A charging temperature control method of a battery module

Abstract

The application relates to the technical field of battery thermal management, and discloses a charging temperature control method for a battery module, which comprises the following steps: monitoring the real-time temperature of each single battery cell, constructing a three-dimensional temperature field model and identifying low-temperature areas and high-temperature areas, establishing an electro-thermal-aging coupling model based on the model, calculating heating power, cooling power and a target balanced temperature with the minimum charging time, temperature rise energy consumption and aging rate as the target, inputting the power as a feedforward control amount into the coupling model for simulation deduction to obtain a predicted temperature change trajectory, dynamically adjusting the working duty cycle of the heating and cooling devices through a model predictive control algorithm based on the deviation between the real-time temperature and the trajectory, and making the temperature converge to the target balanced temperature. The application realizes accurate identification and differential regulation of local temperature in the battery module, effectively shortens the charging time under the premise of taking into account energy consumption and service life, and improves the energy efficiency and intelligent level of the thermal management system.

Description

Technical Field

[0001] This invention relates to the field of battery thermal management technology, and in particular to a method for controlling the charging temperature of a battery module. Background Technology

[0002] As the core energy storage unit of new energy vehicles, lithium-ion battery modules have received widespread attention for their charging efficiency, safety and service life. During high-rate charging, the temperature of the battery cells inside the battery rises rapidly. At the same time, due to the asymmetry of the internal structure of the battery module, the differences between individual cells and the unevenness of the cooling channel design, it is easy to cause uneven temperature distribution inside the battery module. This not only limits the further improvement of charging power, but may also accelerate battery aging and even cause the risk of thermal runaway.

[0003] In existing battery thermal management systems, control strategies mostly rely on preset fixed thresholds or simple feedback control logic, lacking targeted intervention for local abnormal temperature areas. This fails to effectively solve the problem of uneven temperature inside the module. Low-temperature areas will still restrict the charging rate due to excessive internal resistance, while high-temperature areas will continuously bear thermal stress, resulting in limited performance of the entire battery module.

[0004] Most existing technologies only use temperature as a single control target, such as only pursuing the control of battery temperature within a certain ideal range, without incorporating charging time, the additional energy consumed by temperature rise, and the resulting rate of battery capacity decay into a unified optimization framework. As a result, in practical applications, the temperature control system is unable to dynamically adjust the control strategy according to the user's real-time needs (such as emergency fast charging or routine maintenance).

[0005] Because the internal thermal environment of the battery module has strong coupling and nonlinear characteristics, when the low temperature region and the high temperature region coexist, the natural heat conduction between them will generate internal interference. If the control system cannot identify and utilize this interference, it may cause a lot of energy waste. Traditional control algorithms lack the ability to predict the future state of the system, and often lag in adjustment when facing the thermal inertia caused by the large heat capacity of the battery. This results in large fluctuations and slow convergence in the temperature regulation process, affecting the efficiency and stability of the charging process. Summary of the Invention

[0006] The purpose of this invention is to provide a method for controlling the charging temperature of a battery module in order to solve the problems in the prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for controlling the charging temperature of a battery module, comprising the following steps: S1, monitor the real-time temperature of each individual cell in the battery module, construct a three-dimensional temperature field model containing temperature distribution field and local hot spot location information based on the real-time temperature, and identify a first low temperature region with a temperature lower than a preset first temperature threshold and a second high temperature region with a temperature higher than a preset second temperature threshold. S2, Based on the three-dimensional temperature field model, establish an electro-thermal-aging coupling model with the goal of minimizing charging time, temperature rise energy consumption and battery aging rate as the comprehensive optimization objective. Calculate the first heating power for the first low-temperature region, the second cooling power for the second high-temperature region, and the target equilibrium temperature at the current moment using a multi-objective genetic algorithm. S3, the first heating power and the second cooling power are used as feedforward control quantities and input to the three-dimensional temperature field model for simulation and deduction to obtain the predicted temperature change trajectory during the pre-adjustment process; S4. During the actual pre-adjustment process, the real-time temperature of each individual cell is monitored. Based on the deviation between the real-time temperature and the expected temperature value at the corresponding moment on the predicted temperature change trajectory, the working duty cycle of the heating device and the cooling device is dynamically adjusted through the model predictive control algorithm so that the temperatures of the first low-temperature region and the second high-temperature region converge to the target equilibrium temperature within a preset time window.

[0008] The beneficial effects of the technical solution provided by this invention include at least the following: This invention reconstructs a three-dimensional temperature field model containing temperature distribution field and local hot spot location information by sensor data fusion and finite element model correction, and accurately identifies the first low temperature region and the second high temperature region that need intervention, effectively solving the pain point of existing technologies that cannot perform targeted intervention on local abnormal temperature regions.

[0009] This invention establishes an electro-thermal-aging coupled model and combines it with a multi-objective genetic algorithm. It takes three conflicting indicators—charging time, temperature rise energy consumption, and battery aging rate—as comprehensive optimization objectives to achieve global multi-objective optimization of temperature control strategies. This method can search for Pareto optimal solutions in a three-dimensional decision space composed of heating power, cooling power, and target temperature, and dynamically select the most suitable temperature control strategy for the current needs based on the user's preset preference weights. This significantly improves the intelligence level of the thermal management system and solves the problem of difficulty in decision-making under multiple conflicting objectives in existing technologies.

[0010] This invention proposes a feedforward + feedback composite control architecture, and in particular, introduces a model predictive control algorithm based on a reference trajectory. The duty cycle of the heating and cooling devices is dynamically adjusted based on the deviation between the real-time temperature and the expected value. This predictive control strategy effectively overcomes the thermal inertia and hysteresis of the battery system, ensuring that the temperature of each region can smoothly and accurately converge to the target equilibrium temperature along the preset trajectory, significantly improving the stability and accuracy of temperature control.

[0011] This invention fully considers and utilizes the thermal coupling effect between regions within the battery module. By analyzing the spatial adjacency of the first low-temperature region and the second high-temperature region, the coupled heat flux density between them is calculated. This allows for full utilization of internal heat conduction to achieve energy transfer, significantly reducing energy consumption during temperature control and improving the overall energy efficiency of the thermal management system. It also helps to achieve temperature balance within the battery module more quickly. Attached Figure Description

[0012] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of a method provided in an embodiment of the present invention. Detailed Implementation

[0014] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a battery module charging temperature control method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0016] The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0017] The following description, in conjunction with the accompanying drawings, details a specific scheme for a battery module charging temperature control method provided by the present invention.

[0018] Please see Figure 1The diagram illustrates a method flowchart for controlling the charging temperature of a battery module according to an embodiment of the present invention. The method includes the following steps: S1, monitor the real-time temperature of each individual cell in the battery module, construct a three-dimensional temperature field model containing temperature distribution field and local hot spot location information based on the real-time temperature, and identify a first low temperature region with a temperature lower than a preset first temperature threshold and a second high temperature region with a temperature higher than a preset second temperature threshold. S2, an electro-thermal-aging coupled model is established based on a three-dimensional temperature field model. The comprehensive optimization objectives are to minimize charging time, temperature rise energy consumption and battery aging rate. A multi-objective genetic algorithm is used to calculate the first heating power for the first low temperature region, the second cooling power for the second high temperature region and the target equilibrium temperature at the current moment. S3, the first heating power and the second cooling power are used as feedforward control inputs to the three-dimensional temperature field model for simulation and deduction, and the predicted temperature change trajectory during the pre-adjustment process is obtained; S4. During the actual pre-adjustment process, the real-time temperature of each individual cell is monitored. Based on the deviation between the real-time temperature and the expected temperature value at the corresponding moment on the predicted temperature change trajectory, the working duty cycle of the heating device and the cooling device is dynamically adjusted through the model predictive control algorithm so that the temperature of the first low temperature region and the second high temperature region converges to the target equilibrium temperature within the preset time window.

[0019] As one embodiment of the present invention, the step of monitoring the real-time temperature of each individual cell in the battery module and constructing a three-dimensional temperature field model containing temperature distribution field and local hot spot location information based on the real-time temperature includes: The real-time temperature of each individual cell at multiple sampling times is obtained and processed by Kalman filtering to estimate the predicted temperature of each individual cell at the next time step. Based on the real-time temperature, predicted temperature, and finite element heat transfer geometry model of the battery module, the boundary conditions and thermal property parameters of the finite element heat transfer geometry model are corrected by the extended Kalman filter algorithm to generate a dynamic thermal parameter set. The dynamic thermal parameter set is substituted into the finite element heat transfer geometric model for iterative solution to reconstruct the temperature value of the internal spatial point of the battery module at the current moment. The temperature distribution field and local hot spot location information are then generated to construct a three-dimensional temperature field model.

[0020] It should be noted that in this embodiment, a patch-type temperature sensor is pre-embedded on the surface of each individual cell, specifically in the central area of ​​the side of the cell and at the positive and negative electrode tabs. The battery management system reads the real-time temperature from the temperature sensor at a fixed frequency (e.g., 10 times per second). Since there is electromagnetic interference inside the battery pack, the directly read real-time temperature is prone to noise spikes. Therefore, the time series of real-time temperatures collected by each sensor is processed using a Kalman filter algorithm. The algorithm logic is as follows: based on the filtered real-time temperature of the previous moment and the real-time temperature of the current moment, a temperature value that is closest to the actual situation is dynamically estimated through iterative calculation. At the same time, based on the temperature change trend of the current moment, the predicted temperature of the point at the next moment is initially estimated, providing a basis for subsequent model correction.

[0021] In this embodiment, a finite element thermal geometry model of the battery module is pre-stored in the storage chip of the battery management system. This geometry model is constructed based on the actual three-dimensional dimensions (length, width, and height) of the battery module and the physical properties of each layer of materials (such as the cell core, shell, insulating sheet, and structural adhesive). Initial thermal properties (such as thermal conductivity and specific heat capacity) and boundary conditions (such as contact thermal resistance with the cooling plate and convective heat transfer coefficient with the air) are preset. As the battery ages or the ambient temperature changes during use, these preset parameters will gradually become mismatched. To solve this problem, this embodiment uses an extended Kalman filter algorithm for online correction. Specifically, the system uses the real-time temperature obtained by filtering in the aforementioned steps as the observation target, compares the temperature value calculated by the finite element thermal geometry model with the measured value, and adjusts the key parameters in the geometry model that have a significant impact on the results and are prone to change, such as the equivalent thermal conductivity and contact thermal resistance, based on the error between the two. The adjustment is iterated multiple times to bring the error between the model-calculated temperature and the measured temperature to a minimum, generating a dynamic thermal parameter set that best reflects the true physical state of the battery module at the current moment.

[0022] The dynamic thermal parameter set obtained from the above correction is substituted into the finite element heat transfer geometry model. At the same time, the heat generation power of each individual cell is calculated in real time based on the current charging current and the cell internal resistance. This heat generation power is then applied as a thermal load to the finite element heat transfer geometry model of the corresponding individual cell. The grid points inside the entire battery module are iteratively solved according to the basic physical laws of heat conduction to calculate the temperature value of each grid point at the current moment. After the calculation is completed, the system integrates and renders the temperature data of all grid points in a virtual three-dimensional space to generate a three-dimensional temperature distribution cloud map with color gradient. The coordinates of points with significantly higher temperatures than the adjacent areas are automatically scanned and marked as local hotspot location information.

[0023] In one embodiment of the present invention, the step of identifying a first low-temperature region with a temperature lower than a preset first temperature threshold and a second high-temperature region with a temperature higher than a preset second temperature threshold includes: The sliding window clustering algorithm is used to traverse the spatial points of the three-dimensional temperature field model. The set of spatial points that are continuously distributed and whose temperature is lower than the preset first temperature threshold is marked as the first low temperature candidate area, and the set of spatial points that are continuously distributed and whose temperature is higher than the preset second temperature threshold is marked as the second high temperature candidate area. Calculate the regional volume, temperature variance, and heat flux density with adjacent regions of the first low-temperature candidate region and the second high-temperature candidate region, respectively. If the volume of a certain low-temperature candidate area exceeds a preset first volume threshold and its temperature variance is less than a preset uniformity threshold, it is confirmed as a first low-temperature area. If the volume of a certain high-temperature candidate area exceeds a preset second volume threshold and its heat flux density exceeds a preset heat flux threshold, it is confirmed as a second high-temperature area. Output the boundary contours and centroid coordinates of the first low-temperature region and the second high-temperature region.

[0024] It should be noted that in this embodiment, a fixed-size three-dimensional sliding window is set on the existing three-dimensional temperature field model. The size of the window is determined according to the actual structure of the battery module. For example, a cube containing adjacent 3×3×3 grid points (covering approximately the volume of a single battery cell) is set. The three-dimensional sliding window is controlled to start from one corner of the three-dimensional space of the battery module and traverse the entire grid space row by row and layer by layer along the X-axis, Y-axis, and Z-axis according to a preset step size (e.g., the distance between moving one grid point each time). The system reads the temperature values ​​of all grid points in the sliding window each time the sliding window stops and performs the following judgment: If there is a group of consecutive adjacent grid points within the sliding window whose temperature values ​​are all lower than a preset first temperature threshold (e.g., 0°C), then these grid points are marked as part of the first low-temperature candidate area and assigned a temporary number. If there is a group of consecutive adjacent grid points within the sliding window whose temperature values ​​are all higher than a preset second temperature threshold (e.g., 45°C), these grid points are marked as part of the second high-temperature candidate area and assigned another temporary number. If the marked portions of two adjacent windows overlap and are both the first low-temperature candidate area or the second high-temperature candidate area, they are merged and their boundary ranges are updated. Finally, all spatially connected points in the entire temperature field that satisfy the same temperature condition are merged into several independent candidate areas.

[0025] After completing the above steps, the system obtains several first low-temperature candidate areas and second high-temperature candidate areas. Subsequently, quantitative characteristic calculations need to be performed on each candidate area, specifically including: (1) Calculate the volume of the region: The system counts the total number of grid points contained in each candidate region, and then multiplies it by the actual physical volume represented by a single grid point (this volume is determined by the meshing accuracy of the finite element heat transfer geometry model, for example, each grid point represents a cube space of 2mm×2mm×2mm) to obtain the approximate volume value of the candidate region. (2) Calculate temperature variance: The system reads the temperature values ​​of all grid points in each candidate area and calculates the variance of these temperature values. The smaller the variance, the more uniform the temperature distribution in the candidate area. The larger the variance, the more intense the temperature fluctuation in the area, which may contain abnormal hot spots or cold spots and is not suitable as an overall intervention target.

[0026] (3) Calculate the heat flux density with adjacent areas: The system identifies the boundary grid points of each candidate area and reads the temperature values ​​of the adjacent grid points located outside the candidate area that are close to these boundary points. According to the basic principle of heat conduction, the greater the temperature difference and the closer the distance, the greater the heat flux density. In this embodiment, the temperature gradient at the boundary is obtained by calculating the temperature difference between the grid points inside and outside the boundary and dividing it by the grid spacing. Then, the heat flux density between the candidate area and the external environment is estimated by multiplying it by the thermal conductivity of the material. For high-temperature areas, the greater the heat flux density, the higher the rate at which heat is dissipated outward, and the greater the risk of thermal runaway.

[0027] After completing the above feature calculations, the system confirms the results according to preset filtering rules, specifically including: (1) For the first low-temperature candidate area, the system determines whether its area volume exceeds the preset first volume threshold (for example, the first volume threshold is the total volume of three standard single cells). If the volume is too small, it may be due to local minor fluctuations or sensor noise, and there is no need to start the heating device. At the same time, the system determines whether its temperature variance is less than the preset uniformity threshold (for example, 2℃). 2 This means that only when both conditions are met—that the volume is large enough and the internal temperature is uniform enough—can the first low-temperature candidate area be identified as the first low-temperature area requiring heating intervention. This ensures that the heating device only works on large, continuous low-temperature areas, avoiding frequent start-ups and shutdowns. (2) For the second high-temperature candidate area, the system determines whether its area volume exceeds the preset second volume threshold (for example, the second volume threshold is the total volume of two standard single cells), and at the same time determines whether its boundary heat flux density exceeds the preset heat flux threshold (for example, 500W / m). 2 This means that only if the candidate area is not an isolated point in volume and its high temperature has a certain degree of danger can the second high temperature candidate area be identified as the second high temperature area requiring cooling intervention, so as to ensure that cooling resources are given priority to deal with high-risk hot spots that are transferring heat to the surrounding area and may trigger chain reactions.

[0028] If a candidate area is identified as the first low-temperature area or the second high-temperature area requiring intervention, the boundary contour and centroid coordinates of the area are further extracted. The boundary contour refers to the set of three-dimensional coordinates of all grid points that constitute the edge of the area, and the centroid coordinates refer to the average value of the coordinates of all grid points in the area, representing the geometric center of the area.

[0029] As one embodiment of the present invention, the step of establishing an electro-thermal-aging coupling model based on a three-dimensional temperature field model includes: Based on the equivalent circuit model and three-dimensional temperature field model of the battery module, an electro-thermal coupling sub-model is established to output the real-time temperature distribution. An aging sub-model is established based on a pre-defined semi-empirical aging model to quantify the battery aging rate under different thermal conditions, including the battery capacity decay rate and the battery internal resistance increase rate. The real-time temperature distribution output by the electro-thermal coupling sub-model is used as the input to the aging sub-model, and the battery aging rate output by the aging sub-model is fed back to the electro-thermal coupling sub-model to correct its internal resistance parameter, thus forming an electro-thermal-aging coupling model.

[0030] It should be noted that this embodiment describes the electrical behavior of a single cell based on the equivalent circuit model of the battery. In a specific embodiment, a second-order Thevenin equivalent circuit model is used to simulate the polarization characteristics of the battery module under dynamic operating conditions. For each single cell in the battery module, the system pre-calibrates the model parameters under different temperatures and different states of charge through mixed pulse power characteristic tests, including ohmic internal resistance, polarization internal resistance, polarization capacitance, and the correspondence curve between open circuit voltage and state of charge. These parameters are stored in the battery management system in the form of a lookup table.

[0031] During actual operation, the system collects the charging and discharging current and terminal voltage of the battery module in real time, substitutes the current at the current moment into the equivalent circuit model, combines the state of charge and real-time temperature of each individual cell at the current moment, looks up the corresponding internal resistance and polarization parameters from the table, and calculates the current Joule heat generation power of each individual cell. In addition, in a preferred embodiment, the reversible reaction heat generation power is calculated based on the current direction, temperature and entropy thermal coefficient, and then added to the Joule heat generation power to obtain the total heat generation power of each cell at the current moment.

[0032] The three-dimensional temperature field model uses the total heat generation power as a heat source load and applies it to the mesh nodes of the corresponding individual battery cell. Combining the current thermal property parameters and boundary conditions, it calculates the temperature distribution of the battery module at the next moment and then feeds the calculated temperature distribution back to the equivalent circuit model to update the temperature value used when looking up the table at the next moment, thus forming a two-way real-time coupling between electrical behavior and thermal behavior.

[0033] In this embodiment, the system uses a semi-empirical aging model to quantify the aging rate of the battery module under different thermal conditions. This model is an empirical formula obtained by fitting a large amount of battery cycle life test data. In a specific embodiment, the model includes two output dimensions: capacity decay rate and internal resistance increase rate. During real-time operation, the system uses the real-time temperature distribution of each individual cell calculated by the electro-thermal coupling sub-model as the input of the aging sub-model to estimate the capacity decay rate and internal resistance increase rate of the battery module under the current thermal conditions in real time.

[0034] In this embodiment, the internal resistance increase rate calculated by the aging sub-model is also fed back to the electro-thermal coupling sub-model in real time to correct the internal resistance parameters in the equivalent circuit model. Specifically, the system maintains an internal resistance correction coefficient, which gradually increases with the increase of the cumulative aging time. When calculating the heat generation power at the next moment, the system multiplies the basic internal resistance value obtained by looking up the table by the internal resistance correction coefficient to obtain the actual internal resistance value at the current moment for heat generation calculation, so as to form a closed-loop electro-thermal-aging coupling model.

[0035] In one embodiment of the present invention, the step of calculating the first heating power for the first low-temperature region, the second cooling power for the second high-temperature region, and the target equilibrium temperature at the current moment using a multi-objective genetic algorithm includes: Initialize a population containing multiple individuals and generate multiple combinations of candidate temperature control strategies, including a first heating power, a second cooling power, and a target equilibrium temperature. Each candidate temperature control strategy combination is input into the electro-thermal-aging coupled model for simulation and deduction, and the predicted charging time, predicted temperature rise energy consumption and predicted battery aging rate are obtained for each candidate temperature control strategy combination. The Pareto front is used to stratify all individuals in the current population based on the non-dominated sorting algorithm, and the crowding distance of each individual in the same layer is calculated. All individuals are sorted according to their layer and crowding distance. Select a preset number of individuals from the sorted individuals as parents, and perform selection, crossover and mutation to generate offspring population. Repeat the iteration until a preset termination condition is reached to obtain the Pareto front solution set. The optimal solution is selected from the Pareto front solution set according to the preset multi-objective preference weights, and the corresponding first heating power, second cooling power and target equilibrium temperature are output.

[0036] It should be noted that the core of this step is to use a multi-objective genetic algorithm to search for the optimal solution that can simultaneously optimize charging time, temperature rise energy consumption, and battery aging rate in a decision space consisting of heating power, cooling power, and target equilibrium temperature. The genetic algorithm itself (including population initialization, crossover mutation, non-dominated sorting, etc.) is a common technical means in this field for handling multi-objective optimization problems.

[0037] In this embodiment, the range of the first heating power depends on the upper limit of the physical power of the heating film or heating element arranged in the battery module, for example, 0 to 500 watts; the range of the second cooling power depends on the maximum working capacity of the cooling fan or liquid cooling pump arranged in the battery module, for example, 0 to 300 watts; and the range of the target equilibrium temperature is limited by the safe temperature window of the battery material, for example, set to 15°C to 40°C. The above three variables together constitute the three-dimensional search space for the operation of the genetic algorithm.

[0038] In the simulation environment, the system simulates the process of preheating or precooling the battery module according to this set of temperature control strategies, starting from the current moment, and then charging the battery module until it is fully charged using a standard charging strategy. During the simulation, the coupled model records three key output indicators in real time: (1) Predicted charging time: the total time from the start of preheating to full charge; (2) Predict the temperature rise energy consumption, that is, the total electrical energy consumed by the heating and cooling devices during the entire preheating and charging process; (3) Predict the battery aging rate, that is, the capacity decay calculated by the aging sub-model based on the temperature history experienced by each cell during the simulation.

[0039] Since the three optimization objectives mentioned above are physically conflicting—for example, to shorten the charging time, it is often necessary to use a large heating power to quickly heat the battery to the optimal charging temperature window, but this will lead to increased energy consumption due to temperature rise, and although a higher temperature is beneficial to the charging speed, it may accelerate battery aging, etc.—a multi-objective genetic algorithm is used to search for a series of Pareto optimal solutions among this set of conflicting objectives, forming a Pareto front solution set.

[0040] The multiple candidate solutions contained in the Pareto front solution set are dynamically changing in actual application scenarios. Therefore, it is necessary to preset multi-objective preference weights to select the optimal solution for final execution based on actual needs. In a specific embodiment, if an urgent travel need of a user is detected, the weight of charging time is set to the highest, for example, 0.6, and the solution with the shortest charging time is selected from the Pareto front. In the scenario of daily charging at night, the weight of aging rate is set to the highest, for example, 0.5. By selecting weights, the same algorithm can be adapted to different usage scenarios.

[0041] In one embodiment of the present invention, before inputting the first heating power and the second cooling power as feedforward control quantities into the three-dimensional temperature field model for simulation, the method further includes a step of correcting the first heating power and the second cooling power: Based on the three-dimensional temperature field model, the spatial adjacency relationship and heat conduction path between the first low-temperature region and the second high-temperature region are determined; If there is an overlapping area between the first low-temperature region and the second high-temperature region, or if the distance between them is less than a preset thermal interference threshold, the coupled heat flux density generated at the overlapping area is calculated based on the spatial adjacency relationship. Based on the coupled heat flux density, when it represents the direction of coupled heat flux from the high temperature region to the low temperature region, the first heating power is reduced and the second cooling power is increased simultaneously according to the preset compensation coefficient. The modified first heating power and modified second cooling power were used as feedforward control variables to perform simulation.

[0042] It should be noted that in this embodiment, the system reads the coordinates of all grid points on the boundary of the first low temperature region and the coordinates of all grid points on the boundary of the second high temperature region, calculates the minimum Euclidean distance between these two sets of coordinate points. If it is zero, it means that the two regions overlap or border each other. If it is greater than zero but less than the preset thermal interference threshold (for example, less than 5 mm, which is equivalent to the thickness of the cell casing or the gap between adjacent cells), it is considered that there is a risk of thermal interference between the two. Meanwhile, the system determines the main heat conduction path based on the direction of the line connecting the centroid coordinates of each adjacent first low-temperature region and second high-temperature region, combined with the thermal conductivity characteristics of the materials inside the battery module. For example, if the low-temperature region is located directly above the high-temperature region and there is an insulating sheet between them, the main heat conduction path is considered to be in the vertical direction, with relatively low heat transfer resistance.

[0043] When the system determines that there is a risk of thermal interference, it locates the boundary or the nearest interface between the two regions with the risk in the three-dimensional temperature field model. It reads the temperature values ​​of the boundary grid points on the high-temperature and low-temperature sides of this interface and calculates the temperature difference between them. Simultaneously, the system reads the thermal conductivity of the material at this interface (this parameter has been obtained in the previous steps and dynamically corrected using extended Kalman filtering). Based on the fundamental physical laws of heat conduction, it divides the temperature difference by the actual distance between the two boundaries to obtain the temperature gradient, and then multiplies it by the thermal conductivity to estimate the coupled heat flux density from the high-temperature region to the low-temperature region. This means that even without any active heating or cooling, heat from the high-temperature region will flow to the low-temperature region through natural conduction. Therefore, the original heating and cooling power needs to be corrected, specifically including: The total coupled heat flux is obtained by multiplying the calculated coupled heat flux density by the area of ​​the overlapping region. Then, according to the preset compensation coefficient, the heat provided by natural conduction is subtracted from the first heating power, and the heat that should have been borne by the cooling device but was actually carried away by natural conduction is added to the second cooling power. The compensation coefficient is usually set between 0.3 and 0.7 to avoid overcompensation that would cause reverse temperature fluctuations. For example, if the calculated coupled heat flux is 20 watts and the compensation coefficient is set to 0.5, the first heating power is reduced by 10 watts and the second cooling power is increased by 10 watts.

[0044] After the above corrections, the system obtains a new set of heating and cooling power values. These corrected power values ​​will be used as feedforward control inputs to the three-dimensional temperature field model for the next step of simulation and deduction, in order to generate the predicted temperature change trajectory. Through the above corrections, the initial control values ​​on which the simulation is based have taken into account the natural thermal coupling between regions, thus making the subsequently generated predicted trajectory closer to the actual physical process.

[0045] In one embodiment of the present invention, the predicted temperature change trajectory is generated through the following steps: A time window is preset, and the target equilibrium temperature, the starting temperature of the first low temperature region and the starting temperature of the second high temperature region at the start time of the time window are obtained; Based on the principle of minimizing energy dissipation, a reference trajectory function is constructed with time as the independent variable. The reference trajectory function satisfies the following boundary conditions: at the initial moment, the trajectory values ​​of the low-temperature region and the high-temperature region are equal to their initial temperatures, and at the final moment, the trajectory values ​​of the low-temperature region and the high-temperature region are both equal to the target equilibrium temperature. The ideal temperature change rate of each region at each moment within the time window is determined based on the reference trajectory function. If the ideal temperature change rate exceeds the maximum control capability threshold of the corresponding heating or cooling device, the length of the preset time window is adjusted, and a new reference trajectory function that satisfies the physical constraints of the actuator is reconstructed based on the principle of minimizing energy dissipation, until a feasible predicted temperature change trajectory is generated.

[0046] It should be noted that in this embodiment, before performing trajectory planning, the system needs to determine an expected time window length, i.e., how long it will take to adjust the temperatures of both the low-temperature and high-temperature regions to the target equilibrium temperature. In a specific embodiment, the system presets this time window based on the urgency of the current charging demand. For example, if the user selects the "fast charging" mode, the system presets a shorter time window, such as 300 seconds; if the user selects the "standard charging" mode, the preset time window is 600 seconds; and if the user selects the "maintenance charging" mode, the preset time window is 900 seconds. The system reads the target equilibrium temperature output by the multi-objective genetic algorithm in the previous steps as the endpoint temperature (T_target), and reads the average temperature of all grid points in the first low-temperature region and the second high-temperature region at the current moment from the three-dimensional temperature field model as the starting temperature (first low-temperature region: T_low_start, second high-temperature region: T_high_start), thereby generating the boundary conditions for trajectory planning.

[0047] In this embodiment, the system constructs a reference trajectory function based on the principle of minimizing energy dissipation. In a specific embodiment, the system uses a curve with smooth characteristics (e.g., a curve with first-order exponential decay) to construct the trajectory. For the first low-temperature region, its starting temperature T_low_start < ending temperature T_target, so its reference trajectory function is a curve that gradually rises and eventually converges to the target value. The rate of rise is faster at the initial moment and gradually slows down as it approaches the target value. For the high-temperature region, its starting temperature T_high_start > ending temperature T_target, so its reference trajectory function is a curve that gradually falls and eventually converges to the target value. The rate of fall is faster at the initial moment and then gradually slows down. This curve form conforms to the basic laws of heat transfer: that is, the greater the temperature difference between the temperature and the environment, the faster the heat exchange rate. As the temperature difference decreases, the rate naturally decreases. The two reference trajectory functions have different starting points and opposite directions of change, but they converge to the same target equilibrium temperature at the end. This means that temperature uniformity should be achieved throughout the entire battery module.

[0048] After constructing the reference trajectory function, the ideal temperature change rate at each moment on the trajectory can be obtained by taking its time derivative. For low-temperature regions, this rate represents the required heating rate; for high-temperature regions, this rate represents the required cooling rate. The system compares these two rates with the maximum heating capacity threshold of the heating device and the maximum cooling capacity threshold of the cooling device, respectively. The maximum heating capacity threshold of the heating device and the maximum cooling capacity threshold of the cooling device can be obtained through pre-calibration experiments. For example, tests show that the heating film (heating device) can provide a heating rate of up to 5°C per minute to the battery cell at maximum power, and the liquid cooling system (cooling device) can provide a cooling rate of up to 3°C per minute at maximum flow rate. If the calculated ideal heating rate exceeds 5°C per minute, it indicates that the heating device is insufficient; if the ideal cooling rate exceeds 3°C per minute, it indicates that the cooling device is insufficient. Both of these situations mean that the current preset time window is too short, resulting in overly stringent trajectory requirements.

[0049] When the ideal rate of change is determined to exceed the maximum controllability threshold of the heating or cooling device, the system initiates an iterative adjustment mechanism. This involves extending the time window by a preset step size (e.g., by 60 seconds each time). After each extension, the reference trajectory function is reconstructed based on the new time window and the principle of minimizing energy dissipation. The newly calculated ideal rate of temperature change is then compared with the maximum controllability threshold of the heating or cooling device. This process is repeated until a time window length is found that ensures the ideal rate of temperature change at all times along the entire trajectory does not exceed the maximum controllability of the heating and cooling devices.

[0050] After the above iterative adjustments, the system obtains a smooth and achievable time-temperature curve, i.e., the predicted temperature change trajectory. In the subsequent model predictive control, the controller reads the expected temperature value corresponding to the current moment and compares it with the actual monitored temperature value. Based on the deviation, it dynamically adjusts the duty cycle of the heating and cooling devices to guide the actual temperature to run along this preset optimal trajectory, and finally converges precisely to the target equilibrium temperature at the end of the time window.

[0051] As one embodiment of the present invention, before dynamically adjusting the duty cycle of the heating device and the cooling device through the model predictive control algorithm, the method further includes a step of constructing the state vector of the model predictive control algorithm: The real-time temperature of each individual cell is collected within multiple consecutive sampling periods to form a temperature time series of each individual cell. If a single battery cell is located in the first low temperature region or the second high temperature region, the sliding window linear fitting algorithm is used to calculate its temperature change rate. If the single cell is located in the thermally affected transition zone between the first low temperature region and the second high temperature region, the Kalman filter algorithm is used to estimate the state of its temperature time series, separate the heating component driven by the heating device and the cooling component driven by the cooling device, and calculate the first rate component corresponding to the heating component and the second rate component corresponding to the cooling component respectively, and sum them to obtain its temperature change rate. The real-time temperature of each individual cell and the calculated rate of temperature change are used as the state vector of the model predictive control algorithm.

[0052] It should be noted that in this embodiment, the system reads the real-time temperature of each individual cell in the battery module at a fixed sampling period (e.g., once every 100 milliseconds) and maintains a fixed-length first-in-first-out queue in memory to store the historical records of the temperature read by each individual cell in the recent period, forming a temperature time series of each individual cell. In a specific embodiment, the queue length is set to 20 sampling points, that is, storing the temperature data of the most recent 2 seconds.

[0053] Before performing rate calculation, the system first needs to determine which region each individual cell is currently located in. Based on the boundary contour information of the first low-temperature region and the second high-temperature region identified in the previous steps, it determines whether the three-dimensional coordinate position of each individual cell falls within the boundary contour of the first low-temperature region or the second high-temperature region. If not, it further determines whether it is located in the middle zone between these two regions, i.e., the heat-affected zone. This embodiment provides a judgment criterion as follows: if the temperature of an individual cell is between the first temperature threshold and the second temperature threshold, and its distance from the centroid of the first low-temperature region and its distance from the centroid of the second high-temperature region are both less than a certain preset value, for example, both less than the thickness of 5 standard individual cells, then the individual cell is considered to be located in the heat-affected zone.

[0054] In this embodiment, for a single battery cell that is clearly located within the first low-temperature region or the second high-temperature region, it is assumed that its main temperature change is driven solely by the heating or cooling device. In this case, the system uses a sliding window linear fitting algorithm to calculate its temperature change rate. Specifically, this includes: taking the temperature time series of the most recent N (e.g., N=10) sampling periods of the single battery cell, using time as the abscissa and temperature as the ordinate, performing a univariate linear regression fitting on these N temperature data points to obtain a straight line that approximates these points, and the slope of the line is the temperature change rate of the battery cell at the current moment.

[0055] For a single battery cell located in the heat-affected zone, it is assumed that it is affected by both sides simultaneously. Therefore, the actual measured temperature change is the result of the superposition of heating and cooling effects. Directly calculating the combined rate of change cannot accurately reflect the individual effects of the heating and cooling devices. In this embodiment, a Kalman filter algorithm is used to estimate the state of the temperature time series of a single battery cell located in the heat-affected zone. Specifically, the system constructs a Kalman filter containing two state variables. The first state variable represents the heating component driven by the heating device, and the second state variable represents the cooling component driven by the cooling device. The superposition of these two state variables equals the actual measured temperature value. The state transition equation of the Kalman filter is constructed based on the basic laws of heat transfer: that is, the trend of the heating component is related to the current duty cycle of the heating device, and the trend of the cooling component is related to the current duty cycle of the cooling device.

[0056] In actual operation, the system inputs the historical duty cycle records of the heating device and cooling device, as well as the temperature time series of the individual battery cell located in the heat-affected zone, into the Kalman filter. The filter continuously adjusts the estimated values ​​of the heating and cooling components through iterative calculations, so that the sum of the two estimated components is as close as possible to the actual measured temperature value. After multiple iterations and convergence, the actual measured temperature value at the current moment is decomposed into two independent heating and cooling components. Then, the time series of these two components are differentiated to obtain the first rate component corresponding to the heating component and the second rate component corresponding to the cooling component. Finally, the two rate components are added together to obtain the comprehensive temperature change rate of the individual battery cell at the current moment.

[0057] After the above calculations, the system obtains two key values ​​for each individual cell: the real-time temperature at the current moment and the rate of temperature change at the current moment. These two key values ​​of all individual cells are combined to form a high-dimensional state vector. For example, for a battery module containing 96 cells, the dimension of the state vector is 192, which contains the current thermal state and thermal trend of the entire battery module. This state vector will be passed to the model predictive control algorithm as the initial condition for rolling optimization.

[0058] In one embodiment of the present invention, the model predictive control algorithm uses the duty cycle of the heating device and the duty cycle of the cooling device as control variables, the real-time temperature and temperature change rate of each individual cell as state variables, the tracking deviation minimization term and the control action smoothing term as objective functions, and the duty cycle amplitude limit and the duty cycle change rate limit as constraints.

[0059] The steps involved in dynamically adjusting the duty cycle of the heating and cooling devices using model predictive control algorithms include: The real-time temperature of each individual cell at the current moment is compared with the expected temperature value at the corresponding moment on the predicted temperature change trajectory, and the temperature deviation at the current moment is calculated. Using the state vector as the initial condition and the temperature deviation as the driving force, the optimal control sequence is solved in a rolling manner within the preset prediction time domain to minimize the deviation between the predicted temperature value and the expected temperature value at each future time in the prediction time domain. The first control variable in the optimal control sequence, namely the duty cycle of the heating device and the duty cycle of the cooling device in the current cycle, is output to the corresponding heating device and cooling device, respectively.

[0060] It should be noted that in this embodiment, at the beginning of each control cycle (e.g., every 100 milliseconds), the system first reads the state vector and obtains the real-time temperature value of each individual cell at the current moment. It also reads the expected temperature value at the current moment from the predicted temperature change trajectory. For each individual cell, the expected value is subtracted from its real-time temperature to obtain the temperature deviation at the current moment. If the real-time temperature is higher than the expected value, the temperature deviation is positive, indicating that cooling needs to be strengthened or heating needs to be weakened. If the real-time temperature is lower than the expected value, the temperature deviation is negative, indicating that heating needs to be strengthened or cooling needs to be weakened. This serves as the driving force for subsequent optimization calculations, guiding the controller to act in the direction of reducing the temperature deviation.

[0061] Subsequently, the system uses the current state vector as the initial condition and the current temperature deviation as the driving force, calls the previously established electro-thermal-aging coupling model, and performs forward extrapolation within the preset prediction time domain. In a specific embodiment, the prediction time domain is set to 20 steps, i.e., the next 2 seconds (the control cycle is 100 milliseconds). The system generates a set of candidate control sequences, i.e., the duty cycle settings of the heating and cooling devices within the next 20 control cycles. Then, this set of candidate sequences is input into the electro-thermal-aging coupling model to calculate the predicted temperature value of each individual cell at each moment within the next 2 seconds. These predicted temperature values ​​are compared with the expected values ​​at the corresponding moments on the predicted temperature change trajectory to evaluate the quality of the set of candidate control sequences.

[0062] The setting of the objective function and constraints specifically includes: The objective function consists of two parts: the first part is the tracking deviation minimization term, which is the sum of the squares of the deviations between the predicted temperature and the expected temperature of all cells at all times in the future prediction time domain. This is used to ensure that the main purpose of the control is to make the actual temperature follow the preset trajectory. The second part is the control action smoothing term, which is the sum of the squares of the duty cycle changes between adjacent control cycles. This is used to ensure the smoothness of the control action and avoid damage to the actuator or temperature oscillation caused by drastic duty cycle jumps. These two parts are multiplied by their respective weighting coefficients and then added together to form the total objective function value.

[0063] The constraints consist of two parts: the first constraint is the duty cycle amplitude limit, which means that the duty cycle of the heating and cooling devices can only be between 0% and 100%, serving as the physical limit for each device; the second constraint is the duty cycle change rate limit, which means that the change in duty cycle between two adjacent control cycles cannot exceed a certain preset value, for example, each change cannot exceed 10%, in order to protect the actuator and avoid drastic temperature fluctuations.

[0064] After the objective function and constraints are set, the system performs optimization in the prediction time domain. For example, the particle swarm optimization algorithm is used to quickly solve for the optimal control sequence in each control cycle, that is, the duty cycle sequence of the heating device and the duty cycle sequence of the cooling device that minimize the objective function in the next N control cycles. Each value in the sequence should satisfy the amplitude constraint and the rate of change constraint.

[0065] The system extracts the duty cycle of the first heating device and the duty cycle of the first cooling device from the optimal control sequence, and sends them to the corresponding driving circuits of the heating and cooling devices via PWM signals or CAN bus commands. After the duty cycle of the current control cycle is output and executed, the system waits for the next control cycle to arrive. At that time, the latest real-time temperature is read again and the temperature deviation is recalculated, the prediction model is called, the optimal control sequence is solved in the prediction time domain, and the control value is output again. The above process is repeated in each control cycle, forming a continuously rolling closed-loop control.

[0066] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for controlling the charging temperature of a battery module, characterized in that, The method includes: S1, monitor the real-time temperature of each individual cell in the battery module, construct a three-dimensional temperature field model containing temperature distribution field and local hot spot location information based on the real-time temperature, and identify a first low temperature region with a temperature lower than a preset first temperature threshold and a second high temperature region with a temperature higher than a preset second temperature threshold. S2, Based on the three-dimensional temperature field model, establish an electro-thermal-aging coupling model with the goal of minimizing charging time, temperature rise energy consumption and battery aging rate as the comprehensive optimization objective. Calculate the first heating power for the first low-temperature region, the second cooling power for the second high-temperature region, and the target equilibrium temperature at the current moment using a multi-objective genetic algorithm. S3, the first heating power and the second cooling power are used as feedforward control quantities and input to the three-dimensional temperature field model for simulation and deduction to obtain the predicted temperature change trajectory during the pre-adjustment process; S4. During the actual pre-adjustment process, the real-time temperature of each individual cell is monitored. Based on the deviation between the real-time temperature and the expected temperature value at the corresponding moment on the predicted temperature change trajectory, the working duty cycle of the heating device and the cooling device is dynamically adjusted through the model predictive control algorithm so that the temperatures of the first low-temperature region and the second high-temperature region converge to the target equilibrium temperature within a preset time window.

2. The charging temperature control method for a battery module according to claim 1, characterized in that: The steps of monitoring the real-time temperature of each individual cell within the battery module and constructing a three-dimensional temperature field model containing temperature distribution field and local hot spot location information based on the real-time temperature include: The real-time temperature of each individual cell at multiple sampling times is obtained and processed by Kalman filtering to estimate the predicted temperature of each individual cell at the next time step. Based on the real-time temperature, the predicted temperature, and the finite element heat transfer geometry model of the battery module, the boundary conditions and thermal property parameters of the finite element heat transfer geometry model are corrected by the extended Kalman filter algorithm to generate a dynamic thermal parameter set. The dynamic thermal parameter set is substituted into the finite element heat transfer geometric model for iterative solution to reconstruct the temperature value of the internal spatial point of the battery module at the current moment, and further generate the temperature distribution field and local hot spot location information to construct a three-dimensional temperature field model.

3. The charging temperature control method for a battery module according to claim 1, characterized in that: The step of identifying a first low-temperature region with a temperature below a preset first temperature threshold and a second high-temperature region with a temperature above a preset second temperature threshold includes: The sliding window clustering algorithm is used to traverse the spatial points of the three-dimensional temperature field model. The set of spatial points that are continuously distributed and whose temperature is lower than a preset first temperature threshold is marked as the first low temperature candidate area, and the set of spatial points that are continuously distributed and whose temperature is higher than a preset second temperature threshold is marked as the second high temperature candidate area. Calculate the regional volume, temperature variance, and heat flux density with adjacent regions for the first low-temperature candidate region and the second high-temperature candidate region, respectively. If the volume of a certain low-temperature candidate area exceeds a preset first volume threshold and its temperature variance is less than a preset uniformity threshold, it is confirmed as a first low-temperature area. If the volume of a certain high-temperature candidate area exceeds a preset second volume threshold and its heat flux density exceeds a preset heat flux threshold, it is confirmed as a second high-temperature area. Output the boundary contours and centroid coordinates of the first low-temperature region and the second high-temperature region.

4. The charging temperature control method for a battery module according to claim 1, characterized in that: The steps for establishing the electro-thermal-aging coupled model based on the three-dimensional temperature field model include: Based on the equivalent circuit model of the battery module and the three-dimensional temperature field model, an electro-thermal coupling sub-model is established to output the real-time temperature distribution. An aging sub-model is established based on a pre-defined semi-empirical aging model to quantify the battery aging rate under different thermal conditions, including the battery capacity decay rate and the battery internal resistance increase rate. The real-time temperature distribution output by the electro-thermal coupling sub-model is used as the input of the aging sub-model, and the battery aging rate output by the aging sub-model is fed back to the electro-thermal coupling sub-model to correct its internal resistance parameter, thus forming an electro-thermal-aging coupling model.

5. The charging temperature control method for a battery module according to claim 1, characterized in that: The step of calculating the first heating power for the first low-temperature region, the second cooling power for the second high-temperature region, and the target equilibrium temperature at the current moment using a multi-objective genetic algorithm includes: A population containing multiple individuals is initialized, and multiple combinations of candidate temperature control strategies are generated, including the first heating power, the second cooling power, and the target equilibrium temperature. Each candidate temperature control strategy combination is input into the electro-thermal-aging coupling model for simulation and deduction to obtain the predicted charging time, predicted temperature rise energy consumption and predicted battery aging rate for each candidate temperature control strategy combination. The Pareto front is used to stratify all individuals in the current population based on the non-dominated sorting algorithm, and the crowding distance of each individual in the same layer is calculated. All individuals are then sorted according to their layer and the crowding distance. Select a preset number of individuals from the sorted individuals as parents, and perform selection, crossover and mutation to generate offspring population. Repeat the iteration until a preset termination condition is reached to obtain the Pareto front solution set. The optimal solution is selected from the Pareto front solution set according to the preset multi-objective preference weights, and the corresponding first heating power, second cooling power and target equilibrium temperature are output.

6. The charging temperature control method for a battery module according to claim 1, characterized in that: Before inputting the first heating power and the second cooling power as feedforward control quantities into the three-dimensional temperature field model for simulation, the method further includes a step of correcting the first heating power and the second cooling power: Based on the three-dimensional temperature field model, the spatial adjacency relationship and heat conduction path between the first low-temperature region and the second high-temperature region are determined; If there is an overlapping area between the first low-temperature region and the second high-temperature region, or if the distance between them is less than a preset thermal interference threshold, then the coupled heat flux density generated at the overlapping area is calculated based on the spatial adjacency relationship. Based on the coupled heat flux density, when it represents the direction of coupled heat flux from the high temperature region to the low temperature region, the first heating power is reduced and the second cooling power is increased simultaneously according to a preset compensation coefficient. The modified first heating power and modified second cooling power are used as the feedforward control variables to perform simulation.

7. The charging temperature control method for a battery module according to claim 1, characterized in that: The predicted temperature change trajectory is generated through the following steps: A preset time window is used to obtain the target equilibrium temperature, the starting temperature of the first low temperature region, and the starting temperature of the second high temperature region at the start time of the time window. Based on the principle of minimizing energy dissipation, a reference trajectory function is constructed with time as the independent variable. The reference trajectory function satisfies the following boundary conditions: at the initial moment, the trajectory values ​​of the low-temperature region and the high-temperature region are equal to their initial temperatures, and at the final moment, the trajectory values ​​of the low-temperature region and the high-temperature region are both equal to the target equilibrium temperature. The ideal temperature change rate of each region at each moment within the time window is determined based on the reference trajectory function. If the ideal temperature change rate exceeds the maximum control capability threshold of the corresponding heating or cooling device, the length of the preset time window is adjusted, and a new reference trajectory function that satisfies the physical constraints of the actuator is reconstructed based on the principle of minimizing energy dissipation, until a feasible predicted temperature change trajectory is generated.

8. The charging temperature control method for a battery module according to claim 1, characterized in that: Before dynamically adjusting the duty cycle of the heating device and the cooling device using the model predictive control algorithm, the method further includes the step of constructing the state vector of the model predictive control algorithm: The real-time temperature of each individual cell is collected within multiple consecutive sampling periods to form a temperature time series of each individual cell. If a single battery cell is located in the first low-temperature region or the second high-temperature region, its temperature change rate is calculated using a sliding window linear fitting algorithm. If the single cell is located in the thermally affected transition zone between the first low temperature region and the second high temperature region, the Kalman filter algorithm is used to estimate the state of its temperature time series, separate the heating component driven by the heating device and the cooling component driven by the cooling device, and calculate the first rate component corresponding to the heating component and the second rate component corresponding to the cooling component respectively, and sum them to obtain its temperature change rate. The real-time temperature of each individual cell and the calculated rate of temperature change are used as the state vector of the model predictive control algorithm.

9. The charging temperature control method for a battery module according to claim 8, characterized in that: The model predictive control algorithm uses the duty cycle of the heating device and the duty cycle of the cooling device as control variables, the real-time temperature and temperature change rate of each individual cell as state variables, the tracking deviation minimization term and the control action smoothing term as objective functions, and the duty cycle amplitude limit and the duty cycle change rate limit as constraints.

10. The charging temperature control method for a battery module according to claim 8, characterized in that: The step of dynamically adjusting the duty cycle of the heating device and the cooling device using a model predictive control algorithm includes: The real-time temperature of each individual cell at the current moment is compared with the expected temperature value at the corresponding moment on the predicted temperature change trajectory, and the temperature deviation at the current moment is calculated. Using the state vector as the initial condition and the temperature deviation as the driving force, the optimal control sequence is solved in a rolling manner within the preset prediction time domain to minimize the deviation between the predicted temperature value and the expected temperature value at each future time in the prediction time domain. The first control variable in the optimal control sequence, namely the duty cycle of the heating device and the duty cycle of the cooling device in the current cycle, is output to the corresponding heating device and cooling device, respectively.

Images