Battery pack quality judgment and charging optimization method

Abstract

The application provides a battery pack quality judgment and charging optimization method, which realizes real-time judgment of battery health state through data fusion and a support vector machine model, generates a preliminary charging curve in combination with historical scene data when performance attenuation is found, introduces a neural network to predict overcharging risk, dynamically adjusts charging parameters based on risk evaluation values by iterative optimization and fusion of temperature data, forms an optimized charging scheme, analyzes trend matching degrees of execution feedback to ensure consistency of the scheme, maintains system stability by using a cyclic updating mechanism, and finally uses a particle swarm optimization algorithm to expand the strategy, continuously enhances performance parameter sets and refreshes health state indicators by using updated data, so that adaptive optimal charging within a safe boundary of the battery pack and closed-loop unified health management in a whole life cycle are realized.

Description

Technical Field

[0001] This invention relates to a method for judging battery pack quality and optimizing charging. Background Technology

[0002] Battery pack quality assessment and charging optimization are crucial not only for equipment stability and safety but also for energy efficiency and lifespan. With the rapid development of industries such as electric vehicles and energy storage power stations, accurately evaluating battery pack performance and optimizing the charging process has become a key area of ​​technological innovation. Existing technologies often struggle to adapt to the complex changes in battery pack performance under different usage scenarios. Especially during long-term operation, battery performance is affected by multiple factors such as environment and load, leading to significant discrepancies between assessment results and actual performance. This discrepancy not only affects the accurate understanding of battery status but also makes the formulation of charging strategies less targeted, thereby reducing the overall system's operating efficiency.

[0003] Currently, the dynamic matching problem between battery pack quality assessment and charging strategy has not been effectively solved. On the one hand, the accuracy of quality assessment directly determines whether the charging scheme is reasonable. If the assessment is wrong, the charging process will deviate from the optimal path, charging at a high load even when the actual battery capacity has decreased, thus accelerating battery aging. On the other hand, this mismatch will further exacerbate the contradiction between the assessment basis and the actual effect, forming a vicious cycle, making it impossible for the system to adjust the strategy in a timely manner according to the actual state of the battery.

[0004] Therefore, if the system's assessment of the battery pack's health status is flawed, it could lead to overcharging or undercharging, affecting the battery pack's lifespan and potentially creating safety hazards. This disconnect between judgment and execution is precisely the key challenge this research aims to overcome. Establishing a highly consistent dynamic coordination mechanism between quality assessment and charging execution is a critical issue that urgently needs to be addressed. Summary of the Invention

[0005] This invention proposes a battery pack quality judgment and charging optimization method, which solves the collaborative management problem of integrating dynamic assessment of battery pack health status and safe optimization of charging strategy, ensuring that the battery is in the best charging state and extending battery pack life.

[0006] A method for battery pack quality assessment and charging optimization includes: Step S1: Obtain real-time voltage, current, and temperature data from the battery pack using sensors, and process the data using a data fusion method to obtain an initial set of performance parameters; Step S2: Based on the initial performance parameter set, the battery pack is classified and analyzed using a support vector machine model to determine the degree of performance degradation and obtain a health status index. If the health status index shows that the degradation exceeds a preset threshold, charging records under similar scenarios are obtained from historical data, and adjustment coefficients are determined through comparative analysis to obtain a preliminary charging curve. Step S3: Based on the preliminary charging curve, a neural network model is used to predict the response behavior of the battery pack in the current scenario, determine the potential overcharge risk, and obtain a risk assessment value. If the risk assessment value is higher than a preset threshold, the charging current and voltage upper limit are adjusted through an iterative optimization method, and real-time temperature data is fused to obtain an optimized charging parameter set. Step S4: Based on the optimized charging parameter set, obtain the voltage change trend after execution from the battery pack feedback, use the trend analysis method to determine the matching degree, and obtain the coordination consistency index. If the coordination consistency index meets the preset requirements, then the newly acquired data is incorporated through the cyclic update mechanism to determine the overall system stability and obtain the final dynamic coordination scheme for the continuous management of the battery pack. Step S5: Based on the final dynamic coordination scheme, the parameter range of the optimized charging parameter set is adjusted using the particle swarm optimization algorithm to generate an extended charging strategy. Through the extended charging strategy, updated real-time data is obtained from the battery pack. The updated real-time data is processed using a fusion method to obtain an enhanced performance parameter set and update the health status index.

[0007] In another aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the method described thereon.

[0008] In another aspect, the present invention also provides a computer program product, comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method described herein.

[0009] The beneficial effects of this invention are as follows: It uses data fusion and a support vector machine model to determine the battery health status in real time, and generates a preliminary charging curve based on historical scenario data when performance degradation is detected. Simultaneously, it introduces a neural network to predict overcharge risk, and dynamically adjusts charging parameters based on the risk assessment value and fused temperature data through iterative optimization to form an optimized charging scheme. Subsequently, it ensures the consistency of the scheme by analyzing the trend matching degree of the execution feedback, and maintains system stability using a cyclic update mechanism. Finally, it employs a particle swarm optimization algorithm to continuously enhance the performance parameter set and refresh health status indicators using updated data, achieving a closed-loop unification of adaptive optimal charging within the battery pack's safety boundary and full lifecycle health management. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0012] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0013] like Figure 1 The method for judging battery pack quality and optimizing charging, as shown, includes the following steps:

[0014] Step S1: Real-time voltage, current, and temperature data are obtained from the battery pack through sensors, and the data is processed using a data fusion method to obtain an initial performance parameter set.

[0015] Specifically, the real-time voltage, current, and temperature sequences of the acquired battery pack are filtered using a Kalman filter algorithm to obtain filtered data. An instantaneous power sequence is calculated based on the filtered voltage and current data, and a temperature change rate sequence is calculated based on the filtered temperature data. The instantaneous power sequence and the temperature change rate sequence are time-aligned and weighted to generate a comprehensive performance index sequence. If the value in the comprehensive performance index sequence exceeds a preset threshold, an online estimation algorithm for the equivalent resistance inside the battery pack is triggered. Based on the obtained equivalent resistance value and the filtered voltage and current data, the battery health state estimate in the initial performance parameter set is updated.

[0016] In one implementation, data is acquired in real time through a sensor network deployed on the battery pack. Voltage sensors are connected to each individual cell in the battery pack to monitor the output voltage of each cell; current sensors are installed in the charge and discharge circuit of the battery pack to capture current intensity; and temperature sensors are distributed at key locations in the battery pack, such as the surface of the battery cells or the cooling system, to obtain temperature readings. These sensors use wired or wireless transmission modules to send the data to a central processing unit in real time, ensuring the synchronization and accuracy of data acquisition.

[0017] A data fusion method is employed to process and integrate the collected voltage, current, and temperature data into a unified information framework to improve the reliability of parameter estimation. The specific process includes first preprocessing the raw data, such as filtering out noise and outliers; then applying a fusion algorithm to merge the data; and finally using a weighted average fusion method to adjust the weights of the voltage, current, and temperature data based on historical experience, ensuring that the sum of the weights of the three is 1 to balance the influence of each parameter and reduce bias caused by errors from a single sensor.

[0018] Understandably, the resulting initial set of performance parameters is used for subsequent battery management. This set includes the fused average voltage, current fluctuation rate, and temperature distribution. Through this fusion, the parameter set can reflect the real-time health status of the battery pack.

[0019] Step S2: Based on the initial performance parameter set, the battery pack is classified and analyzed using a support vector machine model to determine the degree of performance degradation and obtain a health status index. If the health status index shows that the degradation exceeds a preset threshold, charging records under similar scenarios are obtained from historical data, and adjustment coefficients are determined through comparative analysis to obtain a preliminary charging curve.

[0020] Specifically, the process involves obtaining multi-dimensional feature vectors from the initial performance parameter set, using a support vector machine model to perform high-dimensional space mapping and decision boundary delineation on the multi-dimensional feature vectors to obtain a classification label sequence, calculating a performance degradation score based on the classification label sequence, and extracting a charging record dataset from a historical data warehouse that matches the current battery pack environment if the performance degradation score exceeds a preset threshold. Time series alignment and differencing are then performed on the charging record dataset to determine an adjustment coefficient matrix. The degradation parameters from the health status indicators are then fused with the adjustment coefficient matrix to generate a preliminary charging curve parameter set. Finally, an interpolation algorithm is applied to fill in missing points using the preliminary charging curve parameter set to obtain the preliminary charging curve.

[0021] In one implementation, a support vector machine (SVM) model is used to classify and analyze the battery pack based on an initial set of performance parameters. In battery health monitoring, this model uses the initial set of performance parameters as input features. The fused parameters, such as the mean voltage, current fluctuation rate, and temperature distribution, are mapped to a high-dimensional space to handle nonlinear relationships. The model training process uses historical battery data as the training set, with labels including categories such as normal, healthy degradation, and severe degradation. Support vectors are calculated using kernel functions such as radial basis functions, and the classification boundary is determined based on these vectors.

[0022] After inputting the parameter set, the model outputs the probability of belonging to the health degradation category, thereby determining the degree of performance degradation and obtaining a health status index, such as a value between 0 and 1 representing the remaining lifespan proportion. The calculation of the health status index involves the model's decision function. First, the parameter set is standardized, then input into a trained support vector machine model. The maximum margin is solved through an optimization problem. During classification, the distance from the test sample to the hyperplane is calculated; if the distance indicates severe degradation, the index value is reduced. If the health status index shows degradation exceeding a preset threshold, charging records under similar scenarios are retrieved from historical data. Historical data is stored in a database, including voltage curves, current patterns, and environmental conditions from past charging cycles. A similarity matching algorithm is used to retrieve records similar to the current parameter set, such as matching historical charging data within the average voltage and temperature range.

[0023] The process of determining the adjustment coefficients involves aligning the current parameter set with selected historical charging records. First, key points such as the charging initiation voltage and peak current are extracted. Then, differences, such as the voltage deviation rate, are calculated. The adjustment coefficients are generated based on these differences; for example, the coefficient value is the ratio of the historical curve to the current deviation, used to scale the standard charging curve. If historical records show that charging time is prolonged at high temperatures, a charging duration factor is added to the adjustment coefficients to obtain a preliminary charging curve. This curve is then shifted and adjusted based on historical data to ensure safe charging.

[0024] The obtained preliminary charging curve modifies the parameters of the standard curve by adjusting the coefficient. The standard curve is defined as the constant current and constant voltage stage. After adjustment, the inflection point voltage or current value of the curve changes. After calculating the health status index, if it exceeds the threshold, an adjustment coefficient is determined to extend the constant voltage stage by, for example, by 1.2 times. The preliminary curve is thus formed and used to optimize the next charging.

[0025] Step S3: Based on the preliminary charging curve, a neural network model is used to predict the response behavior of the battery pack in the current scenario, determine the potential overcharge risk, and obtain a risk assessment value. If the risk assessment value is higher than a preset threshold, the charging current and voltage upper limit are adjusted through an iterative optimization method, and real-time temperature data is fused to obtain an optimized charging parameter set.

[0026] Specifically, a multi-dimensional time-series feature sequence is extracted based on the preliminary charging curve. A convolutional neural network model is used to perform convolution and pooling operations on the multi-dimensional time-series feature sequence to obtain a predicted response behavior vector. If the abnormal peak value in the predicted response behavior vector exceeds a preset threshold, historical overcharge event data is fused to determine a potential overcharge risk indicator. Monte Carlo simulation is performed on the potential overcharge risk indicator to generate a risk distribution probability, resulting in a risk assessment value sequence. If the maximum value in the risk assessment value sequence is higher than a preset threshold, a temperature data sequence is collected from a real-time sensor, and the temperature fluctuation frequency is analyzed using Fourier transform to determine the temperature influence factor. An iterative optimization framework is constructed based on the temperature influence factor, and the upper limit of the charging current is iteratively adjusted multiple times using a gradient descent algorithm. The upper limit of the voltage constraint is fused to obtain an intermediate charging parameter set. A particle swarm optimization algorithm is applied to the intermediate charging parameter set to search for the global optimal solution. The potential overcharge risk indicator and the temperature influence factor are fused, and an optimized charging parameter set is generated through matrix decomposition and reconstruction.

[0027] In one implementation, a neural network model predicts the battery pack's response behavior in the current scenario based on the initial charging curve. The voltage, current, and time series of the initial charging curve are used as input features, fed into the network for forward propagation calculations. Model training uses historical battery response data to learn the mapping relationship between the curve and actual behavior, for example, by adjusting weights through a backpropagation algorithm to minimize prediction error. This prediction process simulates the battery's voltage response and capacity changes during charging, and the output includes the expected temperature rise and charge accumulation curves, thereby evaluating the overall behavior. The neural network model's prediction involves activation functions such as ReLU to handle nonlinear relationships. First, the initial charging curve is sampled, extracting key points such as the initial voltage and peak current, which are then input into the trained model. The model generates a response behavior vector through layer-by-layer calculations, representing the battery's potential state changes in the current environment.

[0028] Neural networks can incorporate recurrent structures, such as RNN variants, when processing time-series data to capture dynamic patterns in charging curves. The model predicts the response, outputting behavioral indicators for subsequent judgment, avoiding direct reliance on empirical formulas. Based on these predictions, potential overcharging risks are assessed, yielding a risk evaluation value. Overcharging risk refers to the possibility of thermal runaway caused by the battery voltage exceeding the safe upper limit. The risk value is calculated by comparing the predicted response behavior with a standard threshold as the basis for assessment.

[0029] The risk assessment value is a numerical value between 0 and 1. The calculation process includes quantifying the difference between the area under the curve of the predicted curve and the ideal curve; a larger difference results in a higher value. If the risk assessment value is higher than a preset threshold, the charging current and voltage upper limit are adjusted using an iterative optimization method. This process involves gradient descent algorithms, iteratively updating from initial parameters. First, the objective function is set to minimize the combination of risk value and temperature influence. Then, in each iteration, the gradient is calculated and the current value and voltage upper limit are adjusted, gradually reducing the current to observe the simulated response until the risk value drops below the threshold. It can be understood that the core of the iterative optimization method is a loop feedback mechanism. In each iteration, the charging response is simulated using the current parameters to assess the new risk. If the target is not met, updates are made based on the deviation direction. For example, if the voltage upper limit is too high, leading to an increase in risk, the upper limit is reduced by 0.1V and recalculated. Stable parameters are generated through multiple iterations, avoiding errors from a single adjustment.

[0030] By fusing real-time temperature data, an optimized charging parameter set is obtained. Specifically, real-time temperature data is collected from sensors and incorporated into the iterative results. A weighted average is used, with temperature as a correction factor to adjust the final current and voltage. For example, the upper limit is further reduced at high temperatures. This fusion ensures the parameter set adapts to environmental changes and supports balanced charging. The entire process begins with a predicted response, iteratively adjusting if the risk is high. The parameter set after temperature fusion includes optimized current and voltage for safe management during actual charging.

[0031] Step S4: Based on the optimized charging parameter set, obtain the voltage change trend after execution from the battery pack feedback, use trend analysis method to determine the matching degree, and obtain the coordination consistency index. If the coordination consistency index meets the preset requirements, then the newly acquired data is incorporated through a cyclic update mechanism to determine the overall system stability and obtain the final dynamic coordination scheme for continuous management of the battery pack.

[0032] Specifically, based on the optimized charging parameter set, a voltage time series is obtained. The first derivative of the voltage time series is calculated using a sliding window method to obtain a voltage change trend vector. If the cosine similarity between the voltage change trend vector and a preset reference trend template is lower than a preset threshold, the dynamic time warping distance between the trend vector and the reference template is calculated to obtain a coordination consistency index. If the coordination consistency index meets preset requirements, the monitoring data sequence of the current charging cycle is aligned and spliced ​​with historical data from similar scenarios to form an extended time series dataset. A long short-term memory network model is used to extract the temporal features of the extended time series dataset. The Mahalanobis distance of the feature vectors under a continuous time window is calculated to determine the overall system stability state and generate a stability state identifier. The stability state identifier is fused with the optimized charging parameter set, and the optimal parameter adjustment vector is calculated using a Markov decision process based on state transition probabilities to obtain the final dynamic coordination scheme. The final dynamic coordination scheme is used to update the control instruction set of the battery management system.

[0033] In one implementation, the battery pack is charged according to an optimized set of charging parameters. Simultaneously, the voltage sensors built into the battery management system continuously collect terminal voltage data from representative cells or modules within the battery pack at a fixed sampling frequency. The data collection process covers the entire charging phase, from the initial voltage until the charging cutoff condition is triggered, forming an actual voltage curve that changes over time. Based on the acquired actual voltage curve, a trend analysis method is used to determine its matching degree with the expected change pattern. The matching degree is calculated by comparing the difference between the actual curve and the expected curve over the same time series. The expected change pattern originates from the theoretical charging curve corresponding to the optimized charging parameter set or the typical voltage trajectory of historical safe charging cases. Trend analysis first smooths the actual voltage curve to eliminate measurement noise, and then extracts its key trend characteristics, such as the voltage rise rate, the location and shape of the curve inflection points, and the voltage convergence characteristics after entering the constant voltage charging phase.

[0034] The consistency index can be defined as a normalized similarity score, mapping the calculated error to a range of 0 to 1, where 1 represents a perfect match and 0 represents a complete mismatch. During this process, the actual charging voltage curve is dynamically time-warped and aligned with a reference curve simulated based on optimized parameters. The correlation coefficient of the aligned sequence is then calculated and used as the consistency index. If the calculated consistency index is higher than a preset threshold, it indicates that the current charging process closely matches the optimization expectations, and the system is operating stably. At this point, a cyclical update mechanism integrates the newly acquired actual charging data into the historical database or model parameters. This cyclical update mechanism can be a sliding time window-based data pool management strategy. The system maintains a dataset of voltage curves containing the N most recent successful charging (i.e., the index met). When new compliant charging data is generated, this mechanism adds it to the data pool while removing the oldest historical data, maintaining a constant data pool capacity.

[0035] The integration process includes not only the original voltage sequence but also trend characteristic parameters extracted from this charge, such as the average charging rate and the duration of the constant voltage phase. This updated data will be used to fine-tune the benchmark model or reference curves relied upon for subsequent charging parameter optimization, making them more reflective of the latest aging state of the battery pack and the actual operating environment.

[0036] Based on the updated data pool, the overall system stability is assessed. This stability assessment relies on the analysis of the consistency index sequence from multiple charging processes within the data pool. The mean and variance of this index are calculated over a past period. If the mean remains consistently high and the variance remains low, the system is considered to be in a stable and coordinated state. Conversely, if the index shows a downward trend or increased volatility, it suggests potential battery performance degradation or changes in external conditions. After obtaining the stability assessment, a final dynamic coordination scheme is generated. This scheme is essentially a set of adaptively adjustable rules or parameters.

[0037] The dynamic coordination scheme includes: dynamically adjusting the initial value of the charging current for the next cycle based on the typical voltage rise rate calculated from the latest data pool; or dynamically fine-tuning the charging cut-off voltage threshold based on the recent voltage stabilization time during the constant voltage phase. If analysis reveals that the recent charging voltage curves have all reached the plateau period earlier than expected, the dynamic coordination scheme may suggest slightly increasing the equalization voltage for the next stage to compensate for any slight increase in battery internal resistance and ensure full charge. Ultimately, this dynamic coordination scheme is output to the charging control unit of the battery management system to guide the charging process of the battery pack in subsequent cycles, achieving continuous and adaptive energy management. The entire process forms a closed loop from parameter execution, feedback acquisition, trend analysis, data updates to scheme generation, enabling the charging strategy to be dynamically optimized as the battery state evolves.

[0038] Step S5: Based on the final dynamic coordination scheme, the parameter range of the optimized charging parameter set is adjusted using the particle swarm optimization algorithm to generate an extended charging strategy. Through the extended charging strategy, updated real-time data is obtained from the battery pack. The updated real-time data is processed using a fusion method to obtain an enhanced performance parameter set and update the health status index.

[0039] Specifically, key variables are extracted from the charging parameter set based on the dynamic coordination scheme. A particle swarm optimization algorithm is used to iteratively calculate the boundary values ​​of these key variables to generate an adjusted parameter range sequence. An extended charging strategy model is constructed using this adjusted parameter range sequence. This extended charging strategy model is then executed on the battery pack, and updated real-time data streams are collected from it. A fusion method is used to integrate the updated real-time data streams from multiple sources, and the integrated data feature vector is calculated to obtain an enhanced performance parameter set. If the feature vector of the enhanced performance parameter set matches a preset health threshold, the corresponding entry in the health status indicator database is updated. Continuous management instructions are generated from the updated health status indicator database, and the enhanced performance parameter set is integrated to achieve dynamic updates of the health status indicators.

[0040] In one implementation, based on the parameter adaptation rules established by the final dynamic coordination scheme, a particle swarm optimization algorithm is used to adjust and expand the parameter range of the optimized charging parameter set in order to further explore better charging performance boundaries. The particle swarm optimization algorithm simulates the foraging behavior of a flock of birds, treating each possible combination of charging parameters as a "particle" in the search space. Each particle has two attributes: position and velocity. Position represents a specific set of charging parameter values, such as the current value in the constant current stage, the voltage value in the constant voltage stage, and the state of charge threshold for stage transitions. Velocity represents the direction and magnitude of parameter value adjustment in the next iteration. During initialization, a certain number of particles are randomly generated near the current parameter values ​​given by the dynamic coordination scheme according to a preset fluctuation ratio, forming an initial particle swarm. The parameter set represented by each particle is substituted into a battery charging process simulation model or directly applied to an experimental charging cycle to evaluate its fitness. The design of the fitness function is crucial. This function typically considers charging speed, charging efficiency, and the impact on battery health, and can be defined as a weighted combination of total charging time, energy conversion efficiency, and the smoothness of the charging terminal voltage curve. During the algorithm iteration process, each particle updates its velocity and position based on its own historical best position and the historical best position found by the entire particle swarm. Through multiple iterations, the particle swarm gradually gathers towards the parameter region with higher fitness. Ultimately, the parameter range corresponding to the globally optimal position output by the algorithm constitutes the search space basis of the extended charging strategy.

[0041] The system extracts multiple estimates from different data sources or time segments for the same performance indicator, and then obtains an enhanced value through weighted fusion. Taking the key performance parameter of the battery's internal equivalent ohmic resistance as an example, the system can calculate an estimated resistance value R1 from the voltage jump data at the beginning of charging, another estimated value R2 from the slope of the voltage-current linear relationship in the middle of the constant current stage, and a reference value R3 predicted from the historical health state model. During the fusion process, a dynamic weight is assigned to each estimate. The weight depends on the signal-to-noise ratio and freshness of the data on which the estimate is based. For example, R1 is based on instantaneous data and may be greatly affected by noise, so it has a lower weight; R2 is based on a period of stable data and has a higher weight; R3 is based on long-term trends and provides a benchmark when the battery state does not change abruptly, so it has a moderate weight. Through weighted averaging, a fused enhanced internal resistance value R_fused is finally obtained. Similarly, this type of fusion estimation can be performed on performance parameters such as battery capacity and polarization time constant to form a set of enhanced performance parameters. This fusion method can effectively smooth out the random errors of a single estimation and improve the reliability of parameter estimation by comprehensively utilizing real-time and historical information.

[0042] The health status index is typically a comprehensive evaluation value used to quantify the degree of performance degradation of a battery relative to its brand-new state. The health status index is updated using the enhanced performance parameter set. It is primarily calculated using two core parameters: capacity retention rate and internal resistance growth rate. The actual usable capacity C_actual and ohmic internal resistance R_actual are extracted from the enhanced performance parameter set based on the current charging cycle. C_actual is compared to the battery's rated capacity C_nominal to obtain the capacity retention rate; R_actual is compared to the battery's initial internal resistance baseline R_initial to obtain the internal resistance growth rate. Then, through a pre-calibrated mapping relationship or empirical formula, the capacity retention rate and internal resistance growth rate are combined into a health status index value between 0 and 100%. For example, a weighted geometric average can be used, giving higher weight to the capacity retention rate because it has a more direct impact on range. Once new enhanced performance parameters are obtained, they are used to refresh the calculation input of the health status index, thus achieving dynamic updates. This update process allows the battery management system to more accurately grasp the aging progress of the battery pack, providing crucial information for subsequent maintenance decisions.

[0043] After updating the health status index, the value is fed back into the decision logic of the extended charging strategy. If the health status index indicates that the battery pack has entered an accelerated degradation phase, the extended charging strategy will automatically favor a gentler charging current and a lower cutoff voltage in subsequent parameter selections to slow down the degradation rate. When the health status index calculated from the fused data shows a downward trend for three consecutive updates, the system automatically triggers strategy adjustment, lowering the maximum allowable charging current limit in the extended strategy by a fixed percentage and generating an alarm message to prompt manual inspection. The entire process achieves closed-loop adaptive management from strategy optimization, data acquisition, fusion processing to status assessment and strategy readjustment, enabling charging control to closely follow the microscopic changes in the internal state of the battery pack.

[0044] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0045] In another aspect, the present invention also proposes a computer program product, comprising a computer program, characterized in that the computer program implements the above-described method when executed by a processor.

[0046] In particular, according to some embodiments of this disclosure, the processes described above can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0047] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a task data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated task data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0048] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol, such as HTTP (Hypertext Transfer Protocol), and can interconnect with digital task data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0049] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: determine the network connection status of the switch production line management application in response to detecting a query operation on a production collaboration document in the switch production line management application; replace the webpage entry information corresponding to the production collaboration document with target entry file information and load target webpage resource information in response to determining that the network connection status of the switch production line management application indicates an offline state, so as to display the webpage of the production collaboration document offline in the switch production line management application, wherein the target entry file information is the file information of the entry file corresponding to the webpage of the production collaboration document downloaded in advance, and the target webpage resource information is the resource information corresponding to the webpage stored locally; in response to determining that the network connection status of the switch production line management application indicates an online state and that the webpage resource information corresponding to the production collaboration document is not stored locally, download the webpage resource information of the webpage from the production line document server, wherein the webpage resource information includes an entry file and resource information; display the webpage of the production collaboration document in the switch production line management application according to the webpage resource information, and store the webpage resource information in a local database.

[0050] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including product-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0051] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for judging battery pack quality and optimizing charging, characterized in that, include: Step S1: Obtain real-time voltage, current, and temperature data from the battery pack using sensors, and process the data using a data fusion method to obtain an initial set of performance parameters; Step S2: Based on the initial performance parameter set, the battery pack is classified and analyzed using a support vector machine model to determine the degree of performance degradation and obtain a health status index. If the health status index shows that the degradation exceeds a preset threshold, charging records under similar scenarios are obtained from historical data, and adjustment coefficients are determined through comparative analysis to obtain a preliminary charging curve. Step S3: Based on the preliminary charging curve, a neural network model is used to predict the response behavior of the battery pack in the current scenario, determine the potential overcharge risk, and obtain a risk assessment value. If the risk assessment value is higher than a preset threshold, the charging current and voltage upper limit are adjusted through an iterative optimization method, and real-time temperature data is fused to obtain an optimized charging parameter set. Step S4: Based on the optimized charging parameter set, obtain the voltage change trend after execution from the battery pack feedback, use the trend analysis method to determine the matching degree, and obtain the coordination consistency index. If the coordination consistency index meets the preset requirements, then the newly acquired data is incorporated through the cyclic update mechanism to determine the overall system stability and obtain the final dynamic coordination scheme for the continuous management of the battery pack. Step S5: Based on the final dynamic coordination scheme, the parameter range of the optimized charging parameter set is adjusted using the particle swarm optimization algorithm to generate an extended charging strategy. Through the extended charging strategy, updated real-time data is obtained from the battery pack. The updated real-time data is processed using a fusion method to obtain an enhanced performance parameter set and update the health status index.

2. The battery pack quality judgment and charging optimization method as described in claim 1, characterized in that, Step S1 specifically includes: The real-time voltage, real-time current, and real-time temperature sequences of the acquired battery pack are filtered using a Kalman filter algorithm to obtain the filtered data. The instantaneous power sequence is calculated based on the filtered voltage data and the filtered current data, and the temperature change rate sequence is calculated based on the filtered temperature data. The instantaneous power sequence and the temperature change rate sequence are time-aligned and weighted to generate a comprehensive performance index sequence. If the value in the comprehensive performance index sequence exceeds a preset threshold, an online estimation algorithm for the equivalent resistance inside the battery pack is triggered. Based on the obtained equivalent resistance value and the filtered voltage and current data, the battery health state estimate in the initial performance parameter set is updated.

3. The battery pack quality judgment and charging optimization method as described in claim 1, characterized in that, Step S2 specifically includes: Obtain the multidimensional feature vectors in the initial performance parameter set, and use the support vector machine model to perform high-dimensional space mapping and decision boundary division on the multidimensional feature vectors to obtain a classification label sequence. Calculate the performance degradation score based on the classification label sequence. If the performance degradation score exceeds a preset threshold, extract the charging record dataset under similar scenarios that match the current battery pack environment from the historical data warehouse. Time series alignment and difference calculation are performed on the charging record dataset to determine the adjustment coefficient matrix. The attenuation parameter in the health status index is fused with the adjustment coefficient matrix to generate a preliminary charging curve parameter set. The missing points are filled by an interpolation algorithm using the preliminary charging curve parameter set to obtain the preliminary charging curve.

4. The battery pack quality judgment and charging optimization method as described in claim 1, characterized in that, Step S3 specifically includes: Based on the preliminary charging curve, a multidimensional time-series feature sequence is extracted. A convolutional neural network model is used to perform convolution and pooling operations on the multidimensional time-series feature sequence to obtain a predicted response behavior vector. If the abnormal peak value in the predicted response behavior vector exceeds a preset threshold, historical overcharge event data is fused to determine potential overcharge risk indicators. Monte Carlo simulation is performed on the potential overcharge risk index to generate a risk distribution probability and obtain a risk assessment value sequence. If the maximum value in the risk assessment value sequence is higher than a preset threshold, a temperature data sequence is collected from a real-time sensor, and the temperature fluctuation frequency is analyzed by Fourier transform to determine the temperature influence factor. Based on the temperature influence factor, an iterative optimization framework is constructed. The upper limit of the charging current is adjusted iteratively multiple times using the gradient descent algorithm. The upper limit of the voltage constraint is integrated to obtain an intermediate charging parameter set. The particle swarm optimization algorithm is applied to the intermediate charging parameter set to search for the global optimal solution. The potential overcharge risk index and the temperature influence factor are integrated. The optimized charging parameter set is generated through matrix decomposition and reconstruction.

5. The battery pack quality judgment and charging optimization method as described in claim 1, characterized in that, Step S4 specifically includes: The voltage time series is obtained based on the optimized charging parameter set. The first derivative of the voltage time series is calculated using the sliding window method to obtain the voltage change trend vector. If the cosine similarity between the voltage change trend vector and the preset reference trend template is lower than a preset threshold, the dynamic time warping distance between the trend vector and the reference template is calculated to obtain the coordination consistency index. If the coordination consistency index meets the preset requirements, the monitoring data sequence of the current charging cycle is aligned and spliced ​​with historical data of the same scenario to form an extended time series dataset. The temporal features of the extended time series dataset are extracted using a long short-term memory network model. The Mahalanobis distance of the feature vectors under continuous time windows is calculated to determine the overall system stability state and generate a stability state identifier. By integrating the stability state identifier with the optimized charging parameter set, the optimal parameter adjustment vector is calculated through a Markov decision process based on state transition probabilities to obtain the final dynamic coordination scheme. The final dynamic coordination scheme is used to update the control instruction set of the battery management system.

6. The battery pack quality judgment and charging optimization method as described in claim 1, characterized in that, Step S5 specifically includes: Based on the dynamic coordination scheme, key variables are extracted from the charging parameter set. The boundary values ​​of the key variables are iteratively calculated using the particle swarm optimization algorithm to generate an adjusted parameter range sequence. An extended charging strategy model is constructed using the adjusted parameter range sequence. The extended charging strategy model is executed for the battery pack, and updated real-time data streams are collected from it. The updated real-time data stream is processed by multi-source integration using a fusion method. The integrated data feature vector is calculated to obtain an enhanced performance parameter set. If the feature vector of the enhanced performance parameter set matches a preset health threshold, the corresponding entry in the health status index database is updated. Continuous management instructions are generated using the updated health status indicator database, and the enhanced performance parameter set is integrated to achieve dynamic updates of health status indicators.

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

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images