A battery life cycle state monitoring and power estimation system and method

Abstract

The application discloses a battery life cycle state monitoring and power estimation system and method, and the system comprises a cloud big data platform and an intelligent battery terminal. The cloud platform receives operation data uploaded by the terminal, fits a personalized equivalent circuit model parameter table, predicts a maximum available capacity of the battery, generates a strategy configuration file and issues the strategy configuration file. An embedded battery management master control unit of the terminal stores a preset noise covariance matrix. The master control unit loads the configuration file, obtains an ohmic resistance, a polarization resistance and a polarization capacitance from the parameter table according to real-time working conditions through interpolation. The master control unit constructs an extended Kalman filter state observer based on the maximum available capacity of the battery, the above parameters and the preset noise covariance matrix. The system performs multi-source fusion processing on collected voltage, current and temperature signals, and outputs a real-time state of charge estimation value.

Description

Technical Field

[0001] This invention belongs to the field of battery management technology and relates to a battery life cycle state monitoring and power estimation system and method. Background Technology

[0002] Lithium-ion batteries are widely used in electric vehicles and energy storage systems due to their high energy density and long cycle life. The battery management system (BMS) is a core component ensuring the safe operation of the battery pack. The accuracy of state-of-charge (SOC) estimation directly determines the accuracy of the vehicle's range display. Existing BMS systems mostly use the ampere-hour integration method or the open-circuit voltage method for SOC estimation. The ampere-hour integration method suffers from the problem of accumulated current measurement errors, which gradually increase the estimation deviation over long periods of operation. The open-circuit voltage method requires the battery to be stationary for an extended period to obtain an accurate voltage measurement. During vehicle operation, the battery is constantly in a dynamic state, making this method unsuitable for real-time requirements.

[0003] Some existing technologies introduce equivalent circuit models combined with Kalman filtering algorithms for state estimation. This method corrects estimation errors by establishing a mathematical relationship between battery terminal voltage and internal state. However, battery internal parameters exhibit significant nonlinear characteristics. Ohmic resistance and polarization parameters vary drastically with state of charge and temperature. Traditional methods typically employ fixed parameters or simple lookup tables. Fixed parameters cannot adapt to the aging characteristics throughout the battery's entire lifespan. Simple lookup tables struggle to cover the complex and variable actual operating conditions. This leads to a significant decrease in estimation accuracy during high-current charging / discharging or extreme temperatures.

[0004] Furthermore, battery performance degrades with increased usage. Decreased maximum usable capacity and increased internal resistance are the main signs of aging. Existing local battery management systems have limited computing power, making it difficult to run complex system identification algorithms or machine learning models. Local systems cannot accurately predict battery health based on historical big data and lack a dynamic update mechanism for battery lifecycle parameters. When abnormalities such as sudden increases in internal resistance or capacity drops occur, local systems often react slowly and cannot generate targeted safety protection strategies in a timely manner. This poses a threat to the long-term safe operation of the battery system. Therefore, there is an urgent need for a system that can combine cloud-based big data analysis with high-precision local estimation to achieve predictive health management throughout the battery's lifecycle. Summary of the Invention

[0005] To address the problems existing in the background technology, this invention proposes a battery life cycle state monitoring and power estimation system and method.

[0006] The first aspect of this application provides a battery life cycle state monitoring and power estimation system, including: a cloud big data platform and a smart battery terminal; The cloud-based big data platform receives and stores the battery identification and multi-dimensional time-series operation data uploaded by the smart battery terminal, uses a system identification algorithm to fit a personalized equivalent circuit model parameter table, uses a machine learning regression algorithm to predict the maximum usable capacity of the battery, generates a strategy configuration file containing the parameter table and the capacity, and distributes it. The intelligent battery terminal includes a data acquisition module, a communication module, and an embedded battery management main control unit. The data acquisition module synchronously acquires battery pack voltage, current, and temperature signals. The communication module receives the strategy configuration file. The embedded battery management main control unit interpolates the ohmic internal resistance, polarization resistance, and polarization capacitance values ​​from the parameter table based on the real-time state of charge estimate and the measured temperature value. Based on the battery's maximum usable capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and a preset noise covariance matrix, it constructs an extended Kalman filter state observer and performs multi-source fusion processing on the acquired signals to output a real-time state of charge estimate.

[0007] Optionally, the personalized equivalent circuit model parameter table is a lookup table of ohmic internal resistance, polarization resistance, and polarization capacitance as they change with state of charge and temperature; the embedded battery management main control unit uses bilinear interpolation to look up the ohmic internal resistance, polarization resistance, and polarization capacitance values ​​corresponding to the current operating condition from the parameter table in each estimation cycle.

[0008] Optionally, the state variables of the extended Kalman filter state observer include the state of charge and polarization voltage; the preset noise covariance matrix includes the process noise covariance matrix and the observation noise covariance matrix; when the embedded battery management main control unit performs the state prediction step, it calculates the predicted value of the state of charge and the predicted value of the polarization voltage based on the state equation of the ampere-hour integral principle and the equivalent circuit model, wherein the calculation of the predicted value of the state of charge introduces a compensation coefficient obtained by looking up a table based on real-time temperature and discharge rate to correct the maximum usable capacity of the battery.

[0009] Optionally, when the extended Kalman filter state observer performs the observation update step, it constructs a terminal voltage observation equation, which is obtained by subtracting the product of the current and the ohmic internal resistance from the open-circuit voltage function and then subtracting the polarization voltage prediction value. The embedded battery management main control unit compares the calculation result of the terminal voltage observation equation with the measured terminal voltage to calculate the observation error value, and calculates the Kalman gain based on the process noise covariance matrix and the observation noise covariance matrix, thereby correcting the state of charge prediction value and the polarization voltage prediction value to obtain the final real-time state of charge estimate value.

[0010] Optionally, the cloud-based big data platform also performs an anomaly detection process. By calculating the deviation between the ohmic internal resistance value of the current cycle and the average ohmic internal resistance value of the past few cycles, or by comparing the decrease ratio of the average capacity of the most recent few cycles with the average capacity of the previous few cycles, when the deviation or decrease ratio exceeds a preset threshold, it is determined that the battery has experienced a sudden increase in internal resistance or a drop in capacity, and an updated strategy configuration file containing dynamically adjusted safety thresholds is generated and sent to the smart battery terminal.

[0011] Optionally, the embedded battery management main control unit also performs local real-time early warning. When it detects that the real-time voltage drop slope exceeds the preset power overload threshold, or the lowest cell voltage is lower than the dynamic hard cutoff voltage threshold issued in the updated strategy configuration file, it immediately performs forced protection action and cuts off the discharge circuit.

[0012] A second aspect of this application provides a method for monitoring battery life cycle state and estimating battery capacity, including: The cloud-based big data platform receives battery identification data and multi-dimensional time-series operational data uploaded by the smart battery terminal; The cloud-based big data platform uses the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table based on the data, and uses a machine learning regression algorithm to predict the maximum usable capacity of the battery. The cloud-based big data platform packages the parameter table and the capacity into a strategy configuration file and sends it to the smart battery terminal. The embedded battery management main control unit of the intelligent battery terminal loads the strategy configuration file and the preset noise covariance matrix, and obtains the ohmic internal resistance, polarization resistance and polarization capacitance from the parameter table by interpolation based on the real-time collected voltage, current and temperature signals. The embedded battery management main control unit constructs an extended Kalman filter state observer, using the battery's maximum available capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and the preset noise covariance matrix as model parameters, and performs multi-source fusion calculation on the collected signals to output a real-time state of charge estimate.

[0013] Optionally, the step of using the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table includes: extracting the features of the charge and discharge curves, calculating the average internal resistance and polarization voltage time constant of a specific state of charge interval, and fitting the mapping relationship between the ohmic internal resistance, polarization resistance and polarization capacitance and the state of charge and temperature using the recursive least squares method.

[0014] Optionally, the step of performing multi-source fusion calculation on the acquired signal includes: calculating the predicted state of charge based on the ampere-hour integration principle, calculating the predicted polarization voltage based on the equivalent circuit model, constructing a terminal voltage observation equation that includes the product of open-circuit voltage, current and ohmic internal resistance and the predicted polarization voltage, correcting the observation equation using the measured terminal voltage to eliminate the instantaneous voltage drop interference caused by the large current load, and outputting the corrected real-time state of charge estimate.

[0015] Optionally, the method further includes a predictive maintenance step: the cloud big data platform continuously analyzes the uploaded multi-dimensional time-series operating data, and when it detects that the ohmic internal resistance growth trend or the capacity decay rate meets the preset abnormal judgment conditions, it generates a health status assessment report and remaining service life prediction results, and sends early warning information to the operation and maintenance platform or issues updated safety protection thresholds to the smart battery terminal.

[0016] Compared with the prior art, the present invention has the following beneficial effects: First, this invention organically combines cloud-based big data analysis with local high-precision estimation. The cloud platform utilizes a system identification algorithm to fit a personalized equivalent circuit model parameter table. The cloud platform also uses machine learning regression algorithms to predict the maximum usable battery capacity. This architecture solves the problem of insufficient computing power in the locally embedded battery management control unit. The local terminal can obtain accurate battery parameters without running complex training models. This significantly reduces hardware costs and improves system response speed.

[0017] Secondly, this invention significantly improves the accuracy and robustness of state-of-charge (POC) estimation. The embedded battery management control unit interpolates the ohmic internal resistance, polarization resistance, and polarization capacitance from a parameter table based on real-time operating conditions. The system constructs an extended Kalman filter state observer using a preset noise covariance matrix. This method effectively corrects the accumulated error of the ampere-hour integration method. It also eliminates instantaneous voltage drop interference caused by high-current loads. Even under extreme temperature or battery aging conditions, the system can still output accurate real-time POC estimates.

[0018] Furthermore, this invention possesses predictive health management capabilities throughout the entire battery lifecycle. The cloud platform continuously monitors multi-dimensional time-series operational data of the battery. The system can promptly identify abnormal trends such as sudden increases in internal resistance or capacity drops. The cloud can dynamically generate updated policy configuration files containing adjusted safety thresholds and distribute them to the terminal. This mechanism achieves adaptive updates to the protection strategy. It avoids the false protection or protection lag problems caused by traditional fixed thresholds.

[0019] Finally, this invention enhances the safety and reliability of the battery system. The embedded battery management control unit performs local real-time early warnings. When an abnormal voltage drop rate or a cell voltage below the dynamic hard cutoff voltage threshold is detected, the system immediately disconnects the discharge circuit. This cloud-edge collaborative protection mechanism ensures safe operation of the battery throughout its entire lifespan. This invention effectively extends the battery pack's lifespan and reduces maintenance costs. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of a battery life cycle status monitoring and power estimation system according to an embodiment of the present invention; Figure 2 This is a flowchart of a battery life cycle status monitoring and power estimation method in one embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In one embodiment, such as Figure 1 As shown, a battery life cycle status monitoring and power estimation system is provided. This system corresponds one-to-one with the battery life cycle status monitoring and power estimation methods in the above embodiments. The battery life cycle status monitoring and power estimation system includes a cloud big data platform and an intelligent battery terminal. The functional modules are described in detail below: The cloud-based big data platform receives and stores the battery identification and multi-dimensional time-series operating data uploaded by the smart battery terminal. It uses a system identification algorithm to fit a personalized equivalent circuit model parameter table, and a machine learning regression algorithm to predict the maximum usable battery capacity. It then generates a strategy configuration file containing the parameter table and the capacity and distributes it. The smart battery terminal includes a data acquisition module, a communication module, and an embedded battery management main control unit. The data acquisition module synchronously acquires battery pack voltage, current, and temperature signals. The communication module receives the strategy configuration file. The embedded battery management main control unit interpolates the ohmic internal resistance, polarization resistance, and polarization capacitance values ​​from the parameter table based on the real-time state of charge estimate and the measured temperature value. Based on the maximum usable battery capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and a preset noise covariance matrix, it constructs an extended Kalman filter state observer and performs multi-source fusion processing on the acquired signals to output a real-time state of charge estimate.

[0023] The cloud-based big data platform first establishes communication connections with multiple smart battery terminals. The platform receives and stores the unique battery identifier uploaded by each terminal. Simultaneously, the platform receives and stores the corresponding multi-dimensional time-series operational data. This multi-dimensional time-series operational data includes real-time voltage, real-time current, real-time temperature, and cumulative charge / discharge capacity. The data is stored in a distributed database in timestamp order to form a historical dataset.

[0024] The platform invokes the system's identification algorithm for each battery identifier. The algorithm extracts parameters based on dynamic operating condition segments from historical datasets. It calculates the equivalent circuit model parameters using either the least squares method or recursive least squares. These parameters include ohmic internal resistances for different states of charge and temperature ranges, as well as polarization resistance and capacitance. The platform then fits the calculation results into a personalized equivalent circuit model parameter table. This parameter table covers the aging characteristics throughout the battery's entire lifespan.

[0025] The platform also utilizes machine learning regression algorithms to process historical data. The algorithm's input features include the number of iterations, average operating temperature, depth of discharge distribution, and voltage decay curve. The algorithm outputs a predicted maximum usable battery capacity at the current moment. The regression model, trained on a massive dataset, is capable of identifying capacity decay trends.

[0026] The platform packages the generated personalized equivalent circuit model parameter table and the predicted maximum usable battery capacity. The platform then generates a policy configuration file containing this data. This file is encrypted and transmitted wirelessly to the corresponding smart battery terminal. Upon receiving the file, the terminal parses it and updates its locally stored parameter mapping table and security thresholds.

[0027] This embodiment leverages the powerful computing resources of the cloud to perform complex system identification and machine learning tasks, avoiding the need to run high-load algorithms on resource-constrained embedded terminals. Personalized parameter tables accurately reflect the actual aging state of individual battery cells. Machine learning prediction eliminates estimation errors caused by fixed capacity nominal values. Dynamically distributed policy configuration files enable adaptive adjustments to the protection mechanism. This method significantly improves the accuracy of power estimation and extends the battery pack's lifespan.

[0028] The intelligent battery terminal, serving as the system's local execution unit, integrates a data acquisition module, a communication module, and an embedded battery management control unit. The data acquisition module synchronously acquires the battery pack's total voltage, loop current, and temperature signals at key measurement points via a high-precision analog-to-digital converter. The acquisition process employs multi-channel synchronous sampling technology to ensure strict alignment of voltage, current, and temperature data on the time axis, providing an accurate input source for subsequent state estimation.

[0029] The communication module is responsible for establishing a secure connection with the cloud-based big data platform. This module receives the policy configuration file from the cloud and performs integrity verification. After successful verification, the communication module writes the personalized equivalent circuit model parameter table and the predicted maximum usable battery capacity from the file into the non-volatile memory of the embedded battery management main control unit. This process enables dynamic updates of local parameters, adapting to changes in battery characteristics as it ages without requiring hardware replacement.

[0030] The embedded battery management main control unit is the core processing component of the terminal. The main control unit first reads the current real-time estimated state of charge (SOC) value and the measured temperature value. Then, it performs multi-dimensional interpolation calculations in the loaded personalized equivalent circuit model parameter table. Based on the current SOC and temperature ranges, the interpolation process accurately extracts the corresponding ohmic internal resistance, polarization resistance, and polarization capacitance values. This lookup-based interpolation method avoids the high computational cost of online identification, ensuring real-time performance.

[0031] The main control unit constructs an extended Kalman filter state observer based on the acquired maximum usable battery capacity, ohmic internal resistance, polarization resistance, polarization capacitance, and a preset noise covariance matrix. The preset noise covariance matrix contains the statistical characteristics of system process noise and measurement noise, and is used to optimize the filter gain. The observer uses a coarse state of charge obtained by the ampere-hour integration method as a priori estimate and the acquired voltage signal as the observation correction. Through iterative prediction and update steps, the observer performs multi-source fusion processing on the acquired signal. This process effectively eliminates the accumulated error caused by current sensor drift and suppresses voltage fluctuation interference caused by load abrupt changes, ultimately outputting a high-precision real-time state of charge estimate.

[0032] This embodiment significantly reduces the computational load on the embedded battery management main control unit by offloading the complex parameter identification task to the cloud. Only efficient lookup table interpolation and filtering calculations need to be performed locally, enabling high-precision power estimation even with a low-cost microcontroller. Dynamically updated equivalent circuit model parameters can track the battery aging trajectory in real time, solving the problem of large estimation errors in the later stages of the battery's lifespan caused by traditional fixed-parameter models. The introduction of an extended Kalman filter state observer enhances the system's robustness to noise and nonlinear factors, ensuring the accuracy and stability of power display under various complex operating conditions, thereby improving the overall safety and reliability of the battery system.

[0033] The personalized equivalent circuit model parameter table is a digital mapping of the battery's internal electrochemical characteristics as they change with external operating conditions. This parameter table stores data in a two-dimensional matrix. One dimension of the matrix is ​​the state of charge (SOC), and the other is temperature. Each data point in the matrix contains three key parameters: ohmic internal resistance, polarization resistance, and polarization capacitance. The SOC dimension covers from 0% to 100% in a preset step size. The temperature dimension covers the minimum to maximum allowable operating temperature range of the battery. These data points are pre-calculated and fitted in the cloud using system identification algorithms, accurately reflecting the battery's impedance characteristics at different aging stages.

[0034] The embedded battery management control unit performs parameter acquisition operations in each estimation cycle. The control unit first reads the current real-time state of charge (SOC) estimate and the measured temperature value. Since the real-time operating point usually does not directly fall on the discrete grid nodes of the parameter table, the control unit uses bilinear interpolation for calculation. The interpolation process consists of two steps. The first step, under constant temperature conditions, performs linear interpolation between two adjacent SOC nodes based on the real-time SOC, calculating two intermediate temporary parameter values. The second step, under constant SOC conditions, performs linear interpolation again between two adjacent temperature nodes based on the measured temperature. Through these two consecutive linear calculations, the control unit finally retrieves the precise ohmic internal resistance, polarization resistance, and polarization capacitance values ​​corresponding to the current operating condition.

[0035] This embodiment effectively solves the matching problem between discrete data tables and continuous physical quantities using bilinear interpolation. This method avoids complex real-time curve fitting calculations on the embedded system, significantly reducing the processor load and memory usage of the main control unit. The interpolation algorithm ensures the continuity and smoothness of parameter changes, preventing parameter jumps caused by minor fluctuations in operating conditions, thereby improving the convergence stability of the extended Kalman filter state observer. The equivalent circuit model constructed based on high-precision interpolation parameters can more realistically simulate the voltage response characteristics of the battery under dynamic loads. This refined parameter acquisition method eliminates the systematic errors caused by fixed-parameter models, significantly improving the accuracy of state-of-charge estimation across the entire operating range and extending the effective service life of the battery system.

[0036] The core of the Extended Kalman Filter (EKF) state observer lies in the precise tracking of the battery's internal state. This observer defines two key state variables: the state of charge (SOC) and the polarization voltage. The SOC reflects the current percentage of remaining charge in the battery. The polarization voltage describes the voltage deviation caused by the lag in electrochemical reactions during charging and discharging. These two variables together constitute the state vector describing the battery's dynamic characteristics.

[0037] The preset noise covariance matrix is ​​a key parameter for ensuring filter performance. This matrix is ​​specifically divided into the process noise covariance matrix and the observation noise covariance matrix. The process noise covariance matrix quantifies the uncertainties inherent in the system model itself, such as small drifts in equivalent circuit model parameters or sudden state changes caused by external disturbances. The observation noise covariance matrix characterizes the error statistics of sensor measurement data, including voltage sampling noise and current detection error. The values ​​of these two matrices are determined and distributed during the cloud training phase, enabling the observer to adaptively balance the weights between model predictions and measured data.

[0038] When performing the state prediction step, the embedded battery management main control unit strictly follows the ampere-hour integration principle and the state equations of the equivalent circuit model. The main control unit first uses the state estimate from the previous moment and the current input at the current moment to calculate the predicted state of charge (SOC) and polarization voltage for the current moment. During the calculation of the SOC prediction, the main control unit introduces a dynamic capacity correction mechanism. This mechanism retrieves the corresponding compensation coefficient from a pre-stored compensation coefficient table based on the real-time temperature and real-time discharge rate. The main control unit uses this compensation coefficient to correct the maximum available battery capacity sent from the cloud in real time, obtaining the effective available capacity under the current operating conditions. Subsequently, the main control unit uses the corrected effective available capacity as the denominator, combined with the integral of current over time, to calculate a more accurate SOC prediction.

[0039] This embodiment effectively addresses the issue of dynamically changing battery capacity under varying operating conditions by introducing a capacity correction mechanism based on real-time temperature and discharge rate. Traditional methods typically use nominal capacity or fixed aging capacity for calculation, which can introduce significant errors under low-temperature or high-rate discharge scenarios. This solution uses lookup table compensation to ensure that the denominator of the ampere-hour integral reflects the battery's actual energy storage capacity under the current environment in real time. Combined with refined modeling of process noise and observation noise, the extended Kalman filter state observer can converge quickly under complex dynamic conditions. This method significantly suppresses current accumulation errors, improves the robustness of state-of-charge estimation, and ensures that the battery management system can output stable and reliable charge information under various extreme environments, thereby enhancing the accuracy and safety of the vehicle's range prediction.

[0040] When performing the observation update step, the Extended Kalman Filter (EKF) state observer first constructs the terminal voltage observation equation. This equation is based on the physical characteristics of the battery's equivalent circuit model. The calculation logic of the equation involves subtracting the product of current and ohmic resistance from the open-circuit voltage function, and then further subtracting the polarization voltage prediction value. The open-circuit voltage function describes the nonlinear mapping relationship between the battery's open-circuit voltage and state of charge. The product of current and ohmic resistance represents the ohmic voltage drop inside the battery. The polarization voltage prediction value reflects the voltage loss caused by electrochemical polarization and concentration polarization. Through the above calculations, the observer obtains the theoretical value of the terminal voltage predicted by the model at the current moment.

[0041] The embedded battery management control unit then compares the theoretical value calculated by the terminal voltage observation equation with the measured terminal voltage acquired by the sensor. The difference between the two is the observation error value. This error value reflects the degree of deviation between the model's predicted state and the actual battery state. The control unit dynamically calculates the Kalman gain based on the preset process noise covariance matrix and observation noise covariance matrix. The process noise covariance matrix characterizes the uncertainty of the system model, while the observation noise covariance matrix characterizes the reliability of the measurement data. The Kalman gain, as a weighting coefficient, determines the weight of the observation error's influence on state correction. When the measurement data noise is high, the gain automatically decreases to rely more on the model prediction; when the model uncertainty is high, the gain automatically increases to rely more on the measured data.

[0042] The main control unit uses the calculated Kalman gain to correct the predicted state of charge (SOC) and polarization voltage (PV) values ​​obtained in the previous stage. The correction process involves multiplying the observation error by the Kalman gain and then adding the result to the predicted value. This step corrects the predicted value with the measured data, eliminating accumulated errors and model bias. The corrected SOC prediction is then output as the final real-time SOC estimate to other modules of the battery management system. Simultaneously, the corrected PV value serves as the initial state variable for the next estimation cycle, achieving closed-loop iteration.

[0043] This embodiment establishes a rigorous mathematical connection between the battery's internal, unmeasurable state variables and the externally measurable voltage signal by constructing a precise terminal voltage observation equation. Utilizing a Kalman gain adaptive adjustment mechanism, the system can intelligently balance the weights of model derivations and measured data under different operating conditions. This method effectively suppresses drift errors caused by current integration and quickly converges state estimation deviations caused by sudden load changes. The final real-time state-of-charge estimate has extremely high accuracy and stability, accurately reflecting the battery's remaining charge under complex conditions. This technology significantly improves the intelligence level of battery management, avoids the risk of sudden power outages or overcharging due to inaccurate charge estimation, extends battery pack lifespan, and ensures system operational safety.

[0044] After receiving operational data uploaded by the smart battery terminal, the cloud-based big data platform automatically executes an anomaly detection process. This process aims to monitor sudden changes in the battery's health status in real time. The platform first extracts the ohmic internal resistance value calculated during the current charging or discharging cycle. Then, it retrieves the average ohmic internal resistance value recorded over several past cycles as a benchmark. The platform calculates the deviation between the current ohmic internal resistance value and the historical average ohmic internal resistance value. If this deviation exceeds a preset threshold for sudden internal resistance changes, the platform determines that the battery may have internal faults such as micro-short circuits, loose connections, or electrolyte drying, marking it as an abnormal sudden increase in internal resistance.

[0045] The platform simultaneously monitors the rate of capacity degradation. The system statistically analyzes the average capacity data estimated over the most recent cycles and compares it with the historical average capacity data from previous cycles. The platform calculates the percentage decrease between the two. If this percentage decrease exceeds a preset capacity drop threshold, it indicates that the battery may have experienced a drastic loss of active material or structural collapse within a short period, and the platform classifies it as an abnormal capacity drop. This dual detection mechanism covers both gradual degradation and sudden failure modes during battery aging.

[0046] Once a sudden increase in internal resistance or a drop in capacity is detected, the cloud-based big data platform immediately initiates a dynamic adjustment program for the safety strategy. The platform recalculates the battery's safe operating boundaries based on the severity of the anomaly. These boundaries include the maximum charging voltage limit, the maximum allowable discharge current limit, and the cutoff discharge voltage limit. The platform encapsulates these dynamically adjusted safety thresholds into a new strategy configuration file. This file also contains updated equivalent circuit model parameters to reflect the battery's current characteristics. After generation, the platform distributes the updated strategy configuration file to the corresponding smart battery terminal via the communication network.

[0047] After receiving and applying the new strategy configuration file, the intelligent battery terminal immediately manages the battery's charging and discharging according to the updated safety thresholds. This embodiment leverages cloud-based big data analytics to achieve early identification and rapid response to sudden battery failures. Traditional local management systems often struggle to distinguish between noise interference and genuine faults, leading to missed or false alarms. This solution utilizes historical data trend analysis to significantly improve the accuracy of fault diagnosis. The mechanism of dynamically adjusting safety thresholds ensures safe operation of the battery even in the early stages of performance degradation, avoiding the risks of overcharging or over-discharging due to fixed parameters. This method effectively prevents thermal runaway accidents, extends the battery system's usability in fault conditions, provides a valuable window for vehicle maintenance or replacement, and greatly improves the safety and reliability of the battery throughout its entire lifecycle.

[0048] The embedded battery management main control unit continuously executes a local real-time early warning function during operation. This function, acting as the last line of defense for battery safety management, operates independently of cloud-based strategies. The main control unit collects the total voltage data of the battery pack and the minimum voltage data of individual cells in real time at a high-frequency sampling rate. The system has pre-stored power overload thresholds and dynamic hard cutoff voltage thresholds. The dynamic hard cutoff voltage threshold is derived from an updated strategy configuration file sent from the cloud and can be dynamically adjusted according to battery aging and temperature environment.

[0049] The main control unit calculates the voltage drop slope in real time. The drop slope reflects the rate at which the voltage decreases per unit time. When the real-time voltage drop slope exceeds the preset power overload threshold, it indicates that the battery is currently experiencing a large current surge beyond its physical limits. This situation typically occurs during an external load short circuit or a momentary malfunction of the motor controller. At this time, the internal polarization effect of the battery is severe, and if not intervened in time, it will lead to a sudden voltage collapse and trigger thermal runaway. Upon recognizing this anomaly, the main control unit immediately triggers a forced protection action.

[0050] The main control unit continuously compares the lowest cell voltage with the dynamic hard cutoff voltage threshold. The lowest cell voltage directly reflects the weakest link in the battery pack. When this voltage value falls below the dynamic hard cutoff voltage threshold, it indicates that the cell has entered the over-discharge danger zone. Continued discharge will lead to the dissolution of the copper current collector inside the cell or irreversible capacity loss. Once this condition is detected, the main control unit immediately executes a forced protection action. The forced protection action specifically involves sending a disconnect command to the high-voltage relay or power switch, physically cutting off the discharge circuit.

[0051] This embodiment achieves millisecond-level response to extreme operating conditions through a local real-time early warning mechanism. While cloud-based strategies can optimize long-term operating parameters, they are limited by communication latency and cannot handle sudden electrical faults. This solution pushes key protection logic down to the embedded end, ensuring safety even in the event of communication interruption or delayed cloud commands. By monitoring the voltage drop slope, the system can identify power overload risks before the absolute voltage value reaches the bottom line, achieving preventative protection. The application of a dynamic hard cutoff voltage threshold avoids the false protection or underprotection problems of traditional fixed thresholds in low-temperature or aging scenarios. This dual-criteria mechanism significantly improves the survivability of the battery system under complex electromagnetic environments and extreme loads, effectively preventing permanent battery damage and fire accidents caused by overcurrent or over-discharge, and ensuring the overall safety of the vehicle's electrical system.

[0052] Specific limitations regarding the battery life cycle state monitoring and capacity estimation system can be found in the limitations of the battery life cycle state monitoring and capacity estimation method below, and will not be repeated here. Each module in the aforementioned battery life cycle state monitoring and capacity estimation system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0053] In one embodiment, such as Figure 2 As shown, a method for battery life cycle state monitoring and capacity estimation is provided, which is then applied to... Figure 2 Taking China as an example, the following specific steps will be used: S10: The cloud-based big data platform receives battery identification and multi-dimensional time-series operation data uploaded by the smart battery terminal.

[0054] S20: The cloud-based big data platform uses the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table based on the data, and uses a machine learning regression algorithm to predict the maximum usable capacity of the battery.

[0055] S30: The cloud-based big data platform packages the parameter table and the capacity into a strategy configuration file and sends it to the smart battery terminal.

[0056] S40: The embedded battery management main control unit of the intelligent battery terminal loads the strategy configuration file and the preset noise covariance matrix, and obtains the ohmic internal resistance, polarization resistance and polarization capacitance from the parameter table by interpolation based on the real-time collected voltage, current and temperature signals.

[0057] S50: The embedded battery management main control unit constructs an extended Kalman filter state observer, using the maximum available battery capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and the preset noise covariance matrix as model parameters, and performs multi-source fusion calculation on the collected signals to output a real-time state of charge estimate.

[0058] Specifically, the cloud-based big data platform, as the computing core of the entire battery management system, first receives data packets uploaded from the smart battery terminal. This data packet contains a unique battery identifier and multi-dimensional time-series operational data. The multi-dimensional time-series operational data covers the battery's voltage curve, current waveform, temperature changes, and historical charge / discharge records throughout its entire lifespan. The platform uses the battery identifier to match the newly uploaded data with files in the historical database, constructing a complete operational profile for that specific battery. Based on this massive amount of data, the cloud platform initiates a system identification algorithm. This algorithm uses least squares or recursive identification techniques to fit the battery's dynamic response characteristics. The goal of the fitting is to generate a personalized equivalent circuit model parameter table. This parameter table records detailed values ​​for ohmic internal resistance, polarization resistance, and polarization capacitance at different states of charge, different temperature points, and different degrees of aging. This parameter table accurately reflects the unique electrochemical characteristics of this individual battery, distinguishing it from a general model.

[0059] The cloud-based big data platform simultaneously utilizes machine learning regression algorithms to process historical capacity degradation data. Algorithm inputs include feature variables such as cumulative charge / discharge duration, average operating temperature, and number of deep discharge cycles. Through a trained regression model, the platform predicts the battery's maximum usable capacity at the current moment. This prediction process considers non-linear aging factors and is more accurate than the traditional ampere-hour integration method in assessing the battery's actual energy storage capacity. Subsequently, the platform packages the generated personalized equivalent circuit model parameter table and the predicted maximum usable battery capacity. This data is encapsulated into a standard strategy configuration file. The platform then distributes this strategy configuration file to the corresponding smart battery terminal via a wireless network. This enables cloud-based iteration and local updates of model parameters, ensuring that the terminal always uses the latest and most accurate battery model.

[0060] After receiving the strategy configuration file, the intelligent battery terminal's embedded battery management main control unit immediately performs a loading operation. The main control unit writes the personalized equivalent circuit model parameter table and the battery's maximum usable capacity from the file into non-volatile memory. Simultaneously, the main control unit loads a preset noise covariance matrix. This matrix contains the process noise covariance and observation noise covariance, used to define the confidence weights of the filter. During real-time operation, the main control unit frequently acquires the battery's real-time voltage, real-time current, and real-time temperature signals. Based on the current real-time temperature and the estimated state of charge, the main control unit uses interpolation to look up the corresponding model parameters from the loaded personalized equivalent circuit model parameter table. Through linear or bilinear interpolation calculations, the main control unit obtains the accurate values ​​of the ohmic internal resistance, polarization resistance, and polarization capacitance under the current operating conditions. This method eliminates the step-like errors caused by table lookups, ensuring parameter continuity.

[0061] The embedded battery management control unit constructs an extended Kalman filter (EDF) state observer using acquired parameters. This observer uses the battery's maximum usable capacity, ohmic internal resistance, polarization resistance, polarization capacitance, and a preset noise covariance matrix as core model parameters. The observer establishes state equations and observation equations that include the state of charge (SOC) and polarization voltage. The control unit inputs real-time acquired voltage, current, and temperature signals into the observer for multi-source fusion calculation. The EDF algorithm corrects the SOC estimation error in real-time through prediction and update steps. Finally, the observer outputs a high-precision real-time SOC estimate. This estimate integrates the physical constraints of the mechanistic model with data-driven statistical characteristics, effectively overcoming the limitations of a single method.

[0062] This embodiment achieves personalized and adaptive battery state estimation through the collaborative work of cloud-based big data analysis and local embedded computing. Traditional methods often use fixed parameters or general models, which cannot adapt to individual battery differences and parameter drift throughout the entire battery lifecycle. This solution utilizes the powerful computing power of the cloud for system identification and capacity prediction, generating a dedicated, personalized equivalent circuit model parameter table. The local terminal obtains accurate parameters in real time through interpolation, significantly improving the model's fit. Combined with the multi-source fusion capability of extended Kalman filtering, the system can effectively suppress interference caused by sensor noise and model uncertainty. The effect of this technology is reflected in a significant improvement in the accuracy and robustness of real-time state of charge estimation, especially maintaining stable estimation performance in the later stages of battery aging or under extreme temperature conditions. This not only eliminates users' range anxiety but also prevents overcharging or over-discharging accidents caused by power estimation errors, extending the battery pack's lifespan and improving the overall vehicle safety.

[0063] In one embodiment, step S20, namely the step of using the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table, includes: extracting charge and discharge curve features, calculating the average internal resistance and polarization voltage time constant of a specific state of charge interval, and fitting the mapping relationship between ohmic internal resistance, polarization resistance and polarization capacitance as a function of state of charge and temperature using the recursive least squares method.

[0064] Specifically, when executing the system identification algorithm, the cloud-based big data platform first preprocesses the historical operating data uploaded by the smart battery terminal. The platform extracts complete charge-discharge curve features from massive amounts of multi-dimensional time-series operating data. The extraction process includes identifying current step change points, voltage relaxation stages, and steady-state charge-discharge intervals. The platform divides the continuous charge-discharge process into several specific state-of-charge intervals. Within each interval, the platform calculates the average internal resistance value for that time period. The average internal resistance is calculated based on the ratio of the instantaneous voltage jump to the corresponding current change. Simultaneously, the platform analyzes the voltage relaxation recovery process after current cutoff and calculates the time constant of the polarization voltage. The time constant reflects the rate of elimination of electrochemical polarization and concentration polarization within the battery and is a key indicator characterizing dynamic properties.

[0065] The platform then uses recursive least squares to mathematically fit the extracted feature data. Recursive least squares is a recursive parameter estimation method that can continuously update model parameters as new data is added, without reprocessing all historical data. The algorithm uses state of charge and temperature as independent variables, and ohmic internal resistance, polarization resistance, and polarization capacitance as dependent variables. Through iterative calculation, the algorithm finds the parameter combination that minimizes the sum of squared errors in the model output. The fitting process constructs a three-dimensional mapping relationship between ohmic internal resistance and state of charge and temperature. Similarly, the algorithm also generates mapping relationships between polarization resistance and state of charge and temperature, as well as between polarization capacitance and state of charge and temperature. These mapping relationships are finally compiled into a high-precision, personalized equivalent circuit model parameter table. This parameter table covers the entire operating temperature range and the entire state of charge range of the battery, and can accurately describe the dynamic response characteristics of the battery under different operating conditions.

[0066] This embodiment achieves precise quantification of the complex electrochemical characteristics inside the battery by extracting features from the charge-discharge curves and combining them with the recursive least squares method. Traditional methods often use fixed parameters or simple linear interpolation, which cannot accurately reflect the nonlinear laws of battery parameters changing with operating conditions. This scheme utilizes the recursive least squares method to effectively eliminate measurement noise interference and extract the true physical parameters from actual operating data. The generated personalized equivalent circuit model parameter table fully considers the influence of individual battery differences and aging levels. This technology significantly improves the fitting accuracy of the equivalent circuit model, enabling the state observer based on the model to more accurately simulate the battery's true voltage response. This lays a solid data foundation for subsequent high-precision estimation of real-time state of charge, effectively solves the estimation divergence problem caused by model parameter mismatch, and improves the adaptability and estimation reliability of the battery management system under complex dynamic operating conditions.

[0067] In one embodiment, step S50, namely the step of multi-source fusion calculation of the acquired signal, includes: calculating the predicted state of charge based on the ampere-hour integral principle, calculating the predicted polarization voltage based on the equivalent circuit model, constructing a terminal voltage observation equation that includes the product of open-circuit voltage, current and ohmic internal resistance and the predicted polarization voltage, correcting the observation equation using the measured terminal voltage to eliminate the instantaneous voltage drop interference caused by the large current load, and outputting the corrected real-time state of charge estimate.

[0068] When performing multi-source fusion calculations, the embedded battery management main control unit first calculates the predicted state of charge (SOC) based on the ampere-hour integration principle. The system reads the previous real-time SOC estimate and performs accumulation or subtraction operations based on the currently acquired real-time current signal and time interval. This step provides a trend prediction of SOC changes, reflecting the accumulation or consumption of charge within the battery. However, simple ampere-hour integration is susceptible to the effects of current sensor zero drift and accumulated errors; therefore, it needs to be corrected by incorporating model observations.

[0069] The main control unit simultaneously calculates the predicted polarization voltage based on a personalized equivalent circuit model. The system uses the state variables from the previous moment and the current real-time current signal to iteratively calculate the current polarization voltage component through the state equation. This component characterizes the voltage deviation caused by the hysteresis effect of the internal electrochemical reaction of the battery. Subsequently, the main control unit constructs the terminal voltage observation equation. This equation consists of four key physical quantities: open-circuit voltage, the product of real-time current and ohmic internal resistance, the predicted polarization voltage, and the model residual. The open-circuit voltage is obtained from the mapping relationship based on the current predicted state of charge, while the ohmic internal resistance is obtained from the parameter table of the personalized equivalent circuit model. The terminal voltage observation equation comprehensively describes the external electrical characteristics of the battery under dynamic operating conditions.

[0070] The system uses measured terminal voltage to correct the aforementioned observation equations in real time. The main control unit compares the actual terminal voltage collected by high-precision sensors with the theoretical terminal voltage calculated by the observation equations to obtain the voltage residual. This residual reflects the deviation between the model prediction and the actual physical state, including transient voltage drop interference caused by high current loads and uncertainties in model parameters. The extended Kalman filter algorithm calculates the Kalman gain based on a preset noise covariance matrix and uses this gain to weight the voltage residual. The algorithm feeds the weighted residual back to the state vector, simultaneously correcting the predicted state of charge and polarization voltage. This process effectively eliminates the misleading effect of sudden ohmic voltage drop caused by high current surges on charge estimation and removes transient interference from state estimation.

[0071] After correction, the main control unit outputs the corrected real-time state of charge (SOC) estimate. This embodiment achieves complementary advantages between the ampere-hour integration method and the equivalent circuit model method through a multi-source fusion solution mechanism. The ampere-hour integration method provides smooth tracking capability in the short term, while the equivalent circuit model observer provides long-term absolute reference correction. An observation equation incorporating ohmic internal resistance voltage drop and polarization voltage is constructed, enabling the system to accurately distinguish between the battery's steady-state open-circuit voltage and dynamic load voltage drop. This technique significantly improves estimation accuracy under complex operating conditions, especially in scenarios with large current fluctuations such as rapid acceleration or deceleration of the vehicle, effectively suppressing the problem of charge estimation jumps caused by instantaneous voltage jumps. This method ensures the continuity and stability of the real-time SOC estimate, avoids false low-charge alarms or overcharge risks caused by instantaneous interference, and improves the battery management system's response capability to dynamic loads and overall control accuracy.

[0072] Optionally, the battery life cycle status monitoring and power estimation method further includes a predictive maintenance step: the cloud big data platform continuously analyzes the uploaded multi-dimensional time-series operating data, and when it detects an ohmic internal resistance growth trend or a capacity decay rate that meets preset abnormal judgment conditions, it generates a health status assessment report and a remaining service life prediction result, and sends an early warning message to the operation and maintenance platform or issues an updated safety protection threshold to the smart battery terminal.

[0073] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0074] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A battery life cycle state monitoring and capacity estimation system, characterized in that, include: Cloud-based big data platform and intelligent battery terminal; The cloud-based big data platform receives and stores the battery identification and multi-dimensional time-series operation data uploaded by the smart battery terminal, uses a system identification algorithm to fit a personalized equivalent circuit model parameter table, uses a machine learning regression algorithm to predict the maximum usable capacity of the battery, generates a strategy configuration file containing the parameter table and the capacity, and distributes it. The intelligent battery terminal includes a data acquisition module, a communication module, and an embedded battery management main control unit. The data acquisition module synchronously acquires battery pack voltage, current, and temperature signals. The communication module receives the strategy configuration file. The embedded battery management main control unit interpolates the ohmic internal resistance, polarization resistance, and polarization capacitance values ​​from the parameter table based on the real-time state of charge estimate and the measured temperature value. Based on the battery's maximum usable capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and a preset noise covariance matrix, it constructs an extended Kalman filter state observer and performs multi-source fusion processing on the acquired signals to output a real-time state of charge estimate.

2. The battery life cycle state monitoring and capacity estimation system according to claim 1, characterized in that, The personalized equivalent circuit model parameter table is a lookup table for the changes in ohmic internal resistance, polarization resistance, and polarization capacitance with respect to state of charge and temperature. The embedded battery management main control unit uses bilinear interpolation to retrieve the ohmic internal resistance, polarization resistance, and polarization capacitance values ​​corresponding to the current operating condition from the parameter table in each estimation cycle.

3. The battery life cycle state monitoring and capacity estimation system according to claim 1, characterized in that, The state variables of the extended Kalman filter state observer include the state of charge and polarization voltage; the preset noise covariance matrix includes the process noise covariance matrix and the observation noise covariance matrix; when the embedded battery management main control unit executes the state prediction step, it calculates the predicted value of the state of charge and the predicted value of the polarization voltage based on the state equation of the ampere-hour integral principle and the equivalent circuit model, wherein the calculation of the predicted value of the state of charge introduces a compensation coefficient obtained by looking up a table based on real-time temperature and discharge rate to correct the maximum usable capacity of the battery.

4. The battery life cycle state monitoring and capacity estimation system according to claim 1, characterized in that, When the extended Kalman filter state observer performs the observation update step, it constructs the terminal voltage observation equation, which is obtained by subtracting the product of the current and the ohmic internal resistance from the open-circuit voltage function and then subtracting the polarization voltage prediction value. The embedded battery management main control unit compares the calculation result of the terminal voltage observation equation with the measured terminal voltage to calculate the observation error value, and calculates the Kalman gain based on the process noise covariance matrix and the observation noise covariance matrix, thereby correcting the state of charge prediction value and the polarization voltage prediction value to obtain the final real-time state of charge estimate value.

5. The battery life cycle state monitoring and capacity estimation system according to claim 1, characterized in that, The cloud-based big data platform also performs an anomaly detection process. By calculating the deviation between the current cycle's internal resistance value and the average internal resistance value of the past few cycles, or by comparing the decrease ratio of the average capacity of the most recent few cycles with the average capacity of the previous few cycles, when the deviation or decrease ratio exceeds a preset threshold, it determines that the battery has experienced a sudden increase in internal resistance or a drop in capacity, and generates an update strategy configuration file containing dynamically adjusted safety thresholds and sends it to the smart battery terminal.

6. The battery life cycle state monitoring and capacity estimation system according to claim 5, characterized in that, The embedded battery management main control unit also performs local real-time early warning. When it detects that the real-time voltage drop slope exceeds the preset power overload threshold, or the lowest cell voltage is lower than the dynamic hard cutoff voltage threshold issued in the updated strategy configuration file, it immediately performs forced protection action and cuts off the discharge circuit.

7. A method for monitoring battery lifecycle status and estimating battery capacity, comprising: The cloud-based big data platform receives battery identification data and multi-dimensional time-series operational data uploaded by the smart battery terminal; The cloud-based big data platform uses the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table based on the data, and uses a machine learning regression algorithm to predict the maximum usable capacity of the battery. The cloud-based big data platform packages the parameter table and the capacity into a strategy configuration file and sends it to the smart battery terminal. The embedded battery management main control unit of the intelligent battery terminal loads the strategy configuration file and the preset noise covariance matrix, and obtains the ohmic internal resistance, polarization resistance and polarization capacitance from the parameter table by interpolation based on the real-time collected voltage, current and temperature signals. The embedded battery management main control unit constructs an extended Kalman filter state observer, using the battery's maximum available capacity, the ohmic internal resistance, the polarization resistance, the polarization capacitance, and the preset noise covariance matrix as model parameters, and performs multi-source fusion calculation on the collected signals to output a real-time state of charge estimate.

8. The battery life cycle state monitoring and capacity estimation method according to claim 7, characterized in that, The steps of using the system identification algorithm to fit and generate a personalized equivalent circuit model parameter table include: extracting the features of the charge and discharge curves, calculating the average internal resistance and polarization voltage time constant of a specific state of charge interval, and fitting the mapping relationship between the ohmic internal resistance, polarization resistance and polarization capacitance and the state of charge and temperature using the recursive least squares method.

9. The battery life cycle state monitoring and capacity estimation method according to claim 7, characterized in that, The steps for multi-source fusion and solution of the acquired signals include: calculating the predicted state of charge based on the ampere-hour integral principle, calculating the predicted polarization voltage based on the equivalent circuit model, constructing a terminal voltage observation equation that includes the product of open-circuit voltage, current and ohmic internal resistance and the predicted polarization voltage, correcting the observation equation using the measured terminal voltage to eliminate the instantaneous voltage drop interference caused by the large current load, and outputting the corrected real-time state of charge estimate.

10. The battery life cycle state monitoring and capacity estimation method according to claim 7, characterized in that, The process of performing multi-source fusion calculations on the collected signals to output real-time state of charge estimates also includes: the cloud-based big data platform continuously analyzing the uploaded multi-dimensional time-series operating data; when the ohmic internal resistance growth trend or capacity decay rate is detected to meet the preset abnormal judgment conditions, a health status assessment report and remaining service life prediction results are generated, and early warning information is sent to the operation and maintenance platform or updated safety protection thresholds are issued to the smart battery terminal.

Images