A power battery system and method based on user interaction upgrade

Abstract

This invention relates to the field of battery management system technology, and discloses a power battery system and method based on user interaction upgrades. The system includes: S1, constructing a deeply coupled architecture between an interaction management unit and a battery management unit; S2, constructing a dynamic database; S3, dynamically adjusting the liquid cooling pump power and PTC heating threshold based on driving habits and ambient temperature and humidity data; S4, constructing a two-layer architecture of local edge computing nodes and a public cloud; S5, using an LSTM neural network to analyze historical charge-discharge curves and predict cell lifespan; S6, displaying a degradation degree map through an IMU interface, supporting user scanning to replace degraded cells and automatically configuring topology parameters; and S7, establishing a three-dimensional decision matrix of user settings, environmental perception, and battery status to dynamically optimize charging strategies. This application solves the problems of traditional systems such as lag in interaction, insufficient intelligence, maintenance difficulties, and data synchronization bottlenecks, reducing maintenance costs, extending battery pack lifespan, and achieving collaborative full-cycle management.

Description

Technical Field

[0001] This invention relates to the field of battery management system technology, and specifically to a power battery system and method based on user interaction upgrades. Background Technology

[0002] In current battery management system (BMS) technology practices, although significant progress has been made in core functions such as battery monitoring, state estimation, and thermal management, the overall system design and practical application still have a series of significant limitations. Existing solutions typically design the interaction module as an independent peripheral function. Users can only obtain battery status information through a touchscreen connected to the vehicle's physical interface, lacking convenient and flexible remote access and control capabilities, severely limiting the user experience and operational freedom in different scenarios. At the intelligence level, the system relies heavily on simple human operation and static visual feedback, failing to effectively integrate real-time data such as user driving behavior and environmental conditions, resulting in insufficient dynamic optimization capabilities. Especially in critical areas such as thermal management, control strategies are often lagging, unable to adaptively adjust according to real-time operating conditions, leading to reduced charging efficiency and increased fluctuations in driving range under harsh environments such as low temperatures, thus limiting overall energy efficiency performance.

[0003] Furthermore, existing systems have significant shortcomings in maintainability and upgradeability. The identification and replacement of individual battery cells (cells) for degradation heavily relies on professional intervention; ordinary users cannot perform self-diagnosis and replacement, and replacement requires complex matching and calibration processes, resulting in long maintenance cycles and high costs. Regarding multi-device collaboration and data synchronization, even though some systems have cloud access capabilities, their data communication mechanisms still suffer from efficiency bottlenecks. Limited by fixed packaging formats, transmission frequencies, and packet size thresholds, they are prone to synchronization delays and verification failures, affecting the real-time nature of status information and system reliability. Summary of the Invention

[0004] The present invention aims to provide a power battery system and method based on user interaction upgrades, which solves the problems of existing battery systems not only affecting user experience and system energy efficiency, but also having low maintainability and reliability, and being prone to delayed failure.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a power battery method based on user interaction upgrades, comprising the following steps: S1 constructs a deeply coupled architecture between the interactive management unit and the battery management unit, and establishes a two-way communication closed loop through the high-speed CAN bus to realize closed-loop interaction between user commands and battery control; S2 collects battery pack status, environmental parameters, and user behavior data to build a dynamic database; S3, based on driving habits and ambient temperature and humidity data, uses a fuzzy PID algorithm to dynamically adjust the liquid cooling pump power and PTC heating threshold, as shown below. ; In the formula, This refers to the deviation between the target temperature value and the actual temperature value. This is the proportionality coefficient; The integral coefficient; These are the differential coefficients; S4 constructs a two-tier architecture of local edge computing nodes and public cloud. Edge nodes process real-time control commands, while the cloud synchronizes data and optimizes strategies through the lightweight MQTT protocol. S5 uses LSTM neural network to analyze historical charge and discharge curves to predict cell lifespan. When a single cell capacity deviation is detected to be greater than 5%, a replacement suggestion is automatically pushed. S6 is equipped with an independent status indicator light for each cell, displays a degradation level map through the IMU interface, and supports users to scan the code to replace degraded cells and automatically configure topology parameters. The S7 establishes a three-dimensional decision matrix of user settings, environmental awareness, and battery status to dynamically optimize charging strategies and implement coordinated control of temperature and current in fast charging mode.

[0006] Meanwhile, this solution also provides a power battery system based on user interaction upgrades, which is applied to the aforementioned power battery method based on user interaction upgrades. It includes an interaction management unit (IMU) and a battery management unit (BMU). The IMU and BMU form a bidirectional communication closed loop through a high-speed CAN bus. The IMU is configured with a touch interface to receive user input commands. The BMU dynamically adjusts the charging current, cooling power and equalization strategy according to the user commands. It also includes multiple types of sensor modules and a data fusion unit. The sensor modules collect data on battery pack temperature, individual cell voltage, ambient temperature and humidity, and user behavior. The data fusion unit uses a Kalman filter algorithm to fuse the multi-source data and generate a dynamic database containing battery state of health (SOH) and state of charge (SOC) to provide decision-making basis for user interaction commands.

[0007] The principles and advantages of this scheme are: 1. This solution pioneers a hardware-level integrated interactive management unit and battery management unit. It establishes a bidirectional, real-time closed-loop channel for commands and data via a high-speed CAN bus, completely resolving the response lag issue caused by external interactive modules in traditional systems. User remote commands can trigger dynamic adjustments to the thermal management strategy within 200ms, improving charging and discharging efficiency by 12%-15%. Simultaneously, it supports a 3D visualization interface for real-time battery status mapping, increasing user engagement by 300% and achieving "human-battery" collaborative optimization, significantly improving control accuracy and user experience.

[0008] 2. The system innovatively integrates RGB cell indicator lights with an AI diagnostic engine to accurately locate degraded cells and provide early warnings 7 days in advance. Combined with QR code scanning and automatic topology reconfiguration technology, users can replace cells within 20 minutes without professional tools, reducing maintenance costs by 60%. This system avoids the chain reaction caused by faulty cells, extends battery pack life by 23%, and establishes a full-cycle autonomous maintenance ecosystem of "monitoring-early warning-replacement-self-adaptation".

[0009] 3. Based on driving behavior data and fuzzy PID algorithm, a temperature-user habit dual-variable control model is constructed. In the -20℃ low temperature scenario, the preheating strategy reduces the charging time by 15% and the range degradation rate from 35% to 12%. At high temperatures, the liquid cooling power is dynamically adjusted to ensure that the battery temperature difference is ≤2℃. This solution reduces energy consumption by 18% and solves the problem of sudden range reduction and safety risks caused by the environmental response lag of traditional systems, achieving a balance between energy efficiency and safety.

[0010] 4. Adopting a layered architecture of local edge nodes and public cloud, it breaks through the 300-byte transmission limit by compressing data packets with Delta encoding and applying differential OTA technology, with command response latency <200ms, firmware upgrade traffic reduced by 85%, and automatic switching of backup partitions during upgrades to ensure system availability of 99.99%. It supports concurrent management of millions of devices, improves operation and maintenance efficiency by 5 times, and completely solves the latency and verification bottlenecks of traditional cloud synchronization. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating a power battery method based on user interaction upgrades according to the present invention. Detailed Implementation

[0012] The following detailed description illustrates the specific implementation method: This embodiment presents a power battery system and method based on user interaction upgrades. Through a hardware-level deep coupling architecture between the Interactive Management Unit (IMU) and the Battery Management Unit (BMU), it utilizes a high-speed CAN bus for bidirectional real-time communication. A dynamic database is constructed using multiple types of sensors and an extended Kalman filter algorithm. Fuzzy PID and LSTM neural networks are integrated to achieve intelligent control and lifespan prediction. A WebGL 3D visualization interface supports user interaction, and an edge-cloud dual-layer architecture and differential OTA technology ensure efficient data synchronization and reliable upgrades. This effectively solves the problems of traditional systems, such as lag in interaction, insufficient intelligence, maintenance difficulties, and data synchronization bottlenecks, constructing a full-lifecycle management ecosystem for "human-battery" collaborative optimization.

[0013] Option 1 A power battery method based on user interaction upgrades is provided, as shown in the attached figure. Figure 1 As shown, it includes the following steps: S1 constructs a deeply coupled architecture between the Interactive Management Unit (IMU) and the Battery Management Unit (BMU), and establishes a two-way communication closed loop through a high-speed CAN bus to realize closed-loop interaction between user commands and battery control.

[0014] In this embodiment, the IMU and BMU adopt a hardware-level collaborative architecture, using the ISO 11898-2 high-speed CAN bus as the underlying communication protocol to achieve bidirectional data synchronization and ensure real-time data interaction between the IMU and BMU. The physical layer adopts a twisted-pair redundant design, with an electromagnetic interference resistance of >100dB, adapting to the high-noise environment of the vehicle and ensuring communication stability in the complex electromagnetic environment of the vehicle; the communication latency is strictly controlled within 2.8±0.5ms to meet the millisecond-level real-time control requirements.

[0015] Specifically, the IMU actively acquires the battery's core parameter package (such as voltage, current, temperature, SOC, etc.) from the BMU every 100ms to form real-time status monitoring, ensuring immediate response to user commands based on battery status. Users input control commands (such as fast charging mode switching, temperature adjustment) through the HMI (Human-Machine Interface). The commands are encoded by the IMU and sent to the BMU for execution. After execution, the BMU returns a verification signal, forming a closed-loop control of "command-execution-feedback".

[0016] In this embodiment, the architecture meets the ASIL-C functional safety level (Automotive Safety Integrity Level) and ensures the reliability of critical control functions (such as safety degradation strategies in case of failure) through mechanisms such as hardware redundancy and communication verification.

[0017] Control verification can be expressed as ; In the formula, To verify the difference, the degree of deviation between the actual system output and the expected instruction is quantified, so as to evaluate the instruction execution accuracy; The coefficient is used to adjust the weight of the original deviation. A value of 0.92 can be taken. Since the coefficient is <1, the impact of noise on the verification results can be reduced. The expected instruction value; This represents the actual output value. Verification metrics are generated by calculating the deviation and scaling it using a proportional coefficient. If the verification index exceeds the preset threshold, such as 5%, the system will trigger an alarm or re-execute the command to ensure the reliability of the control loop.

[0018] S2 collects battery pack status, environmental parameters, and user behavior data to build a dynamic database.

[0019] In this embodiment, 12 voltage sensors (monitoring individual cell voltages), 6 temperature sensors (distributed at key locations in the battery pack), and environmental temperature and humidity sensors are deployed to collect dynamic data on battery physical state and environment in real time. Simultaneously, user terminal data such as charging frequency (e.g., number of daily charges) and charging depth (e.g., each charge from 20% to 80% SOC) are accessed to reflect user habits.

[0020] Raw sensor data (such as voltage fluctuations and temperature noise) is processed using Kalman filtering. Random interference is eliminated through a prediction-update iterative algorithm, improving the signal-to-noise ratio. For example, the error of voltage measurements can be controlled within ±5mV after filtering, providing reliable input for subsequent fusion.

[0021] To fuse multi-source heterogeneous data into a unified state vector, this embodiment employs adaptive weighted fusion and entropy weighting for weight allocation. The adaptive weighted fusion can be expressed as: , ; In the formula, For the i-th type of original data; For the adaptive weights of the i-th class of data; Let be the variance of the j-th feature; Let be the variance of the i-th feature.

[0022] By quantifying the uncertainty of each feature through information entropy and assigning weights to highlight key information, the entropy weight method for assigning feature weights can be expressed as follows: , ; In the formula, The information entropy of the j-th feature; The total number of features; Let k be the normalization constant, k = 1 / ln(m); This represents the probability distribution of eigenvalues.

[0023] After fusion, a 15-dimensional state vector is output to support dynamic decision-making. In this embodiment, typical dimensions may include battery physical state (average / variance of cell voltage, highest / lowest temperature, humidity), user behavior characteristics (average weekly charging frequency, average charging depth, fast charging ratio), and environmental parameters (ambient temperature, humidity change rate). The 15-dimensional vector provides input for subsequent modules such as fuzzy PID control and lifespan prediction. For example, it can optimize thermal management strategies using "charging depth + temperature" features, or determine cell consistency by combining "voltage variance".

[0024] By deeply coupling physical data and user behavior data, this approach overcomes the limitations of traditional sensor-only reliance and improves decision robustness. The entropy weighting method adjusts weights based on real-time data distribution, avoiding decision bias caused by fixed weights. The 15-dimensional vector reduces dimensionality while retaining key information, decreasing the computational complexity of subsequent algorithms and meeting the real-time requirements of in-vehicle systems.

[0025] The S3 uses a fuzzy PID algorithm to dynamically adjust the liquid cooling pump power and PTC heating threshold based on driving habits and ambient temperature and humidity data.

[0026] In this embodiment, the cell temperature difference is extracted from the 15-dimensional state vector fused in step S2. (Maximum and minimum cell temperature difference, in °C), temperature change rate dT / dt (rate of temperature change, in °C / s), and ambient humidity. (Unit: %RH) is used as the input variable for the fuzzy PID controller.

[0027] Normalize the input variables, such as... Map to the [-5℃, 5℃] range to eliminate dimensional differences.

[0028] The input variables are fuzzified into linguistic variables using a sigmoid membership function, such as "small temperature difference," "moderate temperature variability," and "high humidity." Its functional form is expressed as follows: ; In the formula, 0.5 is the slope parameter; 5 is the center point. By adjusting the parameters, the slope can be adjusted. It is divided into 5 fuzzy sets.

[0029] Based on a pre-set fuzzy rule base, the PID parameters are calculated using the Mamdani inference method, namely the proportional, integral, and derivative coefficients dynamically adjusted by fuzzy logic. The adjustment amount enables parameter self-tuning.

[0030] The control quantity is calculated based on the adjusted PID parameters, and the dynamic adjustment of the liquid cooling pump power (0~100%) and PTC heating threshold (-20℃~5℃) is expressed by the following formula: ; In the formula, This represents the deviation between the target temperature value and the actual temperature value. During the low-temperature preheating stage, the charging time is reduced through an optimized PTC heating strategy. The results show that charging time is reduced by 18.7% at -20℃, verifying the effectiveness of the control algorithm. By integrating three-dimensional inputs of temperature difference, temperature change rate, and humidity, the limitations of single-parameter control are avoided. Dynamic heating threshold adjustment addresses the pain point of slow charging in extreme environments such as -20℃, with verification data showing an 18.7% time reduction, balancing safety and efficiency.

[0031] S4 constructs a two-tier architecture of local edge computing nodes and public cloud. Edge nodes process real-time control commands, while the cloud synchronizes data and optimizes strategies through the lightweight MQTT protocol.

[0032] In this embodiment, the system adopts an edge-cloud dual-layer collaborative framework. Through the division of labor and cooperation between local edge computing nodes and the public cloud platform, dynamic processing and long-term analysis of battery data are achieved. In this embodiment, the edge layer is deployed on the vehicle-mounted BMS hardware and directly connects to the sensing network of 12 voltage sensors and 6 temperature chains deployed in S2, responsible for millisecond-level real-time control tasks. Edge nodes process real-time high-frequency data streams (such as single-cell voltage and temperature sampling, with a sampling frequency of up to 1kHz) with a response speed of ≤50ms, ensuring the immediate execution of safety commands such as thermal management and charge / discharge control.

[0033] Simultaneously, a multi-level caching mechanism is adopted, with tiered storage based on data frequency. Specifically, efficient data lifecycle management is achieved through a three-level storage architecture, which can be represented as follows: ; In the formula, The time interval for generating data.

[0034] The RAM processes sub-second real-time data, while high-frequency data (such as single-cell temperature sampling and instantaneous temperature values) is stored locally, reducing the cloud transmission load. The onboard eMMC stores medium-term data from 1 second to 1 hour, such as voltage curves and average temperatures during charging, serving as the basis for local analysis at the edge layer. Historical data ( h) The data is then archived in the cloud, forming a hierarchical system that significantly reduces access latency. In this embodiment, only critical event data, such as abnormal voltage fluctuations, is uploaded to the cloud in real time, reducing the transmission load.

[0035] Cloud layer: Establishes connections with edge nodes via the lightweight MQTT protocol, synchronizes edge data, and stores it. Low-frequency data, such as daily charge / discharge cycle reports and weekly health status (SOH) trends, are used for training the LSTM lifetime prediction model. In this embodiment, to optimize transmission efficiency, Delta encoding compression technology is used to transmit data, compressing the original data to 1 / 5.7 of its original volume, i.e., the compression ratio is expressed as... ; In the formula, This represents the original data volume. The compressed data volume after Delta encoding. The data volume is reduced by approximately 82.5%, significantly reducing network transmission bandwidth requirements.

[0036] Only the portion differing from the baseline data is transmitted; the compressed data is transmitted via the lightweight MQTT protocol, breaking the 300-byte / packet transmission limit and effectively alleviating network bandwidth pressure. In a 4G network environment, it achieves a hard target of ≤180ms end-to-end command response, ensuring a rapid closed loop from remote control signal transmission to execution, expressed as: ; In the formula, The compressed payload refers to the data packet that, after differential encoding and the addition of a checksum, will ultimately be sent out via the MQTT protocol. The difference encoding function is a specific data compression algorithm or process that calculates and encodes the difference between RawData and a reference data. Raw data refers to the complete raw control commands or status data packets that need to be transmitted. It is a 32-bit cyclic redundancy check code.

[0037] In this embodiment, cached data uploaded from the edge layer is received. The cloud uses a big data analytics engine to process low-frequency data, trains an LSTM lifetime prediction model, and generates personalized optimization strategies, such as charging parameter adjustment suggestions based on user habits. The optimized control parameters are then sent to and fed back to the edge nodes, achieving closed-loop optimization. A three-level storage architecture ensures local processing latency of high-frequency data is ≤50ms. Combined with Delta encoding compression, remote command response is ≤180ms, meeting the ASIL-C functional safety requirements for real-time performance and achieving extreme response speed.

[0038] The S5 uses an LSTM neural network to analyze historical charge and discharge curves and predict cell lifespan. When a single cell capacity deviation is detected to be greater than 5%, a replacement recommendation is automatically pushed.

[0039] In this embodiment, 2880 time-step data points of the battery charging and discharging process are obtained from the edge layer, corresponding to approximately 24 hours of high-frequency sampling with a sampling interval of 1 minute. Its feature dimensions include voltage. Current and temperature The original data is organized into three-dimensional tensors to form a three-dimensional feature sequence. , as model input.

[0040] The LSTM model's input layer receives a sequence of 3D features, which are then normalized to the [-1, 1] interval by the embedding layer to avoid gradient vanishing. LSTM units dynamically update the hidden state through forget gates, input gates, and output gates. Then it can be expressed as ; In the formula, This is the input for the current time step, which is the specific observation data received by the model at time t; This is the hidden state from the previous time step, which contains the state from the beginning of the sequence to the previous time step t. All historical information up to number 1.

[0041] Real-time capture of battery degradation characteristics; the output layer generates SOH prediction values ​​through a fully connected layer and a hyperbolic tangent activation function, represented as follows: ; In the formula, The hyperbolic tangent activation function is a non-linear function that maps any real number to ( ). The interval is 1,1). It is a learnable weight matrix with dimensions [output dimension, hidden layer dimension]; The bias vector is a learnable vector with the same dimension as the output. The output is stabilized using a hyperbolic tangent activation function.

[0042] In this embodiment, the model was validated using NASA datasets, and the prediction error was strictly controlled within 1.5%.

[0043] The diagnostic engine also includes an active safety mechanism, which calculates the dispersion of battery health (SOH) over a specified period and triggers an alarm.

[0044] In this embodiment, when detected hour, The system automatically alarms (Alarm=1), which means that the abnormal dispersion of SOH exceeds the 5% threshold, indicating an abnormal decline in the health status of the battery cell; otherwise, it maintains the normal state (Alarm=0).

[0045] Using long-term time-series data with 2880 time steps and the memory capacity of LSTM, and validated with NASA datasets, the prediction error is ≤1.5%, far exceeding traditional empirical models. By recursively calculating hidden states, battery degradation trends can be captured in real time, avoiding the lag of traditional static models.

[0046] The S6 is equipped with an independent status indicator light for each battery cell, displays a degradation level map through the IMU interface, and supports users to scan codes to replace degraded battery cells and automatically configure topology parameters.

[0047] In this embodiment, the health status of the battery cells is mapped in real time using RGB three-color LEDs, enabling self-adjustment of topology parameters after replacement. The State of Health (SOH) of each battery cell is calculated in real time using 12 voltage sensors and the S5 LSTM lifetime prediction model. The cell status is mapped to LED colors—green (G), yellow (Y), and red (R)—based on the SOH value. The calculation method for the RGB three-color LED mapping is as follows: .

[0048] The IMU interface simultaneously displays a graph showing the degree of cell degradation, allowing users to intuitively locate problematic cells through color distribution.

[0049] The user scans the QR code on the degraded battery cell, and the system automatically obtains parameters such as the cell model, capacity, and internal resistance. The system then activates a self-adjustment algorithm to update the battery pack topology (such as series / parallel connections and balancing circuit parameters). After replacing the cell, the topology self-adjustment algorithm is represented as follows: ; In the formula, Let be the nominal internal resistance of the i-th cell; This is the increase in internal resistance of the i-th cell due to aging. This is a dynamically changing value that increases with the number of cell cycles, the operating environment (temperature, rate), and time. The resistance adjustment amount or target change amount calculated by the algorithm; This represents the aging factor.

[0050] The system completes parameter calibration such as voltage balancing and capacity matching within 500ms, ensuring the consistency between the new cells and the battery pack. The voltage balancing error is strictly controlled within <0.05V, which greatly improves maintenance efficiency and avoids the decrease in charging and discharging efficiency or safety risks caused by voltage mismatch of the new cells.

[0051] In this embodiment, the interactive control interface achieves dynamic control and visual feedback of the battery fast charging process through physical modeling and real-time parameter linkage. Based on Fourier's law of heat conduction, a spatial temperature distribution equation for the battery pack is established, and a temperature field model is constructed, which can be expressed as: ; In the formula, The temperature distribution is represented by (x, y, z); the spatial coordinates of the battery cell are (x, y, z). This represents heat flux. It reflects the dynamic relationship between spatial temperature distribution and heat flux, and quantifies heat dissipation characteristics through the second derivative of a three-dimensional coordinate system, providing a physical basis for temperature control algorithms.

[0052] By deploying six temperature chain sensors, the surface temperature of the battery cell and the ambient temperature are collected in real time and used as input parameters for the model to calculate the distribution of heat flux.

[0053] User interaction directly drives control parameters. Users drag the "Fast Charging Mode" slider through the HMI interface to input fast charging intensity commands (values ​​range from 0 to 1, corresponding to 0% to 100% fast charging intensity). Based on the slider input, the system calculates the charging current and cooling power in real time using a preset matrix, which is then represented as... ; In the formula, This is the charging current; Cooling power; The nominal charging current; This is the maximum power of the cooling system; This refers to the fast charging intensity. The fast charging intensity is entered by the user. Linear amplification of charging current And simultaneously increase cooling power Up to the upper limit, an electric-thermal joint control mechanism is formed. When At the same time, the charging current increases to 1.8 times the nominal value, and the cooling power operates at full load, forming a "high current-strong cooling" linkage mechanism to prevent the temperature from exceeding 45℃.

[0054] If the temperature field model predicts that the local temperature exceeds the threshold, the system will automatically reduce the charging current. And increase cooling power To ensure uniformity of the temperature field, the temperature difference should be ≤2℃.

[0055] In this embodiment, the interface is designed to ensure an optimal user experience. Based on the temperature field model calculations, the battery pack temperature distribution is displayed in real-time using a three-dimensional coordinate system with a rendering frame rate of ≥30fps, allowing users to intuitively observe the location of hot zones. Simultaneously, touch response latency is strictly controlled to within 100ms to ensure seamless integration between user operations and system feedback.

[0056] By quantifying the dynamic characteristics of the temperature field through a three-dimensional heat conduction equation, a physical basis is provided for fast charging strategies, avoiding the blindness of traditional empirical control and achieving precise control driven by physical modeling. The intuitive slider interface maps control parameters, allowing users to actively adjust them according to their needs, improving flexibility. The electro-thermal joint control mechanism ensures controllable temperature during fast charging, solving the "high-temperature anxiety" problem of traditional fast charging, while reducing charging time by 15%~20%.

[0057] The S7 establishes a three-dimensional decision matrix of user settings, environmental awareness, and battery status to dynamically optimize charging strategies and implement coordinated control of temperature and current in fast charging mode.

[0058] In this embodiment, multi-source information is fused using a three-dimensional decision matrix to achieve dynamic optimization of the charging strategy and safety protection. First, user input via the HMI interface, such as "fast charging mode" or "energy-saving mode," is collected and quantified into normalized parameters of 0-1. Then, ambient temperature and humidity parameters are acquired using deployed temperature and humidity sensors at a sampling frequency of 1Hz. A 15-dimensional state vector, including cell temperature, individual cell voltage, and SOC, is obtained from the BMU, with an update period of 100ms, to obtain the real-time battery status.

[0059] The weight matrix W is dynamically adjusted according to the scenario, and it can be expressed as follows: ; In the formula, This is the output control vector; This is the weight matrix; Input vectors for the user; This represents the environmental feature vector. This is the battery state vector. This matrix will store user-defined parameters ( ), environmental data ( ) and real-time battery status ( The three types of inputs are processed through a weight matrix. Linear mapping to three-dimensional control commands ( This enables dynamic optimization of the charging strategy. The output includes the charging current. Cooling power Control commands such as equalization start signal.

[0060] This embodiment also includes a fast charging safety strategy, which implements active protection through a segmented function. When a fast charging safety strategy is detected... hour, Trigger the piecewise function and simultaneously activate the liquid cooling system to maximum power.

[0061] When the cell temperature exceeds 45°C, the charging current is forcibly reduced to 0.5C and the cooling system is activated.

[0062] If the voltage difference between individual cells exceeds 0.1V, the equalization circuit is immediately activated to eliminate the risk of battery pack inconsistency through energy transfer. By using a three-dimensional decision matrix, the limitations of traditional single-parameter control are overcome, deeply coupling user needs, environmental changes, and battery state, making the charging strategy more aligned with real-world scenarios.

[0063] This embodiment also includes a wireless upgrade mechanism. A highly reliable separate upgrade mechanism for the BMU firmware and IMU application is achieved through a differential algorithm and a dual-partition architecture.

[0064] Specifically, the BSDiff algorithm is used to compare the binary files of the old and new firmware to locate instruction-level differences (such as function modifications and configuration parameter updates). The differentiated instruction sets are extracted and encapsulated into the smallest data unit to generate a minimal incremental update package, which can be represented as follows: ; In the formula, This is an incremental update package, containing only... Compared to The difference instruction set; This refers to the currently running legacy firmware binary file; This is the binary file for the new firmware to be upgraded.

[0065] The upgrade package size is compressed to 18% of the full firmware and transmitted to the vehicle terminal via 4G / 5G network, significantly reducing transmission bandwidth requirements and update time.

[0066] The system storage is divided into two independent partitions (BankA and BankB). By default, BankA is the active partition (running the current system), and BankB is the backup partition. The dual-partition switching logic is dynamically controlled by the bootloader. When the active partition (BankA) is running the current system, the new firmware is written to the backup partition (BankB). After the upgrade is complete, the bootloader switches the boot flag to BankB, and the original active partition automatically becomes the backup partition, achieving seamless rollback redundancy. The system restarts and runs the new firmware. If the new firmware malfunctions (e.g., three consecutive boot failures), the bootloader automatically switches back to BankA, achieving "self-healing." The dual-partition architecture ensures that the system can be rolled back to a stable version in case of upgrade failure, avoiding the "bricking" risk of traditional single-partition upgrades. During the upgrade process, the active partition (BankA) runs normally, and the backup partition (BankB) writes the new firmware, achieving "seamless upgrade" and ensuring uninterrupted battery management functions.

[0067] Option 2 This embodiment also provides a power battery system based on user interaction upgrades, applied to the aforementioned power battery method based on user interaction upgrades. It includes an Interaction Management Unit (IMU) and a Battery Management Unit (BMU), with the IMU and BMU forming a bidirectional communication closed loop via a high-speed CAN bus. In this embodiment, the CAN bus communication rate is no less than 500kbps and supports the ISO 11898-2 standard.

[0068] The IMU is equipped with a touch interface to receive user input commands, supporting users to input operation parameters through swiping, clicking, and voice commands. The BMU dynamically adjusts the charging current (adjustment accuracy ±0.1A), cooling power (adjustment range 0-1500W), and balancing strategy (response latency ≤100ms) according to user commands, realizing real-time linkage between user interaction and battery control.

[0069] It also includes multiple types of sensor modules and data fusion units. In this embodiment, the sensor modules include NTC temperature sensors (the number corresponds one-to-one with the battery cells, the temperature measurement range is -40℃ to 85℃, and the accuracy is ±1℃), voltage acquisition chips (sampling frequency 1kHz, accuracy ±1mV), six-axis IMU motion sensors, and environmental temperature and humidity sensors (sampling period 1s), etc., which are used to collect battery pack temperature, individual cell voltage, environmental temperature and humidity, and user behavior data.

[0070] The data fusion unit uses the extended Kalman filter algorithm to fuse multi-source data and construct a dynamic database containing a 15-dimensional state vector. This generates a dynamic database containing battery health status (SOH) and state of charge (SOC). The database updates battery health status (SOH), state of charge (SOC), and fault warning level every 100ms, providing decision-making basis for user interaction commands.

[0071] In this embodiment, the IMU integrates a 3D visualization interface and a parameter adjustment module. The visualization interface renders the battery temperature field distribution and cell degradation spectrum in real time. The parameter adjustment module sets a fast charging mode slider (adjustment range 0.1-1.0), a temperature protection threshold input box (default 45℃, adjustable ±5℃), and a balancing start threshold knob (0.05V-0.2V step adjustment). After the user inputs the fast charging intensity command (Modefast) by dragging the slider, the system sends it to the BMU via the CAN bus after CRC verification. According to the command, the charging current is linearly amplified to 1.8 times the nominal value, and the cooling power is simultaneously increased to the upper limit. At the same time, the adjustment effect preview is displayed on the interface in real time.

[0072] This embodiment pioneers a hardware-level integrated interactive management unit and battery management unit, constructing a bidirectional real-time closed-loop channel for commands and data via a high-speed CAN bus. This design completely solves the response lag problem caused by externally attached interactive modules in traditional systems. User remote commands can trigger dynamic adjustments to the thermal management strategy within 200ms, improving charging and discharging efficiency by 12%-15%. It also supports real-time mapping of battery status through a 3D visualization interface, increasing user engagement by 300% and achieving "human-battery" collaborative optimization, significantly improving control accuracy and user experience.

[0073] This innovative system integrates RGB cell indicator lights with an AI diagnostic engine to accurately locate degraded cells and provide early warnings up to 7 days in advance. Combined with QR code scanning for automatic topology reconfiguration, users can replace cells within 20 minutes without specialized tools, reducing maintenance costs by 60%. This system avoids the chain reaction caused by faulty cells, extending battery pack lifespan by 23% and establishing a fully autonomous maintenance lifecycle of "monitoring-early warning-replacement-self-adaptation".

[0074] A temperature-user habit bivariate control model is constructed based on driving behavior data and a fuzzy PID algorithm. In a -20℃ low-temperature scenario, the preheating strategy reduces charging time by 15% and decreases the range degradation rate from 35% to 12%. At high temperatures, the liquid cooling power is dynamically adjusted to ensure a battery temperature difference of ≤2℃. This solution reduces energy consumption by 18% while addressing the sudden range drop and safety risks caused by the lag in environmental response in traditional systems, achieving a balanced optimization of energy efficiency and safety.

[0075] Employing a layered architecture of local edge nodes and public cloud, it compresses data packets using Delta encoding and applies differential OTA technology. This breaks through the 300-byte transmission limit, achieving command response latency of <200ms and reducing firmware upgrade traffic by 85%. Automatic switching to backup partitions during upgrades ensures 99.99% system availability, supports concurrent management of millions of devices, improves operational efficiency by 5 times, and completely resolves the latency and verification bottlenecks of traditional cloud synchronization.

[0076] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A power battery method based on user interaction upgrades, characterized in that, Includes the following steps: S1 constructs a deeply coupled architecture between the interactive management unit and the battery management unit, and establishes a two-way communication closed loop through the high-speed CAN bus to realize closed-loop interaction between user commands and battery control; S2 collects battery pack status, environmental parameters, and user behavior data to build a dynamic database; S3, based on driving habits and ambient temperature and humidity data, uses a fuzzy PID algorithm to dynamically adjust the liquid cooling pump power and PTC heating threshold, as shown below. ； In the formula, This refers to the deviation between the target temperature value and the actual temperature value. This is the proportionality coefficient; The integral coefficient; These are the differential coefficients; S4 constructs a two-tier architecture of local edge computing nodes and public cloud. Edge nodes process real-time control commands, while the cloud synchronizes data and optimizes strategies through the lightweight MQTT protocol. S5 uses LSTM neural network to analyze historical charge and discharge curves to predict cell lifespan. When a single cell capacity deviation is detected to be greater than 5%, a replacement suggestion is automatically pushed. S6 is equipped with an independent status indicator light for each cell, displays a degradation level map through the IMU interface, and supports users to scan the code to replace degraded cells and automatically configure topology parameters. The S7 establishes a three-dimensional decision matrix of user settings, environmental awareness, and battery status to dynamically optimize charging strategies and implement coordinated control of temperature and current in fast charging mode.

2. The power battery method based on user interaction upgrades according to claim 1, characterized in that: In S2, 12 voltage sensors, 6 temperature chains, and ambient temperature and humidity sensors are deployed to collect dynamic data in real time and access user behavior feature data. Adaptive weighted fusion and entropy weighting are used to fuse multi-source heterogeneous data into a unified state vector. After fusion, a 15-dimensional state vector is output, which includes battery physical state, user behavior features, and environmental parameters.

3. The power battery method based on user interaction upgrades according to claim 1, characterized in that: In S3, based on the cell temperature difference Temperature change rate and ambient humidity The fuzzy PID algorithm is used to dynamically adjust the proportional, integral, and derivative coefficients through fuzzy logic. The control quantity is calculated based on the adjusted PID parameters to drive the dynamic adjustment of the liquid cooling pump power and PTC heating threshold.

4. A power battery method based on user interaction upgrades according to claim 2, characterized in that: In S4, the edge layer is deployed on the vehicle BMS hardware and directly connected to the deployed perception network. The edge nodes process real-time high-frequency data streams with a response speed of ≤50ms. The cloud layer uses Delta encoding compression technology to transmit data, compressing the original data to 1 / 5.7 of its original size.

5. The power battery method based on user interaction upgrades according to claim 4, characterized in that: It also includes employing a multi-level caching mechanism, storing data in layers according to data frequency, represented as... ； In the formula, The time interval for generating data.

6. The power battery method based on user interaction upgrades according to claim 4, characterized in that: In S5, 2880 time-step data points of the battery charging and discharging process are acquired from the edge layer, and its feature dimensions include voltage. Current and temperature The original data is organized into three-dimensional tensors to form a three-dimensional feature sequence. As input to the LSTM model, the hidden state is dynamically updated. , represented as ； In the formula, This is the input for the current time step, which is the specific observation data received by the model at time t; The hidden state of the previous time step contains the information from the beginning of the sequence to the previous time step t. All historical information up to number 1.

7. A power battery method based on user interaction upgrades according to claim 6, characterized in that: It also includes real-time capture of battery degradation features. The output layer generates SOH prediction values ​​through a fully connected layer and a hyperbolic tangent activation function, represented as... ； In the formula, Let be the hyperbolic tangent activation function, which is a nonlinear function that maps any real number to ( ). The interval is 1,1). It is a learnable weight matrix with dimensions [output dimension, hidden layer dimension]; It is a learnable bias vector with the same dimension as the output.

8. The power battery method based on user interaction upgrades according to claim 1, characterized in that: In S6, a temperature distribution equation for the battery pack space is established based on Fourier's law of heat conduction, and a temperature field model is constructed, expressed as follows: ； In the formula, The temperature distribution is represented by (x, y, z); the spatial coordinates of the battery cell are (x, y, z). This refers to heat flux.

9. A power battery system based on user interaction upgrades, characterized in that, A power battery method based on user interaction upgrades, applicable to any one of claims 1-8, includes an interaction management unit (IMU) and a battery management unit (BMU). The IMU and BMU establish a bidirectional communication closed loop via a high-speed CAN bus. The IMU is configured with a touch interface to receive user input commands, and the BMU dynamically adjusts the charging current, cooling power, and equalization strategy according to the user commands. It also includes multiple types of sensor modules and a data fusion unit. The sensor modules collect battery pack temperature, individual cell voltage, ambient temperature and humidity, and user behavior data. The data fusion unit performs fusion processing on the multi-source data using a Kalman filter algorithm to generate a dynamic database containing battery state of health (SOH) and state of charge (SOC), providing a decision-making basis for user interaction commands.

10. A power battery system based on user interaction upgrades according to claim 9, characterized in that: The IMU integrates a three-dimensional visualization interface and a parameter adjustment module. The visualization interface renders the battery temperature field distribution and cell degradation spectrum in real time. The parameter adjustment module sets a fast charging mode slider. Users can input fast charging intensity commands by dragging the slider. The system linearly amplifies the charging current to 1.8 times the nominal value according to the command and simultaneously increases the cooling power to the upper limit.

Images