curves such as IC curves and dQ / dV curves obtained from the above data processing. However, the prediction method based on voltage dynamic curves is difficult to collect, analyze and predict for each battery product due to the difficulty in collecting voltage information in the formation and capacity division processes, and cannot be applied to the production line; the prediction method based on linear regression models and nonlinear regression models is not applicable to the production line because the algorithm can only obtain empirical formula parameters and cannot effectively identify specific battery cells; the data-driven capacity prediction model lacks the input of temperature parameters, and capacity and temperature are strongly correlated. Models without temperature indicators cannot effectively predict batteries at various temperatures in the production line.

[0004] Therefore, there is an urgent need for a cell capacity prediction method that can be applied to production lines and is more applicable and efficient. Summary of the Invention

[0005] Based on this, it is necessary to provide a battery cell capacity prediction method, device, computer equipment and storage medium to address the above technical problems.

[0006] A method for predicting battery cell capacity, comprising:

[0007] Collect full capacity test process data of sample batteries;

[0008] Dividing the full capacity test process data into a plurality of simplified capacity data;

[0009] Input each simplified capacity data into a pre-established capacity forecasting model to perform capacity forecasting and obtain multiple forecast capacities;

[0010] Comparing each of the predicted capacities with the actual capacities to obtain an error value between each of the predicted capacities and the actual capacities;

[0011] The simplified capacity data corresponding to the predicted capacity with the smallest error value is selected as the simplified data to be selected, and the test process corresponding to the simplified data to be selected is determined as the simplified capacity test process.

[0012] In one embodiment, the full-capacity test process data includes capacity data and temperature data; and the step of collecting the full-capacity test process data of the sample battery includes:

[0013] Collect full capacity test process data of sample batteries;

[0014] A preset temperature compensation calculation formula is used to perform temperature compensation on the capacity data in the full-capacity test process data, and the full-capacity capacity of the battery at a preset temperature is corrected and used as the capacity data in the full-capacity test process data.

[0015] In one embodiment, before the step of collecting full capacity test process data of the sample battery, the step further includes:

[0016] Collect process data, formation process data and full-capacity test process data of sample batteries;

[0017] The process data, formation process data and full capacity test process data of the sample battery are input into a prediction learning model for training to obtain the capacity prediction model.

[0018] In one embodiment, the process data, the formation process data, the full capacity test process data, and the simplified capacity data include at least one of the following:

[0019] The process data includes: core weight and battery weight;

[0020] The formation process data include: formation start time, formation end time, formation start voltage, formation end voltage, formation start capacity, formation end capacity, formation start temperature and formation end temperature;

[0021] The full-capacity test process data includes: test start time, test end time, test start voltage, test end voltage, test start capacity, test end capacity, test start temperature and test end temperature;

[0022] The simplified capacity data includes the start time of the simplified process, the end time of the simplified process, the start voltage of the simplified process, the end voltage of the simplified process, the start capacity of the simplified process, the end capacity of the simplified process, the start temperature of the simplified process, and the end temperature of the simplified process.

[0023] In one embodiment, the step of inputting the process data, formation process data, and full-capacity test process data of the sample battery into a prediction learning model for training to obtain the capacity prediction model includes:

[0024] Dividing the full capacity test process data into a plurality of simplified capacity data;

[0025] The process data, formation process data, and simplified capacity data of the sample batteries were cleaned, and samples with missing values ​​and duplicate samples were deleted.

[0026] Calculate the capacity value, voltage difference, average temperature, and average current of each process flow to obtain a data set for each process flow;

[0027] Randomly divide the dataset into training and test sets;

[0028] The training sets and test sets of different units are standardized to obtain standardized training sets and standardized test sets;

[0029] Detecting anomalies in the standardized training set and the standardized test set using statistical methods or machine learning algorithms to obtain outliers, and processing the outliers;

[0030] Inputting the standardized training set into a prediction learning model for training to obtain a capacity prediction model;

[0031] The capacity prediction model is tested using the standardized test set, and parameters of the capacity prediction model are calibrated according to the test results.

[0032] In one embodiment, it further includes:

[0033] Obtaining each process flow in the full-capacity test process data;

[0034] Calculating the correlation between the test results of each process flow and the full-capacity test process data;

[0035] Different process flows are selected for combination, and parameters of each process flow are modified according to the correlation.

[0036] In one embodiment, it further includes:

[0037] Obtaining each process flow in the full-capacity test process data;

[0038] Select different process flows for combination;

[0039] Calculate the time and energy consumed by each combination of process flows;

[0040] Inputting the simplified capacity data corresponding to each of the combined process flows into a pre-established capacity prediction model to perform capacity prediction, thereby obtaining a plurality of combined predicted capacities;

[0041] Comparing the predicted capacity of each combination with the actual capacity to obtain a comparison result;

[0042] According to the comparison results, the time and energy consumed by each combination of process flows, a simplified process flow is determined from each process flow in the full-capacity test process data.

[0043] The step of dividing the full-capacity test process data into a plurality of simplified capacity data includes:

[0044] The full-capacity test process data is divided into a plurality of simplified capacity data according to the simplified process.

[0045] A battery cell capacity prediction device, comprising:

[0046] Full capacity data acquisition module, used to collect full capacity test process data of sample batteries;

[0047] A simplified capacity division module, configured to divide the full capacity test process data into a plurality of simplified capacity data;

[0048] A capacity prediction module is used to input each simplified capacity data into a pre-established capacity prediction model to perform capacity prediction and obtain multiple predicted capacities;

[0049] a comparison module, configured to compare each of the predicted capacities with the actual capacities to obtain an error value between each of the predicted capacities and the actual capacities;

[0050] The process determination module is used to select the simplified capacity data corresponding to the predicted capacity with the smallest error value as the candidate simplified data, and determine the test process corresponding to the candidate simplified data as the simplified capacity test process.

[0051] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of the battery cell capacity prediction method described in any of the above embodiments are implemented.

[0052] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the battery cell capacity prediction method described in any of the above embodiments.

[0053] The above-mentioned battery cell capacity prediction method, device, computer equipment and storage medium divide the full capacity test process data into multiple simplified capacity data according to the process flow, perform capacity prediction on the simplified capacity data through a capacity prediction model, and compare the predicted capacity with the actual capacity. According to the comparison results, a simplified capacity test process is selected, and then the battery capacity can be tested according to the selected simplified capacity test process. In this way, there is no need to use the full capacity test process for testing, and there is no need to fully charge and discharge the battery, which greatly saves the testing process and effectively improves the capacity testing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 1 is a flow chart of a method for predicting cell capacity in one embodiment;

[0055] Figure 2 1 is a flow chart of a method for predicting cell capacity according to another embodiment;

[0056] Figure 3 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0058] Example 1

[0059] In this embodiment, Figure 1 As shown, a method for predicting battery cell capacity is provided, which includes:

[0060] Step 110 : Collect full capacity test process data of the sample battery.

[0061] In this embodiment, the full-capacity test process data is a test process that fully charges and discharges the battery. This full-capacity test process data can also be referred to as complete capacity test process data. Since the data collection process for the full-capacity test process data requires fully charging the battery and then discharging it from a fully charged state to exhaustion, this process takes a long time. The test process in the full-capacity test process data includes multiple test processes, each of which is also referred to as a process flow. In the following embodiments, a process flow is also referred to as a work step.

[0062] In one embodiment, the full-capacity test process data includes: test start time, test end time, test start voltage, test end voltage, test start capacity, test end capacity, test start temperature, and test end temperature.

[0063] Step 120: Divide the full-capacity test process data into multiple simplified capacity data.

[0064] As mentioned above, the process of collecting full-capacity test process data involves fully charging and discharging the battery, which is a lengthy process. Therefore, in this embodiment, according to preset rules, the full-capacity test process is divided into multiple simplified process flows. Simplified capacity data is extracted from the full-capacity test process data based on the simplified process flows. In other words, each simplified process flow corresponds to a piece of simplified capacity data.

[0065] In one embodiment, the simplified capacity data includes the start time of the simplified process, the end time of the simplified process, the start voltage of the simplified process, the end voltage of the simplified process, the start capacity of the simplified process, the end capacity of the simplified process, the start temperature of the simplified process, and the end temperature of the simplified process.

[0066] If the voltage changes from V1 to V2, the corresponding capacity is Cp1; if the voltage changes from V1 to V3, the corresponding capacity is Cp2; if the voltage changes from V1 to V4, the corresponding capacity is Cp3, and so on. The extracted process data characteristic parameters (simplified capacity data) and the full capacity test capacity form a data set.

[0067] Step 130: Input each simplified capacity data into a pre-established capacity prediction model to perform capacity prediction and obtain multiple predicted capacities.

[0068] In this embodiment, the capacity prediction model is used to predict the battery capacity. The capacity prediction model is a model that is pre-learned and trained and can be used for battery capacity prediction.

[0069] In this embodiment, a machine learning algorithm is used to establish capacity prediction models for different simplified process flows, and each simplified capacity data input is input into the corresponding capacity prediction model for prediction to obtain multiple predicted capacities.

[0070] Step 140: Compare each predicted capacity with the actual capacity to obtain an error value between each predicted capacity and the actual capacity.

[0071] In this embodiment, the actual capacity is the actual capacity of the sample battery, and the full-capacity test process data includes the actual capacity. The error value is the difference between the predicted capacity and the actual capacity.

[0072] Step 150 : Select the simplified capacity data corresponding to the predicted capacity with the smallest error value as the simplified data to be selected, and determine the test process corresponding to the simplified data to be selected as the simplified capacity test process.

[0073] In the embodiment, the predicted capacity with the minimum error value is closest to the actual capacity, so the capacity test result of the test process corresponding to the predicted capacity is the most accurate, and therefore, the test process with the minimum error between the predicted capacity and the actual capacity is selected as the simplified capacity test process for subsequent full capacity prediction. The test process is one or more process flows in the full capacity test process.

[0074] In one embodiment, the simplified capacity test process is used to test the battery in the capacity test of the battery. Thus, in the capacity test of the battery, the full capacity test process is not required for testing, and the battery does not need to be fully charged and discharged, greatly saving the testing process and effectively improving the capacity test efficiency.

[0075] In the embodiment, the full capacity test process data is divided into multiple simplified capacity data according to the process flow, the capacity prediction model is used to predict the capacity of the simplified capacity data, the predicted capacity is compared with the actual capacity, the simplified capacity test process is selected according to the comparison result, and then the capacity of the battery can be tested according to the selected simplified capacity test process. Thus, the full capacity test process is not required for testing, and the battery does not need to be fully charged and discharged, greatly saving the testing process and effectively improving the capacity test efficiency. The embodiment can be embedded in a production system and has stronger applicability. The capacity sorting is performed in cooperation with the production system, real-time sorting can be realized in the production process, and the production efficiency is improved. In the embodiment, the prediction model based on the capacity test data and the simplified test process are combined to comprehensively improve the capacity test efficiency, accuracy and economy of the lithium battery.

[0076] In one embodiment, the full capacity test process data includes capacity data and temperature data; and the step of collecting the full capacity test process data of the sample battery includes:

[0077] collecting the full capacity test process data of the sample battery;

[0078] using a preset temperature compensation calculation formula to perform temperature compensation on the capacity data in the full capacity test process data, correcting to obtain the full capacity of the battery at a preset temperature, and taking the full capacity of the battery at the preset temperature as the capacity data in the full capacity test process data.

[0079] In the embodiment, the environment temperature is different for different batches of tests when the capacity of the battery is tested, which may result in that the capacity cannot be compared under a unified temperature condition. Therefore, in the embodiment, a preset temperature compensation calculation formula is used to perform temperature compensation on the collected capacity data to obtain the full capacity of the battery at a preset temperature. For example, the preset temperature is 25(±2)℃. In this way, the capacity data under a unified temperature condition can be obtained, the environmental interference is eliminated, and the capacity comparison accuracy is improved.

[0080] In one embodiment, before the step of collecting full capacity test process data of the sample battery, the step further includes:

[0081] Collect process data, formation process data and full-capacity test process data of sample batteries;

[0082] The process data, formation process data and full capacity test process data of the sample battery are input into a prediction learning model for training to obtain the capacity prediction model.

[0083] In this embodiment, the predictive learning model is also a machine learning algorithm. In one embodiment, the process data includes: core weight and battery weight; the formation process data includes: formation start time, formation end time, formation start voltage, formation end voltage, formation start capacity, formation end capacity, formation start temperature, and formation end temperature.

[0084] In this embodiment, the full-capacity test process data is divided into multiple simplified capacity data sets. The process data, formation process data, and each simplified capacity data set of the sample battery are input into the prediction learning model for training to obtain a capacity prediction model. In this embodiment, the input features cover process data (core weight, battery weight), formation data (time, voltage, temperature), and simplified test data (step voltage difference, temperature compensation capacity). The multi-dimensional features enhance the model's generalization ability and improve the model's prediction accuracy.

[0085] In one embodiment, the trained capacity prediction model is embedded in the MES (Manufacturing Execution System) system, and the battery cell characteristics of the simplified process flow are obtained through the MES system. The battery cell characteristics include process data, formation process data and each simplified capacity data. The battery cell characteristics are transmitted to the capacity prediction model, and the capacity prediction model is used to calculate and output the predicted capacity of the corresponding battery cell; in the MES system, the battery of the simplified process flow is capacity sorted according to the predicted capacity.

[0086] In this embodiment, the capacity prediction model is embedded in the MES system, which can predict the capacity of battery cells in real time during the production process, and then perform capacity sorting of battery cells in real time, thereby improving production efficiency.

[0087] In one embodiment, the predictive learning model includes a fully connected neural network, a support vector machine model, a decision tree model, a random forest model, a linear regression model, a convolutional neural network model, and a gradient boosting regression model.

[0088] In this embodiment, the prediction learning model can adopt any one of a connectionist neural network (DNN), a support vector machine model (SVR), a decision tree model (DT), a random forest model (RF), a linear regression model (LR), a convolutional neural network model (CNN), and a gradient boosting regression model (GBR).

[0089] In one embodiment, the step of inputting the process data, formation process data, and full-capacity test process data of the sample battery into a prediction learning model for training to obtain the capacity prediction model includes:

[0090] Dividing the full capacity test process data into a plurality of simplified capacity data;

[0091] The process data, formation process data, and simplified capacity data of the sample batteries were cleaned, and samples with missing values ​​and duplicate samples were deleted.

[0092] Calculate the capacity value, voltage difference, average temperature, and average current of each process flow to obtain a data set for each process flow;

[0093] Randomly divide the dataset into training and test sets;

[0094] The training sets and test sets of different units are standardized to obtain standardized training sets and standardized test sets;

[0095] Detecting anomalies in the standardized training set and the standardized test set using statistical methods or machine learning algorithms to obtain outliers, and processing the outliers;

[0096] Inputting the standardized training set into a prediction learning model for training to obtain a capacity prediction model;

[0097] The capacity prediction model is tested using the standardized test set, and parameters of the capacity prediction model are calibrated according to the test results.

[0098] In this example, by cleaning the sample data, removing missing values ​​and duplicate samples, and combining it with standardization, we can eliminate dimensional differences in process parameters, enhancing the convergence speed and stability of the model. Random partitioning ensures a balanced data distribution, avoids evaluation distortion caused by data bias, and enhances the model's generalization ability. Using statistical methods or machine learning algorithms to identify outliers can prevent noise data from interfering with the model and improve the reliability of prediction results.

[0099] In one embodiment, the method further comprises:

[0100] Obtaining each process flow in the full-capacity test process data;

[0101] Calculating the correlation between the test results of each process flow and the full-capacity test process data;

[0102] Different process flows are selected for combination, and parameters of each process flow are modified according to the correlation.

[0103] In this embodiment, the correlation between the test results corresponding to different steps of different full-capacity processes and the full-capacity data is calculated, and the step parameter settings are modified according to the test results of different combinations of steps to ensure smooth operation on the production line. The full-capacity test process is decomposed into multiple independent steps (such as voltage stage division in the charge and discharge cycle, temperature monitoring nodes, etc.), and the test data of each step is obtained (such as voltage difference, capacity increment, temperature change rate, etc.), and the correlation between the test results of each step and the final full-capacity data is calculated, so as to determine the parameters with higher correlation with capacity in each step, and then adjust and optimize the parameters with higher correlation.

[0104] In one embodiment, the method further comprises:

[0105] Obtaining each process flow in the full-capacity test process data;

[0106] Select different process flows for combination;

[0107] Calculate the time and energy consumed by each combination of process flows;

[0108] Inputting the simplified capacity data corresponding to each of the combined process flows into a pre-established capacity prediction model to perform capacity prediction, thereby obtaining a plurality of combined predicted capacities;

[0109] Comparing the predicted capacity of each combination with the actual capacity to obtain a comparison result;

[0110] According to the comparison results, the time and energy consumed by each combination of process flows, a simplified process flow is determined from each process flow in the full-capacity test process data.

[0111] The step of dividing the full-capacity test process data into a plurality of simplified capacity data includes:

[0112] The full-capacity test process data is divided into a plurality of simplified capacity data according to the simplified process.

[0113] In this embodiment, the collection of full-capacity test process data involves multiple process flows (steps), and each process flow corresponds to a simplified capacity data. In this embodiment, at least one of these process flows is selected for combination, that is, the process in the process flow of the combination can be one. The time and energy consumed by the process flows of different combinations are calculated respectively, and the simplified capacity data corresponding to the process flows of different combinations are input into the capacity prediction model for prediction to obtain the prediction results. In this way, according to the prediction results and the time and energy consumed by the process flows of different combinations, the process that consumes less time and energy and whose predicted capacity is closer to the actual capacity is selected as the simplified process. In this way, when the capacity prediction model is used for prediction later, the full-capacity test process data can be divided into multiple data based on the determined simplified process, and then the simplified capacity data can be extracted from the full-capacity test process data for prediction. In this embodiment, by dynamically combining different process flows and calculating their time and energy consumption costs, and combining the capacity prediction model to verify the reliability of the simplified process, the full-capacity test cycle can be significantly shortened and energy consumption can be effectively reduced.

[0114] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0115] Example 2

[0116] This embodiment proposes a simplified process and online capacity prediction method for the lithium battery capacity test phase. Figure 2 , including the following steps:

[0117] S01. Collect the process data, formation process data and complete capacity test process data of the sample battery, extract and process the full capacity data, and divide it into several simplified capacity process data, such as V1~V2, corresponding to capacity Cp1, V1~V3, corresponding to capacity Cp2, V1~V4, corresponding to capacity Cp3, and so on. The extracted process data characteristic parameters and the full capacity test capacity form a data set.

[0118] S02. Use machine learning algorithms to establish full-capacity prediction models for different simplified processes, and select the process with the smallest error between the predicted capacity and the actual capacity as the simplified capacity test process for subsequent full-capacity prediction.

[0119] S03. Embed the capacity prediction model into the MES system, obtain the simplified capacity process result features through the MES system, transfer the features to the model, and the model calculates and outputs the capacity of the corresponding battery cell;

[0120] S04. In the MES system, the capacity of batteries with simplified processes is sorted according to the capacity prediction results.

[0121] The method adopted by the present invention combines simplified capacity testing with a capacity prediction model. The model predicts capacity with high accuracy and can replace traditional capacity testing methods. It reduces the time of the capacity testing process from about 3 to 4 hours to about 0.5 to 1 hour, reduces the company's production energy consumption, improves the company's rhythm and production capacity, and achieves cost reduction and efficiency improvement.

[0122] In a more specific technical solution, in S01, the process data includes but is not limited to: the core weight m1, the battery weight m2; the data of the formation process includes but is not limited to: the formation start time t1, the formation end time t2, the formation start voltage v1, the formation end voltage v2, the formation start capacity q1, the formation end capacity q2, the formation start temperature tem1, the formation end temperature tem2; the data of the complete capacity test process includes but is not limited to: the start time tm, the end time tn, the start voltage vm, the end voltage vn, the start capacity qm, the end capacity qn, the start temperature temm, and the end temperature temn of each step of the complete capacity test.

[0123] In a more specific technical solution, in S01, data set processing and partitioning includes:

[0124] S011. Data set cleaning, deleting samples with missing values ​​and duplicate samples;

[0125] S012, data processing, calculation of capacity value, voltage difference, average temperature, current size, etc. of each step;

[0126] S013, capacity data temperature optimization, using the temperature compensation empirical formula to compensate the capacity obtained from the complete capacity test and correct it to the full capacity of the battery at a temperature of 25 (± 2) ° C;

[0127] S014, data set division, randomly divide the data set into training set and test set with a ratio of 7:3 or 8:2;

[0128] S015. Data standardization: standardize the data of different units to improve the training effect of the model;

[0129] S016. Outlier detection: Use statistical methods or machine learning algorithms to detect and process outliers.

[0130] In a more specific technical solution, in S01, the simplified process includes:

[0131] S017. Calculate the correlation between the test results of different steps in different full-capacity processes and the full-capacity data;

[0132] S018. Select different combinations of process step test results and modify the process step parameter settings to ensure smooth operation on the production line;

[0133] S019. Calculate the time and energy consumption of different work step combinations.

[0134] In a more specific technical solution, in S02, establishing a capacity prediction model includes:

[0135] S021. Model establishment. Model selection includes but is not limited to fully connected neural network (DNN), support vector machine model (SVR), decision tree model (DT), random forest model (RF), linear regression model (LR), convolutional neural network model (CNN), gradient boosting regression model (GBR), etc.

[0136] S022. Model optimization: using randomly initialized hyperparameters and hyperparameter optimization algorithms to optimize the model. Optimization algorithms include but are not limited to: Bayesian optimization, grid search optimization, random search optimization, etc.

[0137] S023. Model evaluation: Use model evaluation indicators to evaluate the performance of the capacity prediction model. Model evaluation indicators include but are not limited to mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R-suqared), etc.

[0138] S024, model saving and calling, saving the best-performing full-capacity prediction model in the model evaluation as a callable file.

[0139] The present invention uses the full-capacity capacity data after temperature compensation as the model training target, and adds the average temperature of each step as a feature in the model input feature to realize the temperature normalization of the capacity prediction model and accurately predict the full-capacity capacity of fresh batteries at 25 (± 2) ℃.

[0140] In a more specific technical solution, in S02, the selection of simplified processes includes:

[0141] S025. Record the results of the model evaluation described in S023 and the time and energy consumption of the current simplified process described in S017 under different simplified processes;

[0142] S026. Comprehensive comparison of model evaluation results and time and energy consumption;

[0143] S027. Use the process with the best overall results as the simplified capacity testing process.

[0144] In a more specific technical solution, in S03, the data transmitted to the model includes:

[0145] The process data and formation data of S031 are the same as those in S01 above;

[0146] S032 Simplified capacity test process data includes but is not limited to: start time tm, end time tn, start voltage vm, end voltage vn, start capacity qm, end capacity qn, start temperature temm, end temperature temn, etc. of each step of the simplified capacity test;

[0147] S033. Add barcode and cell model data to the data and pass the data into the model;

[0148] S034. The model automatically processes and inputs data, completes capacity prediction, and obtains predicted capacity;

[0149] S035. Add the barcode and cell model data to the predicted capacity, transmit it back to the MES system, and save it in the MES system.

[0150] In a more specific technical solution, in S032, the simplified capacity test process step is:

[0151] S0321, let it stand;

[0152] S0322, constant current discharge to the first step cut-off voltage;

[0153] S0323, let stand;

[0154] S0324, constant current discharge to the second step cut-off voltage;

[0155] S0325, let it stand;

[0156] S0326, constant current charging to the third step cut-off voltage;

[0157] S0327, let it stand;

[0158] S0328, constant current discharge to the fourth step cut-off voltage.

[0159] In a more specific technical solution, in S04, the capacity sorting basis includes: predicted capacity, OCV1, OCV2, K value, and internal resistance.

[0160] The present invention can accurately predict the full capacity of a battery cell on a cloud server while reducing the time and energy consumption of capacity testing. The cloud server can store and process a large amount of historical data, and use big data analysis and machine learning technology to continuously optimize the prediction model and improve the monthly accuracy of the capacity. The predicted capacity is the temperature-compensated capacity at 25 (± 2) ° C. The temperature-compensated capacity enables capacity data under different temperature conditions to be uniformly compared, thereby achieving capacity standardization and improving the accuracy of subsequent capacity sorting.

[0161] The online capacity prediction model proposed in this embodiment can complete data processing, model prediction, and result feedback in a short period of time, thereby reducing the cost of deploying the capacity prediction model locally on the production line while meeting the time requirements of the production line process.

[0162] This embodiment uses the full-capacity capacity data after temperature compensation as the model training target, and adds the average temperature of each step as a feature in the model input feature to achieve temperature normalization of the capacity prediction model and accurately predict the full-capacity capacity of fresh batteries at 25 (± 2) ° C.

[0163] This embodiment can accurately predict the full capacity of the battery cell while reducing the time and energy consumption of capacity testing. At the same time, it can store and process a large amount of historical data, and use big data analysis and machine learning technology to continuously optimize the prediction model and improve the accuracy of capacity prediction. The predicted capacity is the temperature compensated capacity at 25 (± 2) ° C. The temperature-compensated capacity enables capacity data under different temperature conditions to be uniformly compared, thereby achieving capacity standardization and improving the accuracy of subsequent capacity sorting.

[0164] Example 3

[0165] In this embodiment, a battery cell capacity prediction device is provided, comprising:

[0166] Full capacity data acquisition module, used to collect full capacity test process data of sample batteries;

[0167] A simplified capacity division module, configured to divide the full capacity test process data into a plurality of simplified capacity data;

[0168] A capacity prediction module is used to input each simplified capacity data into a pre-established capacity prediction model to perform capacity prediction and obtain multiple predicted capacities;

[0169] a comparison module, configured to compare each of the predicted capacities with the actual capacities to obtain an error value between each of the predicted capacities and the actual capacities;

[0170] The process determination module is used to select the simplified capacity data corresponding to the predicted capacity with the smallest error value as the candidate simplified data, and determine the test process corresponding to the candidate simplified data as the simplified capacity test process.

[0171] For the specific definition of the battery cell capacity prediction device, please refer to the definition of the battery cell capacity prediction method above, which will not be repeated here. Each unit in the above-mentioned battery cell capacity prediction device can be implemented in whole or in part by software, hardware and their combination. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units.

[0172] Example 4

[0173] In this embodiment, a computer device is provided. Its internal structure diagram can be shown as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and the non-volatile storage medium is deployed with a database, which is used for capacity prediction models and sample data. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices that have deployed application software. When the computer program is executed by the processor, a method for predicting the capacity of a battery cell is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0174] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0175] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the battery cell capacity prediction method described in any of the above embodiments when executing the computer program.

[0176] Example 5

[0177] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the battery cell capacity prediction method described in any of the above embodiments are implemented.

[0178] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0179] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, 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.

[0180] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for predicting battery cell capacity, characterized in that: include: Collect full capacity test process data of sample batteries; Dividing the full capacity test process data into a plurality of simplified capacity data; Input each simplified capacity data into a pre-established capacity forecasting model to perform capacity forecasting and obtain multiple forecast capacities; Comparing each of the predicted capacities with the actual capacities to obtain an error value between each of the predicted capacities and the actual capacities; The simplified capacity data corresponding to the predicted capacity with the smallest error value is selected as the simplified data to be selected, and the test process corresponding to the simplified data to be selected is determined as the simplified capacity test process.

2. The method according to claim 1, characterized in that The full-capacity test process data includes capacity data and temperature data; the step of collecting the full-capacity test process data of the sample battery includes: Collect full capacity test process data of sample batteries; A preset temperature compensation calculation formula is used to perform temperature compensation on the capacity data in the full-capacity test process data, and the full-capacity capacity of the battery at a preset temperature is corrected and used as the capacity data in the full-capacity test process data.

3. The method according to claim 1, characterized in that Before the step of collecting full capacity test process data of the sample battery, the step further includes: Collect process data, formation process data and full-capacity test process data of sample batteries; The process data, formation process data and full capacity test process data of the sample battery are input into a prediction learning model for training to obtain the capacity prediction model.

4. The method according to claim 3, characterized in that The process data, the formation process data, the full capacity test process data, and the simplified capacity data include at least one of the following: The process data includes: core weight and battery weight; The formation process data include: formation start time, formation end time, formation start voltage, formation end voltage, formation start capacity, formation end capacity, formation start temperature and formation end temperature; The full-capacity test process data includes: test start time, test end time, test start voltage, test end voltage, test start capacity, test end capacity, test start temperature and test end temperature; The simplified capacity data includes the start time of the simplified process, the end time of the simplified process, the start voltage of the simplified process, the end voltage of the simplified process, the start capacity of the simplified process, the end capacity of the simplified process, the start temperature of the simplified process, and the end temperature of the simplified process.

5. The method according to claim 3, characterized in that The step of inputting the process data, formation process data and full capacity test process data of the sample battery into the prediction learning model for training to obtain the capacity prediction model includes: Dividing the full capacity test process data into a plurality of simplified capacity data; The process data, formation process data, and simplified capacity data of the sample batteries were cleaned, and samples with missing values ​​and duplicate samples were deleted. Calculate the capacity value, voltage difference, average temperature, and average current of each process flow to obtain a data set for each process flow; Randomly divide the dataset into training and test sets; The training sets and test sets of different units are standardized to obtain standardized training sets and standardized test sets; Detecting anomalies in the standardized training set and the standardized test set using statistical methods or machine learning algorithms to obtain outliers, and processing the outliers; Inputting the standardized training set into a prediction learning model for training to obtain a capacity prediction model; The capacity prediction model is tested using the standardized test set, and parameters of the capacity prediction model are calibrated according to the test results.

6. The method according to any one of claims 1 to 5, characterized in that Also includes: Obtaining each process flow in the full-capacity test process data; Calculating the correlation between the test results of each process flow and the full-capacity test process data; Different process flows are selected for combination, and parameters of each process flow are modified according to the correlation.

7. The method according to claim 1, characterized in that Also includes: Obtaining each process flow in the full-capacity test process data; Select different process flows for combination; Calculate the time and energy consumed by each combination of process flows; Inputting the simplified capacity data corresponding to each of the combined process flows into a pre-established capacity prediction model to perform capacity prediction, thereby obtaining a plurality of combined predicted capacities; Comparing the predicted capacity of each combination with the actual capacity to obtain a comparison result; Determine a simplified process from each process flow in the full-capacity test process data based on the comparison results and the time and energy consumed by each process flow of each combination; The step of dividing the full-capacity test process data into a plurality of simplified capacity data includes: The full-capacity test process data is divided into a plurality of simplified capacity data according to the simplified process.

8. A battery cell capacity prediction device, characterized in that: include: Full capacity data acquisition module, used to collect full capacity test process data of sample batteries; A simplified capacity division module, configured to divide the full capacity test process data into a plurality of simplified capacity data; A capacity prediction module is used to input each simplified capacity data into a pre-established capacity prediction model to perform capacity prediction and obtain multiple predicted capacities; a comparison module, configured to compare each of the predicted capacities with the actual capacities to obtain an error value between each of the predicted capacities and the actual capacities; The process determination module is used to select the simplified capacity data corresponding to the predicted capacity with the smallest error value as the candidate simplified data, and determine the test process corresponding to the candidate simplified data as the simplified capacity test process.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images