AI-based charging process dynamic temperature control method and system

Abstract

The invention belongs to an AI-based charging process dynamic temperature control method and system, and the method comprises the steps: collecting various data, and obtaining the heat dissipation difference data of a battery according to the arrangement channel difference data and the aging state of a battery pack structure. And the attenuation influence degree of the cooling medium on the battery performance is analyzed, the charging related data is analyzed to obtain transient electromagnetic interference, and transient heat dissipation influence factors are obtained in combination with the heat dissipation difference data of the battery. And analyzing the interaction between the battery heat dissipation difference data and the battery performance data, constructing an AI temperature prediction model in combination with the attenuation influence degree, inputting real-time battery data, and outputting a predicted temperature. Adjusting the temperature regulation and control strategy according to the transient heat dissipation influence factors to obtain a dynamic temperature adjustment strategy; according to the invention, the safety and service life of the battery can be improved, the charging efficiency and physical examination are optimized, and intelligent and adaptive management is realized.

Description

Technical Field

[0001] This invention belongs to the field of temperature control technology, specifically an AI-based dynamic temperature control method and system for the charging process. Background Technology

[0002] With the rapid development of new energy and other fields, lithium-ion batteries have become core energy storage components due to their advantages such as high energy density and long cycle life. The charging process, as a critical link in the entire battery lifecycle, directly determines battery safety, cycle life, and charging efficiency through temperature control. Excessive temperature can easily lead to thermal runaway risks, uneven temperature can accelerate localized aging, while low temperature or delayed temperature control can cause a sharp drop in charging efficiency. Therefore, precise and dynamic charging temperature control technology has become a core research and development direction in the new energy field.

[0003] Current battery charging temperature control technologies are mostly based on preset thresholds or simplified thermal models, resulting in incomplete quantification of heat dissipation differences and a lack of targeted temperature control strategies. Traditional methods often use fixed heat dissipation parameters, ignoring the structural differences of the battery itself and its degradation characteristics throughout its life cycle: different cell arrangements create vastly different heat diffusion paths, and long-term use of the battery pack leads to issues such as casing aging and dust accumulation in the heat dissipation channels, causing the actual heat dissipation efficiency to decrease compared to newer batteries. This "one-size-fits-all" quantification approach makes temperature control strategies unsuitable for different battery types and aging stages. For example, overcooling tightly packed cell groups results in wasted energy, while insufficient heat dissipation assessment in aging batteries can lead to localized overheating.

[0004] The sensitivity to transient electromagnetic interference (EMI) during charging is too low, resulting in lag and misjudgment in temperature control response. In battery charging systems, transient EMI can be generated by grid surges, high-frequency switching of charging piles, and the bouncing of charging guns during insertion and removal. Although short in duration, these transient EMI can cause temperature sensor signal distortion, CAN bus communication interruption, or actuator malfunction. Existing temperature control technologies lack a mechanism linking EMI to the heat dissipation system, relying solely on lagging, measured temperature adjustment strategies.

[0005] In summary, in view of the problems in the existing technology, there is an urgent need for an AI-based dynamic temperature control method and system for the charging process, so as to achieve safe, efficient and intelligent temperature control during the charging process. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention proposes an AI-based dynamic temperature control method and system for the charging process. This invention primarily addresses the problems of incomplete quantification of heat dissipation differences and insensitivity to transient disturbances during charging.

[0007] The technical solution adopted by this invention to solve its technical problem is: the AI-based dynamic temperature control method for the charging process provided by this invention, comprising: We collect battery type data, battery performance data, and charging-related data. We analyze the battery type data from the cell arrangement and heat diffusion path to obtain arrangement channel difference data, and combine it with the battery pack structure aging state to obtain battery heat dissipation difference data.

[0008] The influence of the cooling medium on the degradation of battery performance was analyzed, and transient electromagnetic interference was obtained by analyzing charging-related data. The factors affecting transient heat dissipation were obtained by combining battery heat dissipation difference data.

[0009] The interaction between battery heat dissipation difference data and battery performance data is analyzed using the heat dissipation effect feedback method. An AI temperature prediction model is constructed by combining the degree of degradation influence. Real-time battery data is input and predicted temperature is output.

[0010] A temperature control strategy is formulated based on battery heat dissipation difference data and predicted temperature. The temperature control strategy is then adjusted based on transient heat dissipation influencing factors to obtain a dynamic temperature regulation strategy.

[0011] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for obtaining channel difference data: The cell arrangement is determined based on the battery type, the arrangement geometry parameters are quantified, the gap design details are recorded to obtain the cell arrangement design parameters, and the thermal management channel parameters are collected based on the thermal management type and channel structure.

[0012] The arrangement density is calculated based on the cell arrangement design parameters, the arrangement symmetry and gap uniformity are extracted, and the arrangement difference index is obtained by correlating the tab arrangement.

[0013] A battery pack thermal model is constructed based on cell arrangement design parameters and thermal management channel parameters. The dominant heat diffusion path is identified, the thermal resistance distribution of the path is quantified, and the heat diffusion uniformity is analyzed to obtain the hot spot focusing characteristics.

[0014] By combining the flow velocity distribution, fluid resistance coefficient, and the matching degree between the flow velocity distribution and the thermal diffusion requirements of the battery cell under different arrangement methods, and by combining the hot spot focusing characteristics, the channel arrangement matching differences are obtained.

[0015] The permutation difference index and the channel permutation matching difference are integrated to obtain the permutation channel difference data.

[0016] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for obtaining battery heat dissipation difference data: The structure that affects heat dissipation is defined as the aging heat dissipation structure. The aging index of each component of the aging heat dissipation structure is calculated with the decay of heat dissipation function as the core, and the aging level is generated.

[0017] For different aging levels of heat dissipation structures, the mechanism that damages the heat dissipation logic of the arrangement-thermal management channel is dissected, and the mapping relationship between aging parameters and heat dissipation degradation is obtained by adapting and correcting the battery type.

[0018] The ideal heat dissipation efficiency without aging is calculated based on the difference data of the arrangement channels, and the actual heat dissipation efficiency under aging is calculated by combining the component aging index.

[0019] Based on the same type of battery and no aging state, the heat dissipation difference index under different aging levels is calculated to obtain battery heat dissipation difference data.

[0020] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for determining the degree of temperature degradation: Determine the target medium type, record the corresponding basic medium attributes, and define battery performance degradation indicators and set control variables.

[0021] Different control groups were designed in groups to simulate actual usage scenarios, setting cyclic operating conditions, test cycles, and media control parameters.

[0022] Capacity, internal resistance, cycle life, and rate performance of each experiment are collected as performance parameters, and medium state and temperature are collected as medium environment correlation data.

[0023] Based on the performance parameters, the capacity decay impact coefficient, internal resistance growth suppression rate, and lifetime extension factor are calculated to construct a comprehensive index of decay impact.

[0024] The correlation between media environment data and the comprehensive index of attenuation impact was analyzed using Pearson correlation coefficient, and the degree of attenuation impact was obtained by combining the impact mechanism.

[0025] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for obtaining transient heat dissipation influencing factors: Electrical parameters, communication data, sensor signals, and transient radiation are collected as charging-related data and standardized to obtain standard charging data.

[0026] When standard charging data reaches a preset multi-dimensional trigger threshold, it is determined to be a transient event, and classified as a transient feature event based on the interference source and characteristics.

[0027] For each transient characteristic event, the peak intensity, duration, rise rate, energy density, communication interference rate, and sensor distortion rate of transient electromagnetic interference are extracted as transient electromagnetic interference.

[0028] The interference path of transient electromagnetic interference on the heat dissipation system is determined from the perception layer, control layer and execution layer. Combined with battery heat dissipation difference data, the dynamic heat dissipation impact is calculated.

[0029] Based on the dynamic heat dissipation impact analysis, multiple transient heat dissipation factors are obtained from the analysis of heat dissipation efficiency coupling attenuation, temperature perception distortion, actuator response lag and thermal path blockage. Combined with the battery heat dissipation safety requirements, the factor weights are set to obtain the transient heat dissipation influencing factors.

[0030] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps in constructing an AI temperature prediction model: Analysis of battery heat dissipation differences shows that they have a positive impact on battery performance data, specifically on the effects of heat dissipation efficiency on capacity decay, the effects of blocked thermal paths on internal resistance growth, and the effects of heat dissipation response delay on rate performance.

[0031] Based on the degree of attenuation, the influence of internal resistance growth on heat generation power, the correction of heat dissipation demand by increased heat generation, and the amplification factor of performance degradation on heat dissipation difference are analyzed inversely.

[0032] By integrating positive and negative influences, we can obtain mutual influence effects and construct sliding window features, trend features, and operating condition switching features.

[0033] A basic framework model based on LSTM and attention mechanism is constructed. The interaction effects are divided into training set, test set and validation set according to a preset ratio. The basic framework model is then trained with parameters, and the AI ​​temperature prediction model is obtained by optimizing the transfer learning and generalization ability.

[0034] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for obtaining the predicted temperature: Real-time battery data, real-time heat dissipation system data, and real-time environmental condition data are collected as real-time battery data. Based on the statistical parameters during model training, lightweight standardization is performed and converted into tensor format required by the AI ​​temperature prediction model.

[0035] Based on the preprocessed real-time battery data, heat dissipation performance coupling features, time-series trend features, and operating condition labeling features are extracted as the model input feature set.

[0036] The real-time model input feature set is input into the AI ​​temperature prediction model, triggering the inference mechanism and outputting the predicted value. The temperature of the heat dissipation structure, surface dimension and hot spot peak temperature at a preset time interval in the future are extracted as the predicted temperature.

[0037] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps in formulating a temperature regulation strategy: The control targets and constraints are determined based on the cell's safe temperature, target maximum charging efficiency, and heat dissipation adaptation.

[0038] The control scenarios are divided based on the predicted temperature state, heat dissipation attenuation level and operating condition type, and the priority index is calculated according to the temperature risk priority, heat dissipation attenuation priority and operating condition risk priority.

[0039] The temperature control strategy is obtained by determining the control direction and control strategy corresponding to different control scenarios based on the priority index.

[0040] The AI-based dynamic temperature control method for the charging process provided by this invention includes the following steps for adjusting the dynamic temperature regulation strategy: Based on real-time collected charging-related parameters, the transient heat dissipation influencing factors are updated, and the factor priority coefficients are calculated using a weighted summation method.

[0041] Compare the actual temperature change after the temperature control strategy is implemented with the expected target to determine whether the preset safety threshold has been reached; otherwise, formulate adjustment directions based on the factors affecting transient heat dissipation.

[0042] For transient interference scenarios, dynamic adjustment rules are formulated based on the adjustment direction and factor priority coefficients to adjust the temperature control strategy and obtain the temperature regulation strategy.

[0043] The present invention provides an AI-based dynamic temperature control system for the charging process, comprising: The heat dissipation difference construction module is used to collect battery type data, battery performance data, and charging-related data. It analyzes the cell arrangement and heat diffusion path of the battery type to obtain the arrangement channel difference data, and combines the battery pack structure aging state to obtain battery heat dissipation difference data.

[0044] The transient impact analysis module is used to analyze the degree of impact of the cooling medium on the degradation of battery performance. It analyzes charging-related data to obtain transient electromagnetic interference and combines it with battery heat dissipation difference data to obtain transient heat dissipation influencing factors.

[0045] The AI ​​temperature prediction module is used to analyze the interaction between battery heat dissipation difference data and battery performance data using the heat dissipation effect feedback method. It constructs an AI temperature prediction model by combining the degree of degradation influence, inputs real-time battery data, and outputs the predicted temperature.

[0046] The temperature control strategy formulation module is used to formulate a temperature control strategy based on battery heat dissipation difference data and predicted temperature, and to adjust the temperature control strategy according to transient heat dissipation influencing factors to obtain a dynamic temperature regulation strategy.

[0047] The beneficial effects of this invention are as follows: 1. This invention effectively reduces the risk of battery overheating and thermal runaway by predicting peak temperatures in advance and actively controlling them. Combined with management of the performance degradation coupling relationship, it slows down the rate of capacity and internal resistance degradation, extending the overall service life. By considering practical factors such as transient electromagnetic interference and sensor distortion, and designing corresponding adjustment mechanisms, the reliability and fault tolerance of the system under complex electromagnetic environments and imperfect measurement conditions are improved. It can dynamically adjust management strategies according to individual differences in battery packs, aging states, and real-time operating conditions, achieving truly intelligent thermal management. Attached Figure Description

[0048] The invention will now be further described with reference to the accompanying drawings.

[0049] Figure 1 This is one of the flowcharts of the AI-based dynamic temperature control method for the charging process provided in this embodiment of the invention; Figure 2 This is the second flowchart of the AI-based dynamic temperature control method for the charging process provided in this embodiment of the invention. Figure 3 This is a schematic diagram of a module of an AI-based dynamic temperature control system for the charging process provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0051] like Figures 1 to 3 As shown in the figure, the AI-based dynamic temperature control method for the charging process provided in this embodiment of the invention includes: We collect battery type data, battery performance data, and charging-related data. We analyze the battery type data from the cell arrangement and heat diffusion path to obtain arrangement channel difference data, and combine it with the battery pack structure aging state to obtain battery heat dissipation difference data.

[0052] The steps to obtain the channel difference data include: The cell arrangement is determined based on the battery type, the arrangement geometry parameters are quantified, the gap design details are recorded to obtain the cell arrangement design parameters, and the thermal management channel parameters are collected based on the thermal management type and channel structure.

[0053] Battery types can include cylindrical cells, prismatic cells, or pouch cells, etc.

[0054] The arrangement can include: cylindrical cells arranged in a honeycomb / matrix / series-parallel combination; square cells arranged in a parallel / stacked / modular configuration; and pouch cells arranged in a bonded / staggered / gap configuration.

[0055] The geometric parameters of the arrangement are quantified based on the center-to-center distance / surface spacing between adjacent cells, the number of layers and columns, the distribution density of cells in the battery pack, and the relative position of the tabs to the center of the cells.

[0056] Based on the type of filling material, filling thickness, and gap uniformity design of the cells, the gap design details are recorded.

[0057] Thermal management types can be categorized as: air cooling, liquid cooling, phase change cooling, or hybrid cooling systems.

[0058] Corresponding channel structure parameters: Air-cooled: cross-sectional area of ​​the air duct, number and location of air outlets / inlets, alignment of the air duct with the battery cell, and fan installation position.

[0059] Liquid cooling: width, depth, number of corners and radius of curvature of liquid cooling plate flow channels, location of coolant inlet and outlet, and flow channel distribution density.

[0060] Phase change cooling: PCM filling area, thickness, packaging structure and contact method with the cell.

[0061] The arrangement density is calculated based on the cell arrangement design parameters, the arrangement symmetry and gap uniformity are extracted, and the arrangement difference index is obtained by correlating the tab arrangement.

[0062] Calculate the arrangement density: Arrangement density = total effective volume of cells / volume of battery pack × 100%, and calculate the volume utilization rate of different arrangement methods.

[0063] Analyze gap uniformity: Measure the gap size between all adjacent cells, calculate the gap standard deviation, and mark potential heat accumulation areas with gaps <1mm.

[0064] Extracting the symmetry of the arrangement: Using the geometric center of the battery pack as a reference, the symmetry deviation of the cell arrangement is calculated to evaluate the impact of the symmetrical design on the uniformity of heat diffusion.

[0065] The steps for aligning associated electrodes may include: Calculate the current path length: Based on the position of the electrode and the arrangement of the cells, calculate the current conduction path length from the cell to the electrode in each region, and obtain the standard deviation of the path length.

[0066] Analyze the uniformity of current distribution: Measure the current distribution ratio of each tab in the multi-tab structure through simulation or experiment, and calculate the current distribution deviation rate.

[0067] Correlation and Current Heat Generation: Combining the differences in current path length and resistance distribution, the differences in local heat generation rates under different arrangement methods are quantified.

[0068] A battery pack thermal model is constructed based on cell arrangement design parameters and thermal management channel parameters. The dominant heat diffusion path is identified, the thermal resistance distribution of the path is quantified, and the heat diffusion uniformity is analyzed to obtain the hot spot focusing characteristics.

[0069] A full-size battery pack thermal model can be built using tools such as ANSYS / COMSOL. Parameters such as the thermal conductivity, specific heat capacity, and heat dissipation coefficient of the cells, casing, thermal management channels, and filling materials can be clearly defined.

[0070] Boundary conditions are set for actual operating conditions such as heat generation rate, ambient temperature, and thermal management system operating power during the simulated charging process.

[0071] By simulating the heat flow distribution, we can distinguish between three paths: "cell-to-cell direct conduction", "cell-to-thermal management channel conduction", and "cell-to-casing-environment conduction". We can calculate the heat flow ratio of each path and identify the dominant heat diffusion path.

[0072] The contact thermal resistance, conduction thermal resistance, and convection thermal resistance of each heat diffusion path are calculated segment by segment to obtain the total thermal resistance and the proportion of each segment, and the heat diffusion bottleneck area with thermal resistance >5K / W is marked.

[0073] Extract the surface temperature distribution cloud map of the battery cell, calculate the maximum temperature difference and average temperature difference within the battery pack, and evaluate the differences in the uniformity of heat diffusion under different arrangement methods.

[0074] Areas with temperatures exceeding 3°C above the average cell temperature were identified as hotspots. Simulations were used to obtain the location, area percentage, and temperature rise rate of these hotspots.

[0075] Comparing the hotspot risks of different arrangements: Under the same charging conditions, the hotspot aggregation probability of cylindrical, square, and pouch cells with different arrangements is simulated to form hotspot focusing characteristics.

[0076] By combining the flow velocity distribution, fluid resistance coefficient, and the matching degree between the flow velocity distribution and the thermal diffusion requirements of the battery cell under different arrangement methods, and by combining the hot spot focusing characteristics, the channel arrangement matching differences are obtained.

[0077] The permutation difference index and the channel permutation matching difference are integrated to obtain the permutation channel difference data.

[0078] The data on channel arrangement differences can include arrangement density, gap standard deviation, symmetry deviation, current path length standard deviation, hot spot aggregation probability, volume utilization rate, channel coverage area ratio, flow velocity uniformity, fluid resistance coefficient, heat exchange efficiency, and adaptation deviation index.

[0079] The steps to obtain battery heat dissipation difference data include: The structure that affects heat dissipation is defined as the aging heat dissipation structure. The aging index of each component of the aging heat dissipation structure is calculated with the decay of heat dissipation function as the core, and the aging level is generated.

[0080] The aging heat dissipation structure may include heat insulation / thermal conductive filling materials, sealing structures, heat dissipation brackets / fixtures, thermal management channel bodies, etc.

[0081] The quantitative parameters of each aging component are normalized and weighted according to the component's influence on heat dissipation, such as 0.3 for the thermal management channel body, 0.2 for the sealing structure, 0.2 for the filling material, 0.15 for the bracket, and 0.15 for the actuator, to obtain the aging index of a single component.

[0082] Based on the summation of component-level aging indices, mild aging (total index ≤ 0.3), moderate aging (0.3 < total index ≤ 0.6), and severe aging (total index > 0.6) are classified, and the core attenuation parameters corresponding to the aging levels are recorded.

[0083] For different aging levels of heat dissipation structures, the mechanism that damages the heat dissipation logic of the arrangement-thermal management channel is dissected, and the mapping relationship between aging parameters and heat dissipation degradation is obtained by adapting and correcting the battery type.

[0084] Filler material displacement / aging: disrupts the heat conduction path between cells, leading to increased local thermal resistance → core impact "thermal diffusion uniformity", quantified as thermal resistance increment coefficient = 1 + (displacement area ratio × 0.8 + thermal conductivity deviation rate × 0.2).

[0085] Cracks in the sealing structure: External environment intrusion blocks the gaps in the heat dissipation channels and disrupts the stability of the internal thermal environment of the battery pack → The core impact is on the "channel heat dissipation efficiency", which is quantified as the channel blockage increment coefficient = 1 + (average crack gap × 5 + sealing leakage rate × 0.1).

[0086] Loose brackets: This leads to misalignment of the battery cells and poor contact between the thermal management channel and the battery cells → the core impact is "contact thermal resistance", which is quantified as the contact thermal resistance increment coefficient = 1 + (bracket looseness amount × 0.6 + battery cell fixing deviation × 0.4).

[0087] Aging of thermal management channels / actuators: directly reduces the rated capacity of the heat dissipation system → core impact "heat exchange power", quantified as heat dissipation capacity attenuation coefficient = 1 - (flow channel blockage rate × 0.4 + dust accumulation coverage rate × 0.3 + speed / flow rate attenuation rate × 0.3).

[0088] For cylindrical cells: narrow gaps and sealing cracks lead to more significant dust accumulation and blockage, resulting in a channel blockage increment coefficient of 1.2.

[0089] Square cells: Larger contact area, the increase in contact thermal resistance caused by loose support has a more significant impact → contact thermal resistance increment coefficient ×1.3.

[0090] Soft-pack battery cells: The impact of heat conduction interruption caused by aging of filling materials is more significant → thermal resistance increment coefficient ×1.1.

[0091] The "fit deviation index" is based on the arrangement and thermal management channels. The larger the fit deviation, the more obvious the heat dissipation degradation caused by aging. The correction coefficient is 1 + 0.5 × fit deviation index.

[0092] The ideal heat dissipation efficiency without aging is calculated based on the difference data of the arrangement channels, and the actual heat dissipation efficiency under aging is calculated by combining the component aging index.

[0093] Ideal heat dissipation efficiency = channel heat exchange efficiency × (1 - adaptation deviation index) × arrangement heat dissipation gain coefficient.

[0094] Actual heat dissipation efficiency = ideal heat dissipation efficiency × heat dissipation capacity attenuation coefficient × (1 / thermal resistance increment coefficient) × (1 / contact thermal resistance increment coefficient) × (1 / channel blockage increment coefficient).

[0095] Calculate the actual heat dissipation parameters after aging: Actual temperature rise rate = ideal temperature rise rate × (1 + temperature rise amplification factor corresponding to aging level).

[0096] Actual core-surface temperature difference = ideal core-surface temperature difference × (1 + thermal resistance increment coefficient × 0.6).

[0097] Actual heat exchange path obstruction coefficient = (ideal thermal resistance - actual thermal resistance) / ideal thermal resistance × 100% (actual thermal resistance = ideal thermal resistance × thermal resistance increment coefficient × contact thermal resistance increment coefficient).

[0098] Based on the same type of battery and no aging state, the heat dissipation difference index under different aging levels is calculated to obtain battery heat dissipation difference data.

[0099] Heat dissipation efficiency difference rate = (baseline heat dissipation efficiency - actual heat dissipation efficiency) / baseline heat dissipation efficiency × 100%.

[0100] Temperature rise rate difference rate = (actual temperature rise rate - reference temperature rise rate) / reference temperature rise rate × 100%.

[0101] Temperature difference rate = (actual core - surface temperature difference - reference core - surface temperature difference) / reference core - surface temperature difference × 100%.

[0102] The influence of the cooling medium on the degradation of battery performance was analyzed, and transient electromagnetic interference was obtained by analyzing charging-related data. The factors affecting transient heat dissipation were obtained by combining battery heat dissipation difference data.

[0103] The steps to determine the degree of attenuation effect include: Determine the target medium type, record the corresponding basic medium attributes, and define battery performance degradation indicators and set control variables.

[0104] Target media types may include: air cooling (air), liquid cooling (ethylene glycol aqueous solution, silicone oil, special coolant), and phase change cooling (paraffin-based / composite phase change materials PCM, etc.). Basic media properties may include thermal conductivity, specific heat capacity, viscosity, phase change temperature / latent heat, and chemical stability.

[0105] Battery performance degradation metrics may include capacity degradation rate, internal resistance growth rate, cycle life degradation rate, and rate performance degradation rate.

[0106] Control variables settings: fix the battery type, cell specifications, charging rate, ambient temperature, and heat dissipation system structure, and only change the type or state of the cooling medium.

[0107] Different control groups were designed in groups to simulate actual usage scenarios, setting cyclic operating conditions, test cycles, and media control parameters.

[0108] Cyclic operating condition settings: A "charge-rest-discharge" cycle is adopted. The charging mode is constant current constant voltage (CC-CV), and the discharging mode is constant current (CC). Specific parameters: Charging: 0.05C precharge to 3.0V → target rate charge to cutoff voltage → constant voltage charge to current ≤0.02C.

[0109] Let stand for 30 minutes.

[0110] Discharge: Discharge at the target rate to the cutoff voltage.

[0111] Test cycle: Total number of cycles ≥ 1000, pause the test every 100 cycles to collect performance parameters.

[0112] Media control parameters: Air cooling: fixed wind speed (2m / s), unchanged air duct structure.

[0113] Liquid cooling: fixed flow rate (10L / min), inlet and outlet temperature difference controlled within 5℃.

[0114] Phase change cooling: With a fixed PCM filling amount, the phase change temperature matches the battery's optimal operating temperature.

[0115] Capacity, internal resistance, cycle life, and rate performance of each experiment are collected as performance parameters, and medium state and temperature are collected as medium environment correlation data.

[0116] Capacity test: After every 100 cycles, charge and discharge at a low rate of 0.2C, record the full charge capacity, and calculate the capacity decay rate = (initial capacity - current capacity) / initial capacity × 100%.

[0117] Internal resistance test: The AC impedance method (frequency 1kHz) is used to test the internal resistance of the positive and negative electrodes and the polarization internal resistance of the battery. The internal resistance growth rate is calculated as (current internal resistance - initial internal resistance) / initial internal resistance × 100%.

[0118] Cyclic life assessment: When the capacity decay rate reaches 20%, stop the cycle test, record the cumulative number of cycles, and calculate the life decay rate = (standard life - actual life) / standard life × 100% (standard life is the number of cycles in the blank control group).

[0119] Rate performance test: After every 200 cycles, discharge at rates of 0.5C, 1C, 2C and 3C respectively, and record the discharge capacity retention rate at different rates.

[0120] Thermophysical properties: Thermal conductivity, specific heat capacity, and viscosity of the medium are tested every 200 cycles.

[0121] Chemical stability: Detects whether the medium exhibits oxidation, decomposition, or corrosion products.

[0122] Temperature data: Real-time acquisition of battery core temperature, surface temperature, and inlet and outlet temperatures of cooling medium; calculation of maximum temperature difference within the battery pack; medium heat exchange efficiency = (exit heat of medium - inlet heat) / battery heat generation × 100%.

[0123] Based on the performance parameters, the capacity decay impact coefficient, internal resistance growth suppression rate, and lifetime extension factor are calculated to construct a comprehensive index of decay impact.

[0124] The formula for calculating the capacity decay index is expressed as: In the formula, It is the capacity decay effect coefficient. This is the capacity decay rate of the blank control group. It is the capacity decay rate of the cooling medium group.

[0125] The formula for calculating the internal resistance growth inhibition rate is expressed as: In the formula, It is the internal resistance growth inhibition rate. This is the internal resistance growth rate of the blank control group. It is the growth rate of the internal resistance of the cooling medium group.

[0126] The formula for calculating the lifespan extension factor is expressed as follows: In the formula, This is the cycle life of the standard control group. It is the cycle life of the cooling medium assembly. It is the lifespan extension factor.

[0127] The correlation between media environment data and the comprehensive index of attenuation impact was analyzed using Pearson correlation coefficient, and the degree of attenuation impact was obtained by combining the impact mechanism.

[0128] The mechanisms of influence can be divided into direct influence paths and indirect influence paths.

[0129] Direct impact path analysis: Impact of temperature uniformity: Calculate the maximum temperature difference of battery packs with different dielectric groups, and establish a linear regression model between temperature difference and capacity decay rate.

[0130] Impact of media aging: Compare the differences in the comprehensive impact index between new media and aged media, and calculate the rate of decrease in the attenuation inhibition effect caused by media aging.

[0131] Indirect impact path analysis: Impact of dielectric corrosion: Detect the degree of corrosion of battery tabs and casing, and correlate it with the growth rate of internal resistance.

[0132] Impact of medium leakage / phase change failure: Record the amount of medium leakage or the latent heat decay rate of PCM, analyze its probability of inducing local overheating, and then associate it with the risk of sudden battery degradation.

[0133] Impact Level Classification Table:

[0134] The steps to obtain the influencing factors of transient heat dissipation include: Electrical parameters, communication data, sensor signals, and transient radiation are collected as charging-related data and standardized to obtain standard charging data.

[0135] Electrical parameter acquisition: Continuously acquire charging voltage and current, with a focus on capturing transient signals such as voltage spikes and current pulses.

[0136] Communication data acquisition: Parse CAN bus data frames to extract heat dissipation system control commands and sensor data transmission frames.

[0137] Raw sensor signal acquisition: Acquire the raw analog signal from the temperature sensor, preserving the signal distortion characteristics caused by EMI interference.

[0138] EMI radiation acquisition: The electromagnetic radiation spectrum around the battery pack is acquired using an EMI receiver, and the frequency, peak value, and duration of transient interference are recorded.

[0139] A wavelet threshold denoising algorithm is employed to separate "normal operating signals" from "transient EMI interference signals" in the charging data, retaining the abnormal fluctuation components caused by EMI. Using the rising edge of the charging current as the reference time point, the timestamps of voltage, EMI signals, temperature data, and actuator commands are uniformly calibrated to the same time axis to ensure the temporal correlation of transient events. Non-EMI outliers in the charging data are removed based on the 3σ criterion, and missing data segments are supplemented using linear interpolation.

[0140] When standard charging data reaches a preset multi-dimensional trigger threshold, it is determined to be a transient event, which is further classified into transient characteristic events based on the interference source and characteristics. Transient characteristic events can be categorized into grid surge type, charging equipment switching type, and interface contact type.

[0141] For each transient characteristic event, the peak intensity, duration, rise rate, energy density, communication interference rate, and sensor distortion rate of transient electromagnetic interference are extracted as transient electromagnetic interference.

[0142] The interference path of transient electromagnetic interference on the heat dissipation system is determined from the perception layer, control layer and execution layer. Combined with battery heat dissipation difference data, the dynamic heat dissipation impact is calculated.

[0143] Based on the dynamic heat dissipation impact analysis, multiple transient heat dissipation factors are obtained from the analysis of heat dissipation efficiency coupling attenuation, temperature perception distortion, actuator response lag and thermal path blockage. Combined with the battery heat dissipation safety requirements, the factor weights are set to obtain the transient heat dissipation influencing factors.

[0144] The interaction between battery heat dissipation difference data and battery performance data is analyzed using the heat dissipation effect feedback method. An AI temperature prediction model is constructed by combining the degree of degradation influence. Real-time battery data is input and predicted temperature is output.

[0145] The steps to build an AI temperature prediction model include: Analysis of battery heat dissipation differences shows that they have a positive impact on battery performance data, specifically on the effects of heat dissipation efficiency on capacity decay, the effects of blocked thermal paths on internal resistance growth, and the effects of heat dissipation response delay on rate performance.

[0146] Based on the degree of attenuation, the influence of internal resistance growth on heat generation power, the correction of heat dissipation demand by increased heat generation, and the amplification factor of performance degradation on heat dissipation difference are analyzed inversely.

[0147] By integrating positive and negative influences, we can obtain mutual influence effects and construct sliding window features, trend features, and operating condition switching features.

[0148] Construct sliding window features: Using 50ms as the window, extract the mean, variance, maximum temperature rise rate, and fluctuation range of heat dissipation efficiency within the window.

[0149] Constructing trend characteristics: By linear fitting, the slope of the change in the effect of heat dissipation efficiency on capacity decay in the past 1 second is calculated to predict the evolution trend of performance decay and heat dissipation difference.

[0150] Construct operating condition switching features: mark charging mode switching, sudden changes in ambient temperature, and EMI interference events as operating condition adaptive trigger signals for the model.

[0151] A basic framework model based on LSTM and attention mechanism is constructed. The interaction effects are divided into training set, test set and validation set according to a preset ratio. The basic framework model is then trained with parameters, and the AI ​​temperature prediction model is obtained by optimizing the transfer learning and generalization ability.

[0152] The training set, validation set, and test set are divided in a 7:2:1 ratio.

[0153] Optimizer: AdamW.

[0154] Learning rate: Initial learning rate = 1e-3, using cosine annealing strategy.

[0155] Batch size: 32 (adapted to edge computing unit memory).

[0156] Number of iterations: 500 epochs.

[0157] Transfer learning and generalization optimization can include: Pre-training: The base model is trained based on a large amount of standard laboratory data to ensure baseline accuracy.

[0158] Fine-tuning: Fine-tune the model parameters using real-world scenario data, focusing on optimizing prediction accuracy under extreme conditions.

[0159] Domain Adaptation: Adversarial training is employed to reduce the distributional differences between laboratory data and real-world scenario data, thereby improving the model's generalization ability under different battery types and different cooling systems.

[0160] The steps to obtain the predicted temperature include: Real-time battery data, real-time heat dissipation system data, and real-time environmental condition data are collected as real-time battery data. Based on the statistical parameters during model training, lightweight standardization is performed and converted into tensor format required by the AI ​​temperature prediction model.

[0161] Based on the preprocessed real-time battery data, heat dissipation performance coupling features, time-series trend features, and operating condition labeling features are extracted as the model input feature set.

[0162] The real-time model input feature set is input into the AI ​​temperature prediction model, triggering the inference mechanism and outputting the predicted value. The temperature of the heat dissipation structure, surface dimension and hot spot peak temperature at a preset time interval in the future are extracted as the predicted temperature.

[0163] A temperature control strategy is formulated based on battery heat dissipation difference data and predicted temperature. The temperature control strategy is then adjusted based on transient heat dissipation influencing factors to obtain a dynamic temperature regulation strategy.

[0164] The steps involved in developing a temperature control strategy include: The control targets and constraints are determined based on the cell's safe temperature, target maximum charging efficiency, and heat dissipation adaptation.

[0165] The control scenarios are divided based on the predicted temperature state, heat dissipation attenuation level and operating condition type, and the priority index is calculated according to the temperature risk priority, heat dissipation attenuation priority and operating condition risk priority.

[0166] The temperature control strategy is obtained by determining the control direction and control strategy corresponding to different control scenarios based on the priority index.

[0167] Emergency control strategy, priority index ≥ 0.7: The core objective is to quickly bring the temperature back to a safe range. First, the charging power is urgently reduced, ensuring the power is no lower than 0.3C after the reduction, with a single reduction not exceeding 50% of the current power to avoid sudden voltage fluctuations. Second, the cooling system is activated at full load. In air-cooling mode, the fan speed is increased to 100% of the rated speed and the auxiliary heat dissipation channel is opened. In liquid-cooling mode, the cooling pump flow rate is adjusted to maximum. The hybrid cooling mode is immediately switched to a dual mode of "liquid cooling + phase change material," and the phase change material trigger threshold is lowered by 5°C. If the short-term predicted core temperature exceeds the threshold by more than 5°C, a dual mechanism of "power reduction + alarm" is triggered. If the temperature does not decrease within 100ms, the power is further reduced to 0.3C until the temperature is 2°C below the threshold.

[0168] A dynamic adaptive control strategy, with a priority index of 0.4 ≤ priority index < 0.7, balances safety and efficiency. It employs a stepped adjustment mode for charging power, with each adjustment step being 5% to 10% of the current power, adjusted every 50ms to avoid sudden power fluctuations. The cooling system is adaptively optimized according to the degree of heat dissipation degradation. For mild degradation, the cooling power is 70% to 80% of the rated value, increasing by 5% for every 1°C increase. For moderate degradation, the cooling power is 80% to 90%, with a 20% to 30% increase in flow rate to hot spots. Simultaneously, considering the response lag of the actuator, cooling pre-adjustment is initiated 100ms in advance. For scenarios where the core-surface temperature difference increment is no less than 5°C, the cooling medium distribution is optimized to reduce the temperature difference between cells. For multi-tab batteries, the current distribution of each tab is dynamically adjusted to ensure a current balance index of no less than 0.9.

[0169] Efficiency-first control strategy, with a priority index <0.4: Maximizes charging efficiency while maintaining the current charging rate, only slightly increasing power in 5% increments when the predicted temperature is more than 5°C below the threshold. The cooling system dynamically adjusts based on ambient temperature and heat dissipation differences. When the ambient temperature is not lower than 35°C, the cooling power is 60% to 70% of the rated value; when it is not higher than 10°C, it is 30% to 40%, avoiding excessive heat dissipation and energy waste. In cases of slight heat dissipation degradation, heat dissipation efficiency is improved by optimizing the uniformity of the cooling medium flow rate, without reducing power. In low-temperature environments with a core temperature below 10°C, the heating system is activated to preheat to 10 to 15°C before maintaining fast charging power, preventing capacity degradation caused by low-temperature fast charging.

[0170] Predictive control strategies: Early intervention prevents long-term overheating accumulation. If the predicted peak temperature of hot spots is no less than 2°C below the threshold, a combined strategy of "small power reduction + pre-cooling boost" is initiated in advance, reducing power by 5% to 8% of the current power and increasing cooling power by 10% to 15%. Based on battery aging status, when the State of Harmony (SOH) is below 0.8, the temperature threshold is predictively lowered by 2°C to allow for greater safety margin. For high-rate fast charging scenarios, based on long-term temperature prediction, a "constant current - stepped current reduction" curve is planned 10 seconds in advance, reducing power by 5% every 5 seconds to avoid emergency power reduction near the threshold.

[0171] The steps to adjust and obtain a dynamic temperature regulation strategy include: Based on real-time collected charging-related parameters, the transient heat dissipation influencing factors are updated, and the factor priority coefficients are calculated using a weighted summation method.

[0172] Compare the actual temperature change after the temperature control strategy is implemented with the expected target to determine whether the preset safety threshold has been reached; otherwise, formulate adjustment directions based on the factors affecting transient heat dissipation.

[0173] For transient interference scenarios, dynamic adjustment rules are formulated based on the adjustment direction and factor priority coefficients to adjust the temperature control strategy and obtain the temperature regulation strategy.

[0174] High-priority adjustments: For scenarios with strong transient interference, implement precise adjustments based on the dominant influencing factors. If the coupling attenuation of heat dissipation efficiency is the dominant factor: add to the original power adjustment to ensure that the total reduction does not exceed 30% of the current power. Increase the cooling system power to 90%~100% of the rated value, increase the flow rate of hot spots by an additional 30% in liquid cooling mode, and allow the fan speed to enter the high speed range earlier in air cooling mode.

[0175] If temperature sensing distortion is the dominant factor: adjust the temperature safety threshold and reduce its magnitude to avoid insufficient protection due to sensing distortion. Apply a dual processing method of "wavelet filtering + trend smoothing" to the temperature sensor data to reduce the impact of signal distortion caused by EMI.

[0176] If actuator response lag is the dominant factor: extend the pre-adjustment time of the cooling / heating system from 100ms to 150ms, and issue execution commands in advance based on AI short-term temperature prediction. Optimize the communication protocol to compress the control command transmission delay to within 20ms, compensating for actuator response lag.

[0177] If thermal path blockage is the primary factor: activate the thermal path unblocking auxiliary strategy, temporarily increase the pump flow rate by 15%~20% in liquid cooling mode, switch to "pulse air delivery" mode in air cooling mode, and slightly reduce the charging power to alleviate local thermal blockage.

[0178] Medium-priority adjustments: Implement localized optimization adjustments to balance security and efficiency.

[0179] Power adjustment: The step size for each adjustment is kept at 5%~8%, and iterates once every 50ms to avoid sudden changes.

[0180] Cooling system adaptation: controlled at 70%~90% of the rated value. To address response delay caused by actuator lag, a "delay compensation coefficient" is added to the command to dynamically adjust the actuator start-up threshold.

[0181] Temperature sensing correction: The "predicted temperature + measured temperature correction" mode is adopted. The correction value = temperature sensing distortion × (predicted temperature - measured temperature) to ensure that the temperature control decision is based on the actual temperature state.

[0182] Low-priority adjustments: Primarily minor compensations to avoid excessive adjustments that could negatively impact charging efficiency. The power output remains unchanged, except when the heat dissipation efficiency coupling attenuation is ≥0.2, in which case the power output is slightly reduced in steps of 3%.

[0183] The cooling system operates in its original mode, compensating for the slight impact of thermal path blockage by optimizing the uniformity of medium flow rate.

[0184] The sensor data is filtered using a simple sliding window to reduce minor distortions caused by temperature sensing distortion.

[0185] Based on the same general inventive concept, this invention also protects an AI-based dynamic temperature control system for the charging process, the system comprising: The heat dissipation difference construction module is used to collect battery type data, battery performance data, and charging-related data. It analyzes the cell arrangement and heat diffusion path of the battery type to obtain the arrangement channel difference data, and combines the battery pack structure aging state to obtain battery heat dissipation difference data.

[0186] The transient impact analysis module is used to analyze the degree of impact of the cooling medium on the degradation of battery performance. It analyzes charging-related data to obtain transient electromagnetic interference and combines it with battery heat dissipation difference data to obtain transient heat dissipation influencing factors.

[0187] The AI ​​temperature prediction module is used to analyze the interaction between battery heat dissipation difference data and battery performance data using the heat dissipation effect feedback method. It constructs an AI temperature prediction model by combining the degree of degradation influence, inputs real-time battery data, and outputs the predicted temperature.

[0188] The temperature control strategy formulation module is used to formulate a temperature control strategy based on battery heat dissipation difference data and predicted temperature, and to adjust the temperature control strategy according to transient heat dissipation influencing factors to obtain a dynamic temperature regulation strategy.

[0189] In summary, the AI-based dynamic temperature control method and system for the charging process provided in this embodiment improves the characterization accuracy of heat dissipation difference data by coupling the arrangement-thermal management channel matching analysis with aging, and can cover different battery types (cylindrical, prismatic, and pouch cells) and the entire aging process. Through precise temperature control and cooling medium attenuation compensation, the battery capacity decay rate is reduced, cycle life is extended, and the problem of accelerated battery aging caused by high-temperature fast charging is alleviated. This improves battery safety and lifespan, optimizes charging efficiency and performance monitoring, and achieves intelligent and adaptive management.

[0190] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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; and these 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 the present invention.

Claims

1. An AI-based dynamic temperature control method for charging process, comprising: characterized by: collecting battery type data, battery performance data, and charging-related data, analyzing the battery type data from cell arrangement and heat diffusion path to obtain arrangement channel difference data, and combining battery pack structure aging state to obtain battery heat dissipation difference data; analyzing the attenuation influence degree of the cooling medium on the battery performance, analyzing the charging-related data to obtain transient electromagnetic interference, and combining the battery heat dissipation difference data to obtain transient heat dissipation influencing factors; using a mutual feedback method to analyze the mutual influence of the battery heat dissipation difference data and the battery performance data, combining the attenuation influence degree to construct an AI temperature prediction model, inputting real-time battery data, and outputting predicted temperature; formulating a temperature regulation strategy according to the battery heat dissipation difference data and the predicted temperature, and adjusting the temperature regulation strategy according to the transient heat dissipation influencing factors to obtain a dynamic temperature regulation strategy.

2. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of obtaining the arrangement channel difference data comprises: determining the cell arrangement mode according to the battery type, quantifying the arrangement geometric parameters, recording the gap design details to obtain the cell arrangement design parameters, and collecting the heat management channel parameters according to the heat management type and channel structure; calculating the arrangement density according to the cell arrangement design parameters, extracting the arrangement symmetry and gap uniformity, and obtaining the arrangement difference index in association with the tab arrangement; constructing a battery pack thermal model based on the cell arrangement design parameters and the heat management channel parameters, identifying the heat diffusion dominant path, quantifying the path thermal resistance distribution, and analyzing the heat diffusion uniformity to obtain the heat spot focusing characteristics; combining the flow velocity distribution, the fluid resistance coefficient, and the matching degree of the flow velocity distribution under different arrangement modes with the cell heat diffusion demand, and combining the heat spot focusing characteristics to obtain the channel arrangement matching difference; integrating the arrangement difference index and the channel arrangement matching difference to obtain the arrangement channel difference data. 3.The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of obtaining the battery heat dissipation difference data comprises: defining the structure affecting heat dissipation as an aging heat dissipation structure, calculating the component aging index of each aging heat dissipation structure with heat dissipation function attenuation as the core, and generating an aging level; for different aging heat dissipation structure aging levels, disassembling the damage mechanism to the arrangement-heat management channel heat dissipation logic, adapting and correcting in combination with the battery type to obtain the mapping relationship between the aging parameters and the heat dissipation attenuation; calculating the ideal heat dissipation efficiency without aging based on the arrangement channel difference data, and combining the component aging index to calculate the actual heat dissipation efficiency with aging; taking the same type of battery and the non-aging state as the benchmark, calculating the heat dissipation difference index under different aging levels to obtain the battery heat dissipation difference data.

4. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of obtaining the attenuation influence degree comprises: determining the target medium type, recording the corresponding medium basic attributes, defining the battery performance attenuation index, and setting the control variable; grouping design of different control groups, simulating actual use scenarios to set up cycle working conditions, test periods, and medium control parameters; collecting capacity, internal resistance, cycle life, and rate performance as performance parameters, and collecting medium state and temperature as medium environment correlation data; According to the performance parameters, a capacity attenuation influence coefficient, an internal resistance growth inhibition rate, and a life extension multiple are calculated, and a comprehensive attenuation influence index is constructed; Pearson correlation coefficient is used to analyze the correlation between the medium environment correlation data and the comprehensive attenuation influence index, and the influence mechanism is combined to obtain the attenuation influence degree.

5. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of obtaining the transient heat dissipation influencing factor includes: Collecting electrical parameters, communication data, sensor signals, and transient radiation as the charging-related data, and performing standardization processing to obtain standard charging data; When the standard charging data reaches a preset multi-dimensional trigger threshold, it is determined as a transient event, and is divided into a transient characteristic event based on the interference source and characteristics; For each transient characteristic event, the transient electromagnetic interference peak intensity, duration, rising edge rate, energy density, communication interference rate, and sensor distortion rate are extracted as the transient electromagnetic interference; The interference path of the transient electromagnetic interference on the heat dissipation system is determined from the perception layer, the control layer, and the execution layer, and the dynamic heat dissipation influence is calculated in combination with the battery heat dissipation difference data; According to the dynamic heat dissipation influence, a plurality of transient heat dissipation factors are obtained from heat dissipation efficiency coupling attenuation, temperature perception distortion, execution mechanism response lag, and heat path blockage analysis, and the factor weight is obtained by combining the battery heat dissipation safety requirement setting factor to obtain the transient heat dissipation influencing factor.

6. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of constructing the AI temperature prediction model includes: Analyzing the influence of the battery heat dissipation difference data on the battery performance data from the heat dissipation efficiency on the capacity attenuation, the heat path obstruction on the internal resistance growth, and the heat dissipation response delay on the rate performance to obtain a positive influence; According to the attenuation influence degree, the influence of the internal resistance growth on the heat generation power, the correction of the heat generation increase on the heat dissipation demand, and the amplification coefficient of the performance attenuation on the heat dissipation difference are obtained to obtain a reverse influence; Integrating the positive influence and the reverse influence to obtain the mutual influence, and constructing sliding window features, trend features, and working condition switching features; A basic framework model based on LSTM and attention mechanism is constructed, the mutual influence is divided into a training set, a test set, and a validation set according to a preset proportion, and the basic framework model is subjected to parameter training, transfer learning, and generalization ability optimization to obtain the AI temperature prediction model.

7. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of obtaining the predicted temperature includes: Collecting battery body real-time data, heat dissipation system real-time data, and environmental working condition real-time data as the real-time battery data, performing lightweight standardization based on the statistical parameters during model training, and converting to a tensor format required by the AI temperature prediction model; Based on the preprocessed real-time battery data, heat dissipation performance coupling features, time series trend features, and working condition marker features are extracted as model input feature sets; The real-time model input feature set is input into the AI temperature prediction model, triggering the inference mechanism, outputting the predicted value, and extracting the heat dissipation structure temperature, surface dimension, and hotspot peak temperature of the future preset time interval as the predicted temperature.

8. The AI-based charging process dynamic temperature control method of claim 1, wherein: The step of formulating the temperature regulation strategy includes: According to the cell safety temperature, the target maximum charging efficiency, and the heat dissipation adaptation, the regulation target and the constraint condition are determined; The control system comprises: a heat dissipation difference modeling module, configured to collect battery type data, battery performance data and charging related data, analyze the battery type from cell arrangement and heat diffusion path to obtain arrangement channel difference data, and obtain battery heat dissipation difference data in combination with the aging state of the battery pack structure; 9. The AI-based charging process dynamic temperature control method of claim 1, wherein: a transient influence analysis module, configured to analyze the attenuation influence degree of the cooling medium on the battery performance, analyze the charging related data to obtain transient electromagnetic interference, and obtain transient heat dissipation influence factors in combination with the battery heat dissipation difference data; an AI temperature prediction module, configured to analyze the mutual influence of the battery heat dissipation difference data and the battery performance data by using a mutual feedback method, construct an AI temperature prediction model in combination with the attenuation influence degree, input real-time battery data, and output a predicted temperature; a temperature control strategy formulation module, configured to formulate a temperature control strategy according to the battery heat dissipation difference data and the predicted temperature, and adjust the temperature control strategy according to the transient heat dissipation influence factors to obtain a dynamic temperature adjustment strategy. ​ 10. An AI-based dynamic temperature control system for charging process, applied to the AI-based dynamic temperature control method for charging process according to any one of claims 1 to 9, characterized in that, ​ ​ ​ ​ ​

Images