Safety Management Methods and Systems for Large-Capacity Batteries in RVs

Abstract

This application relates to the field of battery safety management technology, and particularly to a method and system for the safety management of large-capacity batteries in RVs. The proposed solution involves first acquiring operational data from multiple battery cells to construct a measurement matrix. Principal component analysis is used to identify collaboratively changing characteristic values, determining candidate abnormal cells that deviate from these characteristics. Based on trend fitting and disturbance response analysis, the disturbance coupling degree is calculated to determine the final abnormal cell. After identifying the abnormal cell, a hierarchical response management strategy is implemented, including a first scheduling scheme and a second scheduling scheme. The second scheduling scheme calculates the coupling strength between the power-consuming action module and the abnormal cell, and reconstructs the module structure based on electrical feature fingerprints and dependencies. This allows the behavioral unit to be migrated to a more suitable battery pack power supply path, achieving behavior retention and power supply optimization. This method improves the accuracy and scheduling flexibility of the battery system in scenarios involving abnormal individual cells.

Description

Technical Field

[0001] This application relates to the field of battery safety management technology, and in particular to a method and system for the safety management of large-capacity batteries for RVs. Background Technology

[0002] Existing RV battery management systems primarily rely on operating parameters such as voltage, temperature, and current to manage the battery pack holistically. When a battery cell experiences performance abnormalities, the system often struggles to identify the specific location of the anomaly in a timely manner, failing to accurately isolate the malfunctioning cell and thus expanding the scope of fault handling. Furthermore, existing technologies generally employ methods such as whole-pack isolation and average scheduling to respond to malfunctioning cells, failing to finely schedule or preserve critical loads. This results in widespread power outages, low system recovery efficiency, and severely impacts the continuity and stability of power supply in various RV scenarios. To address these issues, this application designs a safety management method and system for large-capacity RV batteries. Summary of the Invention

[0003] The technical problem this application aims to solve is to address the shortcomings of existing technologies by providing a safety management method and system for large-capacity batteries in RVs. First, operational data from multiple battery cells are acquired to construct a measurement matrix. Principal component analysis is used to identify cooperative variation characteristic values, determine candidate abnormal cells deviating from these characteristics, and calculate the disturbance coupling degree based on trend fitting and disturbance response analysis to determine the final abnormal cell. After identifying the abnormal cell, a hierarchical response management strategy is implemented, including a first scheduling scheme and a second scheduling scheme. The second scheduling scheme calculates the coupling strength between the electrical action module and the abnormal cell, and reconstructs the module structure based on electrical feature fingerprints and dependencies, thereby migrating the behavioral unit to a more suitable battery pack power supply path, achieving behavior retention and power supply optimization.

[0004] To achieve the above objectives, this application provides the following technical solution:

[0005] A safety management method for large-capacity batteries in RVs is applied to a BMS management system. The BMS management system includes multiple battery packs, each containing multiple individual battery cells. Multiple user power consumption behaviors constitute a power consumption action module. The battery safety management method includes:

[0006] The operating data of the battery cell is acquired, and a measurement matrix is ​​constructed, wherein the operating data includes voltage signal, temperature signal and transient current wave;

[0007] Based on the measurement matrix, battery cells that deviate from the specified characteristics within a set monitoring period are identified by the cooperative change feature values, and an abnormal cell candidate set is generated, wherein the cooperative change feature values ​​are determined based on the statistical characteristics of the measurement matrix.

[0008] The candidate set of anomalous individuals is modeled for time-series anomalous behavior based on evolutionary trend fitting, and the anomalous individuals are determined according to the perturbation coupling degree.

[0009] If an abnormal unit exists, a hierarchical response management strategy is determined and executed through the BMS system. The hierarchical response management strategy includes a first scheduling scheme and a second scheduling scheme. The second scheduling scheme includes determining the coupling strength between the multiple power-consuming action modules corresponding to the abnormal unit and the abnormal unit, and reconstructing the structure of the power-consuming action modules according to the coupling strength.

[0010] The construction of the measurement matrix includes:

[0011] The voltage signal, temperature signal, and transient current wave are time-aligned, normalized, and denoised to generate an electrical parameter feature sequence.

[0012] The electrical parameter feature sequence is sorted by battery cell as rows and sampling time sequence as columns to generate a measurement matrix, where each row represents the electrical parameter features of a battery cell within the monitoring period.

[0013] The method of identifying battery cells that deviate from their characteristics within a set monitoring period through coordinated change feature values ​​includes:

[0014] Calculate the covariance matrix of the measurement matrix, and perform principal component analysis on the covariance matrix to obtain the principal component vector;

[0015] The electrical parameter characteristics of each battery cell are projected into the principal component space constructed by the principal component vector, and the cooperative change characteristic value is calculated based on the projection value.

[0016] The aforementioned coordinated change characteristic value is compared with the average coordinated change characteristic value of all battery cells within the same monitoring period;

[0017] If the deviation between the cooperative change characteristic value of a battery cell and the average cooperative change characteristic value is greater than or equal to a preset deviation judgment threshold, the corresponding battery cell is judged as an abnormal cell candidate.

[0018] The candidate set of anomalous individuals is used to model the time-series anomalous behavior based on evolutionary trend fitting, and the anomalous individuals are determined according to the perturbation coupling degree, including:

[0019] For each candidate anomalous individual, a time window sliding model is performed on the electrical parameter features to calculate the vector of the change trend of the electrical parameter features over time.

[0020] Calculate the temporal characteristics of the change trend vector, wherein the temporal characteristics include fluctuation frequency, abrupt change amplitude, and offset rate;

[0021] Based on the aforementioned time-series characteristics, the degree of change in the linkage response of the candidate abnormal cell to other cells in the same battery pack within the monitoring period is calculated to obtain the disturbance response matrix.

[0022] Calculate the normalized response offset value of the disturbance response matrix, and obtain the disturbance coupling degree based on the average amplitude of the normalized response offset value;

[0023] If the disturbance coupling degree is greater than or equal to the preset response threshold, the abnormal single-unit candidate is determined to be an abnormal single-unit.

[0024] The determination of the tiered response management strategy includes:

[0025] Obtain the power consumption behavior of the user corresponding to the abnormal unit, and determine the corresponding battery pack based on one or more power consumption action modules corresponding to the user's power consumption behavior;

[0026] Based on the operating status of the individual battery cells in the corresponding battery pack, determine whether a first scheduling scheme exists. If it exists, calculate the first scheduling scheme. The determination of whether a first scheduling scheme exists includes evaluating whether the total discharge capacity meets the operating power requirements of the power-consuming action module. If it does, a first scheduling scheme exists; if it does not, a first scheduling scheme does not exist.

[0027] If not, calculate the second scheduling scheme, wherein the second scheduling scheme includes determining the coupling strength between the corresponding multiple power-consuming action modules and the abnormal unit, reconstructing the structure of the power-consuming action modules according to the coupling strength, and mapping the reconstructed power-consuming action modules to the power supply paths corresponding to other battery packs. The structural reconstruction includes splitting the power-consuming action modules into independently executable sub-behavioral units.

[0028] The calculation of the second scheduling scheme includes:

[0029] Calculate the coupling strength between multiple electrical action modules and the abnormal unit;

[0030] Electrical action modules with coupling strength greater than a preset coupling threshold are designated as modules to be split.

[0031] Electrical feature fingerprints are extracted from the user's electricity consumption behavior of the module to be split. The electrical feature fingerprints include transient current waves, power change rate, and continuous discharge characteristics when the electricity consumption behavior is started.

[0032] Based on the electrical feature fingerprint, the dependencies between user electricity consumption behaviors are determined, and the module to be split is divided into multiple independently executable sub-behavioral units;

[0033] Obtain dynamic load profiles of other battery packs, including the current load capacity, voltage stability, and thermal stability parameters of the battery packs;

[0034] Calculate the degree of compatibility between the electrical feature fingerprint of the sub-behavioral unit and the dynamic load profile of other battery packs, and calculate a second scheduling scheme based on the degree of compatibility.

[0035] The calculation conditions for the coupling strength include one of the following:

[0036] The percentage of current supplied through the abnormal unit power supply path during operation corresponding to the user's power consumption behavior in the power consumption action module;

[0037] The rate of temperature rise of abnormal battery cells during the operation of the electrical action module;

[0038] The power consumption action module includes the power consumption behavior of users powered by abnormal units;

[0039] Correlation between the response of the electrical action module and the abnormal unit during historical operation.

[0040] Determining the dependencies between user electricity consumption behaviors based on the electrical feature fingerprint includes:

[0041] Obtain the electrical feature fingerprint vector of the electricity consumption behavior of all users in the same electricity consumption action module;

[0042] The coupling dependence between user electricity consumption behaviors is calculated by measuring the similarity of transient current wave spectrum, power change rate, and discharge charge ratio.

[0043] Construct a dependency matrix between user electricity consumption behaviors based on the coupling dependency.

[0044] The method for calculating the degree of fit between the electrical feature fingerprint of the sub-behavioral unit and the dynamic load profile of the battery pack includes:

[0045] The load demand vector is generated based on the transient current wave spectrum characteristics, power change rate, and discharge charge of the sub-behavioral unit.

[0046] The Euclidean distance is calculated after standardizing the load demand vector and the dynamic load profile vector of the battery pack.

[0047] The adaptation score is determined based on the calculated Euclidean distance, and the battery pack with the highest adaptation score is selected as the target power supply path for the sub-behavioral unit.

[0048] The RV high-capacity battery safety management system includes:

[0049] Multiple battery packs, each battery pack comprising multiple individual battery cells;

[0050] The BMS management system is used to acquire the operating data of the battery cells and determine whether there are any abnormal battery cells.

[0051] The behavior recognition module is used to identify the corresponding user power consumption behavior based on the power supply path of the abnormal battery cell, and to determine the power consumption action module to which the user power consumption behavior belongs and its corresponding battery pack.

[0052] The scheduling judgment module is used to determine whether a first scheduling scheme exists based on the operating status of other battery cells in the battery pack.

[0053] The first scheduling execution module is used to control the power consumption action module to be executed by the discharge of other battery cells in the battery pack when a first scheduling scheme exists;

[0054] The second scheduling module is used to perform the following operations when the first scheduling scheme is not feasible:

[0055] Calculate the coupling strength between the electrical action module and the abnormal battery cell;

[0056] Structural reconstruction is performed on electrical action modules with coupling strength greater than a threshold, electrical feature fingerprints of user electricity consumption behavior are extracted, and behavioral dependencies are constructed.

[0057] The dynamic load profiles of other battery packs are obtained, and based on the degree of adaptation between the feature fingerprints and the load profiles, the reconstructed sub-behavioral units are mapped to the power supply paths of other battery packs for execution.

[0058] Compared with the prior art, the beneficial effects of this application are:

[0059] 1. This application constructs a measurement matrix and introduces cooperative variation feature values ​​for anomaly identification, which can accurately determine the state of a single battery cell from multiple electrical parameter dimensions, thereby improving the accuracy and stability of anomaly detection.

[0060] 2. This application adopts a hierarchical response management strategy, which differentiates the handling of power consumption modules under abnormal conditions, thus possessing higher scheduling flexibility and adaptability. Through structural reconstruction and behavior decomposition, independently executable power consumption behavior units are mapped to other battery pack power supply paths, enabling continuous operation of critical behaviors under abnormal conditions, enhancing behavior retention capability and system power supply stability. Attached Figure Description

[0061] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0062] Figure 1 This is a schematic diagram illustrating an exemplary application scenario of an embodiment of this application;

[0063] Figure 2 This is a schematic diagram of a safety management method for large-capacity batteries in RVs, as described in an embodiment of this application.

[0064] Figure 3 This is a schematic diagram illustrating the principle of the first scheduling scheme in the embodiments of this application;

[0065] Figure 4 This is a schematic diagram illustrating the process of determining a schedulable battery pack in an embodiment of this application;

[0066] Figure 5 This is a schematic diagram illustrating the module screening principle in an embodiment of this application;

[0067] Figure 6 This is a schematic diagram illustrating the module splitting principle in an embodiment of this application;

[0068] Figure 7 This is a schematic diagram illustrating the subgraph partitioning principle of an embodiment of this application. Detailed Implementation

[0069] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0070] The term "embodiment" as used herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0071] This application applies to mobile energy terminals with compact structures, complex loads, and locally constrained power supply paths, and the application scenarios include, but are not limited to:

[0072] Medium to large-sized motorhomes equipped with multiple distributed battery packs;

[0073] Equipped with a variety of functional electrical devices that start and stop frequently and exhibit significant power variations;

[0074] The power supply device has control habits that are continuous in power consumption scenarios and combined in behavioral chains, that is, users tend to trigger a specific set of power consumption actions continuously in certain typical scenarios.

[0075] The selection of application scenarios is based on the common characteristics of the region. The characteristics that the selected application scenarios should possess include:

[0076] Each battery pack is connected to several electrical loads through a power distribution unit. Although the power supply path can be switched, it has the characteristics of default path priority or engineering wiring dependence in the scenario.

[0077] The electrical equipment on board has the characteristics of high frequency start-stop and transient power change, and some power consumption behaviors are prone to load scheduling conflicts when there is a cell-level abnormality.

[0078] In typical usage scenarios, users often trigger multiple consecutive actions in a chain of behaviors. These consecutive actions have temporal dependencies and behavior aggregation characteristics, which can be abstracted into structured action modules.

[0079] It should be noted that this application does not take the division of power supply into battery pack areas as a premise, but is aimed at application scenarios with default power supply path dependencies under the influence of engineering wiring constraints or path switching complexity. It indirectly determines the correspondence between abnormal cells and current active loads by using the power path information and behavior model shared by the upper-level control system to be compatible with the traditional BMS architecture.

[0080] It is understood that this application does not require each power-consuming module to be permanently bound to a certain battery pack. Instead, it identifies the main power supply battery pack path of the module at the current moment based on the system wiring structure and the current path status collected by the BMS, and uses this as a reference for scheduling and control.

[0081] Please see Figure 1 This figure is a schematic diagram of an exemplary application scenario provided by an embodiment of this application.

[0082] like Figure 1 As shown, the target system of this application is applicable to RV power supply systems equipped with multi-battery pack structures. The RV power supply system includes multiple battery packs (as shown in the figure, battery pack A and battery pack B; the specific number is not limited in this application). Figure 1 (For reference only), the battery packs are connected in parallel via the main power supply bus to provide distributed energy supply to various electrical loads inside the RV. It should be understood that... Figure 1 The simplified illustrations provided are for illustrative purposes only. Specific RV power supply systems may include other battery packs or individual battery cells, other devices, or other unit modules. The BMS management system and application scenarios described in this application are intended to more clearly illustrate the technical solutions of this embodiment and do not constitute a limitation on the technical solutions provided in this application.

[0083] Furthermore, those skilled in the art will understand that, with the evolution of RV architecture and new power supply scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0084] Figure 1The diagram shows that each battery pack consists of multiple battery cells (as shown in the figure: battery cell A, battery cell B, battery cell C, battery cell D, battery cell E, and battery cell F; the specific number and whether the number of battery cells in each battery pack is consistent are not limited in this application). Figure 1 (For reference only).

[0085] Figure 1 The battery pack shows that the individual battery cells inside can be connected in series or partially in parallel according to different design requirements. It should be noted that in order to ensure the manageability, safety and voltage consistency of the system, this application explicitly excludes the connection structure in which all the battery cells in the battery pack are connected in parallel. That is, each battery pack must contain at least a partially series-connected group of cells. In addition, in the embodiments of this application, individual battery cells can also be replaced by cells.

[0086] In one example, the RV power supply system includes multiple power input paths, including solar panels, AC charging interfaces, and onboard engine generators. To ensure dynamic power balance and meet electrical protection requirements, multiple battery packs are configured with independent power management submodules and access control logic in the system structure. Therefore, although the battery packs are connected in parallel to the bus, only some battery packs participate in the discharge task during certain periods or operating conditions, while other battery packs are in standby, dormant, or charging states. In other words, power supply scheduling faces cross-packet limitation issues.

[0087] In another example, the RV power supply system is equipped with user electricity behavior recording and recognition functions to collect user behavior chain patterns in specific scenarios. These behavior chain patterns are abstracted into electricity consumption action modules and stored and scheduled in a structured manner within the system. This application can achieve behavior reconstruction and power supply path reallocation based on the operational relationships of these behavior modules and the battery path status.

[0088] In a further example, the RV control system is composed of a main control unit and an RV power supply system, wherein the RV power supply system in this application may be a battery management system (BMS). The BMS is only responsible for collecting the cell operating status and executing disconnection and protection commands, while strategies such as behavior scheduling, path reconstruction, and behavior module structure reconstruction are executed by the main control unit, which sends control commands to the BMS through standard CAN communication or UART bus.

[0089] Next, with reference to the accompanying drawings, the safety management method for large-capacity batteries in RVs provided in the embodiments of this application will be described. Figure 2 The method shown can be applied to a BMS management system, which includes multiple battery packs, each containing multiple individual battery cells. Multiple user power consumption behaviors constitute a power consumption action module. Figure 2 The method shown includes the following steps S1-S4, and the specific steps are as follows:

[0090] S1: Obtain the operation data of the battery cell and construct a measurement matrix;

[0091] In this embodiment, the operation data includes but is not limited to the voltage signal, temperature signal, and transient current wave of the battery cell.

[0092] Those skilled in the art can understand that the operation data can be collected according to the actual situation, and the specific types can be deleted by themselves, as long as it can minimally meet the requirement of obtaining the operation status based on the operation data.

[0093] S2: According to the measurement matrix, identify the battery cells with deviation characteristics during the set monitoring period through co-variation eigenvalues, and generate a candidate set of abnormal cells;

[0094] In this embodiment, based on the constructed measurement matrix, the principal component analysis (PCA) algorithm is used to extract the principal components and construct a feature space. The operation vector of each battery cell is projected in this space to obtain its co-variation eigenvalue. Set an empirical deviation threshold, and screen out the suspected deviated cells according to the deviation magnitude between the eigenvalue of each cell and the mean value, and generate a candidate set of abnormal cells.

[0095] S3: Perform a time-series abnormal behavior modeling based on the evolution trend fitting on the candidate set of abnormal cells, and determine the abnormal cells according to the perturbation coupling degree;

[0096] In this embodiment, further construct the time evolution trend vectors of the voltage, temperature, and current of each cell in the candidate set. The sliding time window regression modeling technology is used to fit the trend, extract the characteristic parameters such as the fluctuation frequency, mutation amplitude, and offset rate, and record its behavior consistency in multiple monitoring periods.

[0097] Furthermore, taking the abnormal candidate cell as the source, compare the linkage response caused by its fluctuation behavior on other cells in the same battery pack. Calculate the normalized offset amplitude of other cells, and use the overall offset mean value as the perturbation coupling degree index to determine whether the candidate cell has a system-level impact.

[0098] S4: If there are abnormal cells, determine a hierarchical response management strategy, and execute the hierarchical response management strategy through the BMS system;

[0099] In this embodiment, if it is determined that there are abnormal cells, enter the response control process. First, extract the user power consumption behavior supported by this cell, and query the power consumption action module to which it belongs and its corresponding default battery pack. Based on the operation status of other cells in the current battery pack, evaluate whether it can meet the operation power demand of the current module. If feasible, execute the first scheduling scheme, that is, shunt discharge within the pack, and keep the behavior structure unchanged.

[0100] If the assessment is not feasible, a second scheduling scheme is triggered. This scheme includes: determining the coupling strength between the current abnormal unit and the module, selecting highly coupled modules for structural reconstruction; splitting the module into several independently executable sub-behavioral units through electrical feature fingerprint extraction and dependency calculation; calculating the degree of adaptation by combining the dynamic load profiles of other battery packs, and mapping the sub-units to the battery pack path with the highest degree of adaptation to complete the power supply migration.

[0101] In real-world RV operation scenarios, due to the relatively closed nature of the energy system and the prevalence of off-grid power supply configurations, coupled with the unpredictable and structurally dependent nature of user behavior patterns, RVs frequently experience concentrated high-load activation within a given timeframe. For example, upon entering the vehicle, users may activate ceiling lights, electric cabinets, refrigerators, air conditioners, and other devices in quick succession. Compared to the centralized control strategies of the power grid system, which addresses stable load fluctuations and relies on a unified dispatch center, the RV scenario's dispatch structure is closer to an event-driven response.

[0102] The safety management method proposed in this application starts with the identification of abnormal battery cells, reverse maps the user's electricity consumption behavior supported by them, and thereby locates the associated electricity consumption action modules. Since the same user behavior may be reused by multiple modules, and the power supply paths of each module may be distributed across different battery packs in different scenarios, after the first scheduling fails, the entire module is no longer used as a static scheduling unit. Instead, independently executable behavioral units are extracted from its internal behavioral dependency structure, and the execution structure of the action module is reconstructed.

[0103] It should be noted that this application includes two layers of scheduling and allocation logic. The first allocation transfers power supply paths at the granularity of user power consumption behavior, completing local scheduling without disrupting the module structure. The second allocation, on the other hand, aims at module executability, decoupling the structure before establishing new path mapping relationships, thereby balancing operational continuity and battery safety.

[0104] In one example, constructing the measurement matrix includes:

[0105] The voltage signal, temperature signal, and transient current wave are time-aligned, normalized, and denoised to generate an electrical parameter feature sequence.

[0106] The electrical parameter feature sequence is sorted by battery cell as rows and sampling time sequence as columns to generate a measurement matrix, where each row represents the electrical parameter features of a battery cell within the monitoring period;

[0107] In one example, the specific steps of S2 are as follows:

[0108] S2.1: Calculate the covariance matrix of the measurement matrix, perform principal component analysis on the covariance matrix, and obtain the principal component vector;

[0109] Specifically, the calculation of the covariance matrix is ​​performed to capture the joint variation relationship between multiple battery cells under multiple parameter dimensions, essentially to quantify the consistency of cell behavior. When multiple cells are operating simultaneously, even if their individual voltage, temperature, or current fluctuation amplitudes are different, if their fluctuation trends show synchronous or approximately linked characteristics, they will exhibit a strong positive correlation structure in the covariance matrix.

[0110] In this embodiment, the selected measurement matrix contains three main data dimensions: voltage signal, temperature signal, and transient current wave vector after feature compression. Principal component analysis is performed using the covariance matrix to extract the first three principal component vectors, which serve as the feature space characterizing the main coherent trend of the cell's operating state within the current measurement cycle.

[0111] S2.2: Project the electrical parameter characteristics of each battery cell into the principal component space constructed by the principal component vector, and calculate the cooperative change characteristic value based on the projection value;

[0112] Specifically, projecting the battery cells into the principal component space addresses the issues of high redundancy and directional inconsistency among the original electrical parameters of the cells. The original data is physically non-orthogonal; for example, when the temperature of a battery cell slowly increases, its voltage fluctuations may produce a reverse response due to load changes, which is difficult to reflect in traditional Euclidean space. By projecting it into the feature space composed of principal components, the joint performance of each battery cell in the high-dimensional space can be reduced to a one-dimensional or two-dimensional numerical index, ensuring that this value reflects the degree of consistency with the principal cooperative direction.

[0113] In this embodiment, the multidimensional feature vector of each cell is sequentially mapped to the orthogonal space composed of principal components, and its projection value on each principal axis is recorded. The projection on the first principal component axis is taken as the benchmark expression of the cooperative change feature value, which can reflect whether the cell is consistent with the main cooperative trend in the current measurement period, and can also assess its contribution to the group behavior.

[0114] S2.3: Compare the aforementioned coordinated change characteristic value with the average coordinated change characteristic value of all battery cells within the same monitoring period;

[0115] In this embodiment, the principal component projection values ​​of all cells are first statistically analyzed to calculate first-order statistical features, including the mean, standard deviation, and maximum offset. Then, the mean is subtracted from the cooperative variation feature value of each cell to obtain its relative offset value.

[0116] S2.4: If the deviation between the cooperative change characteristic value of a battery cell and the average cooperative change characteristic value is greater than or equal to a preset deviation judgment threshold, the corresponding battery cell is judged as an abnormal cell candidate.

[0117] In this embodiment, the deviation judgment threshold can be modeled based on the standard distribution of feature value deviation under historical normal conditions, and set to a range of 2 standard deviations above and below the mean; alternatively, it can be combined with the amplitude of multi-period trend changes to construct an adaptive boundary and dynamically adjust the judgment sensitivity.

[0118] In one example, the specific steps of S3 are as follows:

[0119] S3.1: Perform time window sliding modeling on the electrical parameter features of each candidate anomalous entity, and calculate the vector of the change trend of electrical parameter features over time;

[0120] S3.2: Calculate the temporal characteristics of the change trend vector, wherein the temporal characteristics include fluctuation frequency, abrupt change amplitude, and offset rate;

[0121] S3.3: Calculate the degree of change in the linkage response of the candidate abnormal cell to other cells in the same battery pack within the monitoring period based on the time-series characteristics, and obtain the disturbance response matrix;

[0122] S3.4: Calculate the normalized response offset value of the disturbance response matrix, and obtain the disturbance coupling degree based on the average amplitude of the normalized response offset value;

[0123] S3.5: If the disturbance coupling degree is greater than or equal to the preset response threshold, the abnormal single-unit candidate is determined to be an abnormal single-unit;

[0124] In an optional embodiment, the criteria for determining whether an abnormal entity exists based on operational data may further include the following two aspects:

[0125] Firstly, the judgment is based on the absolute threshold of the rate of voltage drop of a single battery cell.

[0126] In one scenario, if the voltage drop rate of a single battery cell is greater than or equal to a first threshold change rate, it can be determined that an abnormal cell exists. The first threshold change rate can be a voltage drop rate threshold value preset according to the battery type, ambient temperature, and load power.

[0127] For example, for lithium iron phosphate cells, when discharged at a rate of 10C at a room temperature of 25°C, the first threshold change rate can be set to 0.15V / s. Those skilled in the art will understand that the first threshold change rate can also be obtained by fitting experimental data or statistically analyzing historical operating data under other types of cells or different temperature control strategies.

[0128] Understandably, when a battery cell discharges too quickly, experiences a sharp increase in internal resistance, or shows signs of a micro-short circuit, its voltage drop rate will be higher than the steady-state characteristic during normal discharge. If this is not identified and addressed promptly, it can easily lead to a severe voltage imbalance between the battery cell and the entire battery pack, triggering the BMS's over-discharge protection or abnormal pack tripping operation, resulting in a power outage for the entire pack. Therefore, it can be determined that the corresponding battery cell is in an abnormal state.

[0129] In another scenario, if the voltage drop rate of a single battery cell is less than the first threshold change rate, it indicates that the battery cell does not meet the risk conditions for being identified as an abnormal cell. In this case, the anomaly detection process can skip that battery cell and redirect the scheduling process to other battery cells or maintain the current power supply path.

[0130] Understandably, the discharge process under the current load conditions is within the normal range, and there is no rapid voltage decay. It can also be considered that its electrochemical reaction, conductivity path and thermal stability are all within the controllable range. Therefore, this battery cell is not an abnormal cell in this assessment.

[0131] Secondly, the judgment is based on the absolute threshold of the temperature rise rate of the individual battery cells.

[0132] In one scenario, an abnormal cell can be identified when the temperature rise rate of a single battery cell is greater than or equal to the second threshold change rate. The second threshold change rate can be preset based on factors such as cell type, internal thermal management conditions, load intensity, and heat dissipation structure design.

[0133] For example, for lithium iron phosphate cells without forced cooling structures, the safe temperature rise rate at a typical discharge rate of 10°C should not exceed 2.5°C / s. Those skilled in the art will understand that in scenarios with active thermal management systems or using other types of cathode materials (such as ternary systems), the critical temperature rise rate should be calibrated through thermal simulation or historical statistics based on the specific heat capacity, thermal conductivity, and thermal diffusion path.

[0134] Understandably, when a short-term overcurrent, intensified reaction, or limited heat dissipation path occurs inside the battery cell, the temperature change will be much faster than the thermal stability curve during normal discharge. If the rate of temperature rise exceeds the second threshold change rate, it may indicate a precursor to thermal runaway, increased interface impedance, or uneven heating caused by local pressure difference. If the relevant behavior is not timely managed or isolated, local hot spots may expand and the thermal diffusion inside the battery may become unbalanced.

[0135] In another scenario, if the temperature rise rate of a single battery cell is less than the second threshold change rate, it indicates that the battery cell does not meet the risk conditions for being identified as an abnormal cell. In this case, the anomaly determination process can skip that battery cell and redirect the scheduling process to other battery cells or maintain the current power supply path.

[0136] Furthermore, the first and second aspects can be combined to jointly determine whether it is an anomalous monomer.

[0137] When determining whether a cell is abnormal by combining the first and second aspects, if the results based on the first and second aspects are inconsistent, the results based on the second aspect take precedence. This is because the rate of temperature rise, as a thermal characteristic indicator, typically has a more direct risk indication. When a cell is in an early fault state or has abnormal internal reactions, the temperature rise usually fluctuates significantly before voltage changes. This is especially true in the early stages of latent faults such as short-term overload, micro-short circuits, or electrode debonding, where temperature changes more promptly reflect abnormal energy conversion efficiency and localized heating phenomena.

[0138] For example, in certain high-rate discharge scenarios, if the cell voltage drop rate remains within the normal range, increased local impedance of the electrodes or uneven heat diffusion may cause heat to accumulate rapidly inside the cell in a short period of time, triggering a high-temperature alarm or exceeding the temperature rise rate limit. In this case, relying solely on the voltage drop rate may miss the early identification of the abnormal cell, causing the system to only respond when there is a more severe voltage drop or the battery pack trips, resulting in a delayed risk response.

[0139] In addition to the two aspects mentioned above, this application embodiment can also determine whether a battery cell is an abnormal cell through other methods or in combination with other methods.

[0140] For example, an abnormal cell can be determined based on the voltage difference between a single battery cell and other battery cells in the same battery pack.

[0141] Optionally, if the difference between the static voltage value of a certain battery cell and the average voltage value of other battery cells in the same pack is greater than or equal to a third threshold voltage difference, the battery cell can be determined to be an abnormal cell. The third threshold voltage difference can be set based on factors such as historical statistical analysis, differences in typical state of charge (SOC) distribution, and pack balancing efficiency.

[0142] For example, the temperature difference between a single battery cell and other battery cells in the same battery pack can be used to determine whether a cell is abnormal. The specific determination method is similar to that described above and will not be repeated here.

[0143] In one example, the specific steps of S4 are as follows:

[0144] S4.1: Obtain the power consumption behavior of the user corresponding to the abnormal unit, and determine the corresponding battery pack based on one or more power consumption action modules corresponding to the user's power consumption behavior;

[0145] In this embodiment, once an abnormal battery cell is identified, the user's current power consumption behavior it is supporting is traced. It's easy to understand that user power consumption behavior typically consists of multiple specific devices, and related user power consumption behaviors can be attributed to a specific power consumption action module. Because the actual wiring in RVs tends to have fixed power supply paths, the same battery cell is connected to only a limited range of behavior paths in most scenarios.

[0146] Those skilled in the art will understand that user electricity consumption behavior in this application can be understood as the device usage actions triggered by the user within a specific time and under a specific scenario. This typically manifests as an operation request for a specific electrical component or function, including but not limited to turning on cabin lighting, starting the water heater, and starting the air conditioner. It is readily understood that user electricity consumption behavior, in terms of control logic, manifests as start-up and power adjustment control commands for individual devices. In terms of physical electrical paths, it typically corresponds to a defined set of load nodes and power paths. Theoretically, it also includes device shutdown, but this application does not consider this.

[0147] Those skilled in the art will understand that the electrical action module in this application can be understood as a composite control unit consisting of two or more user electrical behaviors that are time-related, spatially related, or functionally dependent. For example, when a user enters the RV, a combination of actions will be triggered in sequence, such as opening the electric door, turning on the interior lights, opening the roof window, and starting the air conditioner.

[0148] S4.2: Based on the operating status of the individual battery cells in the corresponding battery pack, determine whether a first scheduling scheme exists. If it exists, calculate the first scheduling scheme. The determination of whether a first scheduling scheme exists includes evaluating whether the total discharge capacity meets the operating power requirements of the power-consuming action module. If it does, a first scheduling scheme exists; if it does not, a first scheduling scheme does not exist.

[0149] In this embodiment, the first scheduling scheme can be implemented by comprehensively evaluating the operating status of the remaining cells in the battery pack containing the abnormal cell. The evaluation includes the number of available cells, load balancing capability, current redundancy margin, and local temperature rise trend. If the evaluation results indicate that other cells still have the capacity to share the load, the power supply path can be adjusted within the pack, for example, by using dynamic load balancing or prioritizing the discharge of parallel branches to maintain the operational integrity of the action module.

[0150] S4.3: If not, calculate the second scheduling scheme, wherein the second scheduling scheme includes determining the coupling strength between the corresponding multiple power-consuming action modules and the abnormal unit, reconstructing the structure of the power-consuming action modules according to the coupling strength, and mapping the reconstructed power-consuming action modules to the power supply paths corresponding to other battery packs. The structural reconstruction includes splitting the power-consuming action modules into independently executable sub-behavioral units.

[0151] In one example, taking the presence of an abnormal individual battery as an example, the scheduling algorithm is used to obtain the first scheduling scheme for a certain battery pack, which can be referenced. Figure 3 To understand, Figure 3 The battery pack is shown to contain four individual battery cells: cell A, cell B, cell C, and the faulty cell.

[0152] Battery cell A and the abnormal cell are combined in series, battery cell B and battery cell C are combined in series, and the two combinations are further combined in parallel. It is easy to understand that when there is an abnormal cell in the series state, according to the judgment mechanism of this embodiment, battery cell A and the abnormal cell can no longer supply power normally.

[0153] Figure 3 This application demonstrates that, in response to this abnormal state, the present application does not directly interrupt the output function of the entire battery pack. Instead, based on the state assessment results of other parallel battery cells within the battery pack, it performs a local load readjustment within the pack, distributing the load originally borne by battery cell A and the abnormal cell to battery cell B and battery cell C. This partially avoids the power supply participation of the abnormal cell without disrupting the original power supply structure.

[0154] according to Figure 3 As shown in the content, the first scheduling scheme obtained in the embodiments of this application includes:

[0155] Within the corresponding battery pack, the user's power consumption behavior corresponding to the power consumption module is discharged by other battery cells, wherein the discharge includes parallel current sharing.

[0156] Combination Figure 3 It is understandable that, in the case of an abnormal cell in the circuit of battery cell A, this embodiment assesses whether the abnormal cells (such as battery cells B and C) have the ability to bear the remaining load by identifying their current operating status. When the conditions are met, the discharge task originally undertaken by the abnormal cell is internally transferred by controlling the internal equalization conduction circuit or adjusting the current path weight, thereby achieving a reconciliation between local isolation and functional preservation.

[0157] Furthermore, how to determine whether a first scheduling scheme exists includes:

[0158] Based on the operating status of other battery cells, assess whether the total discharge capacity meets the operating power requirements of the power-consuming module; if it does, a first scheduling scheme exists; if it does not, no first scheduling scheme exists.

[0159] Those skilled in the art will understand that determining whether a first scheduling scheme exists and how to determine if the first scheduling scheme belongs to the prior art is not elaborated here.

[0160] It should be noted that the existence of a first scheduling scheme can only be determined when the battery pack is connected in series with some parallel connections.

[0161] In other words, when the battery pack is connected only in series, there is no need to determine whether a first scheduling scheme exists. This is because in a series structure, all battery cells form the same current path, and the load current must flow continuously through each cell. If any cell malfunctions, it will directly limit the output performance of the entire series branch. In this structure, there is no physical current shunting capability between cells, nor is it possible to bypass a certain cell and have other cells independently supply power. Therefore, there is no basis for implementing intra-pack scheduling. This application will skip the determination of the first scheduling scheme and proceed directly to the second scheduling scheme.

[0162] Please see Figure 4 The figure is a schematic diagram of the process for determining a schedulable battery pack provided in an embodiment of this application. Figure 4 The method shown can be applied to S4.1 of the aforementioned method, and the specific steps are as follows:

[0163] S4.1.1: Obtain the working log of the abnormal unit;

[0164] Specifically, after locating the abnormal unit, the power supply task it undertakes may be in multiple concurrent power consumption behaviors. Therefore, it is necessary to combine the actual data of its operation process to track the power supply path it accesses and the load behavior it participates in at different time periods.

[0165] In this embodiment, the method of obtaining the working log includes extracting the voltage, current, temperature and on / off status data of the abnormal unit from the BMS built-in operation record cache, and combining it with the load distribution log recorded by the controller to establish the correspondence between the abnormal unit and the power supply path within a specific time segment.

[0166] Preferably, the work log can be sliced ​​according to time windows, and time periods in low-power standby mode or without load transfer can be filtered out to ensure the targeting and efficiency of behavior tracing.

[0167] S4.1.2: Based on the power supply path records of the abnormal unit once or multiple times, identify the corresponding user's electricity consumption behavior;

[0168] Specifically, in a motorhome structure with multiple devices and interleaved power supply, a single battery cell may supply power to multiple devices at different times. In an easily understandable scenario, these devices may be triggered continuously by the user in a chain-like operation habit, so it is impossible to determine the supported device object by the change in current.

[0169] It is easy to understand that the BMS management system only allocates power to individual battery cells and does not select specific battery cells to supply power based on specific user power consumption behavior. In other words, the BMS management system does not actively identify user intent or behavioral logic, nor does it bind specific battery cells to specific devices or control operations.

[0170] In this embodiment, the power supply path record is used as an intermediate quantity, combined with the behavior command logs of the controller or load side, to construct a time-domain cross-comparison between battery cell power supply events and user behavior events. Those skilled in the art will understand that by recording the time period during which a battery cell participates in power supply, and by checking the on / off state of the load switch logic or relay group connected to it, it can be determined whether a user-issued control signal command occurred during that time period, such as turning on lights, raising cabinet doors, or activating a multi-function panel. By matching the degree of overlap between these signal triggers and battery cell voltage fluctuations or discharge characteristics, it can be determined that the battery cell assumed power supply responsibility within the corresponding user power consumption cycle.

[0171] It should be noted that, since multiple behaviors inevitably interfere with each other within the same time window in real-world scenarios, this embodiment can improve the discrimination accuracy by adding control channels. For example, redundant confirmation logic can be added to CAN commands, relay status, and power module responses. Those skilled in the art will understand that the RV in this application has default path priority or engineering wiring dependency characteristics, thus enabling indirect identification of upper-level user behaviors involving abnormal units through path records even without explicit battery cell and device binding information.

[0172] S4.1.3: Match the identified user electricity consumption behavior with the electricity consumption action module;

[0173] Specifically, in practical application scenarios, user power consumption behavior often has a reuse relationship in multiple power consumption action modules. The user power consumption behavior determined in the aforementioned steps may appear in multiple power consumption action modules. Therefore, in this application, based on the operation records of these power consumption action modules, it can be deduced whether there are other battery packs that can support the same action module or some action units.

[0174] For example, the same device activation behavior, such as "turning on the lights," may appear as part of the "evening entertainment module" or the "nighttime ventilation module." It should be noted that this application does not restrict the module names, but only combines them based on the actual user behavior.

[0175] S4.1.4: In the historical power supply mapping records of the matched power-consuming action module and the battery pack, identify the path relationship of the current power-consuming action module being powered by the battery pack, and determine that the battery pack is the target power supply for the current power-consuming action module;

[0176] It is easy to understand that in most cases, the power supply management system does not restrict the power supply objects of the battery pack. For example, during the initial configuration of the system, the management logic usually only dynamically selects the single or multiple battery packs most suitable for undertaking the discharge task based on the current SOC state, voltage stability and power supply rationality of the battery pack. It does not require that a specific battery pack can only serve a certain module or a certain type of behavior. The method shown in this application only extracts the alternative battery packs that are different from the current abnormal single unit as the target battery packs to be given priority in subsequent scheduling.

[0177] It is important to note that during the actual operation of a battery management system, as the system runs continuously and records the power supply path corresponding to each action trigger, a historical mapping will objectively form at the log level, showing that a specific power-consuming module has been powered by one or more battery packs. These mappings essentially reflect that, under a given load structure and scheduling strategy, one or more battery packs have successfully supported the operation of a certain power-consuming module, thus possessing the potential to undertake the power supply task of that module again.

[0178] Those skilled in the art will understand that the scheduling process and the specific implementation process may be inconsistent or consistent. The power supply process can be statically or dynamically configured according to the logic configured in the specific main control unit. On this basis, this application achieves optimized dynamic configuration through additional scheduling logic. This is the unique feature of this application's cross-path power supply mapping driven by user electricity consumption behavior.

[0179] Taking the further screening of electrical action modules as an example, you can refer to... Figure 5 To understand, Figure 5 This is a schematic diagram illustrating the module screening principle in an embodiment of this application. Figure 5 This demonstrates that a user's electricity consumption behavior occurs in electricity consumption action modules one, two, and three. The coupling strength between each module and the abnormal unit is calculated sequentially, resulting in coupling strengths one, two, and three, respectively. These coupling strengths are then compared to preset coupling thresholds. Modules exceeding these thresholds are designated as modules to be split. Figure 5 In this process, the second electrical action module was selected as the module to be disassembled.

[0180] Understandable Figure 5 The simplified illustration is for illustrative purposes only. In actual implementation, more power consumption action modules may be included, and the modules to be split may include one or more. This application does not limit the number, but at least one module to be split must be included. The specific number can be limited by the coupling threshold. If there is no module to be split in the actual process, the corresponding user power consumption behavior can be split separately, and the remaining user power consumption behavior will not be further split.

[0181] In one example, the steps for screening using electrically operated modules are as follows:

[0182] Calculate the coupling strength between multiple electrical action modules and the abnormal unit;

[0183] Electrically operated modules with coupling strength greater than a preset coupling threshold are designated as modules to be split.

[0184] In one example, the conditions for calculating the coupling strength include one of the following:

[0185] The percentage of current supplied through the abnormal unit power supply path during operation corresponding to the user's power consumption behavior in the power consumption action module;

[0186] The rate of temperature rise of abnormal battery cells during the operation of the electrical action module;

[0187] The power consumption action module includes the power consumption behavior of users powered by abnormal units;

[0188] Correlation between the response of the electrical action module and the abnormal unit during historical operation.

[0189] Specifically, coupling strength reflects the degree of functional dependence between an electrical action module and a certain abnormal unit in terms of electrical behavior, thermal impact, and response linkage. In other words, when the coupling strength between an electrical action module and a certain abnormal unit is low, it can be understood that the module does not depend on the corresponding abnormal unit, meaning that the user's electrical behavior in the module does not affect the operation of the abnormal unit.

[0190] It is important to note that high coupling strength often means that multiple behavioral nodes within the module rely on that single cell as their primary or sole power source during operation. This dependency can lead to a situation where the entire module structure becomes unusable if a cell becomes unavailable. Furthermore, it's easy to understand that directly migrating the entire power-consuming module to another battery pack could cause other cells to malfunction due to the similarity of their individual cells—a situation undesirable to those skilled in the art. Therefore, it is necessary to split the module and reconstruct the behavioral sequences and path matching relationships.

[0191] Taking module splitting as an example, you can refer to... Figure 6 To understand, Figure 6 This is a schematic diagram illustrating the module splitting principle in an embodiment of this application. Figure 6 The diagram illustrates a module to be split, comprising six user electricity consumption behaviors: User Electricity Consumption Behavior 1, User Electricity Consumption Behavior 2, User Electricity Consumption Behavior 3, User Electricity Consumption Behavior 4, User Electricity Consumption Behavior 5, and User Electricity Consumption Behavior 6. After electrical feature fingerprint extraction and dependency analysis, the module is divided into three sub-units to be executed. User Electricity Consumption Behavior 1 has no strong dependency on the other user electricity consumption behaviors and is executed independently. User Electricity Consumption Behavior 2 and User Electricity Consumption Behavior 6 have a strong dependency and are executed together as a sub-unit. Similarly, User Electricity Consumption Behavior 3, User Electricity Consumption Behavior 4, and User Electricity Consumption Behavior 5 are executed together as a sub-unit.

[0192] In one example, the steps for module splitting are as follows:

[0193] Electrical feature fingerprints are extracted from the user's electricity consumption behavior of the module to be split. The electrical feature fingerprints include transient current waves, power change rate, and continuous discharge characteristics when the electricity consumption behavior is started.

[0194] Specifically, before the module structure is reconfigured, the electricity consumption behavior of each user within the module needs to be quantitatively characterized to determine whether the behaviors have decoupled execution conditions. Since, under the constraints of this application, the electricity consumption module is essentially composed of multiple user electricity consumption behaviors that are logically and electrically related, some of these behaviors inevitably have electrical dependencies, load following, or synchronous activation relationships. In the scheduling algorithm, these related user electricity consumption behaviors are preferentially packaged together and split to meet the scheduling algorithm requirements of behavior execution continuity and power supply path structure consistency.

[0195] It should be noted that the aforementioned scheduling algorithm requirements are only specific requirements of the method in this application and do not mean that behavior should be divided according to these scheduling algorithm requirements in all energy management systems.

[0196] In the scheduling scheme proposed in this application, the scheduling objectives include not only maintaining basic load power supply, but also preserving the integrity of the maximum proportion of behavioral functions under the influence of abnormal individual units, and minimizing the probability of behavioral failure at the user perception level. Therefore, in the module structure reconfiguration stage, this application specifically defines a splitting logic based on the electrical coupling relationship of behavior, that is, grouping multiple user power consumption behaviors with electrical linkage characteristics into a decoupling boundary to ensure that they can still be reused as a whole during the migration process, and avoiding situations where behaviors become unexecutable due to loss of connectivity, power supply path segmentation, or loss of synchronous execution capability after splitting.

[0197] In one example, the extracted electrical fingerprint mainly includes three features:

[0198] The first characteristic is the transient current wave, which is used to identify the type of starting load and determine whether it has impact characteristics. For example, starting a motor or compressor will cause a high-amplitude current spike in a short period of time.

[0199] The second feature is the power change rate during the operation phase, which is used to characterize the load ramp-up characteristics of the behavior from activation to stable operation.

[0200] The third characteristic is the continuous discharge characteristic, which includes the average current value, discharge time window length, and current fluctuation range during the stable phase of the behavior, and is used to identify the stress impact of the behavior on the continuous output capability of the cell.

[0201] Those skilled in the art will understand that electrical fingerprint features are similar to user profiles and can be extracted using existing technologies, which will not be elaborated upon here.

[0202] Based on the electrical feature fingerprint, the dependencies between user electricity consumption behaviors are determined, and the module to be split is divided into multiple independently executable sub-behavioral units;

[0203] Specifically, after obtaining the electrical characteristic fingerprint of the behavior, it is necessary to analyze the electrical coupling relationship between different users' electricity consumption behaviors in the actual operation process. Based on this, the module is decomposed into several sub-behavioral units that can be separated at the circuit level and execution logic.

[0204] In one example, determining the dependencies between user electricity consumption behaviors based on the electrical feature fingerprint includes:

[0205] Obtain the electrical feature fingerprint vector of the electricity consumption behavior of all users in the same electricity consumption action module;

[0206] The coupling dependence between user electricity consumption behaviors is calculated by measuring the similarity of transient current wave spectrum, power change rate, and discharge charge ratio.

[0207] Construct a dependency matrix between user electricity consumption behaviors based on the coupling dependency.

[0208] In one example, it is necessary to analyze the behavioral coupling relationships in the dependency matrix to determine which user power consumption behaviors can be migrated individually to other power supply paths when the current battery pack is unavailable, which user power consumption behaviors need to be retained as a group due to dependencies, or which are postponed due to non-migration.

[0209] In this embodiment, a weighted undirected graph is constructed based on the dependency matrix, with user electricity consumption behavior as nodes and the coupling dependency between behaviors as edge weights.

[0210] For example, regarding subgraph partitioning, please refer to... Figure 7 To understand, Figure 7 This is a schematic diagram illustrating the subgraph partitioning principle of an embodiment of this application. Figure 7 This shows a weighted undirected graph derived from a dependency matrix. Figure 7 The weighted undirected graph shown includes four nodes: node 1, node 2, node 3, and node 4.

[0211] Figure 7 The edge weight between node 1 and node 2 is 0.1, the edge weight between node 2 and node 3 is 0.2, the edge weight between node 1 and node 4 is 0.8, the edge weight between node 1 and node 3 is 0.7, and the edge weight between node 3 and node 4 is 0.9.

[0212] Figure 7 This further illustrates that during the graph partitioning process, the connections between node 2 and node 1, as well as node 3, are first cut off based on the edge weights. It is easy to understand that, relatively speaking, node 2 can exist as an independently executable node.

[0213] Figure 7 This further illustrates that during graph partitioning, nodes 1, 4, and 3 can belong to the same connected component according to the connectivity algorithm, and the dependency between nodes 1 and 3 is relatively small compared to the other two edges in the specific edge weight comparison, so they can also be deleted. It is easy to understand that in the subgraph formed by nodes 1, 3, and 4, the specific dependency connection can be in the order of node 3-node 4-node 1.

[0214] Those skilled in the art will understand that Figure 7 This is just a simple example. Specific graph partitioning algorithms can be selected through connected component algorithms, which will not be elaborated upon in this application.

[0215] Furthermore, the weighted undirected graph is grouped by behavior according to the graph partitioning algorithm. The maximum connected subgraph algorithm is used first to obtain multiple subgraphs with the maximum connectivity in the weighted undirected graph. For edges with smaller weights, they can be cut during the graph partitioning process, and the corresponding nodes are regarded as independent migration behavior units.

[0216] It should be noted that since there may be multiple edges between nodes, the edge with the largest edge weight is retained as the dividing criterion in the specific process of dividing the maximum connectivity.

[0217] In one example, the calculation steps for the second scheduling scheme are as follows:

[0218] Obtain dynamic load profiles of other battery packs, including the current load capacity, voltage stability, and thermal stability parameters of the battery packs;

[0219] Specifically, this application addresses the power supply path allocation problem among multiple healthy battery packs for sub-behavioral units after scheduling reconfiguration by using dynamic load profiling.

[0220] In this embodiment, the dynamic load profile is a multi-dimensional feature vector structure, mainly including three parameters:

[0221] The first parameter is the current load capacity. The sustainable discharge capacity of each battery pack is estimated by the remaining SOC and historical discharge efficiency.

[0222] The second parameter is the voltage stability parameter, which is derived from the statistical combination of the standard deviation of voltage fluctuation and the short-time fluctuation frequency over a number of recent operating cycles. It is used to measure the ability of the voltage response to suppress dynamic changes in the load.

[0223] The third parameter is the thermal stability parameter, which is calculated based on the temperature change gradient of the battery pack and the temperature rise uniformity index of local cells, reflecting the safety margin of the battery pack in continuous operation under the current thermal environment.

[0224] Calculate the degree of compatibility between the electrical feature fingerprint of the sub-behavioral unit and the dynamic load profile of other battery packs, and calculate the second scheduling scheme based on the degree of compatibility;

[0225] The method of this application has now become clear: the adaptation problem between behavior and battery pack is essentially a structural similarity matching process between two multi-dimensional vectors. Therefore, it is necessary to represent the electrical feature fingerprint as a standardized demand vector structure and calculate the matching degree with the dynamic profile of each battery pack to quantify the adaptability of the target battery pack.

[0226] The methods for calculating the degree of fit include:

[0227] In this embodiment, a load demand vector is first generated based on the electrical feature fingerprint extracted from each sub-behavioral unit. This vector includes the transient current spectrum distribution at startup, the power change rate per unit time, and the amount of discharge charge within a complete execution cycle.

[0228] Furthermore, the load demand vector and the dynamic load profile of each healthy battery pack at the current moment are standardized to ensure they are within the same numerical scale range. The Euclidean distance between them is then calculated; a smaller distance indicates that the battery pack is more suitable for carrying the current sub-behavioral unit, exhibiting stronger discharge capacity matching and electrical response margin.

[0229] The adaptation score is determined based on the calculated Euclidean distance, and the battery pack with the highest adaptation score is selected as the target power supply path for the sub-behavioral unit.

[0230] In actual scheduling, a priority list can be constructed based on the adaptation score of each candidate battery pack, and target battery pack paths can be assigned to each behavioral unit based on the priority supplementation method. If a behavioral unit cannot meet the minimum adaptation threshold requirement in all current candidate paths, it enters the scheduling failure feedback process, and the control unit triggers user confirmation, delay waiting, or secondary reconstruction of the module structure.

[0231] Those skilled in the art should understand that after obtaining the specific target power supply path and the corresponding sub-behavioral unit, the remaining calculation steps of the second scheduling scheme are all conventional path configuration and control operation procedures, which can be naturally implemented according to the existing load scheduling framework, and will not be elaborated here.

[0232] In one example, this application embodiment provides a safety management system for a large-capacity battery in a motorhome, the system comprising:

[0233] Multiple battery packs, each battery pack comprising multiple individual battery cells;

[0234] The BMS management system is used to acquire the operating data of the battery cells and determine whether there are any abnormal battery cells.

[0235] The behavior recognition module is used to identify the corresponding user power consumption behavior based on the power supply path of the abnormal battery cell, and to determine the power consumption action module to which the user power consumption behavior belongs and its corresponding battery pack.

[0236] The scheduling judgment module is used to determine whether a first scheduling scheme exists based on the operating status of other battery cells in the battery pack.

[0237] The first scheduling execution module is used to control the power consumption action module to be executed by the discharge of other battery cells in the battery pack when a first scheduling scheme exists;

[0238] The second scheduling module is used to perform the following operations when the first scheduling scheme is not feasible:

[0239] Calculate the coupling strength between the electrical action module and the abnormal battery cell;

[0240] Structural reconstruction is performed on electrical action modules with coupling strength greater than a threshold, electrical feature fingerprints of user electricity consumption behavior are extracted, and behavioral dependencies are constructed.

[0241] The dynamic load profiles of other battery packs are obtained, and based on the degree of adaptation between the feature fingerprints and the load profiles, the reconstructed sub-behavioral units are mapped to the power supply paths of other battery packs for execution.

[0242] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for managing the safety of a large-capacity battery of a recreational vehicle, applied to a BMS management system, the BMS management system comprising a plurality of battery packs, each battery pack comprising a plurality of battery cells, and a plurality of user power consumption behaviors constituting a power consumption action module, characterized in that, The battery safety management method comprises: Obtaining the operation data of the battery monomer, constructing a measurement matrix, wherein the operation data comprises a voltage signal, a temperature signal and a transient current wave; According to the measurement matrix, the battery monomer with deviating characteristics within a set monitoring period is identified by a cooperative change characteristic value, and an abnormal monomer candidate set is generated, wherein the cooperative change characteristic value is determined according to the statistical characteristics of the measurement matrix; The abnormal monomer candidate set is modeled based on the fitting of the evolution trend, and the abnormal monomer is determined according to the disturbance coupling degree; If there is an abnormal monomer, a hierarchical response management strategy is determined, and the hierarchical response management strategy is executed through the BMS system, wherein the hierarchical response management strategy comprises a first scheduling scheme and a second scheduling scheme, and the second scheduling scheme comprises judging the coupling strength between the abnormal monomer and a plurality of power action modules corresponding to the abnormal monomer, and restructuring the power action modules according to the coupling strength.

2. The method of claim 1, wherein the method further comprises: The construction of the measurement matrix comprises: The voltage signal, the temperature signal and the transient current wave are respectively time-aligned, normalized and denoised, and an electrical parameter characteristic sequence is generated; The electrical parameter characteristic sequence is sorted according to the battery monomer as the row and the sampling time sequence as the column, and a measurement matrix is generated, wherein each row represents the electrical parameter characteristics of a battery monomer within a monitoring period.

3. The method of claim 2, wherein the method further comprises: The battery monomer with deviating characteristics within a set monitoring period is identified by a cooperative change characteristic value, comprising: The covariance matrix of the measurement matrix is calculated, and the principal component analysis of the covariance matrix is performed to obtain a principal component vector; The electrical parameter characteristics of each battery monomer are projected in the principal component space constructed by the principal component vector, and the cooperative change characteristic value is calculated according to the projection value; The cooperative change characteristic value is compared with the average cooperative change characteristic value of all battery monomers within the same monitoring period; If the deviation value of the cooperative change characteristic value of a battery monomer from the average cooperative change characteristic value is greater than or equal to a preset deviation judgment threshold, the corresponding battery monomer is determined as an abnormal monomer candidate.

4. The method of claim 3, wherein the step of determining the state of charge of the battery is performed by a battery management system. The abnormal monomer candidate set is modeled based on the fitting of the evolution trend, and the abnormal monomer is determined according to the disturbance coupling degree, comprising: The electrical parameter characteristics of each abnormal monomer candidate are time-window sliding modeled, and the change trend vector of the electrical parameter characteristics with time evolution is calculated; The time sequence characteristics of the change trend vector are calculated, wherein the time sequence characteristics comprise fluctuation frequency, mutation amplitude and offset rate; The linkage response change degree of the abnormal monomer candidate to other battery monomers in the same battery pack within the monitoring period is calculated according to the time sequence characteristics, and a disturbance response matrix is obtained; The normalized response offset value of the disturbance response matrix is calculated, and the disturbance coupling degree is obtained according to the average amplitude of the normalized response offset value; If the disturbance coupling degree is greater than or equal to a preset response threshold, the abnormal monomer candidate is determined as an abnormal monomer.

5. The method of claim 1, wherein the method further comprises: The determination of the hierarchical response management strategy comprises: obtaining a user electricity consumption behavior corresponding to the supply of the abnormal monomer, determining a corresponding battery pack according to one or more electricity action modules corresponding to the user electricity consumption behavior; determining whether there is a first scheduling scheme according to the operating state of the battery monomer of the corresponding battery pack, and calculating the first scheduling scheme if there is, wherein the determination of whether there is the first scheduling scheme includes evaluating whether the total discharge capacity meets the operating power requirement of the electricity action module, and there is the first scheduling scheme if it meets, and there is no first scheduling scheme if it does not meet; calculating a second scheduling scheme if there is no, wherein the second scheduling scheme includes judging the coupling strength between the corresponding multiple electricity action modules and the abnormal monomer, reconstructing the electricity action module according to the coupling strength, and mapping the reconstructed electricity action module to the power supply path corresponding to other battery packs, and the structural reconstruction includes splitting the electricity action module into sub-behavior units that can be independently executed.

6. The method of claim 5, wherein the step of determining the state of charge of the battery is performed by a battery management system. The calculation of the second scheduling scheme includes: calculating the coupling strength between the multiple electricity action modules and the abnormal monomer; taking the electricity action module with the coupling strength greater than the preset coupling threshold as a splitting module; extracting the electrical characteristic fingerprint of the user electricity consumption behavior of the splitting module, the electrical characteristic fingerprint including the transient current wave, power change rate and continuous discharge characteristics when the electricity consumption behavior starts; determining the dependency relationship between the user electricity consumption behaviors according to the electrical characteristic fingerprint, and splitting the splitting module into multiple sub-behavior units that can be independently executed; obtaining the dynamic load image of other battery packs, the dynamic load image including the current load capacity, voltage stability and thermal stability parameters of the battery pack; calculating the adaptation degree of the electrical characteristic fingerprint of the sub-behavior unit and the dynamic load image of other battery packs, and calculating the second scheduling scheme according to the adaptation degree.

7. The method for safe management of large-capacity batteries in RVs according to claim 6, characterized in that, The calculation conditions of the coupling strength include one of the following: the current proportion of the user electricity consumption behavior corresponding to the electricity action module through the abnormal monomer power supply path during operation; the temperature rise rate of the abnormal battery monomer during the operation of the electricity action module; the user electricity consumption behavior in the electricity action module includes the electricity consumption behavior supplied by the abnormal monomer; the response correlation between the electricity action module and the abnormal monomer in the historical operation process.

8. The method of claim 6, wherein the method further comprises: Determining the dependency relationship between the user electricity consumption behaviors according to the electrical characteristic fingerprint includes: obtaining the electrical characteristic fingerprint vector of all user electricity consumption behaviors in the same electricity action module; calculating the coupling dependency between the user electricity consumption behaviors by calculating the transient current wave spectrum similarity, power change rate similarity and discharge charge ratio between the user electricity consumption behaviors; constructing the dependency relationship matrix between the user electricity consumption behaviors according to the coupling dependency.

9. The method of claim 6, wherein the method further comprises: The method for calculating the adaptation degree of the electrical characteristic fingerprint of the sub-behavior unit and the dynamic load image of the battery pack includes: generating a load demand vector according to the transient current wave spectrum characteristics, power change rate and discharge charge amount of the sub-behavior unit; calculating the Euclidean distance after standardizing the load demand vector and the dynamic load image vector of the battery pack; According to the calculated Euclidean distance, a matching score is determined, and a battery pack with the highest matching score is selected as a target power supply path of the sub-behavior unit.

10. A recreational vehicle high capacity battery safety management system for implementing the recreational vehicle high capacity battery safety management method of any one of claims 1-9, characterized by, The system comprises: a plurality of battery packs, each battery pack comprising a plurality of battery cells; a BMS management system configured to acquire operation data of the battery cells and determine whether there is an abnormal battery cell; a behavior recognition module configured to recognize a corresponding user electricity consumption behavior according to a power supply path of the abnormal battery cell, and determine an electricity consumption action module to which the user electricity consumption behavior belongs and a corresponding battery pack of the user electricity consumption behavior; a scheduling judgment module configured to determine whether there is a first scheduling scheme according to an operation state of other battery cells in the battery pack; a first scheduling execution module configured to control the electricity consumption action module to be discharged by the other battery cells in the battery pack for execution in a case where the first scheduling scheme exists; a second scheduling module configured to perform the following operations when the first scheduling scheme is not feasible: calculating a coupling strength between the electricity consumption action module and the abnormal battery cell; reconfiguring the electricity consumption action module with a coupling strength greater than a threshold value, extracting an electrical characteristic fingerprint of the user electricity consumption behavior, and constructing a behavior dependency relationship; acquiring a dynamic load portrait of other battery packs, and mapping the reconfigured sub-behavior unit to a power supply path of the other battery packs for execution based on an adaptation degree of the characteristic fingerprint and the load portrait.

Images