Rare earth lithium battery high-speed fast charging intelligent management system based on rare earth composite material

Abstract

The application relates to the technical field of lithium battery intelligent management, and discloses a rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material. The system comprises a charging track construction module, a multi-scale segmentation module, an event identification and labeling module, a space-time alignment and mapping module, and a control strategy generation module. The charging track construction module forms a complete charging track composed of a charging state snapshot sequence. The multi-scale segmentation module processes to generate charging subspaces with different time spans. The event identification and labeling module identifies and labels key charging and discharging event points representing the characteristics of the rare earth material in each subspace. The space-time alignment and mapping module maps the labeling results with historical reference patterns. The control strategy generation module dynamically outputs real-time charging instructions accordingly. Through multi-scale analysis of the charging track, the evolution trend of the battery state can be predicted. Through accurate identification and mapping of key events of the rare earth characteristics, the charging control strategy is deeply matched with the intrinsic characteristics of the material, so that the fast-charging efficiency is improved under the premise of ensuring the safety and service life of the battery.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent management of lithium batteries, in particular to a rare earth lithium battery high-speed fast charging intelligent management system based on rare earth composite materials. BACKGROUND

[0002] At present, with the increasing demand for charging rate of electric vehicles and large-scale energy storage systems, rare earth lithium batteries with high energy density and excellent fast charging potential have become a research hotspot. The application of rare earth composite materials in battery electrodes can improve the lithium ion migration number and the stability of the electrode structure, but the charging and discharging process also introduces complex electrochemical behavior characteristics different from conventional lithium batteries. The existing battery management system generally adopts a control strategy based on preset thresholds and fixed charging curves. These strategies mainly rely on real-time monitoring and feedback adjustment of single parameters such as voltage, current, temperature, etc.

[0003] Such general management schemes have inherent defects. The control logic is based on parameter points in a short time window for judgment, lacking a continuous and overall perspective of the dynamic evolution of the battery in the entire charging process. The system cannot effectively capture the evolution trajectory and internal rules of the charging state in the time dimension, making it difficult to distinguish between normal electrochemical processes and potential risk precursors, resulting in lagging and rough control behavior. In order to ensure safety, the charging strategy tends to be conservative, and the fast charging capability of rare earth composite material batteries cannot be fully tapped.

[0004] The existing technical solutions are not designed for the characteristics of rare earth composite materials. Conventional battery models and historical databases are mostly based on traditional materials such as lithium iron phosphate and ternary lithium, and cannot identify and interpret the unique electrochemical signals caused by rare earth elements. This leads to a disconnection between the management strategy and the material nature of the battery, making it difficult to perform precise and powerful charging in the most suitable fast charging interval of the material, and also difficult to implement early intervention in the sensitive stage of the material, thereby limiting the full play of the performance advantages of rare earth lithium batteries. SUMMARY

[0005] The present application aims to provide a rare earth lithium battery high-speed fast charging intelligent management system based on rare earth composite materials to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides a rare earth lithium battery high-speed fast charging intelligent management system based on rare earth composite materials, which comprises:

[0007] a charging trajectory construction module for constructing a complete charging trajectory of a battery monomer in the charging process, the complete charging trajectory being composed of a series of charging state snapshots arranged in chronological order;

[0008] a multi-scale segmentation module configured to segment the complete charging trajectory into a plurality of charging subspaces with different time spans;

[0009] an event identification and labeling module configured to identify and label key charging and discharging event points representing characteristics of the rare earth composite battery in each of the charging subspaces;

[0010] a spatio-temporal alignment and mapping module configured to align and map the labeled charging subspaces with reference charging patterns in a battery monomer historical database in a spatio-temporal manner;

[0011] a control strategy generation module configured to dynamically generate a real-time charging control strategy for the battery monomer according to the mapping result and to issue the real-time charging control strategy to a charging execution device.

[0012] Preferably, the complete charging trajectory of the battery monomer during the charging process comprises:

[0013] voltage, current and temperature sequences of the battery monomer during the charging process are collected at a fixed frequency;

[0014] a unique timestamp identifier is assigned to each of the voltage, current and temperature values at each collection time point;

[0015] the voltage, current and temperature values at the same timestamp are combined into a charging state snapshot;

[0016] all the charging state snapshots are linked in the order of the timestamps to form the complete charging trajectory.

[0017] Preferably, the multi-scale segmentation of the complete charging trajectory to generate a plurality of charging subspaces with different time spans comprises:

[0018] a plurality of time window scales are defined, each of which corresponds to a specific time length;

[0019] each of the time window scales is used as a sliding window to slide over the complete charging trajectory, and each time a trajectory segment is intercepted;

[0020] each of the trajectory segments intercepted by the sliding is defined as a charging subspace;

[0021] a space identifier is assigned to each of the charging subspaces, which records the time window scale information corresponding thereto and the start and end timestamps in the complete charging trajectory.

[0022] Preferably, the identification and labeling of key charging and discharging event points representing characteristics of the rare earth composite battery in each of the charging subspaces comprises:

[0023] For each charging sub-space, extract the voltage sequence, current sequence and temperature sequence of all the charging state snapshots contained in it;

[0024] Calculate the first-order difference sequence and second-order difference sequence of the voltage sequence, and identify the voltage inflection point as a potential key event point;

[0025] Calculate the fluctuation feature of the current sequence, and identify the current mutation point as a potential key event point;

[0026] Calculate the rising slope of the temperature sequence, and identify the temperature rise anomaly point as a potential key event point;

[0027] Based on the preset electrochemical characteristics rules of rare earth composite materials, all the identified potential key event points are screened and merged to finally determine the key charging and discharging event points in the charging sub-space and label the event type.

[0028] Preferably, the spatio-temporal alignment and mapping of the labeled multiple charging sub-spaces and the reference charging patterns in the battery cell historical database includes:

[0029] Retrieve all historical charging records with the same identity as the battery cell from the battery cell historical database;

[0030] Extract the reference charging pattern stored in each historical charging record, which contains the type, occurrence time and context information of the historical key event points;

[0031] Match the multiple charging sub-spaces generated in the current charging process and labeled with key charging and discharging event points with the retrieved reference charging patterns respectively;

[0032] The matching process includes adjusting the time axis of the charging sub-space to align with the time axis of the reference charging pattern, and calculating the consistency measure of the type and relative timing of the key event points in the charging sub-space and the historical key event points in the reference charging pattern.

[0033] Preferably, the real-time charging control strategy for the battery cell is dynamically generated according to the mapping result, which includes:

[0034] For each charging sub-space, calculate a matching confidence according to the mapping result of the reference charging pattern;

[0035] If the matching confidence is higher than the preset threshold, extract the corresponding historical charging control parameters from the successfully matched reference charging pattern as the basic parameters;

[0036] Based on the specific numerical characteristics of the key charging and discharging event points in the current charging and discharging space, the basic parameters are fine-tuned to generate real-time charging control parameters applicable to the current time;

[0037] The real-time charging control parameters generated by all charging and discharging spaces are integrated in chronological order to form the real-time charging control strategy.

[0038] Preferably, the fine-tuning of the basic parameters based on the specific numerical characteristics of the key charging and discharging event points in the current charging and discharging space to generate real-time charging control parameters applicable to the current time includes:

[0039] Extracting the voltage measurement value, current measurement value and temperature measurement value of each key charging and discharging event point in the current charging and discharging space;

[0040] Calculating the feature deviation between the measurement value of each key charging and discharging event point and the corresponding reference value in the basic parameters;

[0041] According to the direction and size of the feature deviation, the adjustment direction and adjustment amplitude of the charging current set value, the cutoff voltage threshold and the temperature control threshold in the basic parameters are determined;

[0042] Using a proportional-integral-derivative control algorithm, the adjustment direction and adjustment amplitude are converted into specific correction amounts of the basic parameters;

[0043] The correction amount and the basic parameter are superimposed to obtain real-time charging control parameters applicable to the current charging and discharging space time range;

[0044] Preferably, the method further includes a verification and update step of the real-time charging control strategy:

[0045] After applying the real-time charging control strategy, continue to collect the subsequent charging state snapshots of the battery monomer to form a verification trajectory segment;

[0046] Comparing the verification trajectory segment with the expected charging behavior model to calculate the strategy execution deviation degree;

[0047] If the strategy execution deviation degree exceeds the allowable range, a control strategy update mechanism is triggered, which includes: based on the new key event points identified in the verification trajectory segment, re-executing the mapping step with the reference charging mode and generating an updated real-time charging control strategy.

[0048] Preferably, the method further includes an embedded assessment step of the battery monomer health state:

[0049] During the construction of the complete charging trajectory, a feature vector reflecting the stability of the internal rare earth composite material structure of the battery is extracted synchronously;

[0050] inputting the feature vector into a pre-trained battery health assessment model, the battery health assessment model outputting a current health index;

[0051] integrating the current health index as a hidden variable into the generation logic of the real-time charging control strategy, so that the real-time charging control strategy takes into account the real-time health status of the battery monomer when formulating.

[0052] Preferably, the method further comprises a cooperative management step across the battery monomers:

[0053] When the management target is a battery pack containing a plurality of battery monomers, a complete charging trajectory is independently constructed for each battery monomer in the battery pack and a respective real-time charging control strategy is generated;

[0054] A coordinator at the battery pack level is established, which receives the real-time charging control strategies of all battery monomers;

[0055] The coordinator optimizes the real-time charging control strategies of all battery monomers based on the current health index and real-time charging state of each battery monomer, and generates a set of cooperative charging control instructions to ensure the balance and safety of the overall charging process of the battery pack.

[0056] Compared with the prior art, the present application has the following advantages:

[0057] By constructing a complete charging trajectory composed of a series of charging state snapshots and performing multi-scale segmentation, panoramic state perception from micro-transient to macro-trend is achieved. This method elevates the charging process of the battery from a sequence of discrete data points to a continuous spatiotemporal trajectory object. In the charging subspace of different time spans, the system can simultaneously analyze long-term health state evolution, medium-term thermodynamic characteristics, and short-term dynamic response. This analysis dimension enables the control system to understand the evolution trend behind the instantaneous fluctuations of parameters, identify the early features of abnormal patterns that are masked in a single time scale, and thus realize the transition from passive response to proactive decision-making, making it possible to implement extreme optimization charging within the absolute safety boundary.

[0058] By identifying and labeling the key event points representing the intrinsic characteristics of the rare earth composite material in each charging subspace, and aligning and mapping these time-space sequences with event-specific markers to historical reference patterns, the deep coupling of control strategies and material intrinsic characteristics is achieved. The definition of key event points is rooted in the specific electrochemical behavior of rare earth composite materials. The time-space alignment mapping technology ensures that under different initial conditions and external disturbances, the current state of charge can find the most suitable historical optimal charging path as a reference. This enables the charging control to be tailored to the material's kinetic characteristics, maximizing the charging current within the phase or voltage window where the rare earth composite material exhibits high stability and high ionic conductivity, while avoiding sensitive areas or side reaction prone windows determined by material characteristics, thereby achieving simultaneous improvement of safety and fast charging efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 The working principle diagram of the rare earth lithium battery high-speed fast charging intelligent management system based on rare earth composite materials described in the present application;

[0060] Figure 2 The flowchart for constructing the complete charging trajectory of the battery monomer during the charging process;

[0061] Figure 3 The flowchart for identifying and labeling key charge-discharge event points in each charging subspace;

[0062] Figure 4 The time-space alignment consistency measurement diagram of the rare earth lithium battery fast charging subspace and historical reference patterns;

[0063] Figure 5 The dynamic variation diagram of the real-time charging current of the rare earth lithium battery fast charging. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0065] Please refer to Figure 1The application provides a rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material. The system comprises: a charging trajectory construction module responsible for collecting voltage, current and temperature data at a fixed frequency during battery charging, and assigning a time stamp to each data point, thereby combining the multi-parameter values under the same time stamp into a charging state snapshot, and linking these snapshots in chronological order to form a complete charging trajectory. A multi-scale segmentation module uses multiple predefined time window scales as a sliding window to slide on the complete charging trajectory to intercept trajectory segments of different time spans, each segment being defined as a charging subspace and assigned a unique identifier. An event identification and labeling module analyzes the differential characteristics and fluctuation patterns of the voltage, current and temperature sequences within each charging subspace, identifies key event points such as voltage inflection points or current mutation points that meet the electrochemical characteristics of the rare earth composite material, and performs type labeling. A space-time alignment and mapping module matches the labeled charging subspace with reference charging patterns in a historical database, adjusts the time axis to align and calculates the consistency measure of the event point type and timing. A control strategy generation module calculates the matching confidence based on the mapping results, extracts the basic parameters from the matched historical patterns, fine-tunes the real-time charging control parameters based on the current event point characteristics, and integrates them into a strategy to issue to the charging equipment.

[0066] Embodiment 1: refer to Figure 2 In the process of constructing the battery monomer charging trajectory, the system collects voltage sequence, current sequence and temperature sequence at a fixed frequency during charging. The sampling frequency is set according to the battery characteristics and system accuracy requirements, for example, collecting data once per second or using a higher frequency such as ten times per second. The voltage value, current value and temperature value of each acquisition time corresponding to the sensor reading are assigned a unique time stamp identifier. The time stamp identifier is derived from a high-precision clock chip, ensuring that all data points have a globally consistent and strictly monotonically increasing time sequence throughout the charging process. In specific implementation, the voltage sequence is obtained through a high-precision analog-to-digital converter, the current sequence is sampled through a Hall current sensor, and the temperature sequence is obtained through a thermocouple or multi-point temperature sensor array attached to the surface of the battery monomer. The voltage value, current value and temperature value under the same time stamp are combined into a charging state snapshot. The data structure of the charging state snapshot usually adopts a record form, including a time stamp field, a voltage value field, a current value field and a temperature value field. According to the chronological order of the time stamp, the system links all the charging state snapshots through pointers or array indexes to form a complete charging trajectory, which is stored and managed in the form of a time series database or a ring buffer in memory to support efficient data writing and traversal queries.

[0067] In the multi-scale segmentation stage, the system defines multiple time window scales, the types and specific lengths of which are set based on prior knowledge of the electrochemical relaxation process of the rare earth composite material battery. In specific implementations, the multiple time window scales can include second-level windows, minute-level windows, and ten-minute-level windows. For example, one specific configuration includes three scales: a short-scale window length of ten seconds, used to capture fast transient responses; a medium-scale window length of one minute, used to analyze polarization processes with medium time constants; and a long-scale window length of five minutes, used to observe slow thermal accumulation and material phase transition trends. Using each time window scale as a sliding window, the sliding is performed on the complete charging trajectory. The moving step of the sliding window can be set to a fixed value, such as half the window length, to achieve partial overlapping coverage of the complete charging trajectory, avoiding missing key features that occur at the window boundaries. Each sliding operation extracts a continuous trajectory segment from the complete charging trajectory, and the extraction process is completed by calculating the start and end time stamps in the complete charging trajectory. Each trajectory segment extracted by sliding is defined as a charging subspace, which is an independent data unit that contains complete data copies or reference pointers of all state-of-charge snapshots within the time window.

[0068] A spatial identifier is assigned to each charging subspace, which uses a structured coding method to record at least the corresponding time window scale information and the start and end time stamps in the complete charging trajectory. In some embodiments, the spatial identifier can be a composite key composed of the window scale code, the trajectory start index, and the trajectory end index. The spatial identifier enables subsequent processing modules to quickly locate the spatio-temporal attributes and scale range of the charging subspace. The multi-scale segmentation module performs the above sliding window segmentation operation on all pre-defined time window scales in parallel or series, ultimately generating a plurality of charging subspaces with different time spans, which collectively constitute a multi-granularity perspective representation of the same charging process.

[0069] In specific implementations, the construction of the complete charging trajectory and the multi-scale segmentation process can be performed in a pipelined manner, i.e., when the charging process begins, the charging trajectory construction module generates new state-of-charge snapshots in real time and appends them to the complete charging trajectory, while the multi-scale segmentation module continuously monitors the growth of the complete charging trajectory and dynamically applies the sliding window to generate new charging subspaces. It can be understood that the extraction operation of the sliding window needs to consider data boundary processing, for example, when the sliding window approaches the end of the complete charging trajectory, if the remaining data is insufficient to fill the entire window, a zero padding, hold, or wait strategy can be used. Optionally, the system can attach metadata to each charging subspace, including statistical features such as the mean of the voltage sequence and the variance of the current sequence within the space, for pre-screening by the event recognition module.

[0070] Embodiment 2: Refer to Figure 3 The event identification and labeling module operates on a charging sub-space, extracts voltage sequence, current sequence and temperature sequence of all charging state snapshots contained in the charging sub-space. Voltage sequence is composed of voltage measurement values in the charging sub-space arranged in time stamp order. First order difference sequence of voltage sequence is calculated by backward difference operation on consecutive time stamp voltage values, i.e. subtracting the previous voltage value from the next voltage value to obtain the first order difference sequence. Second order difference sequence of voltage sequence is calculated by performing the same backward difference operation on the first order difference sequence again. The process of identifying voltage inflection points as potential key event points involves setting a difference threshold, when the absolute value of second order difference sequence exceeds the preset threshold, the corresponding voltage sequence point is marked as a potential voltage inflection point. The fluctuation feature of current sequence is evaluated by calculating the standard deviation of current values in a sliding window, the length of the sliding window is configurable, for example, containing five consecutive sampling points. When the calculated standard deviation value is greater than the set fluctuation threshold, the current sampling point at the center of the window is identified as a current mutation point. The rising slope of temperature sequence is calculated by linear fitting method, least squares linear regression is performed on a segment of consecutive temperature data points in the charging sub-space, the slope value of the fitted straight line is the rising slope. When the slope value exceeds the upper limit of the normal range set based on the thermal characteristics of the battery material, the corresponding temperature sequence interval is marked as a temperature rise abnormal point.

[0071] All potential key event points identified are screened and merged based on preset rare earth composite material electrochemical characteristic rules, which specifically reflect the influence of rare earth element doping on the lattice structure and ion diffusion dynamics of electrode materials. In specific implementation, the rule library contains multiple logical judgment conditions, for example, the rule may stipulate that if two voltage inflection points are too close in time and the voltage change direction is consistent, they will be merged into a single event point representing solid solution phase transition; the rule may also stipulate that only when a current mutation point is accompanied by a voltage inflection point and the temperature change is within a certain tolerance range, the current mutation point will be retained as a valid point, which corresponds to the possible interface side reaction in rare earth composite materials. The key charging and discharging event points in the charging sub-space are finally determined and labeled with event types, and the event type labels come from a predefined set, which may include "rare earth lattice ordering transition", "interface lithium deposition initiation", "solid electrolyte interface film reconstruction" and other categories directly related to the characteristics of rare earth composite materials.

[0072] The spatio-temporal alignment mapping module retrieves all historical charging records with the same identity as the battery cell currently being charged from the battery cell historical database. The identity is usually a unique code of the battery cell. The reference charging pattern stored in each historical charging record is extracted, which is a data structure encapsulating key information in the historical charging process, including the type of historical key event point, the absolute occurrence time or relative occurrence time to the start of charging of each historical key event point in the historical charging process, and the context information of the historical key event point, which can include the average current, ambient temperature and other auxiliary parameters at the time of the event. The multiple charging subspaces generated in the current charging process, which are labeled with key charging and discharging event points, are respectively matched with the retrieved reference charging patterns.

[0073] The core of the matching process is the time axis alignment and consistency measure calculation. Adjusting the time axis of the charging subspace means performing a linear transformation on the timestamps of all event points inside the charging subspace, so that the time axis of the charging subspace is aligned with the time axis of the reference charging pattern in length and phase. The scaling factor s is determined by comparing the total time span of the charging subspace with the time span of the corresponding segment of the reference charging pattern, and the translation factor t is determined by finding the time offset between the first key event point in the charging subspace and the first historical key event point in the reference charging pattern. The consistency measure of the key event points in the charging subspace and the historical key event points in the reference charging pattern in type and relative timing is calculated, and the consistency measure value which can be calculated by a weighted function:

[0074]

[0075] wherein the character represents the final consistency measure value; the character represents the weight coefficient assigned to the event type similarity; the character represents the event type similarity score, which is calculated according to the matching degree of the type of the key event point in the charging subspace and the type of the corresponding historical key event point in the reference charging pattern; the character represents the weight coefficient assigned to the event sequence similarity; the character represents the event sequence similarity score, which is calculated based on the matching degree of the relative time interval between the key event points on the aligned time axis. In some embodiments, the dynamic time warping algorithm can be used to calculate , the algorithm can handle the non-linear stretching on the time axis. It can be understood that the matching process can find multiple candidate reference charging modes for the same charging subspace, at which the reference charging mode with the highest consistency measure value C is selected as the final mapping object. Alternatively, the matching process can also introduce the matching of spatial scales, that is, the charging subspace with similar time window scales is preferentially mapped with the reference charging mode with similar analysis granularity. The whole spatio-temporal alignment mapping process provides important historical basis and comparison benchmark for the subsequent control strategy generation.

[0076] Referring to Figure 4 The figure is the core visualization result of the spatio-temporal alignment mapping link. After identifying the key event points of the charging subspace, it needs to be matched with the reference charging mode in the battery historical database. This figure is the quantitative result of this matching process. The horizontal coordinate in the figure is four kinds of historical reference charging modes, and the vertical coordinate is the consistency measure value. The role of this figure is to provide a basis for the subsequent control strategy generation: the system selects the reference mode with the highest consistency measure, extracts its historical charging control parameters as the basic parameters, and then fine-tunes to obtain the current real-time charging strategy. This result also reflects that the system can accurately match the historical charging path that is adapted to the characteristics of rare earth composite materials, and provides data support for the material adaptability of the subsequent fast charging strategy.

[0077] Embodiment 3: The control strategy generation module starts the processing flow for each charging subspace. The system calculates a matching confidence according to the mapping result of the charging subspace and the reference charging mode. The matching confidence is obtained by normalizing the consistency measure value obtained in the mapping process, and the normalization range is set between zero and one. If the calculated matching confidence is higher than the preset threshold, which can be configured as zero point eight or other empirical value, the corresponding historical charging control parameters are extracted from the successfully matched reference charging mode as the basic parameters, which usually include the charging current set value, the cutoff voltage threshold and the temperature control threshold and other key control variables. Based on the specific numerical characteristics of the key charging and discharging event points in the current charging subspace, the basic parameters are fine-tuned, and the voltage measurement value, current measurement value and temperature measurement value of each key charging and discharging event point in the current charging subspace are extracted. These measurement values are directly read from the charging state snapshot. The feature deviation between the measurement value of each key charging and discharging event point and the corresponding reference value in the basic parameters is calculated. The direction of the feature deviation is determined by the size comparison of the measurement value and the reference value, for example, when the voltage measurement value is greater than the reference value, the deviation direction is positive. The size of the feature deviation is calculated by the absolute value difference or the relative difference, and the relative difference uses the percentage difference value of the measurement value and the reference value.

[0078] According to the direction and size of the characteristic deviation, the adjustment direction and adjustment amplitude of the basic parameters of the charging current set value, the cut-off voltage threshold and the temperature control threshold are determined, the adjustment direction is determined according to the deviation sign, and the adjustment amplitude is proportional to the deviation size. The adjustment direction and adjustment amplitude are converted into the specific correction amount of the basic parameters by using the proportional-integral-derivative control algorithm. The proportional term of the proportional-integral-derivative control algorithm processes the characteristic deviation at the current time, the integral term accumulates the sum of historical deviations, and the differential term predicts the trend of the deviation. The calculation of the correction amount can be expressed by the following formula:

[0079]

[0080] Wherein: the character represents the specific correction amount of the basic parameters, including the adjustment amount of the current, voltage and other parameters; the character represents the proportional gain coefficient, controlling the response strength of the current deviation; the character represents the current calculated characteristic deviation value; the character represents the integral gain coefficient, controlling the correction strength of the historical deviation accumulation; the character represents the integral of the characteristic deviation in time, which is obtained by multiplying the deviation value by the time interval ; the character represents the differential gain coefficient, controlling the suppression degree of the deviation rate; the character represents the rate of change of the characteristic deviation, which is calculated by dividing the difference between the current deviation and the deviation at the last time by the time interval. The calculated correction amount is superimposed with the basic parameters, and the superimposition operation is vector addition, obtaining the real-time charging control parameters suitable for the current charging time range. The real-time charging control parameters generated by all charging subspaces are integrated in time sequence, and the integration process is sorted based on the start and end time stamps in the space identifier of the charging subspaces, forming a complete real-time charging control strategy. The real-time charging control strategy is finally converted into a device recognizable instruction sequence and sent to the charging execution device.

[0081] In specific implementation, the calculation of the matching confidence can introduce a weighting factor, for example, the consistency measurement value is weighted and then normalized according to the historical use times of the reference charging mode or the health state of the battery monomer. The extraction of the basic parameters may involve interpolation or extrapolation. When the time scale of the reference charging mode is not completely consistent with the current charging sub-space, the system resamples the historical control parameters on the time axis to match the current context. The calculation of the characteristic deviation can use a multivariate comprehensive deviation index, for example, the deviations of voltage, current and temperature are fused into a characteristic deviation value The proportional-integral-derivative control algorithm is input again. Optionally, the fine-tuning process can introduce fuzzy logic control, combining the fuzzy set of feature deviations with the fuzzy rule base of adjustment amounts to handle nonlinear feature deviations. The integration of real-time charging control parameters needs to consider the problem of time overlap. When there is overlap in the time range of adjacent charging subspaces, the system performs weighted averaging or selects the parameter value with higher confidence in the overlapping region.

[0082] Referring to Figure 5 This figure is a visual representation of the control strategy generation link, which extracts basic parameters based on the spatiotemporal mapping results, fine-tunes them in combination with the characteristics of the current key event point, and generates real-time charging control parameters. This figure is exactly the adjustment process of the charging current parameter. The horizontal axis in the figure represents the charging time, and the vertical axis represents the charging current. The dashed line represents the basic current, and the solid line represents the adjusted current. This figure directly reflects the design of the control strategy that dynamically adapts to the battery state: through fine-tuning, both the efficiency loss of conservative charging and the electrochemical characteristics of rare earth composites are adapted, achieving a balance between fast charging efficiency and safety.

[0083] Example 4: After applying the real-time charging control strategy, the system continues to collect subsequent charging state snapshots of the battery monomer at a fixed frequency. The subsequent charging state snapshots include voltage, current, and temperature readings. These snapshots are arranged in chronological order to form a verification trajectory segment. The time length of the verification trajectory segment covers the complete period from the application of the control strategy to the next evaluation point. The system compares the verification trajectory segment with the expected charging behavior model, which is a reference trajectory trained based on a large amount of historical normal charging data and describes the standard mode of changes in battery voltage, current, and temperature over time under ideal control. The strategy execution deviation degree is calculated, which is a scalar index quantifying the difference between the verification trajectory segment and the expected charging behavior model. The strategy execution deviation degree It can be calculated by the following formula:

[0084]

[0085] Wherein: the character represents the final calculated strategy execution deviation degree; the character represents the total number of charging state snapshots contained in the verification trajectory segment; the character represents the weight coefficient of the voltage sequence; the character represents the voltage measurement value of the i-th snapshot in the verification trajectory segment; the character represents the voltage reference value at the corresponding time point in the expected charging behavior model; the character represents the weight coefficient of the current sequence; the character represents the current measurement value of the i-th snapshot in the verification trajectory segment; and the character a current reference value representing a corresponding time point in the expected charging behavior model; character a weight coefficient representing a temperature sequence; character a temperature measurement value representing the i-th snapshot in the verification trajectory segment; character a temperature reference value representing a corresponding time point in the expected charging behavior model. If the calculated strategy execution deviation degree exceeds the preset allowable range, a control strategy updating mechanism is triggered, which includes a mapping step of re-executing the reference charging mode based on the identified new key event points in the verification trajectory segment, and generating an updated real-time charging control strategy.

[0086] The embedded evaluation step of the battery monomer health state is executed synchronously in the process of constructing the complete charging trajectory, and the system synchronously extracts a feature vector that can reflect the stability of the rare earth composite material structure inside the battery. The feature vector contains multiple dimensional indicators. The extracted feature vector is input into a pre-trained battery health degree evaluation model. The battery health degree evaluation model is trained using a supervised learning algorithm. The battery health degree evaluation model outputs a current health index, which is a value ranging from zero to one. The closer the value is to one, the better the health state. The current health index is integrated into the generation logic of the real-time charging control strategy as a hidden variable. The hidden variable means that the current health index is not directly output as a control parameter, but is an internal input condition of the strategy generation algorithm. The real-time charging control strategy takes into account the real-time health state of the battery monomer when formulating by means of constraint conditions or objective function weighting, for example, when fine-tuning the charging current setting value, a scaling factor based on the current health index is introduced, and the upper limit of the current adjustment is automatically reduced when the health index is low.

[0087] In specific implementation, the expected charging behavior model can be a multi-dimensional time series model, such as a vector autoregressive model or a long short-term memory network model, which can predict the expected evolution path of the battery parameters under a given control strategy. The allowable range of the strategy execution deviation degree can be dynamically set according to the battery type and application scenario, for example, a typical allowable range value can be set to zero point five. The trigger of the control strategy updating mechanism can be immediate trigger or delayed trigger, the immediate trigger means that the update is started immediately once the deviation exceeds the limit, and the delayed trigger allows the deviation to continue for several sampling periods before confirming the update to avoid false judgments caused by noise. Referring to Table 1, the specific composition of the feature vector input into the battery health degree evaluation model.

[0088] Table 1: Composition table of battery health state evaluation feature vector

[0089] Feature name Data source Description Constant current charge internal resistance rate of change Voltage sequence, current sequence Calculate the amount of ohmic internal resistance change per unit time during the constant current phase Charge transfer impedance fit value Electrochemical impedance spectroscopy or relaxation voltage analysis Charge transfer impedance parameters from equivalent circuit model fit Voltage relaxation time constant Voltage recovery curve after charge stop Time constant characterizing the rate of polarization voltage dissipation Differential capacitance curve peak shift Voltage vs. capacity differential curve Analyze the shift in differential capacitance peak relative to new battery state Thermal generation rate steady state value Temperature sequence, current sequence Steady increase in battery temperature per unit time under steady charge current

[0090] It can be understood that the extraction of the feature vector relies on the in-depth analysis of the complete charging trajectory, especially the feature capture of the charging relaxation phase and the specific current step response. The training data of the battery health assessment model comes from a large number of cycle test data of the same type of battery monomer throughout its life cycle. The specific way of integrating the current health index into the control strategy generation logic can be achieved by modifying the gain coefficient of the proportional integral derivative control algorithm. Alternatively, the system can establish a mapping relationship table between the health index and the maximum allowed charging current, directly limiting the upper limit of the current set value in the real-time charging control parameter.

[0091] In embodiment 5, when the management target is a battery pack containing multiple battery monomers, the system independently constructs a complete charging trajectory for each battery monomer in the battery pack and generates a respective real-time charging control strategy. The complete charging trajectory construction process for each battery monomer follows the same sampling and snapshot linking rules, and the real-time charging control strategy generation process for each battery monomer is independent and does not interfere with each other. A coordinator at the battery pack level is established, which runs as an independent software module or hardware logic unit. The coordinator receives the real-time charging control strategies of all battery monomers, including the identifier, real-time charging control parameter sequence, and timestamp information of each battery monomer. Based on the current health index and real-time charging state of all battery monomers, the coordinator performs overall optimization of the real-time charging control strategies of each battery monomer. The current health index comes from the output of the health state assessment model of each battery monomer, and the real-time charging state includes the current voltage, current, and temperature measurements of each battery monomer. The overall optimization process aims to generate a set of coordinated charging control instructions, which will replace the independent real-time charging control strategies of each battery monomer and be sent to the charging execution device to ensure the balance and safety of the overall charging process of the battery pack.

[0092] In specific implementations, the coordinator can use a multi-objective optimization algorithm for overall calculation. The objective function is usually set to minimize the standard deviation of the voltages among all monomers in the battery pack, minimize the difference between the maximum temperature and the average temperature, and maximize the overall charging efficiency of the battery pack. The optimization process is subject to a series of constraints, including the maximum allowed charging current of each battery monomer, the total input power limit of the battery pack, and the cutoff voltage safety threshold of each battery monomer. By solving this constrained multi-objective optimization problem, the coordinator obtains a set of Pareto optimal solutions, then selects an optimal solution according to the priority of the actual application scenario, and determines the adjustment amount of the real-time charging control strategy for each battery monomer. When generating the coordinated charging control instructions, the coordinator will package the optimized parameters, such as the corrected charging current value and voltage threshold for each battery monomer, into a unified instruction format.

[0093] It is understood that inconsistency among battery cells within a battery pack is normal, and the core role of the coordinator is to suppress the expansion of such inconsistency through global optimization. The coordinator receives the current health index of all battery cells, and the battery cell with a lower health index will be assigned more conservative charging parameters in the optimization process, such as reducing the weight of its charging current, to protect the weak cell from accelerating aging. The coordinator also monitors the real-time charging state of all battery cells, and if it detects that the voltage or temperature of a certain battery cell abnormally rises, the coordinator will immediately intervene, dynamically adjust the constraint conditions of the optimization problem, temporarily limit the charging power of the abnormal cell or even start the suspension of the charging instruction, while adjusting the charging strategy of other normal cells to maintain the overall charging progress of the battery pack.

[0094] In some embodiments, the optimization algorithm of the coordinator can be implemented based on a model predictive control framework, which utilizes the equivalent circuit model or electrochemical model of the battery pack to predict the state evolution of each cell in the future period of time, so as to make more forward-looking optimization decisions. The coordinator maintains a battery pack topology mapping table, which records the series-parallel relationship between battery cells, because the current in the series loop must be consistent, and the voltage between parallel branches needs to be balanced, and the topology information is a key basis for feasible region calculation and instruction generation. Optionally, the coordinator can introduce a dynamic weight adjustment mechanism to assign different weights to different sub-goals in the optimization objective function at different stages of charging, for example, paying more attention to efficiency at the beginning of charging and paying more attention to voltage balance at the end of charging. The collaborative charging control instruction is finally sent to the charging execution device corresponding to each battery cell, such as a bidirectional DC-DC converter or a contactor controller, through a communication bus.

[0095] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A high-speed fast-charging intelligent management system for a rare earth lithium battery based on a rare earth composite material, characterized in that, The system comprises: a charging trajectory construction module for constructing a complete charging trajectory of a battery cell in a charging process, the complete charging trajectory being composed of a series of charging state snapshots arranged in time stamp order; a multi-scale segmentation module for multi-scale segmentation of the complete charging trajectory to generate a plurality of charging subspaces with different time spans; an event identification and labeling module for identifying and labeling key charging and discharging event points representing characteristics of rare earth composite material batteries in each of the charging subspaces; a spatio-temporal alignment and mapping module for spatio-temporal alignment and mapping of the labeled plurality of charging subspaces with reference charging patterns in a battery cell historical database; a control strategy generation module for dynamically generating a real-time charging control strategy for the battery cell according to the mapping result and issuing the real-time charging control strategy to a charging execution device; the identification and labeling of key charging and discharging event points representing characteristics of rare earth composite material batteries in each of the charging subspaces comprises: for a charging subspace, extracting voltage sequences, current sequences and temperature sequences of all charging state snapshots contained therein; calculating first-order and second-order differential sequences of the voltage sequences to identify voltage inflection points as potential key event points; calculating fluctuation features of the current sequences to identify current mutation points as potential key event points; calculating rising slopes of the temperature sequences to identify temperature rise abnormal points as potential key event points; based on preset electrochemical characteristics rules of rare earth composite materials, screening and merging all identified potential key event points to finally determine key charging and discharging event points in the charging subspace and label event types thereof; the dynamic generation of a real-time charging control strategy for the battery cell according to the mapping result comprises: for each charging subspace, calculating a matching confidence according to the mapping result thereof with a reference charging pattern; if the matching confidence is higher than a preset threshold, extracting corresponding historical charging control parameters from the matched reference charging pattern as basic parameters; based on specific numerical features of key charging and discharging event points in the current charging subspace, fine-tuning the basic parameters to generate real-time charging control parameters applicable to the current time; integrating all real-time charging control parameters generated by the charging subspaces in time sequence to form the real-time charging control strategy; the fine-tuning of the basic parameters based on specific numerical features of key charging and discharging event points in the current charging subspace to generate real-time charging control parameters applicable to the current time comprises: extracting voltage measurement values, current measurement values and temperature measurement values of each key charging and discharging event point in the current charging space; calculating feature deviations between the measurement values of each key charging and discharging event point and corresponding reference values in the basic parameters; determining adjustment directions and adjustment amplitudes of the charging current set value, the cut-off voltage threshold and the temperature control threshold in the basic parameters according to the directions and sizes of the feature deviations; converting the adjustment directions and adjustment amplitudes into specific correction amounts of the basic parameters by using a proportional-integral-derivative control algorithm; and superimposing the correction amounts and the basic parameters to obtain real-time charging control parameters suitable for the current charging space time range; The correction amount is calculated by the following formula: Wherein: character represents the specific correction amount of the base parameter, including the adjustment amount of current, voltage and temperature; character represents the proportional gain coefficient, which controls the response strength of the current deviation; character represents the characteristic deviation value calculated at present; character represents the integral gain coefficient, which controls the correction strength of the historical deviation accumulation; character represents the integral of the characteristic deviation in time, which is calculated by multiplying the deviation value by the time interval and then summed; character represents the differential gain coefficient, which controls the suppression degree of the deviation change rate; character represents the change rate of the characteristic deviation, which is calculated by dividing the difference between the current deviation and the deviation at the last time by the time interval.

2. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 1, characterized in that, The complete charging trajectory of the battery cell during the charging process is constructed, including: collecting voltage sequences, current sequences and temperature sequences of the battery cell during the charging process at a fixed frequency; allocating a unique timestamp identifier to each voltage value, current value and temperature value at each collection time; combining the voltage value, current value and temperature value under the same timestamp into a charging state snapshot; linking all charging state snapshots in the order of timestamps to form the complete charging trajectory.

3. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 2, characterized in that, The complete charging trajectory is segmented at multiple scales to generate multiple charging spaces with different time spans, including: defining multiple time window scales, each of which corresponds to a specific time length; using each time window scale as a sliding window to slide on the complete charging trajectory, and each time sliding to intercept a trajectory segment; defining each trajectory segment intercepted by each time sliding as a charging space; allocating a space identifier to each charging space, which records the corresponding time window scale information and the start and end timestamps in the complete charging trajectory.

4. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 3, characterized in that, The labeled multiple charging spaces are spatiotemporally aligned and mapped with reference charging patterns in the battery cell historical database, including: retrieving all historical charging records with the same identity as the battery cell from the battery cell historical database; extracting reference charging patterns stored in each historical charging record, which includes the type, occurrence time and context information of the historical key event points; matching the multiple charging spaces generated during the current charging process and labeled with key charging and discharging event points with the retrieved reference charging patterns respectively; The matching process comprises adjusting the time axis of the charging episode to align with the time axis of the reference charging pattern and calculating a measure of consistency in type and relative timing of key event points in the charging episode with historical key event points in the reference charging pattern, the measure of consistency having a value is calculated by a weighting function wherein: character represents the final consistency measure value; character represents the weight coefficient assigned to the event type similarity; character represents the event type similarity score, calculated based on the matching degree of the key event point types within the charge session to the corresponding historical key event point types in the reference charge pattern; character represents the weight coefficient assigned to the event sequence similarity; character represents the event sequence similarity score, calculated based on the matching degree of the relative time intervals between the key event points on the aligned time axis.

5. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 4, characterized in that, The verification and update steps of the real-time charging control strategy are also included: After applying the real-time charging control strategy, continue to collect subsequent charging state snapshots of the battery cell to form a verification trajectory segment; comparing the verification trajectory segment with the expected charging behavior model to calculate the strategy execution deviation degree; if the strategy execution deviation degree exceeds the allowable range, trigger the control strategy update mechanism, which includes: based on the new key event points identified in the verification trajectory segment, re-perform the mapping step with the reference charging pattern and generate an updated real-time charging control strategy.

6. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 5, characterized in that, The embedded evaluation step of the battery cell health state is also included: In the process of constructing the complete charging trajectory, a feature vector reflecting the stability of the rare earth composite material structure inside the battery is synchronously extracted; The feature vector is input into a pre-trained battery health assessment model, which outputs a current health index; The current health index is used as a hidden variable to integrate into the generation logic of the real-time charging control strategy, so that the real-time charging control strategy takes into account the real-time health status of the battery monomer when it is formulated.

7. The rare earth lithium battery high-speed fast-charging intelligent management system based on a rare earth composite material according to claim 6, characterized in that, It also includes a cross-battery monomer collaborative management step: When the management target is a battery pack containing multiple battery monomers, a complete charging trajectory is independently constructed for each battery monomer in the battery pack, and a respective real-time charging control strategy is generated; A coordinator at the battery pack level is established, which receives the real-time charging control strategies of all battery monomers; Based on the current health indices and real-time charging states of all battery monomers, the coordinator optimizes the real-time charging control strategies of individual battery monomers, generates a set of collaborative charging control instructions to ensure the balance and safety of the overall charging process of the battery pack.

Images