A battery formation method integrating rapid pre-charging and capacity prediction

Abstract

The application discloses a battery formation method integrating rapid pre-charging and capacity prediction, and relates to the technical field of battery formation, and comprises the following steps: obtaining a battery dQ / dV curve; performing a charging operation by adopting a staged charging strategy; recording the pre-charging capacity of the same batch of battery cells to establish a reference relationship curve of the pre-charging capacity and temperature; collecting the actual pre-charging capacity and temperature of the battery cells, and obtaining the intrinsic pre-charging capacity by correction according to the reference relationship curve; collecting the actual capacity grading capacity and temperature of the test sample, and solving the calibration intercept of each capacity gradient based on the calibrated capacity grading capacity and the intrinsic pre-charging capacity to establish a capacity prediction curve; calculating the predicted capacity of the battery cells according to the capacity prediction curve for the battery cells which have not been tested; and taking the calculated predicted capacity as the basis for battery grading and grouping, and canceling the independent capacity grading process of the battery cells. The application replaces the capacity grading process with the predicted capacity, forms a formation process integrating pre-charging and capacity prediction, shortens the formation period, and reduces the equipment investment and operation cost.

Description

Technical Field

[0001] This invention relates to the field of battery formation technology, specifically a battery formation method that integrates rapid pre-charging and capacity prediction. Background Technology

[0002] In the manufacturing process of lithium-ion batteries, the formation process is a crucial step that determines battery performance. It involves forming a solid electrolyte interphase (SEI) film on the negative electrode surface during the initial charge, directly impacting the battery's capacity, cycle life, and safety performance. Following the formation process, capacity grading testing is typically performed to determine the actual capacity of the batteries for tiering and grouping. This capacity grading process usually consumes significant amounts of charging and discharging equipment, time, and energy, making it a substantial component of battery manufacturing costs.

[0003] In recent years, the technology of predicting battery capacity using precharge data has gradually become a research hotspot. By using data-driven methods, the capacity can be predicted based on non-fully charged precharge data. This method can greatly reduce the investment in capacity testing equipment, factory occupancy, and charging and discharging costs and time. However, the practical application of capacity prediction requires small prediction errors and strong result stability, which involves several technical challenges such as cleaning process abnormal data, data balancing, and extraction of capacity feature factors.

[0004] In existing technologies, several studies have been conducted focusing on pre-charge process optimization and capacity prediction, but some shortcomings still exist:

[0005] 1. Existing pre-charge process designs mainly serve the formation purpose (i.e., SEI film formation) without simultaneously considering the data collection requirements for capacity prediction. For example, patent CN101783424A discloses a staged pre-charge method that uses at least two charging steps, with the current in the second charging step being greater than that in the first charging step. The total pre-charge amount is 15% to 35% of the total battery capacity, which shortens the pre-charge time and improves production efficiency. However, the stage division of this method is based only on time sequence or capacity ratio and is not related to the internal electrochemical reaction mechanism of the battery. It cannot accurately control key processes such as SEI film formation and phase change, nor can it collect characteristic data that can be used for capacity prediction during the pre-charge process.

[0006] 2. Existing capacity prediction methods mostly employ complex machine learning models. For example, patent CN119001467A discloses a method for calculating capacity based on battery precharge data. This method extracts dQ / dV peak features, precharge step features, and process data, and uses complex models such as LightGBM to achieve capacity prediction. However, this method requires the deployment of a machine learning model runtime environment, places high demands on the computing power of the MES system, and requires a large amount of historical data for model training. It also requires a professional data scientist team for continuous maintenance and parameter tuning. Although some patents, such as CN116522153B and CN116794533B, extract precharge data peak features and use tree models to achieve capacity grading, they do not consider polarization compensation and data balancing issues, and lack detailed demonstrations on the influence mechanism of peak features and feature selection.

[0007] 3. In the existing technology, the pre-filling process and the capacity-dividing process are independent of each other. Even if the capacity prediction method is used, the capacity-dividing process cannot be truly eliminated. Although some patents have proposed integrated formation and capacity-dividing equipment solutions, they are only physical integrations at the equipment level. The formation and capacity-dividing processes are still carried out independently one after the other, which does not fundamentally simplify the process flow.

[0008] In addition, dQ / dV curve analysis technology has been used in battery characteristic research. For example, patent CN111194412A discloses a device, method, battery pack and electrical system for determining battery electrode information. It uses V-dQ / dV curve to detect multiple feature points to determine electrode information. For example, patent CN107634274B discloses a battery pack matching method that uses two peaks of the dQ / dV curve to match battery packs. However, these technologies are mainly applied to battery state assessment and matching, and have not been used to guide the phased design of pre-charge process, nor have they been used for collaborative optimization with capacity prediction models.

[0009] In summary, the existing technology has the following technological gaps:

[0010] 1. Separation of pre-charge process and capacity prediction: The pre-charge process only serves the formation purpose and does not collect the characteristic data required for capacity prediction at the same time, which means that a separate capacity assessment process is still required after pre-charge;

[0011] 2. Lack of mechanistic guidance for staged charging: Existing staged pre-charging methods do not identify key electrochemical reaction regions based on dQ / dV curve analysis, making it difficult to achieve the optimal balance between SEI film quality and charging efficiency;

[0012] 3. Complex capacity prediction models: Existing prediction methods rely on complex machine learning models, which are costly to deploy and maintain, and are not suitable for large-scale mass production scenarios that pursue extreme efficiency and low cost.

[0013] 4. Lack of process integration: There is a lack of technical solutions that systematically integrate pre-filling process optimization, capacity prediction and cancellation of capacity sizing.

[0014] Based on this, a battery formation method integrating rapid pre-charging and capacity prediction is now provided, which can eliminate the drawbacks of existing technical solutions. Summary of the Invention

[0015] The purpose of this invention is to provide a battery formation method that integrates rapid pre-charging and capacity prediction, in order to solve the problems in the prior art such as the separation of pre-charging process and capacity prediction, lack of mechanism guidance for staged charging, reliance on complex models for capacity prediction, and failure to achieve integrated process.

[0016] To achieve the above objectives, the present invention provides the following technical solution:

[0017] A battery formation method integrating rapid pre-charging and capacity prediction specifically includes the following steps:

[0018] Step S1: Obtain the dQ / dV curve of the target battery, and identify the key voltage range based on the characteristic peaks under the dQ / dV curve;

[0019] Step S2: Perform the charging operation using a phased charging strategy. Use the first current to charge in the critical voltage range and the second current to charge in the non-critical range, with the first current being less than the second current.

[0020] Step S3: Precharge the same batch of cells, record the precharge capacity under different temperature gradients, establish a reference relationship curve between precharge capacity and temperature, and obtain the reference precharge capacity at the corresponding temperature.

[0021] Step S4: During the pre-charging process, collect the actual pre-charging capacity and actual pre-charging temperature of the same batch of cells. Based on the reference relationship curve, correct the actual pre-charging capacity to the standard temperature to obtain the intrinsic pre-charging capacity at the standard temperature.

[0022] Step S5: Based on the intrinsic precharge capacity, divide the same batch of cells into multiple capacity gradients. For each capacity gradient, extract several sample cells to complete the full capacity test. Collect the actual capacity and capacity temperature of the test samples. Perform temperature calibration on the actual capacity. Based on the calibrated capacity and intrinsic precharge capacity, solve the calibration intercept of each capacity gradient and establish a capacity prediction curve.

[0023] Step S6: For cells in the same batch that have not undergone full capacity testing, calculate the predicted capacity of the cells based on the capacity prediction curve of the corresponding capacity gradient matched with the intrinsic precharge capacity.

[0024] Step S7: Implement SPC control on the key process parameters of the battery cell during the pre-charging process, set the control window range, and when the key process parameters are within the control window set range, execute steps S4~S6 to calculate the predicted capacity of the corresponding battery cell.

[0025] Step S8: Using the calculated predicted capacity as the basis for battery grading and grouping, except for cells whose key process parameters do not meet the control window setting range and several sample cells used for calibration, the independent capacity grading process for the remaining cells is cancelled, and they enter the subsequent standing and grouping process.

[0026] Furthermore, the key voltage range in step S1 includes at least one of the SEI film formation region, the gas generation region, and the phase transition region.

[0027] Furthermore, the non-critical region in step S2 is the normal lithium intercalation region, and the phased charging strategy specifically includes:

[0028] The SEI film formation region is charged with a small current of 0.01C to 0.1C;

[0029] The gas generation zone is charged with a small constant current of 0.02C to 0.25C.

[0030] A small current of 0.03C to 0.3C is used for charging in the phase transition region;

[0031] High-current charging of 0.1C~0.8C is used in the normal lithium intercalation region.

[0032] Furthermore, the formula for calculating the baseline precharge capacity in step S3 is as follows:

[0033] ;

[0034] in, For temperature The baseline precharge capacity, Pre-charge temperature, As the reference intercept, , represents the fitting coefficient.

[0035] Furthermore, the formula for calculating the intrinsic precharge capacity in step S4 is as follows:

[0036] ;

[0037] in, Pre-charge capacity for intrinsic purposes, This is the actual pre-charged capacity. This is the actual pre-charge temperature. Standard temperature For temperature The baseline precharge capacity, For temperature The baseline precharge capacity.

[0038] Furthermore, in step S5, temperature calibration is performed on the actual capacity, specifically as follows:

[0039] ;

[0040] in, For the calibrated capacity, To satisfy the actual capacity measured after full capacity testing, This is the average temperature during actual volume division. Standard temperature For temperature The baseline precharge capacity, For temperature The baseline precharge capacity, For temperature The baseline precharge capacity;

[0041] The calibration intercepts for each capacity gradient are calculated based on the calibrated full-capacity and the intrinsic pre-charge capacity, and are specifically expressed as follows:

[0042] , ;

[0043] in, To calibrate the intercept, Pre-charge capacity for intrinsic purposes, For the first The calibrated mean of samples with capacity gradients and satisfying , This represents the total number of capacity gradients.

[0044] Furthermore, the formula for calculating the predicted cell capacity in step S6 is as follows:

[0045] ;

[0046] in, It is a constant. To predict capacity, Pre-charge capacity for intrinsic purposes, For the first Mean value after calibration of a capacity gradient sample.

[0047] Furthermore, step S6 also includes: performing a capacity prediction operation on all cells in the same batch according to the capacity gradient, calculating the predicted capacity, performing a capacity grading operation on all cells, recording the actual capacity grading capacity and capacity grading temperature, performing temperature calibration on the actual capacity grading capacity to obtain the calibrated capacity grading capacity, comparing the calibrated capacity grading capacity with the predicted capacity, calculating the prediction error, and verifying the prediction accuracy based on the prediction error.

[0048] Furthermore, step S7 also includes: when the key process parameters exceed the control window setting range, marking the corresponding cell and diverting it to the traditional capacity testing process, wherein the control window range is... , This is the sample standard deviation of key process parameters calculated from historical stable production batches.

[0049] Further, the step of obtaining the dQ / dV curve in step S1 is as follows: pre-charge the battery cell with a small current, continuously collect voltage and charging capacity data, calculate dQ and dV, and then perform exponential filtering based on a sliding window to calculate dQ / dV. Construct the dQ / dV curve based on dQ / dV, where Q is the charging capacity and V is the voltage.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0051] 1. This invention identifies key voltage ranges such as the SEI film formation region and phase transition region through dQ / dV curve analysis. In the key range, a small current is used to ensure film formation quality, while a large current is used in the non-critical range to improve efficiency. Through this differentiated staged charging strategy, the pre-charging time is shortened without reducing the quality of the SEI film.

[0052] 2. This invention simultaneously collects characteristic data of key intervals, such as charging capacity and temperature at specific voltage ranges, during the phased charging process. It expands the pre-charging process from a single formation function to a dual function of formation and data acquisition, providing high-quality input for capacity prediction. It establishes a benchmark relationship curve between pre-charging capacity and temperature, and establishes a single-factor linear regression relationship between intrinsic pre-charging capacity and capacity testing capacity. Ultimately, it realizes the replacement of the capacity testing process with predicted capacity, forming an integrated formation process of pre-charging and capacity prediction, shortening the formation cycle, and reducing equipment investment and operating costs. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0054] Figure 2 This is a schematic diagram of the dQ / dV curve of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0056] Considering that the final capacity of a battery cell is highly linearly correlated with its intrinsic precharge capacity (i.e., the precharge capacity after excluding temperature interference), the measured precharge capacity is significantly affected by temperature. For every 5°C increase in temperature, the measured precharge capacity increases by approximately 10-40mAh. Based on this, the present invention adopts a two-step strategy of "compensation first, prediction later": using the temperature compensation curve, the measured precharge capacity is corrected to the standard temperature to obtain the intrinsic precharge capacity; a single-factor linear regression relationship between the intrinsic precharge capacity and the capacity grading is established; the capacity is predicted using the single-factor linear regression relationship; and the batteries are grouped and matched based on the predicted capacity, replacing the capacity grading process. The specific technical solution is as follows.

[0057] Example 1

[0058] In this embodiment, as Figure 1 As shown, this invention provides a battery formation method integrating rapid pre-charging and capacity prediction, specifically including the following steps:

[0059] Step S1: Obtain the dQ / dV curve of the target battery, and identify the key voltage range based on the characteristic peaks under the dQ / dV curve. The key voltage range includes at least one of the SEI film formation region, gas generation region, and phase transition region.

[0060] Specifically, considering that existing staged pre-charging methods only divide stages according to time sequence or capacity ratio, without relating them to the internal electrochemical reaction mechanism of the battery, this invention uses dQ / dV curve analysis to accurately identify key voltage ranges such as the SEI film formation region and phase transition region. This differentiated charging strategy between key and non-key voltage ranges shortens the pre-charging time. The steps for obtaining the dQ / dV curve are as follows: pre-charge the cell with a small current, continuously collect voltage and charging capacity data, calculate dQ and dV, then perform exponential filtering based on a sliding window to calculate dQ / dV, and construct the dQ / dV curve based on dQ / dV, where Q is the charging capacity and V is the voltage. Figure 2 As shown;

[0061] Step S2: Perform the charging operation using a phased charging strategy. Use the first current to charge in the critical voltage range and the second current to charge in the non-critical range, with the first current being less than the second current.

[0062] Step S3: Precharge the same batch of cells, record the precharge capacity under different temperature gradients, establish a reference relationship curve between precharge capacity and temperature, and obtain the reference precharge capacity at the corresponding temperature.

[0063] Step S4: During the pre-charging process, collect the actual pre-charging capacity and actual pre-charging temperature of the same batch of cells. Based on the reference relationship curve, correct the actual pre-charging capacity to the standard temperature to obtain the intrinsic pre-charging capacity at the standard temperature.

[0064] Step S5: Based on the intrinsic precharge capacity, divide the same batch of cells into multiple capacity gradients. For each capacity gradient, extract several sample cells to complete the full capacity test. Collect the actual capacity and capacity temperature of the test samples. Perform temperature calibration on the actual capacity. Based on the calibrated capacity and intrinsic precharge capacity, solve the calibration intercept of each capacity gradient and establish a capacity prediction curve.

[0065] Step S6: For cells in the same batch that have not undergone full capacity testing, calculate the predicted capacity of the cells based on the capacity prediction curve of the corresponding capacity gradient matched with the intrinsic precharge capacity.

[0066] Step S7: Implement SPC control on the key process parameters of the battery cell during the pre-charging process, set the control window range, and when the key process parameters are within the control window set range, execute steps S4~S6 to calculate the predicted capacity of the corresponding battery cell.

[0067] Step S8: Use the calculated predicted capacity as the basis for battery grading and grouping. Except for cells whose key process parameters do not meet the control window setting range and a number of sample cells used for calibration (e.g., 10 cells per batch), the independent capacity grading process for the remaining cells is cancelled, and they enter the subsequent standing and grouping process.

[0068] In this embodiment, the present invention also achieves continuous monitoring of prediction accuracy and adaptive updating of curve relationships through the following mechanism:

[0069] After each batch prediction is completed, the prediction error of that batch is calculated using the obtained calibration sample volume data. If the error exceeds a preset threshold (e.g., 1%), an alarm is triggered, prompting process personnel to check whether there are any abnormalities in that batch (e.g., temperature control deviation, equipment failure, etc.).

[0070] Furthermore, the calibration intercept value of each batch is recorded to form a historical database. When the calibration intercept values ​​of multiple consecutive batches show systematic drift (such as monotonically increasing for 5 consecutive batches), it prompts process personnel to check whether there are any changes in process trends (such as electrolyte batch replacement, equipment aging, etc.).

[0071] At the same time, when the fitting coefficient is found , When the curve is no longer applicable (e.g., when the cell model is changed or the material system is significantly altered), repeat steps S3-S5 to establish a new baseline relationship curve.

[0072] Furthermore, the prediction results can be traced back to specific characteristic factors and process parameters. When the prediction accuracy fluctuates, process engineers can quickly locate the cause and carry out targeted optimization.

[0073] Specifically, the non-critical section in step S2 is the normal lithium intercalation region, and the phased charging strategy includes:

[0074] A small current of 0.01C~0.1C (charging rate) is used to charge the SEI film in the SEI film formation region to ensure that the SEI film is uniform and dense.

[0075] In the gas generation zone, a small constant current of 0.02C~0.25C is used for charging;

[0076] A small current of 0.03C to 0.3C is used to charge the material in the phase transition region to stabilize its structure;

[0077] High-current charging of 0.1C~0.8C is used in the normal lithium intercalation region to improve pre-charge efficiency;

[0078] This invention shortens the pre-charging time and eliminates the capacity testing process by using staged charging, thereby shortening the overall formation cycle. At the same time, it significantly reduces the investment and energy consumption costs of the capacity testing equipment.

[0079] Specifically, the formula for calculating the baseline precharge capacity in step S3 is as follows:

[0080] ;

[0081] in, For temperature The baseline precharge capacity is in mAh. This is the pre-charge temperature, in °C. As the reference intercept, , represents the fitting coefficient.

[0082] Specifically, the formula for calculating the intrinsic precharge capacity in step S4 is as follows:

[0083] ;

[0084] in, Pre-charge capacity for intrinsic purposes, This refers to the actual pre-charge capacity, in mAh. This is the actual pre-charge temperature. Standard temperature, unit: °C For temperature The baseline precharge capacity, For temperature The baseline precharge capacity.

[0085] Specifically, in step S5, a small sample of the batch of cells is taken and charged according to the intrinsic pre-charge capacity. Stratified sampling is performed, and the samples are actually divided into portions to obtain the full discharge capacity after full filling. Actual test temperature Possibly related to standard reference temperature There is a deviation. To eliminate the impact of the temperature difference in the capacity division on the verification results, the actual capacity division needs to be compared with the baseline relationship curve. Temperature calibration is specifically represented as follows:

[0086] ;

[0087] in, This is the calibrated capacity, in mAh. The capacity measured after full capacity testing is expressed in mAh. The average temperature during actual volume division, in °C. The standard temperature is expressed in °C. It is the standard temperature for cell design or manufacturing, preferably 25 °C. For temperature The baseline precharge capacity, For temperature The baseline precharge capacity, For temperature The baseline precharge capacity;

[0088] Using the calibrated capacity, the reference relationship curve is calibrated, and the intercept parameter is adjusted. Therefore, the calibration intercept of each capacity gradient is calculated based on the calibrated full-capacity and the intrinsic pre-charge capacity, specifically expressed as:

[0089] , ;

[0090] in, To calibrate the intercept, Pre-charge capacity for intrinsic purposes, For the first The calibrated mean of samples with capacity gradients and satisfying , The total capacity gradient is given; different calibration intercepts can be obtained for samples with different capacity gradients. This forms a family of capacity prediction curves.

[0091] Specifically, in step S6, for the remaining cells in the same batch, the predicted capacity is calculated by matching the corresponding calibration curve based on the intrinsic pre-charge capacity. The formula for calculating the predicted capacity of a battery cell is:

[0092] ;

[0093] in, The constant is 1, 2, etc., and the actual subsequent values ​​are calculated according to... Calculated To predict capacity, Pre-charge capacity for intrinsic purposes, For the first Mean value after calibration of each capacity gradient sample;

[0094] Step S6 further includes: performing a capacity prediction operation on all cells in the same batch according to the capacity gradient, calculating the predicted capacity, and simultaneously performing a capacity grading operation on all cells, recording the actual capacity grading capacity and capacity grading temperature, and performing temperature calibration on the actual capacity grading capacity to obtain the calibrated capacity grading capacity. The calibrated capacity is compared with the predicted capacity. Compare and calculate the prediction error. The prediction accuracy is verified based on the prediction error, and the calculation formula is as follows: ;

[0095] This invention not only performs temperature compensation for pre-charge capacity to eliminate the influence of temperature on the predicted input, but also performs temperature calibration on the capacity grading stage during the verification process. This ensures that the pre-charge and grading stages are at the same reference temperature, eliminating the interference of temperature differences on the prediction accuracy assessment and guaranteeing the authenticity and reliability of the verification results. When the cell model or material system changes, it is only necessary to re-execute steps S1 to S6 to quickly complete the process adaptation, demonstrating good adaptability.

[0096] Specifically, step S7 further includes: when the key process parameters exceed the control window setting range, marking the corresponding cell and diverting it to the traditional capacity testing process. For cells diverted to the traditional capacity testing process, as well as a small number of sample cells used for calibration in each batch, capacity testing is still completed in the traditional way, with the control window range being [missing information]. , The sample standard deviation of the key process parameters is calculated from historical stable production batches. The calculation formula adopts the unbiased estimation version commonly used in engineering: , For a single measurement or recorded value in the sample, The sample mean. For the degree-of-freedom correction term, the conventional SPC control limit is usually... However, in order to ensure the consistency of input data for capacity prediction, this invention actively narrows the control window to... Cells that exceed the control window are considered to have precharge data that is significantly affected by process fluctuations and are diverted to the traditional capacity testing process. Only cells that are within the control window are allowed to enter the prediction operation.

[0097] The process stability control involved in step S7 is a prerequisite for the high accuracy of the capacity prediction method of this invention. Only when the key process parameters are stably controlled can the temperature-precharge capacity baseline relationship curve established in the experiment be repeatable, the collected precharge data be consistent, and the calibrated intercept be accurate. Only then can the true differences between batches be accurately reflected.

[0098] To further demonstrate the application value of the battery formation method in this invention, the specific execution process of this invention is illustrated below, taking a 21700 cylindrical battery as an example.

[0099] Step S1: Obtain the dQ / dV curve of the target battery: Take 3 cells, precharge the cells with a small current of 0.05C, continuously collect voltage V and charging capacity Q data, calculate dQ and dV, and then perform exponential filtering according to the sliding window n=10 to calculate dQ / dV. Construct the dQ / dV curve based on dQ / dV.

[0100] Identifying key voltage ranges based on characteristic peaks under the dQ / dV curve: such as... Figure 2 The 0.05C V-dQ / dV curve shows the following characteristic peaks, as shown in Table 1.

[0101] Table 1 - Schematic diagram of characteristic peaks under the dQ / dV curve

[0102]

[0103] Step S2: Based on the key voltage range identified in step S1, design a phased charging strategy and use the phased charging strategy to perform the charging operation, as shown in Table 2 below, where each step corresponds to a specific voltage range.

[0104] Table 2 - Schematic Diagram of Phased Charging Steps

[0105]

[0106] By adopting the above-mentioned phased charging strategy, the pre-charging time is shortened from about 20 hours in the traditional process to about 3.5 hours, saving 16.5 hours, while the quality of the SEI film is comparable.

[0107] Step S3: To further verify the rationality of using only intrinsic pre-charge capacity (the characteristic factor used in this invention), this invention conducted a contribution analysis on other potential characteristic factors, as shown in Table 3 below;

[0108] Table 3 - Schematic diagram of the contribution analysis of different characteristic factors to capacity

[0109]

[0110] Table 3 shows that, through extensive experiments, it was found that the final capacity of the battery cell is highly linearly correlated with its intrinsic pre-charge capacity (i.e., the pre-charge capacity after excluding temperature interference). However, the measured pre-charge capacity is significantly affected by temperature. For every 5°C increase in temperature, the measured pre-charge capacity increases by approximately 10-40 mAh. Therefore, this invention does not directly use the measured pre-charge capacity and temperature as two parallel feature inputs. Instead, it first uses a temperature compensation curve to correct the measured pre-charge capacity to the standard temperature, obtains the intrinsic pre-charge capacity, and establishes a single-factor linear regression relationship (capacity prediction model) between intrinsic pre-charge capacity and capacity. Under this framework, the information of other potential feature factors such as total weight after liquid injection, electrode weight, and SEI film formation peak potential is already implicit in the intrinsic pre-charge capacity (correlation analysis shows that the correlation coefficients between these factors and intrinsic pre-charge capacity are all >0.85). Using the pre-charge temperature as a compensation parameter, the above operation eliminates the interference of temperature on prediction accuracy and maintains the simplicity of the single-factor linear regression relationship (capacity prediction model), reducing data acquisition and model maintenance costs.

[0111] Take the same batch of battery cells and precharge them at different temperature gradients, such as 20℃, 25℃, 30℃, 35℃, and 40℃. Use the staged charging strategy in step S2 to record the precharge capacity and corresponding temperature at different temperature gradients, as shown in Table 4 below.

[0112] Table 4 - Precharge Capacity under Different Temperature Gradients

[0113]

[0114] The baseline relationship curve between precharge capacity and temperature was obtained by fitting the above data: ,in, (Preferred value: -0.3 to -0.1) (Preferred 10-20) (Preferred value: 3000-4000), R²>0.98. R² is the core indicator for measuring the goodness of fit of the model in regression analysis. This value indicates that the temperature and the pre-charge capacity have a good quadratic function relationship.

[0115] Step S4: During the pre-charging process, take 100 cells and pre-charge them at 25°C as in step S2. Record the actual pre-charge capacity and actual pre-charge temperature of each cell. Based on the reference relationship curve, correct the actual pre-charge capacity of this batch of cells to the standard temperature and calculate the intrinsic pre-charge capacity.

[0116] ;

[0117] The pre-charge capacity of the 100 cells is as follows. Approximately 3770~3820mAh, with an average of about 3790mAh, actual pre-charge temperature The temperature range is approximately 26.5~29.5℃, with an average of approximately 28℃. Intrinsic precharge capacity at standard temperature 25℃ The range is approximately 3750~3800mAh, with an average of approximately 3775mAh.

[0118] Step S5: Based on the intrinsic pre-charge capacity Divide the capacity into 3 to 15 levels. The more capacity levels there are, the higher the prediction accuracy will be. The capacity gradients are shown in Table 5 below. Take 3 cells with representative intrinsic precharge capacity from each of the 5 capacity gradients (15 cells in total) and perform three capacity tests. Record the capacity test temperature and capacity test data for each cell and calculate the average value (representative value).

[0119] Table 5 - Intrinsic Precharge Capacity under Different Temperature Gradients

[0120]

[0121] Referring to the baseline relationship curve, temperature compensation was applied to these 15 cells, and the compensated capacity was calculated: ,in, For the calibrated capacity, This refers to the actual capacity measured during capacity testing. This is the average temperature during actual volume division. For temperature The reference precharge capacity, i.e. the precharge capacity-temperature reference curve function shown above, is shown in Table 6 below.

[0122] Table 6 - Schematic diagram of capacity after temperature compensation

[0123]

[0124] Based on the calibrated split capacity and the corresponding intrinsic pre-charge capacity, the baseline relationship curve in step S3 is calibrated, and the calibration intercepts for each gradient are calculated so that the calibrated curve can accurately predict the split capacity of the sample. The expression is: , , , ,in, To calibrate the intercept, Pre-charge capacity for intrinsic purposes, For the first The calibrated mean of samples with capacity gradients and satisfying , The total capacity gradient is given; different calibration intercepts can be obtained for samples with different capacity gradients. The capacity prediction curve is generated, and the specific calibration intercept data are shown in Table 7 below.

[0125] Table 7 - Calibration Intercept Values ​​at Various Temperature Gradients

[0126]

[0127] Step S6: For all cells in this batch, predict the capacity according to the corresponding formula based on the capacity gradient to obtain the predicted capacity. After the actual capacity allocation, compare it with the calibrated capacity allocation to verify the prediction accuracy. The specific data is shown in Table 8 below.

[0128] According to Table 8, the predicted capacity range of the 100-cell battery is approximately 4280~4330mAh, with an average of approximately 4310mAh. The temperature range is approximately 27.0~30.5℃, with an average of approximately 29℃. After temperature compensation, the capacity range is approximately 4270~4340mAh, with an average of approximately 4310mAh. The prediction error range is 0%~1%, and the standard deviation is approximately 0.2%.

[0129] Table 8 - Schematic diagram of cell precharge error data

[0130]

[0131] Then, take about 500 cells from the same batch and perform a pre-charge operation according to step S2. Take about 10 representative cells from each of the different capacity gradients and repeat the above experiment to verify the prediction accuracy. The specific data are shown in Table 9.

[0132] Table 9 - Schematic diagram of error sets for batch verification of battery cells with different precharge capacity gradients

[0133]

[0134] Step S7: Implement SPC control on key process parameters of the battery cell during pre-charging, and set the control window range, such as: narrowing the pre-charge current rate control limit from an error of 0.05C to 0.02C, the pre-charge cutoff voltage from 3900±50mV to 3900±20mV, the pre-charge ambient temperature from 25±5℃ to 25±2℃, and increasing the CPK (process capability index) value from 1.02 to 1.33. When the key process parameters exceed the control window setting range, mark the corresponding battery cell and divert it to the traditional capacity testing process. The remaining battery cells use the capacity prediction of this invention. Compare the formation time of this invention with that of the traditional capacity testing process to obtain specific data as shown in Table 10 below.

[0135] Table 10 - Comparison of formation time between the present invention and traditional processes

[0136]

[0137] Step S8: Using the calculated predicted capacity as the basis for battery grading and grouping, the method of the present invention was applied to 5 production batches for experiments. The specific data is shown in Table 11 below.

[0138] Table 11 - Schematic diagram of capacity prediction error for continuous production of multiple batches

[0139] Batch P 0.52% Batch Q 0.54% Batch R 0.51% Batch S 0.55% Batch T 0.47%

[0140] Based on the above specific execution process and table data, it can be seen that in terms of capacity prediction, based on the single-factor regression model with a high linear correlation between intrinsic precharge capacity and capacity gradation, in the batch verification of 500pcs cells, the mean prediction error of each capacity gradient is less than 0.7%, and the standard deviation of prediction error is controlled within 0.3%, indicating that the method has excellent prediction accuracy and stability. The mean prediction error of five consecutive production batches is stable between 0.47% and 0.55%, proving the consistency and reliability of the method in multi-batch production.

[0141] Furthermore, by adopting the predicted capacity to replace the capacity grading process, the settling time before sorting after pre-charging is shortened from 17 days to 7 days, and the total formation cycle is shortened from about 428 hours to about 171 hours, a reduction of about 60%. Only about 2% of the sample cells need to be actually calibrated for capacity grading, which reduces the investment in capacity grading equipment, energy consumption and factory occupancy, and realizes the integrated formation process of pre-charging and capacity prediction.

[0142] In summary, this invention guides a phased charging strategy through dQ / dV curve mechanism analysis, which significantly shortens the pre-charge time while ensuring the quality of the SEI film. By establishing a pre-charge capacity-temperature reference curve, the measured pre-charge capacity is corrected to the standard temperature to obtain the intrinsic pre-charge capacity, thus eliminating the interference of temperature fluctuations on the predicted input.

[0143] Based on this, the present invention adopts a single-factor linear regression relationship between intrinsic precharge capacity and capacity grading capacity (capacity prediction model), avoiding the high deployment and maintenance costs brought about by complex machine learning models. At the same time, combined with SPC control and stratified sampling calibration, the reliability of prediction results and batch-to-batch consistency are ensured. Finally, the predicted capacity is used as the basis for battery grading and grouping, eliminating the independent capacity grading process for all cells except for a small number of calibration samples. This realizes an integrated formation process of precharge and capacity prediction, which has good application value.

[0144] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A battery formation method integrating rapid pre-charging and capacity prediction, characterized in that, Specifically, the following steps are included: Step S1: Obtain the dQ / dV curve of the target battery, and identify the key voltage range based on the characteristic peaks under the dQ / dV curve; Step S2: Perform the charging operation using a phased charging strategy. Use the first current to charge in the critical voltage range and the second current to charge in the non-critical range, with the first current being less than the second current. Step S3: Precharge the same batch of cells, record the precharge capacity under different temperature gradients, establish a reference relationship curve between precharge capacity and temperature, and obtain the reference precharge capacity at the corresponding temperature. Step S4: During the pre-charging process, collect the actual pre-charging capacity and actual pre-charging temperature of the same batch of cells. Based on the reference relationship curve, correct the actual pre-charging capacity to the standard temperature to obtain the intrinsic pre-charging capacity at the standard temperature. Step S5: Based on the intrinsic precharge capacity, divide the same batch of cells into multiple capacity gradients. For each capacity gradient, extract several sample cells to complete the full capacity test. Collect the actual capacity and capacity temperature of the test samples. Perform temperature calibration on the actual capacity. Based on the calibrated capacity and intrinsic precharge capacity, solve the calibration intercept of each capacity gradient and establish a capacity prediction curve. Step S6: For cells in the same batch that have not undergone full capacity testing, calculate the predicted capacity of the cells based on the capacity prediction curve of the corresponding capacity gradient matched with the intrinsic precharge capacity. Step S7: Implement SPC control on the key process parameters of the battery cell during the pre-charging process, set the control window range, and when the key process parameters are within the control window set range, execute steps S4~S6 to calculate the predicted capacity of the corresponding battery cell. Step S8: Using the calculated predicted capacity as the basis for battery grading and grouping, except for cells whose key process parameters do not meet the control window setting range and several sample cells used for calibration, the independent capacity grading process for the remaining cells is cancelled, and they enter the subsequent standing and grouping process.

2. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, The key voltage range in step S1 includes at least one of the SEI film formation region, gas generation region, and phase transition region.

3. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 2, characterized in that, The non-critical region in step S2 is the normal lithium intercalation region, and the phased charging strategy specifically includes: The SEI film formation region is charged with a small current of 0.01C to 0.1C; The gas generation zone is charged with a small constant current of 0.02C to 0.25C. A small current of 0.03C to 0.3C is used for charging in the phase transition region; High-current charging of 0.1C~0.8C is used in the normal lithium intercalation region.

4. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, The formula for calculating the baseline precharge capacity in step S3 is as follows: ； in, For temperature The baseline precharge capacity, Pre-charge temperature, As the reference intercept, , The fitting coefficients are denoted as .

5. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, The formula for calculating the intrinsic precharge capacity in step S4 is as follows: ； in, For intrinsic pre-charge capacity, This is the actual pre-charged capacity. This is the actual pre-charge temperature. Standard temperature For temperature The baseline precharge capacity, For temperature The baseline precharge capacity.

6. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, In step S5, temperature calibration is performed on the actual capacity, specifically as follows: ； in, For the calibrated capacity, To satisfy the actual capacity measured after full capacity testing, This is the average temperature during actual volume division. Standard temperature For temperature The baseline precharge capacity, For temperature The baseline precharge capacity, For temperature The baseline precharge capacity; The calibration intercepts for each capacity gradient are calculated based on the calibrated full-capacity and the intrinsic pre-charge capacity, and are specifically expressed as follows: ， ； in, To calibrate the intercept, For intrinsic pre-charge capacity, For the first The calibrated mean of samples with capacity gradients and satisfying , This represents the total number of capacity gradients.

7. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, The formula for calculating the predicted capacity of the battery cell in step S6 is as follows: ； in, It is a constant. To predict capacity, For intrinsic pre-charge capacity, For the first Mean value after calibration of a capacity gradient sample.

8. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 7, characterized in that, Step S6 further includes: performing a capacity prediction operation on all cells in the same batch according to the capacity gradient, calculating the predicted capacity, performing a capacity grading operation on all cells, recording the actual capacity grading capacity and capacity grading temperature, performing temperature calibration on the actual capacity grading capacity to obtain the calibrated capacity grading capacity, comparing the calibrated capacity grading capacity with the predicted capacity, calculating the prediction error, and verifying the prediction accuracy based on the prediction error.

9. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, Step S7 further includes: when key process parameters exceed the control window setting range, marking the corresponding battery cell and diverting it to the traditional capacity testing process, wherein the control window range is... , This is the sample standard deviation of key process parameters calculated from historical stable production batches.

10. The battery formation method integrating rapid pre-charging and capacity prediction according to claim 1, characterized in that, The steps for obtaining the dQ / dV curve in step S1 are as follows: pre-charge the battery cell with a small current, continuously collect voltage and charging capacity data, calculate dQ and dV, perform exponential filtering based on a sliding window, calculate dQ / dV, and construct the dQ / dV curve based on dQ / dV, where Q is the charging capacity and V is the voltage.

Images