A new energy vehicle battery management method and system

Abstract

The present application relates to the technical field of power battery detection and control, and particularly relates to a new energy automobile battery management method and system thereof. The method comprises: synchronously collecting voltage, current and temperature data of the battery monomer, and performing frequency scanning to obtain electrochemical impedance spectrum; extracting impedance parameters and calculating the change rate based on an equivalent circuit model; adjusting the noise covariance matrix by using a double-time scale adaptive Kalman filtering algorithm to realize joint estimation of the state of charge and the state of health; determining the target state of charge range of the battery by a multi-objective optimization algorithm based on the current state of charge, road power demand and engine fuel consumption rate; adjusting the operation of the liquid cooling system by using a temperature threshold strategy to maintain the battery temperature in a safe range; and predicting battery attenuation by using a random forest based on historical charging and discharging data and temperature information and issuing an alarm. The method can realize accurate estimation and dynamic management of the battery state, and improve the energy utilization efficiency and operation safety under hybrid working conditions.

Description

Technical Field

[0001] This invention relates to the field of power battery testing and control technology, and in particular to a method and system for managing batteries in new energy vehicles. Background Technology

[0002] With the rapid development of the new energy vehicle industry, the performance and lifespan of the power battery, as a core energy device, directly affect the vehicle's range, safety, and economy. To ensure the efficient and safe operation of the battery, the battery management system (BMS) has become an important component of new energy vehicles. The main functions of a BMS include battery state monitoring, state of charge estimation, health assessment, thermal management, and safety protection.

[0003] In hybrid electric vehicles (HEVs) and plug-in hybrid electric vehicles (PHEVs), the battery and engine work together, resulting in complex energy flow and frequent battery charging and discharging cycles. This significantly increases the difficulty of estimating the battery's State of Charge (SOC) and State of Health (SOH). Traditional SOC estimation methods based on open-circuit voltage or coulomb integration are susceptible to noise accumulation and model errors, leading to insufficient estimation accuracy. Meanwhile, SOH estimation typically relies on long-term historical data, failing to reflect real-time battery performance changes and thus struggling to meet the dynamic response requirements under hybrid operating conditions. Summary of the Invention

[0004] To overcome the above deficiencies, this invention provides a battery management method and system for new energy vehicles, aiming to improve the accuracy of battery status assessment and operational safety, extend battery life, and optimize the energy utilization efficiency of the entire vehicle.

[0005] In a first aspect, the present invention provides the following technical solution: a method for managing batteries in new energy vehicles, comprising:

[0006] Simultaneous acquisition of data from individual battery cells, including voltage, current, and temperature data, and frequency scanning is performed during the acquisition process to obtain electrochemical impedance spectroscopy data;

[0007] The electrochemical impedance spectroscopy data is received, and the ohmic impedance, charge transfer impedance, and diffusion impedance parameters are extracted based on the equivalent circuit model. The rate of change of the impedance characteristic parameters is then calculated.

[0008] A dual-time-scale adaptive Kalman filter algorithm is adopted, and the noise covariance matrix in the algorithm is dynamically adjusted according to the rate of change of impedance characteristic parameters to achieve joint estimation of battery state of charge and health.

[0009] Based on the current state of charge, predicted road power demand, and engine fuel consumption rate, the target state of charge range of the battery is dynamically determined through a multi-objective optimization algorithm.

[0010] Based on the temperature distribution of individual battery cells, the working state of the liquid cooling system is adjusted through a temperature threshold control strategy to maintain the battery temperature within a preset safe range.

[0011] Based on historical battery charge and discharge data and temperature information, a local empirical degradation model is established, and an alarm is issued when the state of charge exceeds a preset threshold to assist in battery status monitoring and lifespan management.

[0012] Preferably, the frequency range of the frequency scan is from 0.01Hz to 10kHz, the number of scan points is not less than 50, and the single scan time is controlled within 30 seconds.

[0013] Preferably, the step of calculating the rate of change of the impedance characteristic parameter includes:

[0014] Frequency domain interpolation and noise filtering were performed on the electrochemical impedance spectroscopy data to obtain a smooth impedance curve.

[0015] At a preset frequency point, the impedance spectrum is fitted based on the equivalent circuit model, and the corresponding ohmic impedance, charge transfer impedance and diffusion impedance parameters are extracted.

[0016] The extracted impedance parameters are compared with the initial parameters under the reference state, and the relative changes of each impedance parameter are calculated.

[0017] The rate of change of the impedance characteristic parameters is calculated based on the relative change value.

[0018] Preferably, the joint estimation step of battery state of charge and state of health includes:

[0019] It receives input data including the rate of change of voltage, current, temperature, and impedance characteristic parameters;

[0020] Within a short timescale, rapid state updates are performed based on real-time voltage and current data to obtain an initial estimate of the battery's state of charge.

[0021] Over a long time scale, the battery health status estimate is corrected based on the rate of change of impedance characteristic parameters and historical operating data.

[0022] The noise covariance matrix in the filtering algorithm is dynamically adjusted based on the battery's operating conditions. The results of short-term state of charge estimation and long-term health estimation are fused to output a joint state estimate of the battery.

[0023] Preferably, the step of dynamically determining the target state of charge range of the battery includes:

[0024] Obtain current battery state of charge, vehicle operating conditions, and road condition prediction information;

[0025] Based on the vehicle power demand prediction model, the predicted road condition power demand for future periods is calculated.

[0026] Determine the system energy efficiency under different states of charge by combining engine fuel consumption rate;

[0027] Construct a multi-objective optimization model with the objectives of minimizing fuel economy, power responsiveness, and battery life degradation;

[0028] The target state of charge range of the battery is dynamically determined based on the optimization results.

[0029] Preferably, the multi-objective optimization model is solved using Pareto front search, and the solution steps include:

[0030] S401: Initialize the optimization population, setting the population size, number of iterations, and multi-objective weight parameters;

[0031] S402: Encode the SOC target range parameters for each individual and perform fitness evaluation based on objective functions of fuel economy, battery life degradation, power responsiveness, and thermal safety;

[0032] S403: Use a non-dominated ordination method to classify individuals into ranks and calculate crowding distance to maintain population diversity;

[0033] S404: Perform selection, crossover, and mutation operations based on individual rank and crowding level to generate a new generation of candidate solutions;

[0034] S405: Repeat steps S402 to S404 until the iteration termination condition is met, and obtain the Pareto optimal solution set that satisfies the multi-objective constraints.

[0035] Preferably, the temperature threshold control strategy is implemented using a PID controller, and the control steps include:

[0036] Measure the temperature of individual battery cells and calculate the temperature deviation;

[0037] The flow rate of the liquid cooling system is adjusted proportionally to directly respond to temperature deviations.

[0038] The temperature deviation is accumulated by the integral element to eliminate steady-state error;

[0039] The cooling flow rate is adjusted according to the rate of temperature change using a differential element to suppress excessively rapid temperature changes;

[0040] The final control flow rate of the liquid cooling system is determined by combining the output signals of the proportional, integral, and derivative components.

[0041] Preferably, the steps for establishing a local empirical decay model include:

[0042] Collect historical battery charge / discharge data, current, voltage, and temperature information;

[0043] A random forest model is trained based on the historical data to predict battery capacity degradation and health status indicators.

[0044] During each charge-discharge cycle, the current battery state data is input into the random forest model, and the predicted capacity decay and health status are output.

[0045] When the predicted capacity or health status indicators exceed the preset threshold, an alarm signal is triggered to assist in battery status monitoring and lifespan management.

[0046] Secondly, the present invention provides the following technical solution: a new energy vehicle battery management system, the system comprising:

[0047] The data acquisition module is used to synchronously acquire data from individual battery cells, including voltage, current, and temperature data, and to perform frequency scanning during the acquisition process to obtain electrochemical impedance spectroscopy data.

[0048] The impedance feature extraction module is used to receive the electrochemical impedance spectroscopy data, extract ohmic impedance, charge transfer impedance and diffusion impedance parameters based on the equivalent circuit model, and calculate the rate of change of the impedance feature parameters.

[0049] The battery state assessment module is used to employ a dual-time-scale adaptive Kalman filter algorithm to dynamically adjust the noise covariance matrix in the algorithm based on the rate of change of impedance characteristic parameters, thereby achieving a joint estimation of the battery's state of charge and health.

[0050] The state of charge range optimization module is used to dynamically determine the target state of charge range of the battery based on the current state of charge, predicted road power demand, and engine fuel consumption rate through a multi-objective optimization algorithm.

[0051] The liquid cooling system control module is used to adjust the working state of the liquid cooling system according to the temperature distribution of the individual battery cells and through a temperature threshold control strategy, so that the battery temperature is maintained within a preset safe range.

[0052] The battery status warning module is used to establish a local empirical degradation model based on the battery's historical charge and discharge data and temperature information, and to issue an alarm when the state of charge exceeds a preset threshold, so as to assist in battery status monitoring and lifespan management.

[0053] The present invention has the following beneficial effects:

[0054] 1. This invention acquires battery electrochemical impedance spectroscopy data and extracts impedance characteristic parameters using an equivalent circuit model. It then employs a dual-time-scale adaptive Kalman filter algorithm to jointly estimate the battery's state of charge (SOC) and state of health (SOH). Compared to traditional single-method estimation based on voltage and current, this approach dynamically adjusts the filter noise covariance matrix, fully integrating short-term operating characteristics and long-term degradation information. This significantly improves the accuracy and stability of SOC and SOH estimations, providing more reliable state support for energy distribution and control in hybrid power systems.

[0055] 2. This invention comprehensively considers battery state of charge (SOC), road power requirements, and engine fuel consumption rate to construct a multi-objective optimization model to dynamically determine the target SOC range. Simultaneously, it incorporates a PID temperature control strategy to maintain battery temperature within a safe range and uses a random forest model to predict battery degradation trends. This solution effectively slows down battery aging while meeting fuel economy and power responsiveness requirements, improving overall system energy efficiency and safety. It is suitable for intelligent battery management and lifespan optimization under complex hybrid operating conditions. Attached Figure Description

[0056] Figure 1 This is a flowchart of a new energy vehicle battery management method proposed in this invention;

[0057] Figure 2 This is a structural diagram of a new energy vehicle battery management system proposed in this invention. Detailed Implementation

[0058] The technical solutions in 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.

[0059] Example 1

[0060] In a first embodiment of the present invention, the present invention provides a method for managing batteries in new energy vehicles, such as... Figure 1 As shown, it includes the following steps:

[0061] S100: Simultaneously acquires data from individual battery cells, including voltage, current, and temperature data, and performs frequency scanning during the acquisition process to obtain electrochemical impedance spectroscopy data;

[0062] Preferably, the frequency range of the frequency scan is from 0.01Hz to 10kHz, the number of scan points is not less than 50, and the single scan time is controlled within 30 seconds.

[0063] Specifically, a high-precision voltage sensor is used, with a sampling frequency of at least 1 kHz, and the measurement range covers the operating voltage range of the battery cells. The error range of voltage data acquisition is controlled within ±0.5%. The current sensor employs the Hall effect or shunt resistor method to provide high-precision current measurement. The sampling frequency is 1 kHz, and the measurement range covers the maximum value of the battery charging and discharging current. The error range is controlled within ±1%. The temperature sensor uses a thermocouple or NTC thermistor to monitor the surface temperature of the battery cells in real time. The sensor accuracy is ±0.5℃, and the sampling frequency is 1 Hz.

[0064] Electrochemical impedance spectroscopy (EIS) involves scanning between the battery electrodes using a small-amplitude alternating current signal. To ensure measurement accuracy, the scan frequency range is set from 0.01 Hz to 10 kHz, which covers the impedance characteristics of various internal physical processes within the battery. During each scan, the system applies an alternating current signal of known amplitude while simultaneously monitoring the battery's voltage response in real time using a voltage sensor. The complex impedance (including real and imaginary parts) at each frequency point is calculated, yielding the battery's EIS. To avoid excessively long response times, a single scan is typically limited to within 30 seconds to ensure the battery management system can acquire impedance data in real time.

[0065] S200: Receive the electrochemical impedance spectroscopy data, extract ohmic impedance, charge transfer impedance and diffusion impedance parameters based on the equivalent circuit model, and calculate the rate of change of the impedance characteristic parameters;

[0066] Preferably, the step of calculating the rate of change of the impedance characteristic parameter includes:

[0067] Frequency domain interpolation and noise filtering were performed on the electrochemical impedance spectroscopy data to obtain a smooth impedance curve.

[0068] At a preset frequency point, the impedance spectrum is fitted based on the equivalent circuit model, and the corresponding ohmic impedance, charge transfer impedance and diffusion impedance parameters are extracted.

[0069] The extracted impedance parameters are compared with the initial parameters under the reference state, and the relative changes of each impedance parameter are calculated.

[0070] The rate of change of the impedance characteristic parameters is calculated based on the relative change value.

[0071] Specifically, electrochemical impedance spectroscopy (EIS) data is often affected by measurement noise and data sparsity. To improve data accuracy, frequency domain interpolation is first performed on the impedance data. Cubic interpolation or spline interpolation is used to supplement the data between frequency points to ensure a smooth transition between frequency scan points. During interpolation, sufficiently dense interpolation points are selected, such as at least 10 points per frequency interval, to ensure the smoothness and accuracy of the impedance spectrum. Since noise may be introduced during sampling, noise filtering is required for the raw impedance data to ensure signal smoothness and accuracy. Kalman filtering or low-pass filtering is used to smooth high-frequency noise. Specifically, the cutoff frequency of the filter is set to remove high-frequency signal fluctuations caused by measurement noise while retaining the actual electrochemical response signal. The filtering process should ensure that the processed impedance data is as close as possible to the actual battery response data. The impedance data after interpolation and filtering will form a smooth impedance curve, which can effectively reveal the electrochemical characteristics of the battery.

[0072] A commonly used equivalent circuit model, such as the Randles model, is selected, which includes components such as series resistance, parallel charge transfer capacitance, charge transfer resistance, and diffusion resistance. This equivalent circuit model can accurately reflect the electrochemical behavior inside the battery and can fit different electrochemical characteristics in the impedance spectrum. The smoothed impedance data is fitted within a preset frequency range. Specifically, multiple frequency points, such as 0.01Hz, 1Hz, 100Hz, and 10kHz, are selected, and the least squares method is used for fitting at these frequency points to minimize the error between the model output and the actual impedance data. During the fitting process, the parameters based on the equivalent circuit model are gradually adjusted until the fitting error is minimized. The corresponding impedance parameters, such as ohmic impedance, charge transfer impedance, and diffusion impedance, are obtained through fitting calculations.

[0073] When the battery is in its initial factory state, record the initial ohmic impedance, charge transfer impedance, and diffusion impedance parameters. These initial values ​​serve as the reference values ​​for subsequent comparisons. Record the impedance parameters of the battery in its current operating state and compare them with the initial impedance parameters in the reference state, calculating the relative change in each impedance parameter.

[0074] By calculating the ratio between the relative change value and the battery usage time, the rate of change of the impedance characteristic parameter can be obtained.

[0075] By interpolating and filtering the data, fitting based on the equivalent circuit model, and comparing with the benchmark value, the impedance characteristic change rate of the battery can be accurately extracted and calculated, providing a reliable basis for battery health status monitoring and prediction.

[0076] S300: Employs a dual-time-scale adaptive Kalman filter algorithm, dynamically adjusting the noise covariance matrix in the algorithm based on the rate of change of impedance characteristic parameters, to achieve joint estimation of battery state of charge and health status.

[0077] Preferably, the joint estimation step of battery state of charge and state of health includes:

[0078] It receives input data including the rate of change of voltage, current, temperature, and impedance characteristic parameters;

[0079] Within a short timescale, rapid state updates are performed based on real-time voltage and current data to obtain an initial estimate of the battery's state of charge.

[0080] Over a long time scale, the battery health status estimate is corrected based on the rate of change of impedance characteristic parameters and historical operating data.

[0081] The noise covariance matrix in the filtering algorithm is dynamically adjusted based on the battery's operating conditions. The results of short-term state of charge estimation and long-term health estimation are fused to output a joint state estimate of the battery.

[0082] Specifically, the input data mainly includes the battery's real-time voltage, current, temperature, and rate of change of impedance characteristic parameters. The battery's voltage, current, and temperature data are obtained by measuring the battery's terminal voltage and current, and by monitoring temperature information through embedded sensors; while the rate of change of impedance characteristic parameters is calculated through electrochemical impedance spectroscopy and impedance analysis.

[0083] Short timescales typically refer to the state changes of a battery within 1-5 seconds, used for rapid updates to the battery's state of charge (SOC). Short timescale estimations focus on changes in real-time voltage and current data, reflecting the battery's dynamic state under instantaneous operating conditions. The SOC is estimated based on the integration of current data over time. The specific calculation method is as follows:

[0084] ;

[0085] in, For real-time current, For the battery's rated capacity, This represents the state of charge (SOC) at the previous moment. If the SOC of the battery is complex, or if the current and voltage of the battery change drastically, the SOC can be updated in real time using extended Kalman filtering or particle filtering. These methods can quickly correct the SOC by modeling and estimating voltage and current changes, avoiding the error accumulation in simple integration methods.

[0086] Longer timescales typically refer to battery state changes over minutes to hours, primarily used to correct battery health. At this scale, the focus is on battery aging, changes in impedance characteristics, and historical operating data. Battery health is estimated by monitoring the rate of change of its impedance characteristic parameters. For example, increased battery impedance indicates aging and a decline in health. The battery health is corrected by comparing historical operating data with current impedance characteristics. For instance, when the rate of change in battery impedance exceeds a set threshold, a historical degradation model can be used to correct the health status. A degradation model (such as linear regression, support vector machine, or random forest) is built based on historical data such as charge / discharge cycle count, temperature, and charging current. This model can be compared with actual data to correct the estimated health status. After each estimation, the health status is corrected against the degradation model to output a more accurate health status.

[0087] Based on the battery's current operating conditions, such as temperature, voltage change rate, and current variation, the noise covariance matrix in the filtering algorithm is dynamically adjusted. When the battery is under relatively stable operating conditions, the process noise covariance Q can be reduced to improve the stability of the health state estimation; while under conditions of large battery load changes or battery anomalies, the observation noise covariance R is appropriately increased to enhance the algorithm's flexibility. By combining short-timescale state of charge estimates with long-timescale health state estimates using extended Kalman filtering or unscented Kalman filtering algorithms, a joint state estimate of the battery is output, which includes both the state of charge and the health state.

[0088] In this fusion algorithm, the short-timescale state of charge (SOC) estimation results are used to provide the battery's current instantaneous SOC, while the long-timescale health state estimation results are used to provide information on battery aging. The two are then fused using a filtering algorithm to obtain a comprehensive joint state estimate of the battery.

[0089] S400: Based on the current state of charge, predicted road power demand, and engine fuel consumption rate, the target state of charge range of the battery is dynamically determined through a multi-objective optimization algorithm.

[0090] Preferably, the step of dynamically determining the target state of charge range of the battery includes:

[0091] Obtain current battery state of charge, vehicle operating conditions, and road condition prediction information;

[0092] Based on the vehicle power demand prediction model, the predicted road condition power demand for future periods is calculated.

[0093] Determine the system energy efficiency under different states of charge by combining engine fuel consumption rate;

[0094] Construct a multi-objective optimization model with the objectives of minimizing fuel economy, power responsiveness, and battery life degradation;

[0095] The target state of charge range of the battery is dynamically determined based on the optimization results.

[0096] Specifically, the battery's state of charge (SOC) can be estimated in real time by monitoring the battery's voltage, current, and temperature data through the battery management system and combining this with a battery model. The most common estimation method is the current integration method. Vehicle operating conditions, including acceleration and braking force, are acquired in real time through onboard sensors. Based on the vehicle's driving path and historical traffic data, combined with GPS systems and online traffic information services, road conditions for the next few minutes or hours can be predicted.

[0097] A power demand prediction model based on vehicle operating conditions and road condition information is employed. This model learns the power consumption patterns of vehicles under different road conditions using historical data. Common methods include regression analysis, neural networks, or physics-based modeling. Using the trained power demand prediction model, the power demand of vehicles in future time periods is predicted based on future road condition information. The parameters involved in this process include: the vehicle's current power consumption, the expected road gradient, traffic conditions, and the vehicle's acceleration and braking characteristics.

[0098] Fuel consumption rate is a crucial indicator of engine efficiency, typically related to engine speed, load, and operating conditions. By modeling the vehicle's powertrain and combining it with engine load and speed data, the corresponding fuel consumption rate can be obtained. This data can be determined experimentally or provided by the engine control unit. Under different states of charge (SOCs), the energy efficiency of the engine and battery system is calculated. At lower SOCs, the engine participates more in driving and recharges by generating electricity, resulting in a higher fuel consumption rate. Conversely, at higher SOCs, the battery bears more of the power output, reducing the engine's workload and lowering the fuel consumption rate. Combining the battery SOC and engine fuel consumption rate, the overall energy efficiency of the system under different SOCs is evaluated.

[0099] A multi-objective optimization model is established to minimize fuel economy, power response, and battery life degradation. The target state of charge (SOC) range for the battery is determined, with the minimum and maximum SOC values ​​falling within the battery's safe operating range to avoid over-discharge or overcharge. The Pareto front search algorithm is used to find the optimal solution. Through iterative iteration, the optimal SOC range is determined to satisfy the comprehensive requirements of fuel economy, power response, and battery life.

[0100] By acquiring information on the battery's current state of charge, vehicle operating conditions, and road condition predictions, and combining this with a power demand prediction model, engine fuel consumption rate, and a multi-objective optimization algorithm, the target state of charge range for the battery is dynamically determined. This process balances fuel economy, power responsiveness, and battery life through multi-objective optimization, ensuring optimal energy management and performance for the vehicle under various operating conditions.

[0101] Preferably, the multi-objective optimization model is solved using Pareto front search, and the solution steps include:

[0102] S401: Initialize the optimization population, setting the population size, number of iterations, and multi-objective weight parameters;

[0103] S402: Encode the SOC target range parameters for each individual and perform fitness evaluation based on objective functions of fuel economy, battery life degradation, power responsiveness, and thermal safety;

[0104] S403: Use a non-dominated ordination method to classify individuals into ranks and calculate crowding distance to maintain population diversity;

[0105] S404: Perform selection, crossover, and mutation operations based on individual rank and crowding level to generate a new generation of candidate solutions;

[0106] S405: Repeat steps S402 to S404 until the iteration termination condition is met, and obtain the Pareto optimal solution set that satisfies the multi-objective constraints.

[0107] Specifically, a population is defined, the size of which is determined by a population size parameter. Typically, the population size can be chosen between 50 and 100 individuals. Each individual represents a potential solution for optimizing the battery state of charge range (SOCmin, SOCmax). Each individual is represented by a chromosome, which can be encoded using binary encoding, real number encoding, or other suitable representation methods. For the battery state of charge range, the length of the chromosome is related to the search range of the target SOCmin and SOCmax. The population size should ensure sufficient diversity, typically chosen between 50 and 200 individuals, to allow for exploration of a larger solution space. A reasonable maximum number of iterations is set based on the complexity of the problem and the required accuracy, typically a value between 100 and 1000. Weights are assigned to each optimization objective (fuel economy, battery life degradation, power responsiveness, thermal safety). The weight coefficients can be dynamically adjusted according to different practical needs, typically using a normalized form.

[0108] Each individual represents a potential solution, consisting of two parameters (SOCmin and SOCmax) representing the target state of charge range of the battery. Based on the chromosome encoding method, the SOCmin and SOCmax values ​​corresponding to that individual are parsed. For each individual, the values ​​of multiple objective functions are calculated based on the current solution (SOCmin and SOCmax). These objective functions include fuel economy, battery life degradation, power responsiveness, and thermal safety. Based on these objective function values, a fitness value is calculated. The fitness value is used to measure the quality of each individual.

[0109] For each individual, a non-dominated sorting algorithm is executed, dividing individuals into multiple ranks based on their objective function values. The basic idea of ​​non-dominated sorting is that if a solution is not inferior to another solution on all objectives and is superior to another solution on at least one objective, then that solution dominates the other solution. Among individuals in the same rank, crowding distance is used to measure the sparsity of solutions. Crowding distance maintains population diversity by calculating the "density" of individuals in the objective space. A smaller crowding value indicates that the individual has many similar solutions nearby, while a larger crowding value indicates that the individual's region is relatively sparse. The population is then sorted according to the non-dominated sorting results, and the crowding distance for each individual is calculated. Individuals with higher crowding distances have higher priority in selection.

[0110] Using tournament selection or roulette wheel selection methods, parent individuals are selected from the current population based on their rank and crowding distance for crossover. A crossover method suitable for real-number encoding or binary encoding is employed to generate the next generation of individuals. For each generation, a random subset of individuals undergoes mutation. Mutation can involve randomly altering a specific position on a chromosome to introduce new solutions, thus preventing the algorithm from getting trapped in local optima. After selection, crossover, and mutation, a new generation of candidate solutions is generated, and these solutions will continue to evolve as the next generation of the population.

[0111] The iterative process will terminate when any of the following conditions are met: the predetermined maximum number of iterations is reached, a Pareto optimal solution set that meets the accuracy requirements is found, or the change in solutions tends to plateau. When the termination condition is met, all non-dominated solutions in the current population are output, which is the Pareto optimal solution set that satisfies the multi-objective constraints.

[0112] Through the above implementation process, the Pareto front search method is used to optimize the multi-objective optimization model for the target state of charge range of the battery. Under the constraints of multiple objectives such as fuel economy, battery life degradation, power response and thermal safety, the optimal solution set can be obtained.

[0113] S500: Based on the temperature distribution of individual battery cells, the working state of the liquid cooling system is adjusted through a temperature threshold control strategy to maintain the battery temperature within a preset safe range.

[0114] Preferably, the temperature threshold control strategy is implemented using a PID controller, and the control steps include:

[0115] Measure the temperature of individual battery cells and calculate the temperature deviation;

[0116] The flow rate of the liquid cooling system is adjusted proportionally to directly respond to temperature deviations.

[0117] The temperature deviation is accumulated by the integral element to eliminate steady-state error;

[0118] The cooling flow rate is adjusted according to the rate of temperature change using a differential element to suppress excessively rapid temperature changes;

[0119] The final control flow rate of the liquid cooling system is determined by combining the output signals of the proportional, integral, and derivative components.

[0120] Specifically, a temperature sensor is used to measure the operating temperature of each battery cell in real time. The measured battery temperature is compared with a preset target temperature to calculate the temperature deviation. The proportional winding directly adjusts the flow rate of the liquid cooling system based on the current temperature deviation to ensure rapid temperature response. When the battery temperature is higher than the target temperature, the proportional winding increases the cooling flow rate to lower the temperature; conversely, it decreases the cooling flow rate. The integral winding accumulates the temperature deviation to eliminate steady-state errors that may occur during long-term operation. If the temperature deviation persists, the integral winding will compensate by increasing the cooling flow rate to ensure long-term stable operation of the system and avoid persistent small deviations. The derivative winding monitors the rate of temperature change to predict and suppress excessively rapid temperature fluctuations. The derivative winding can slow down the rate of temperature change by adjusting the flow rate, preventing rapid temperature rises or falls and protecting the battery from the effects of excessively rapid changes. The PID controller integrates the output signals of the proportional, integral, and derivative windings to generate the final control flow rate of the liquid cooling system.

[0121] Through the above steps, the PID controller can dynamically adjust the operating state of the liquid cooling system based on the temperature deviation, rate of change, and long-term operating data of individual battery cells, comprehensively considering proportional, integral, and derivative control strategies. This method can effectively regulate the cooling flow rate, maintain the battery temperature stable within a preset safe range, and avoid battery performance degradation or safety hazards caused by excessive temperature fluctuations or long-term instability.

[0122] S600: Based on historical battery charge and discharge data and temperature information, a local empirical degradation model is established, and an alarm is issued when the state of charge exceeds a preset threshold to assist in battery status monitoring and lifespan management.

[0123] Preferably, the steps for establishing a local empirical decay model include:

[0124] Collect historical battery charge / discharge data, current, voltage, and temperature information;

[0125] A random forest model is trained based on the historical data to predict battery capacity degradation and health status indicators.

[0126] During each charge-discharge cycle, the current battery state data is input into the random forest model, and the predicted capacity decay and health status are output.

[0127] When the predicted capacity or health status indicators exceed the preset threshold, an alarm signal is triggered to assist in battery status monitoring and lifespan management.

[0128] Specifically, the battery management system collects historical charge and discharge data of the battery in real time, including relevant information such as voltage, current, and temperature during the charging and discharging process.

[0129] The collected historical data will undergo cleaning and normalization to remove outliers and noisy data, ensuring data quality. The processed data will be divided into training and testing sets. Features related to battery health, such as depth of charge, charging rate, and temperature changes during charging and discharging, will be extracted from parameters including battery voltage, current, temperature, and number of charge / discharge cycles. A random forest algorithm will be used to train the processed data. This model constructs multiple decision trees and determines the final prediction result through voting, used to predict battery capacity degradation and health indicators.

[0130] During each charge and discharge cycle, real-time data on the battery's voltage, current, and temperature are acquired and input into a trained random forest model. Based on this input data, the random forest model predicts the battery's current capacity degradation and health indicators, such as the ratio of battery capacity to initial capacity and state of equilibrium (SOH). The prediction results can be used to monitor the battery's health in real time, serving as a basis for the battery management system to determine whether the battery needs maintenance or replacement.

[0131] Based on the battery's health status and capacity degradation prediction results, one or more thresholds are set to indicate that the battery's health status has reached a warning level. If the capacity degradation or health status indicators output by the model exceed the set thresholds, the system will trigger an alarm signal to remind the user or battery management system to further monitor or maintain the battery.

[0132] Through the above steps, the random forest model trained based on historical data can accurately predict the battery's capacity degradation and health status. This model receives real-time data input during each charge and discharge cycle, enabling it to monitor the battery's health status in real time and issue a warning signal when battery degradation exceeds a set threshold.

[0133] Example 2

[0134] Due to the complexity of urban driving environments, existing battery management systems (BMS) cannot accurately manage battery status under different driving modes and environmental conditions. In particular, in high-temperature environments, existing temperature control systems may lack sufficient precision, leading to battery overheating and affecting its performance.

[0135] To address the aforementioned problems, this invention provides a new energy vehicle battery management system, the framework of which is as follows: Figure 2 As shown. The specific implementation process of this system is as follows:

[0136] When the vehicle starts, the battery management system (BMS) synchronously collects battery voltage, current, and temperature data via sensors. During battery operation, the BMS monitors battery voltage, current, and ambient temperature changes in real time. The data acquisition frequency is set to once per second to ensure real-time monitoring. During each charge / discharge cycle, the system periodically performs frequency scans of electrochemical impedance spectroscopy, ranging from 0.01 Hz to 10 kHz. The collected impedance data is used for subsequent battery health analysis.

[0137] Electrochemical impedance spectroscopy (EIS) data undergoes frequency domain interpolation and noise filtering to ensure smoothness and freedom from interference. At preset frequency points, the system fits the impedance data using an equivalent circuit model, extracting characteristic parameters such as ohmic impedance, charge transfer impedance, and diffusion impedance. The current impedance parameters are compared to the initial parameters under baseline conditions, and the rate of change of each impedance characteristic parameter is calculated. This rate of change is used to predict the battery's health status and remaining lifespan.

[0138] Within a short timescale, the BMS performs rapid state updates based on real-time voltage and current data to estimate the battery's state of charge (SOC). Within a long timescale, the battery's SOC estimate is revised based on the rate of change of impedance characteristic parameters and historical operating data. According to the real-time changing battery data, the system uses an adaptive Kalman filter algorithm to dynamically adjust the noise covariance matrix. This algorithm can fuse short-timescale SOC estimates and long-timescale SOH estimates to output a joint state estimate of the battery.

[0139] The Battery Management System (BMS) acquires the current battery state of charge (SOC), predicted vehicle power demand, and road condition information (such as anticipated traffic congestion and uphill driving). Based on the road condition prediction model, the system calculates the power demand for future periods and, combined with the engine's fuel consumption characteristics, determines the system energy efficiency under different SOCs. Through a multi-objective optimization algorithm, considering factors such as fuel economy, battery life, and power response, the system dynamically determines the target SOC range for the battery (e.g., SOCmin and SOCmax). Based on the optimization results, the target SOC range for the battery is adjusted to ensure that the vehicle can maximize battery life while maintaining performance.

[0140] The temperature of each battery cell is monitored in real time by a temperature sensor. If the battery temperature deviates from the set safe range, the PID controller adjusts the cooling flow rate of the liquid cooling system according to the temperature deviation to keep the battery temperature within the safe range.

[0141] The Battery Management System (BMS) collects historical charge / discharge data, current, voltage, and temperature information from the battery, and trains a random forest model using this historical data. During each charge / discharge cycle, the system inputs real-time battery status data (such as SOC, temperature, and historical load) into the trained random forest model to predict the battery's capacity degradation and health status. If the prediction results indicate that the battery capacity or health status exceeds a preset threshold, the system will trigger an alarm, prompting the user to perform maintenance or check the battery to ensure that the battery does not experience performance degradation or safety issues due to excessive degradation.

[0142] Through the above implementation process, the new energy vehicle battery management system can accurately monitor the battery's state of charge, health status, and temperature based on real-time data, optimizing battery lifespan and performance. In hybrid driving mode, the system can dynamically adjust the battery's charging and discharging strategy and ensure stable battery temperature, thereby extending battery lifespan and improving fuel economy.

[0143] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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 method for managing batteries in new energy vehicles, characterized in that, include: Simultaneous acquisition of data from individual battery cells, including voltage, current, and temperature data, and frequency scanning is performed during the acquisition process to obtain electrochemical impedance spectroscopy data; The electrochemical impedance spectroscopy data is received, and the ohmic impedance, charge transfer impedance, and diffusion impedance parameters are extracted based on the equivalent circuit model. The rate of change of the impedance characteristic parameters is then calculated. A dual-time-scale adaptive Kalman filter algorithm is adopted, and the noise covariance matrix in the algorithm is dynamically adjusted according to the rate of change of impedance characteristic parameters to achieve joint estimation of battery state of charge and health. Based on the current state of charge, predicted road power demand, and engine fuel consumption rate, the target state of charge range of the battery is dynamically determined through a multi-objective optimization algorithm. Based on the temperature distribution of individual battery cells, the working state of the liquid cooling system is adjusted through a temperature threshold control strategy to maintain the battery temperature within a preset safe range. Based on historical battery charge and discharge data and temperature information, a local empirical degradation model is established, and an alarm is issued when the state of charge exceeds a preset threshold to assist in battery status monitoring and lifespan management. The joint estimation steps for battery state of charge and state of health include: It receives input data including the rate of change of voltage, current, temperature, and impedance characteristic parameters; Within a short timescale, rapid state updates are performed based on real-time voltage and current data to obtain an initial estimate of the battery's state of charge. Over a long time scale, the battery health status estimate is corrected based on the rate of change of impedance characteristic parameters and historical operating data. The noise covariance matrix in the filtering algorithm is dynamically adjusted based on the battery's operating conditions. The results of short-term state of charge estimation and long-term health estimation are fused to output a joint state estimate of the battery.

2. The battery management method for new energy vehicles according to claim 1, characterized in that, The frequency range of the frequency scan is from 0.01Hz to 10kHz, the number of scan points is not less than 50, and the single scan time is controlled within 30 seconds.

3. The battery management method for new energy vehicles according to claim 1, characterized in that, The steps for calculating the rate of change of impedance characteristic parameters include: Frequency domain interpolation and noise filtering were performed on the electrochemical impedance spectroscopy data to obtain a smooth impedance curve. At a preset frequency point, the impedance spectrum is fitted based on the equivalent circuit model, and the corresponding ohmic impedance, charge transfer impedance and diffusion impedance parameters are extracted. The extracted impedance parameters are compared with the initial parameters under the reference state, and the relative changes of each impedance parameter are calculated. The rate of change of the impedance characteristic parameters is calculated based on the relative change value.

4. The battery management method for new energy vehicles according to claim 1, characterized in that, The steps for dynamically determining the target state of charge range of a battery include: Obtain current battery state of charge, vehicle operating conditions, and road condition prediction information; Based on the vehicle power demand prediction model, the predicted road condition power demand for future periods is calculated. Determine the system energy efficiency under different states of charge by combining engine fuel consumption rate; Construct a multi-objective optimization model with the objectives of minimizing fuel economy, power responsiveness, and battery life degradation; The target state of charge range of the battery is dynamically determined based on the optimization results.

5. A new energy vehicle battery management method according to claim 4, characterized in that, The multi-objective optimization model is solved using Pareto front search, and the solution steps include: S401: Initialize the optimization population, setting the population size, number of iterations, and multi-objective weight parameters; S402: Encode the SOC target range parameters for each individual and perform fitness evaluation based on objective functions of fuel economy, battery life degradation, power responsiveness, and thermal safety; S403: Use a non-dominated ordination method to classify individuals into ranks and calculate crowding distance to maintain population diversity; S404: Perform selection, crossover, and mutation operations based on individual rank and crowding level to generate a new generation of candidate solutions; S405: Repeat steps S402 to S404 until the iteration termination condition is met, and obtain the Pareto optimal solution set that satisfies the multi-objective constraints.

6. The battery management method for new energy vehicles according to claim 1, characterized in that, The temperature threshold control strategy is implemented using a PID controller, and the control steps include: Measure the temperature of individual battery cells and calculate the temperature deviation; The flow rate of the liquid cooling system is adjusted proportionally to directly respond to temperature deviations. The temperature deviation is accumulated by the integral element to eliminate steady-state error; The cooling flow rate is adjusted according to the rate of temperature change using a differential element to suppress excessively rapid temperature changes; The final control flow rate of the liquid cooling system is determined by combining the output signals of the proportional, integral, and derivative components.

7. The battery management method for new energy vehicles according to claim 1, characterized in that, The steps to establish a local empirical decay model include: Collect historical battery charge / discharge data, current, voltage, and temperature information; A random forest model is trained based on historical data to predict battery capacity degradation and health indicators. During each charge-discharge cycle, the current battery state data is input into the random forest model, and the predicted capacity decay and health status are output. When the predicted capacity or health status indicators exceed the preset threshold, an alarm signal is triggered to assist in battery status monitoring and lifespan management.

8. A battery management system for new energy vehicles, characterized in that, The system for implementing the battery management method according to any one of claims 1-7 includes: The data acquisition module is used to synchronously acquire data from individual battery cells, including voltage, current, and temperature data, and to perform frequency scanning during the acquisition process to obtain electrochemical impedance spectroscopy data. The impedance feature extraction module is used to receive the electrochemical impedance spectroscopy data, extract ohmic impedance, charge transfer impedance and diffusion impedance parameters based on the equivalent circuit model, and calculate the rate of change of impedance feature parameters. The battery state assessment module is used to employ a dual-time-scale adaptive Kalman filter algorithm to dynamically adjust the noise covariance matrix in the algorithm based on the rate of change of impedance characteristic parameters, thereby achieving a joint estimation of the battery's state of charge and health. The state of charge range optimization module is used to dynamically determine the target state of charge range of the battery based on the current state of charge, predicted road power demand, and engine fuel consumption rate through a multi-objective optimization algorithm. The liquid cooling system control module is used to adjust the working state of the liquid cooling system according to the temperature distribution of the individual battery cells and through a temperature threshold control strategy, so that the battery temperature is maintained within a preset safe range. The battery status warning module is used to establish a local empirical degradation model based on the battery's historical charge and discharge data and temperature information, and to issue an alarm when the state of charge exceeds a preset threshold, so as to assist in battery status monitoring and lifespan management.

Images