Battery pack equalization charging and discharging intelligent control method and system based on adaptive learning

Abstract

The invention discloses a battery pack equalization charging and discharging intelligent control method and system based on adaptive learning. The method comprises the steps that the SOC change trend of battery monomers is predicted, an SOC-time curve is obtained after correction, and the monomers are clustered and recombined to obtain a target battery pack; monitoring a real-time operation state of the target battery pack, when a voltage difference value, a temperature difference value, an SOC difference value or an internal resistance difference value between at least two batteries exceeds a first preset value, judging that the target battery pack is in an unbalanced state, and determining a regulation and control priority according to an unbalance degree; and when the priority is greater than a second preset value, adaptively adjusting the energy of different battery monomers through energy transfer, and otherwise, adaptively adjusting the charging and discharging rates of different battery monomers. According to the method, the SOC change trend of the single body can be predicted in advance, the consistency of the battery pack is ensured through clustering and regrouping, the battery charging and discharging are balanced by adopting a grading strategy when the battery pack is unbalanced, and the operation energy efficiency and the safety of the battery pack are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery pack charge and discharge management, and in particular to a battery pack balanced charge and discharge intelligent control method and system based on adaptive learning. Background Art

[0002] In modern battery management systems (BMS), ensuring consistency between cells in a battery pack is critical to maintaining system performance, extending service life, and ensuring safety. Traditional methods mainly rely on simple threshold-triggered balancing strategies, such as passive balancing, which dissipates energy through resistance, or active balancing based on fixed rules, such as constant current / voltage regulation. However, these methods have many limitations: 1) They are unable to judge future trends, resulting in delayed reactions, insufficient dynamic responses under different working conditions, and difficulty in flexibly responding to complex and changing working environments. 2) The balancing process is not sophisticated enough and inefficient, further resulting in waste of resources; 3) It is difficult to achieve accurate grouping in large-scale battery pack applications, which is not conducive to the long-term use and regulatory operation and maintenance of battery packs, and is prone to equipment aging, increasing potential fault hazards and safety risks. Summary of the invention

[0003] In order to solve at least one of the above-mentioned technical problems, the present invention provides a battery pack balanced charging and discharging intelligent control method and system based on adaptive learning.

[0004] In a first aspect, the present invention provides a battery pack balanced charging and discharging intelligent control method based on adaptive learning, the method comprising:

[0005] Collect historical operation data of multiple battery packs and input them into a pre-trained prediction model to obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend and obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp;

[0006] Clustering the SOC-time curves of each battery cell to obtain multiple clusters, and recombining the battery cells of the clusters to obtain multiple new target battery packs;

[0007] Monitor the real-time operating status of the target battery pack, and determine that the target battery pack is in an unbalanced state when a voltage difference, a temperature difference, a SOC difference, or an internal resistance difference between at least two batteries exceeds a first preset value;

[0008] The voltage difference, temperature difference, SOC difference or internal resistance difference are sorted, and the control priority is determined according to the degree of difference between the battery cells. The balancing strategy is started according to the control priority to adjust the operating status of different battery cells; wherein,

[0009] When the priority is greater than the second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack;

[0010] When the priority is less than or equal to the second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

[0011] In one embodiment, the adaptively adjusting the energy of different battery cells through energy transfer includes:

[0012] E tra =ΔV·I tra ·t·μ;

[0013] ΔV=V high -V low ;

[0014]

[0015] In the formula, E tra Represents the energy transferred from the high-voltage battery cell to the low-voltage battery cell, and ΔV is the current voltage V of the high-voltage battery cell high The current voltage V of the low voltage battery cell low The voltage difference, I tra represents the current used for energy transfer; t represents the duration of energy transfer, μ represents the conversion efficiency; k is the proportionality coefficient, and its value range is 0 <k<1,SOC high , SOC low Indicates the current SOC of high-voltage battery cells and low-voltage battery cells, SOC tar is the SOC of the target equilibrium point, C nom is the nominal capacity of the battery cell.

[0016] In one embodiment, the adaptively adjusting the charge and discharge rates of different battery cells includes:

[0017]

[0018] In the formula, I adj,i Represents the charge and discharge current adjusted for the i-th battery cell, I max is the maximum charge and discharge current allowed by the battery cell, α is the adjustment coefficient, and the value range is 0<α<1; SOC max is the maximum SOC difference between different battery cells, ΔSOC i is the difference between the ith battery cell and the average SOC, T c is the temperature compensation factor, which is determined according to the temperature-performance curve.

[0019] In one embodiment, after the balancing strategy is started according to the control priority to adjust the operating states of different battery cells, the method further includes:

[0020] Obtaining the energy conversion loss and participation frequency of each battery cell participating in the balancing strategy, and calculating the average value of the energy conversion loss and the average value of the participation frequency respectively;

[0021] Battery cells corresponding to the battery cells having an average value greater than the third preset value are identified and removed from the target battery group and replaced with new battery cells.

[0022] In one embodiment, the battery cells of the clusters are recombined to obtain a plurality of new target battery groups, including:

[0023] For any cluster, determine the relationship between the number of battery cells in the cluster and the number of batteries required by the target battery pack;

[0024] When the number of battery cells in the cluster is greater than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and the battery cells with the closest Euclidean distance are selected from the cluster based on the required number of batteries to form the target battery pack;

[0025] When the number of battery cells in the cluster is equal to the required number of batteries in the target battery group, the battery cells in the current cluster are directly combined into the target battery group;

[0026] When the number of battery cells in the cluster is less than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and new battery cells are added according to the average value of the Euclidean distance to form the target battery pack with the battery cells in the current cluster.

[0027] In one embodiment, the prediction model is trained based on any one of a decision tree, a random forest, a support vector machine or a neural network model.

[0028] In a second aspect, the present invention further provides a battery pack balanced charge and discharge intelligent control system based on adaptive learning, the system comprising:

[0029] A trend prediction unit, used to collect historical operation data of multiple battery packs, input it into a pre-trained prediction model, and obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend and obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp;

[0030] A battery grouping unit, used for clustering the SOC-time curves of each battery cell to obtain a plurality of clusters, and recombining the battery cells of the clusters to obtain a plurality of new target battery groups;

[0031] A balancing judgment unit, used to monitor the real-time operating state of the target battery pack, and judge that the target battery pack is in an unbalanced state when a voltage difference, a temperature difference, a SOC difference or an internal resistance difference between at least two batteries exceeds a first preset value;

[0032] The balancing control unit is used to sort the voltage difference, temperature difference, SOC difference or internal resistance difference, determine the control priority according to the difference between the battery cells, and start the balancing strategy according to the control priority to adjust the operating status of different battery cells; wherein,

[0033] When the priority is greater than the second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack;

[0034] When the priority is less than or equal to the second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

[0035] In a third aspect, the present invention further provides an electronic device comprising a processor and a memory, wherein the memory is used to store computer program code, and the computer program code comprises computer instructions. When the processor executes the computer instructions, the electronic device executes the method as described in the first aspect above and any possible implementation thereof.

[0036] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, wherein the computer program includes program instructions, and when the program instructions are executed by a processor of an electronic device, the processor executes the method as described in the first aspect above and any possible implementation thereof.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1) The present invention collects historical operating data of multiple battery packs and inputs the pre-trained prediction model to obtain the SOC change trend of each battery cell; uses the preset remaining life-SOC curve to correct these trends and generate an accurate SOC-time curve, so that the behavior pattern of the battery cell can be predicted quickly and accurately, providing a reasonable basis for subsequent clustering and regrouping; by clustering the SOC-time curve of each battery cell, multiple clusters are obtained, and the cells in these clusters are recombined into a new target battery pack. This ensures that the cells in each new battery pack have similar performance characteristics, enhances internal consistency, increases the service life of the battery pack equipment, reduces the risk of failure and safety hazards, and also facilitates subsequent refined management and maintenance.

[0039] 2) The present invention monitors the real-time operating status of the target battery pack, including but not limited to voltage difference, temperature difference, SOC difference and internal resistance difference, and sets a first preset value to determine whether it is in an unbalanced state. By comprehensively considering multiple factors, the limitations brought by a single indicator are avoided, and the diagnostic accuracy and decision-making accuracy are improved. Furthermore, the control priority is determined according to the degree of difference. When the priority is greater than the second preset value, energy transfer is used for balancing; otherwise, the purpose is achieved by adjusting the charge and discharge rate. By adopting a hierarchical strategy, the most appropriate balancing method can be selected according to the actual situation, which not only ensures a rapid response but also takes into account energy consumption control, thereby improving the overall balancing efficiency.

[0040] 3) After the balancing operation, the present invention obtains the energy conversion loss and participation frequency of each battery cell participating in the balancing strategy, and calculates the average value of the energy conversion loss and the average value of the participation frequency respectively; identifies the battery cells corresponding to the average value greater than the third preset value and removes them from the target battery pack, and replaces them with new battery cells. Through continuous self-learning and iteration, the system can gradually improve its control logic, better adapt to various application scenarios, and thus improve long-term stability and reliability.

[0041] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0043] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and are used to illustrate the technical solutions of the present disclosure together with the specification.

[0044] Figure 1A flowchart of a battery pack balanced charging and discharging intelligent control method based on adaptive learning provided by an embodiment of the present invention;

[0045] Figure 2 A schematic diagram of the structure of a battery pack balanced charging and discharging intelligent control system based on adaptive learning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1 , Figure 1 The present invention provides a flow chart of a battery pack balanced charging and discharging intelligent control method based on adaptive learning. Figure 1 As shown, the method includes:

[0048] S10. Collect historical operation data of multiple battery packs and input them into a pre-trained prediction model to obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend to obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp.

[0049] Battery SOC, or state of charge, is also called the remaining capacity. It is a parameter used to reflect the remaining capacity of the battery. It is numerically defined as the ratio of the remaining capacity to the battery capacity. Usually, when SOC = 0, it means that the battery is fully discharged, that is, there is no remaining power in the battery. When SOC = 1 (or 100%), it means that the battery is fully charged, that is, the power in the battery has reached its maximum capacity. However, the battery SOC cannot be measured directly. It is usually estimated by using certain external characteristics of the battery, such as internal resistance, open circuit voltage, temperature, and charge and discharge current parameters, using specific algorithms or formulas.

[0050] The SOC change trend usually refers to the way the SOC changes over time during the operation of the battery. Therefore, the SOC change trend is usually a SOC-time curve. When predicting the SOC change trend, this embodiment mainly uses voltage, current, temperature and timestamp data. The voltage directly reflects the energy level and charging state of the battery, and is one of the most intuitive indicators. The current determines the speed at which energy flows in and out, affecting the rate of change of the SOC. The temperature has an important influence on the rate of chemical reactions inside the battery, which in turn affects the performance and life. The timestamp gives the data a time dimension, which facilitates the analysis of trends and developments over time.

[0051] In order to quickly and accurately predict the SOC change trend, it is usually necessary to train a prediction model in advance based on a large amount of historical operation data. The details are as follows:

[0052] 1.1) Collecting data sets:

[0053] Collect multiple sets of voltage, current, temperature and timestamps collected during historical operation. Then process these data, including removing noise points and smoothing outliers, converting data of different dimensions to the same scale, and then extracting features, such as calculating statistics such as mean, standard deviation, maximum and minimum values, and constructing time series features, such as moving average, exponential smoothing, etc.

[0054] 1.2) Choose a suitable model:

[0055] In one embodiment, the prediction model is trained based on any one of a decision tree, a random forest, a support vector machine or a neural network model. The neural network model includes but is not limited to a long short-term memory network (LSTM), a convolutional neural network (CNN), a recurrent neural network (RNN), etc. For the prediction of SOC change trend, the LSTM model is preferably used because LSTM has good processing capabilities for time series data and is more robust.

[0056] 1.3) Training model:

[0057] Use historical operation data as the training set, and adjust hyperparameters through cross-validation and other methods until the model achieves optimal performance on the validation set. During the training process, it is necessary to monitor the changes in the loss function to ensure model convergence. During training, it is preferred to use supervised learning, using known past data as input and actual SOC changes as output labels for training. Continuously optimize the model weights through the back-propagation algorithm until the model can achieve satisfactory results on the test set.

[0058] Through the above three steps, a trained prediction model can be obtained. When it is necessary to predict the SOC change trend of each battery cell in the current battery pack, it is only necessary to collect the voltage, current, temperature and timestamp of each battery cell in the current battery pack and input them into the prediction model to quickly and accurately obtain the SOC change trend. Therefore, this embodiment enables the model to accurately capture the SOC change law of the battery cell through high-quality training, ensuring that the model has strong robustness and generalization ability.

[0059] It should be noted that as time goes by, the battery will inevitably age, resulting in a decrease in performance. If not corrected, the prediction model may ignore this natural phenomenon, resulting in deviations in the prediction results. In order to solve this problem, this embodiment introduces a remaining life-SOC curve, which can quantify the impact of aging, thereby correcting the SOC change trend and improving the accuracy of the SOC-time change curve.

[0060] Specifically, a standard curve describing the relationship between the remaining life of the battery and the SOC can be constructed based on a large amount of experimental data or information provided by the manufacturer. This curve reflects the process of gradual attenuation of the battery capacity as the number of charge and discharge cycles increases. The predicted SOC change trend is combined with the remaining life-SOC curve, and the prediction result is adjusted to take into account the aging effect. Specific methods include but are not limited to weighted average, interpolation or direct application of mathematical formulas for correction. The corrected SOC change trend is closer to the actual situation and fully takes into account the aging characteristics of the battery. Finally, based on the corrected SOC change trend and combined with the timestamp information, a detailed SOC-time curve is produced for each battery cell. These curves intuitively show the SOC evolution of the cell throughout its life cycle. Each curve represents the actual operation history of a battery cell, and the intuitive graphical representation helps to identify abnormally performing cells and guide further operation and maintenance measures.

[0061] S20 , clustering the SOC-time curves of the battery cells to obtain a plurality of clusters, and recombining the battery cells of the clusters to obtain a plurality of new target battery packs.

[0062] It should be noted that battery reorganization refers to the redistribution of monomers within a battery pack to form new combinations or subgroups. During use, due to manufacturing differences, inconsistent aging rates, different working conditions and other factors, the performance of each monomer in the battery pack will gradually differentiate, resulting in significant differences in key parameters such as SOC, voltage, and temperature. If the battery pack is used in this condition for a long time, it will accelerate the aging rate of the battery pack, reduce its service life, and increase safety hazards. To this end, the embodiment aims to cluster the SOC-time curves of each battery monomer to perform battery reorganization, thereby enhancing the matching between monomers within the battery pack, reducing the need for balancing, and reducing unnecessary energy losses. Further improve the safety and reliability of the battery pack and reduce potential risks.

[0063] Specifically, before clustering, first ensure that the SOC-time curves of all battery cells are within the same time range and have the same sampling frequency. If the data lengths of different cells are inconsistent, they can be aligned by interpolation or truncation. Extract key features from each SOC-time curve, such as average SOC, maximum and minimum values, fluctuation amplitude, slope change, etc. These features will be used for subsequent clustering analysis. Then, select a suitable clustering algorithm according to specific needs, such as K-means, hierarchical clustering, DBSCAN, etc. The K-means clustering algorithm is preferably used to speed up clustering. After selecting the clustering algorithm, it is necessary to further determine the number of clusters, that is, the number of clusters, which can be determined according to empirical rules or by methods such as the elbow method and the silhouette coefficient.

[0064] When performing cluster analysis, an appropriate distance measurement method is selected to calculate the similarity between any two SOC-time curves, such as Euclidean distance, dynamic time warping (DTW), etc. In this embodiment, Euclidean distance is preferably used. After clustering, multiple clusters can be obtained, and the monomers in each cluster should have relatively similar SOC change trends and characteristic performances.

[0065] In one embodiment, the battery cells of the clusters are recombined to obtain a plurality of new target battery groups, including:

[0066] For any cluster, determine the relationship between the number of battery cells in the cluster and the number of batteries required by the target battery pack;

[0067] When the number of battery cells in the cluster is greater than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and the battery cells with the closest Euclidean distance are selected from the cluster based on the required number of batteries to form the target battery pack;

[0068] When the number of battery cells in the cluster is equal to the required number of batteries in the target battery group, the battery cells in the current cluster are directly combined into the target battery group;

[0069] When the number of battery cells in the cluster is less than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and new battery cells are added according to the average value of the Euclidean distance to form the target battery pack with the battery cells in the current cluster.

[0070] After the clusters are obtained, because the Euclidean distance is used to determine the similarity division, each cluster may contain a different number of battery cells. However, when the batteries are combined to obtain the target battery pack, the number of battery cells is usually fixed and consistent. Therefore, this embodiment requires further allocation and scheduling for the case where the number of battery cells in the clusters is different.

[0071] For any cluster, first count the number of battery cells A in the cluster, and the required number of batteries for the target battery pack is B. Compare the size relationship between A and B;

[0072] Assuming A is greater than B, it means that there are too many battery cells in the current cluster. In order to maintain consistency and reduce unnecessary complexity, it is necessary to filter out the closest B battery cells. For example, the cluster has 10 batteries, but the battery pack only needs 6 battery cells. In this case, 6 battery cells need to be filtered out from these 10 battery cells. The similarity can be determined according to the Euclidean distance between the 10 battery cells, and then the last 4 battery cells with smaller similarity are eliminated in descending order of similarity, so that the 6 filtered battery cells can be combined.

[0073] Assuming that A is exactly equal to B, there is no need to allocate excess batteries and the batteries can be reorganized directly.

[0074] Assume that A is less than B, which means that the existing number of cells is insufficient to form a complete target battery pack, and new cells need to be added from other clusters or spare pools. In order to ensure the consistency of the newly added cells with the existing cells, the average value of the Euclidean distance is used to guide the selection, including: for each battery cell in the current cluster, calculate the Euclidean distance between it and all other cells and find the average value. Select new cells from other clusters or spare pools that have the smallest average Euclidean distance to the cells in the current cluster until the total number reaches B, and merge the selected supplementary cells with the cells in the current cluster to form a new target battery pack.

[0075] Therefore, by reorganizing the batteries in the above manner, we can ensure that the cells in each new battery pack are as consistent as possible, reducing the performance differences within the pack. By rationally allocating and utilizing the existing battery cell resources, waste can be avoided. This can simplify subsequent balancing operations, improve charging and discharging efficiency, and save costs.

[0076] S30, monitoring the real-time operating status of the target battery pack, and determining that the target battery pack is in an unbalanced state when a voltage difference, a temperature difference, a SOC difference or an internal resistance difference between at least two batteries exceeds a first preset value.

[0077] After obtaining the target battery pack, the real-time operating status of the target battery pack is further monitored. In order to determine whether the target battery pack is in an unbalanced state, this embodiment uses four indicators: voltage difference, temperature difference, SOC difference or internal resistance difference. The voltage difference between battery cells is the most direct indicator to measure battery consistency. When a battery has a voltage difference during the charging and discharging process, this usually means that some battery cells may reach a full charge or empty state earlier than other cells, resulting in insufficient charging or over-discharging, affecting the overall performance and life. At different temperatures, the chemical reaction rate of the battery will be different, which may cause some battery cells to age faster. In addition, high temperature will also accelerate the decay of battery materials. Therefore, the temperature difference can reflect the uniformity of the battery working environment and the effectiveness of the heat dissipation system. SOC is a measure of the remaining power of the battery. Even under the same external conditions, if there is a significant SOC difference between the batteries, it means that their internal states have deviated, and this deviation may be aggravated if it is not controlled. As the use time increases, the aging degree of each battery cell will not be exactly the same, which will cause differences in internal resistance between them. High internal resistance will not only affect the energy transfer efficiency of the battery, but may also cause increased heat generation, further deteriorating the battery condition. Therefore, when it is found that the voltage difference, temperature difference, SOC difference or internal resistance difference between at least two batteries exceeds the first preset value, it can be considered that the target battery group is in an unbalanced state.

[0078] S40, sorting the voltage difference, temperature difference, SOC difference or internal resistance difference, determining the control priority according to the degree of difference between the battery cells, and starting the balancing strategy according to the control priority to adjust the operating status of different battery cells.

[0079] This embodiment aims to prioritize balancing for situations with greater differences. For example, if two groups of differences are found to exceed the first preset value, the first group is the voltage difference, and the second group is the internal resistance difference, and the voltage difference is greater than the internal resistance difference. This means that the voltage difference situation is more urgent and needs to be prioritized for balancing. By identifying which battery cells have the most significant differences, limited balancing resources can be concentrated where they are most needed, ensuring the maximization of the overall system performance.

[0080] Specifically, when starting the balancing strategy to adjust the operating states of different battery cells, the following steps are included:

[0081] S401: When the priority is greater than a second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack.

[0082] In one embodiment, the adaptively adjusting the energy of different battery cells through energy transfer includes:

[0083] E tra =ΔV·I tra ·t·μ;

[0084] ΔV=V high -V low ;

[0085]

[0086] In the formula, E tra Represents the energy transferred from the high-voltage battery cell to the low-voltage battery cell, and ΔV is the current voltage V of the high-voltage battery cell high The current voltage V of the low voltage battery cell low The voltage difference, I tra represents the current used for energy transfer; t represents the duration of energy transfer, μ represents the conversion efficiency; k is the proportionality coefficient, and its value range is 0 <k<1,SOC high , SOC low Indicates the current SOC of high-voltage battery cells and low-voltage battery cells, SOC tar is the SOC of the target equilibrium point, C nom is the nominal capacity of the battery cell.

[0087] E tra The calculation model takes into account the conversion efficiency μ, thus calculating exactly how much energy is actually transferred from one monomer to another under given conditions. It ensures that energy is conserved during the energy transfer process and takes into account the inevitable losses in practical operation.

[0088] I tra The computational model introduces nonlinear terms It is used to enhance the response to large differences while avoiding overly aggressive adjustments for small differences. Specifically, when the SOC difference between the two cells is large, the nonlinear term will increase significantly, prompting faster energy transfer. For smaller differences, the impact of the nonlinear term is weaker, making the adjustment more gentle and smooth. At the same time, the model also combines the nominal capacity C nom and transfer time t, ensuring that the amount of energy transferred matches the actual energy storage capacity and operation time of the monomer.

[0089] So, the E traThe calculation model introduces nonlinear regulation, dynamic adaptation mechanism and consideration of conversion efficiency, which can accurately regulate the energy between battery cells while maintaining high efficiency. This method not only improves the consistency within the battery pack and extends the service life, but also enhances the safety and reliability of the system, and is suitable for battery management systems in various application scenarios.

[0090] S402: When the priority is less than or equal to a second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

[0091] In one embodiment, the adaptively adjusting the charge and discharge rates of different battery cells includes:

[0092]

[0093] In the formula, I adj,i Represents the charge and discharge current adjusted for the i-th battery cell, I max is the maximum charge and discharge current allowed by the battery cell, α is the adjustment coefficient, and the value range is 0<α<1; SOC max is the maximum SOC difference between different battery cells, ΔSOC i is the difference between the ith battery cell and the average SOC, T c is the temperature compensation factor, which is determined according to the temperature-performance curve.

[0094] In the above formula, α is the adjustment coefficient, which ranges from 0<α<1 and is used to control the adjustment intensity. A smaller α indicates a more conservative adjustment strategy, while a larger α means a more aggressive adjustment. max is the maximum SOC difference between different battery cells, that is, the difference between the highest and lowest SOC of all cells. c is the temperature compensation factor, which is determined according to the temperature-performance curve and reflects the effect of temperature on battery performance. When the SOC of a cell deviates greatly from the average value, the adjustment force increases significantly; when the deviation is small or close to the average value, the adjustment is more gentle. Considering the impact of temperature on battery performance, especially charging efficiency and safety, the temperature compensation factor T is used to adjust the battery performance. c To adjust the actual charge and discharge current. This ensures optimal performance and safe operation under different temperature conditions. By setting a maximum charge and discharge current I max , ensuring that the physical limits of the battery cells are not exceeded, thus protecting the battery from the risk of overcharging or over-discharging.

[0095] Therefore, the above formula can accurately adjust the charge and discharge rate between battery cells while maintaining high efficiency by introducing mechanisms such as nonlinear adjustment, temperature compensation, and maximum current limit. This method not only improves the consistency within the battery pack and extends the service life, but also enhances the safety and reliability of the system, and is suitable for battery management systems in various application scenarios. In this way, it is possible to better cope with complex and changing working conditions and ensure the long-term stable operation of the battery system.

[0096] In one embodiment, after the balancing strategy is started according to the control priority to adjust the operating states of different battery cells, the method further includes:

[0097] Obtaining the energy conversion loss and participation frequency of each battery cell participating in the balancing strategy, and calculating the average value of the energy conversion loss and the average value of the participation frequency respectively;

[0098] Battery cells corresponding to the battery cells having an average value greater than the third preset value are identified and removed from the target battery group and replaced with new battery cells.

[0099] Understandably, monomers that frequently participate in balancing and have high energy conversion losses often show asynchrony with other monomers. By identifying and removing these monomers, the consistency between the remaining monomers can be significantly improved, reducing unnecessary balancing needs, thereby simplifying management and improving efficiency.

[0100] In this embodiment, after executing the balancing strategy, the energy conversion loss of each battery cell participating in the balancing and the frequency of its participation in the balancing operation are recorded. These data can be monitored and stored in real time by the BMS. For each battery cell, the average value of its energy conversion loss and participation frequency is calculated respectively. This value can be based on a fixed time period, such as statistical data within a day or a week, or a dynamically updated rolling average.

[0101] Furthermore, a reasonable threshold value, namely the third preset value, is pre-set. It can be determined based on historical data and experimental results to judge which monomers have too high energy conversion loss and participation frequency. The average value of the energy conversion loss and the average value of the participation frequency of each monomer are compared with the corresponding third preset value respectively. If it is found that the third preset value is exceeded, the monomer is considered to be abnormal. Once a monomer is confirmed to be an abnormal monomer, it is immediately removed from the current target battery pack to prevent it from continuing to affect the performance and safety of the entire system. Select a new battery monomer that has passed the test to replace the removed monomer. The new monomer should have similar or better performance characteristics to ensure that no new inconsistency problems are introduced after replacement.

[0102] Therefore, through careful monitoring and analysis of energy conversion loss and participation frequency, abnormal monomers can be quickly located and corresponding measures can be taken. By rationally allocating and utilizing battery monomer resources, unnecessary losses caused by overuse of certain monomers can be avoided, while also ensuring that the system is always in the best configuration state.

[0103] In summary, the method provided by the present invention has at least the following beneficial effects:

[0104] 1) The present invention collects historical operating data of multiple battery packs and inputs the pre-trained prediction model to obtain the SOC change trend of each battery cell; uses the preset remaining life-SOC curve to correct these trends and generate an accurate SOC-time curve, so that the behavior pattern of the battery cell can be predicted quickly and accurately, providing a reasonable basis for subsequent clustering and regrouping; by clustering the SOC-time curve of each battery cell, multiple clusters are obtained, and the cells in these clusters are recombined into a new target battery pack. This ensures that the cells in each new battery pack have similar performance characteristics, enhances internal consistency, increases the service life of the battery pack equipment, reduces the risk of failure and safety hazards, and also facilitates subsequent refined management and maintenance.

[0105] 2) The present invention monitors the real-time operating status of the target battery pack, including but not limited to voltage difference, temperature difference, SOC difference and internal resistance difference, and sets a first preset value to determine whether it is in an unbalanced state. By comprehensively considering multiple factors, the limitations brought by a single indicator are avoided, and the diagnostic accuracy and decision-making accuracy are improved. Furthermore, the control priority is determined according to the degree of difference. When the priority is greater than the second preset value, energy transfer is used for balancing; otherwise, the purpose is achieved by adjusting the charge and discharge rate. By adopting a hierarchical strategy, the most appropriate balancing method can be selected according to the actual situation, which not only ensures a rapid response but also takes into account energy consumption control, thereby improving the overall balancing efficiency.

[0106] 3) After the balancing operation, the present invention obtains the energy conversion loss and participation frequency of each battery cell participating in the balancing strategy, and calculates the average value of the energy conversion loss and the average value of the participation frequency respectively; identifies the battery cells corresponding to the average value greater than the third preset value and removes them from the target battery pack, and replaces them with new battery cells. Through continuous self-learning and iteration, the system can gradually improve its control logic, better adapt to various application scenarios, and thus improve long-term stability and reliability.

[0107] See also Figure 2 In one embodiment, the present invention further provides a battery pack balanced charge and discharge intelligent control system based on adaptive learning, the system comprising:

[0108] The trend prediction unit 100 is used to collect historical operation data of multiple battery packs, input it into a pre-trained prediction model, and obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend and obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp;

[0109] The battery grouping unit 200 is used to cluster the SOC-time curves of each battery cell to obtain a plurality of clusters, and to regroup the battery cells of the clusters to obtain a plurality of new target battery groups;

[0110] The balancing judgment unit 300 is used to monitor the real-time operating state of the target battery pack, and judge that the target battery pack is in an unbalanced state when the voltage difference, temperature difference, SOC difference or internal resistance difference between at least two batteries exceeds a first preset value;

[0111] The balancing control unit 400 is used to sort the voltage difference, temperature difference, SOC difference or internal resistance difference, determine the control priority according to the difference between the battery cells, and start the balancing strategy according to the control priority to adjust the operating status of different battery cells; wherein,

[0112] When the priority is greater than the second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack;

[0113] When the priority is less than or equal to the second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

[0114] It can be understood that the functions or modules included in the system provided in this embodiment can be used to execute the method described in the above method embodiment. Its specific implementation can refer to the description of the above method embodiment. For the sake of brevity, it will not be repeated here.

[0115] The present invention also provides an electronic device, including a processor and a memory, wherein the memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device executes a method as described in any possible implementation manner.

[0116] The present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes a method as described in any possible implementation manner.

[0117] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0118] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. Those skilled in the art can also clearly understand that the descriptions of the various embodiments of the present invention have different focuses. For the convenience and brevity of description, the same or similar parts may not be repeated in different embodiments. Therefore, for parts not described or not described in detail in a certain embodiment, refer to the records of other embodiments.

Claims

1. A battery pack balanced charging and discharging intelligent control method based on adaptive learning, characterized in that: The method comprises: Collect historical operation data of multiple battery packs and input them into a pre-trained prediction model to obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend and obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp; Clustering the SOC-time curves of each battery cell to obtain multiple clusters, and recombining the battery cells of the clusters to obtain multiple new target battery packs; Monitor the real-time operating status of the target battery pack, and determine that the target battery pack is in an unbalanced state when a voltage difference, a temperature difference, a SOC difference, or an internal resistance difference between at least two batteries exceeds a first preset value; The voltage difference, temperature difference, SOC difference or internal resistance difference are sorted, and the control priority is determined according to the degree of difference between the battery cells. The balancing strategy is started according to the control priority to adjust the operating status of different battery cells; wherein, When the priority is greater than the second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack; When the priority is less than or equal to the second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

2. The method for controlling balanced charging and discharging of a battery pack based on adaptive learning according to claim 1, characterized in that: The method of adaptively adjusting the energy of different battery cells by energy transfer includes: E tra =ΔV·I tra ·t·μ; ΔV=V high -V low ; In the formula, E tra Represents the energy transferred from the high-voltage battery cell to the low-voltage battery cell, and ΔV is the current voltage V of the high-voltage battery cell high The current voltage V of the low voltage battery cell low The voltage difference, I tra represents the current used for energy transfer; t represents the duration of energy transfer, μ represents the conversion efficiency; k is the proportionality coefficient, and its value range is 0 <k<1,SOC high , SOC low Indicates the current SOC of high-voltage battery cells and low-voltage battery cells, SOC tar is the SOC of the target equilibrium point, C nom is the nominal capacity of the battery cell.

3. The method for controlling balanced charging and discharging of a battery pack based on adaptive learning according to claim 1, characterized in that: The adaptively adjusting the charge and discharge rates of different battery cells includes: In the formula, I adj,i Represents the charge and discharge current adjusted for the i-th battery cell, I max is the maximum charge and discharge current allowed by the battery cell, α is the adjustment coefficient, and the value range is 0<α<1; SOC max is the maximum SOC difference between different battery cells, ΔSOC i is the difference between the ith battery cell and the average SOC, T c is the temperature compensation factor, which is determined according to the temperature-performance curve.

4. The method for controlling balanced charging and discharging of a battery pack based on adaptive learning according to claim 1, characterized in that: After starting the balancing strategy according to the control priority to adjust the operating states of different battery cells, the method further includes: Obtaining the energy conversion loss and participation frequency of each battery cell participating in the balancing strategy, and calculating the average value of the energy conversion loss and the average value of the participation frequency respectively; Battery cells corresponding to the average values ​​being greater than the third preset value are identified and removed from the target battery group, and replaced with new battery cells.

5. The method for controlling balanced charging and discharging of a battery pack based on adaptive learning according to claim 1, characterized in that: The battery cells of the clustered clusters are recombined to obtain a plurality of new target battery groups, including: For any cluster, determine the relationship between the number of battery cells in the cluster and the number of batteries required by the target battery pack; When the number of battery cells in the cluster is greater than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and the battery cells with the closest Euclidean distance are selected from the cluster based on the required number of batteries to form the target battery pack; When the number of battery cells in the cluster is equal to the required number of batteries in the target battery group, the battery cells in the current cluster are directly combined into the target battery group; When the number of battery cells in the cluster is less than the required number of batteries of the target battery pack, the Euclidean distance of each battery cell in the current cluster is obtained, and new battery cells are added according to the average value of the Euclidean distance to form the target battery pack with the battery cells in the current cluster.

6. The method for controlling balanced charging and discharging of a battery pack based on adaptive learning according to claim 1, characterized in that: The prediction model is obtained by training based on any one of a decision tree, a random forest, a support vector machine or a neural network model.

7. An intelligent control system for balanced charging and discharging of battery packs based on adaptive learning, characterized in that: The system comprises: A trend prediction unit, used to collect historical operation data of multiple battery packs, input it into a pre-trained prediction model, and obtain the SOC change trend of each battery cell in the battery pack; use a preset remaining life-SOC curve to correct the SOC change trend and obtain the SOC-time curve of each battery cell; the historical operation data includes voltage, current, temperature and timestamp; A battery grouping unit, used for clustering the SOC-time curves of each battery cell to obtain a plurality of clusters, and recombining the battery cells of the clusters to obtain a plurality of new target battery groups; A balancing judgment unit, used to monitor the real-time operating state of the target battery pack, and judge that the target battery pack is in an unbalanced state when a voltage difference, a temperature difference, a SOC difference or an internal resistance difference between at least two batteries exceeds a first preset value; The balancing control unit is used to sort the voltage difference, temperature difference, SOC difference or internal resistance difference, determine the control priority according to the difference between the battery cells, and start the balancing strategy according to the control priority to adjust the operating status of different battery cells; wherein, When the priority is greater than the second preset value, the balancing strategy is to adaptively adjust the energy of different battery cells through energy transfer to balance the operation of the battery pack; When the priority is less than or equal to the second preset value, the balancing strategy is to adaptively adjust the charge and discharge rates of different battery cells to balance the operation of the battery pack.

8. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store computer program codes, wherein the computer program codes include computer instructions. When the processor executes the computer instructions, the electronic device executes the intelligent control method for balanced charging and discharging of a battery pack based on adaptive learning as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes the intelligent control method for balanced charging and discharging of a battery pack based on adaptive learning as described in any one of claims 1 to 6.

Images