Lithium ion power battery thermal management online regulation and control method

Abstract

The invention discloses a lithium ion power battery thermal management online regulation and control method comprising the following steps: obtaining battery pack temperature data, and carrying out statistical analysis on the battery pack temperature data to obtain temperature fluctuation distribution characteristics; according to the temperature fluctuation distribution characteristics, cooling power output information is obtained through a feedback control method; obtaining a power allocation weight according to the cooling power output information, and carrying out prediction simulation on the temperature field according to the power allocation weight to obtain predicted temperature field distribution; the heat conductivity coefficient of a variable thermal interface material in the cooling system is adjusted according to the predicted temperature field distribution, a thermal diffusion path optimization scheme is obtained, a dynamic thermal channel network is constructed according to the thermal diffusion path optimization scheme, and a cooling strategy is obtained through matching according to the dynamic thermal channel network in combination with the running state of the battery; hierarchical control parameters are extracted from the cooling strategy, and an energy distribution optimization scheme is obtained; and according to the energy distribution optimization scheme, controlling a cooling system of the lithium ion power battery to realize thermal management regulation and control.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery thermal management, and in particular relates to an online thermal management control method for a lithium-ion power battery. Background Art

[0002] Lithium-ion battery thermal management is a core area of development for electric vehicles and energy storage systems, directly impacting battery performance, safety, and lifespan. Its importance is self-evident. With breakthroughs in battery technology toward higher energy density, the need for thermal management is becoming increasingly urgent. Any temperature runaway can lead to performance degradation and even safety hazards.

[0003] Existing thermal management methods often rely on static cooling strategies or simple temperature threshold controls, making them difficult to adapt to the dynamic thermal behavior of batteries under complex operating conditions. This single control logic often leads to insufficient cooling efficiency or excessive energy consumption. This is particularly true under high loads or extreme environmental conditions, where uneven temperature distribution and frequent hotspots are particularly prominent. Furthermore, balancing system response speed with energy consumption is difficult to achieve. Despite continued industry exploration in the field of thermal management, several core challenges remain. First, precise control of temperature stability is a key challenge. Due to the complex temperature gradients within the battery pack, traditional methods struggle to control temperature fluctuations within an ideal range under dynamic operating conditions. Second, optimization of heat diffusion paths is insufficient. Current solutions offer crude designs for directional heat conduction and uniform heat dissipation, making it difficult to effectively eliminate hotspots. Furthermore, optimizing the full lifecycle of thermal management systems presents significant challenges. Existing technologies lack the ability to intelligently adapt to multiple operating conditions, resulting in a poor match between cooling strategies and actual needs. These unresolved technical issues collectively limit thermal management effectiveness and hinder further improvements in battery performance. Therefore, how to achieve precise and stable control of battery pack temperature under dynamic working conditions, optimize the heat diffusion path to eliminate hot spots, and build an intelligent operation optimization plan for the entire life cycle has become a key issue in the research of online control methods for thermal management of lithium-ion power batteries. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes an online control method for thermal management of lithium-ion power batteries to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above objectives, the present invention provides a method for online thermal management of a lithium-ion power battery, comprising:

[0006] Acquiring battery pack temperature data of a lithium-ion power battery, performing statistical analysis on the battery pack temperature data to obtain temperature fluctuation distribution characteristics; obtaining cooling power output information through a feedback control method based on the temperature fluctuation distribution characteristics; obtaining a power allocation weight based on the cooling power output information, and performing a temperature field prediction simulation based on the power allocation weight to obtain a predicted temperature field distribution;

[0007] The thermal conductivity of the variable thermal interface material in the cooling process is adjusted according to the predicted temperature field distribution to obtain a heat diffusion path optimization scheme. Based on the heat diffusion path optimization scheme, a dynamic heat channel network is constructed. The dynamic heat channel network is combined with the battery operating status to match and obtain a cooling strategy. Hierarchical control parameters are extracted from the cooling strategy, and an energy distribution optimization scheme is obtained through the hierarchical control parameters. Based on the energy distribution optimization scheme, the cooling system of the lithium-ion power battery is controlled to achieve thermal management regulation.

[0008] Optionally, the process of acquiring the temperature fluctuation distribution characteristics includes:

[0009] The battery pack temperature data is divided into units, and the temperature mean of the divided units is statistically calculated. The deviation value is calculated based on the temperature and mean of each unit, and the deviation value is judged. Based on the judgment result, an abnormal deviation distribution is obtained. The fluctuation characteristics of each unit are calculated based on the abnormal deviation distribution. Based on the fluctuation characteristics, a temperature spatial distribution model of each unit is constructed to obtain the temperature distribution characteristics.

[0010] Optionally, the process of acquiring the cooling power output information includes:

[0011] Deviation calculation is performed based on the temperature fluctuation distribution characteristics and the control target value to obtain a deviation value sequence. The cooling power output value is obtained by controlling the calculation through the adaptive PID algorithm based on the deviation value sequence, wherein the gain parameter of the adaptive PID algorithm is updated according to the deviation value sequence.

[0012] Optionally, the process of obtaining the predicted temperature field distribution includes:

[0013] According to the temperature fluctuation distribution characteristics, the cooling power output information is weighted to obtain the power distribution weight; according to the temperature distribution characteristics, a temperature gradient model is constructed, and the power allocated according to the power distribution weight is combined with the temperature gradient model simulation to obtain an optimized temperature distribution. According to the optimized temperature analysis and prediction, the temperature change trend at subsequent moments is obtained, and according to the temperature change trend, the temperature field distribution at the next moment is obtained. The temperature field distribution at the next moment is smoothed and filtered to obtain a predicted temperature field distribution.

[0014] Optionally, the process of obtaining the heat diffusion path optimization solution includes:

[0015] Based on the predicted temperature field distribution, the predicted temperature value of each unit in the space is obtained, and the predicted temperature values are judged and divided to obtain a set of abnormal areas. Based on the temperature control of the abnormal area set, the corresponding variable thermal conductivity material is obtained. The thermal conductivity adjustment value of each material is calculated by using the finite element analysis algorithm to obtain an optimized thermal conductivity set. Based on the optimized thermal conductivity set, the heat diffusion path is obtained. The heat diffusion path is optimized and updated using the path generation algorithm to obtain the final heat diffusion path optimization solution.

[0016] Optionally, the process of acquiring the dynamic heat channel network includes:

[0017] According to the heat diffusion path optimization scheme, material activation points are assigned to the abnormal area set to obtain the heat absorption distribution. Based on the heat absorption distribution and the thermal conductivity of the variable thermal conductivity material, a dynamic heat channel network is constructed. The heat flow direction is extracted from the heat absorption distribution, and the heat conduction path and efficiency of the dynamic heat channel network are obtained based on the heat flow direction.

[0018] Optionally, the cooling strategy acquisition process includes:

[0019] The thermal conductivity efficiency data of the dynamic heat channel network and the battery operating status are fused to obtain fused operating data, the fused operating data are identified by a working condition recognition algorithm to obtain operating condition characteristics, the operating condition characteristics are matched with typical working conditions to obtain a corresponding initial cooling strategy; wherein the working condition recognition algorithm is a clustering method, and the typical working condition is a typical working condition under pre-stored historical data, corresponding to a cooling strategy; the preliminary cooling strategy is optimized according to the thermal conductivity efficiency data of the dynamic heat channel network to obtain a cooling strategy.

[0020] Optionally, the process of obtaining the energy allocation optimization solution includes:

[0021] Obtain the hierarchical control parameters in the cooling strategy, perform feature extraction on the hierarchical control parameters, strategy layer instructions, decompose the strategy layer instructions, obtain execution layer actions, judge the execution layer actions, obtain the action sequence, sort the action sequence, obtain the priority of the action sequence, obtain the initial energy allocation plan based on the action sequence and priority, optimize the initial energy allocation plan, and obtain the energy allocation optimization plan.

[0022] Optionally, the process of controlling the cooling system of the lithium-ion power battery includes:

[0023] According to the energy distribution optimization plan, the parameters of the cooling system are adjusted to obtain adjustment parameters, adjustment instructions are obtained according to the adjustment parameters, and the cooling system controlled by the thermal management system is controlled through the adjustment instructions.

[0024] On the other hand, the present application provides a lithium-ion power battery thermal management online control system for executing the above method.

[0025] Compared with the prior art, the present invention has the following advantages and technical effects:

[0026] The present invention realizes the precise online control of the temperature of lithium-ion power battery packs. Based on the temperature fluctuation distribution characteristics and the adaptive PID algorithm, the cooling power is dynamically adjusted, which effectively solves the problems of delayed response and high energy consumption of traditional static strategies; through the optimization of the thermal conductivity of variable thermal interface materials and the construction of a dynamic heat channel network, heat is directed and local hot spots are eliminated, which significantly improves the uniformity of temperature distribution; combined with the working condition identification algorithm and hierarchical control parameters, it can intelligently match the optimal cooling strategy in multiple scenarios, taking into account the balance between system response speed and energy efficiency. In addition, the closed-loop cycle optimization mechanism continuously corrects temperature deviations to ensure the efficiency and robustness of thermal management throughout the life cycle, and achieves a coordinated improvement in battery performance, safety and service life under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0028] Figure 1 This is a flow chart of an online control method for thermal management of a lithium-ion power battery according to an embodiment of the present invention;

[0029] Figure 2 Schematic diagram of the structure of the online control system for thermal management of lithium-ion power batteries according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0031] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0032] like Figure 1 As shown, the present embodiment of a lithium-ion power battery thermal management online control method and system may specifically include:

[0033] S101. Collect battery pack temperature data in real time through sensors, calculate the deviation between each cell temperature value and the average temperature based on a preset temperature collection frequency, and obtain temperature fluctuation distribution characteristics.

[0034] The temperature data of each battery pack unit is acquired at a preset frequency and stored as a time series dataset to obtain the original temperature dataset. Using the time series dataset, the average temperature of each unit is calculated to obtain the average battery pack temperature. The deviation between each unit temperature and the average temperature is calculated to obtain a set of unit temperature deviations. If any unit in the deviation set exceeds a preset threshold, it is marked as an outlier, and an outlier deviation distribution is obtained. Based on the outlier deviation distribution, the frequency and amplitude of each unit temperature fluctuation are calculated to obtain the temperature fluctuation characteristics. Based on the fluctuation characteristics, a spatial distribution model of each unit temperature is constructed to obtain the temperature distribution characteristics.

[0035] In a battery pack temperature monitoring scenario, sensors collect temperature data from each cell at a fixed frequency, forming a time series dataset. The key to this process is ensuring data continuity and accuracy. Assuming a battery pack contains 10 cells, and the sensor collects temperature once per second for one hour, this generates 3,600 data sets, each containing temperature values for 10 cells. This dataset provides the foundation for subsequent analysis.

[0036] Specifically, the average temperature of each cell is calculated to obtain the average temperature of the battery pack. After collecting data for one hour, the temperatures of the 10 cells at a certain moment are 20°C, 21°C, 19°C, 22°C, 20°C, 21°C, 20°C, 23°C, 19°C, and 20°C, respectively, with an average temperature of 20.5°C. This average reflects the overall thermal state of the battery pack and lays the foundation for deviation analysis. The deviation of each cell temperature from the average temperature is calculated. The deviation reflects the degree to which each cell temperature deviates from the overall temperature. Taking the above data as an example, the deviation of cell 1's temperature of 20°C from the average of 20.5°C is -0.5°C, and the deviation of cell 8's temperature of 23°C is +2.5°C. By iterating through all the data, the deviation set of each cell is obtained.

[0037] Preferably, the deviation set can be used to identify potential outliers. For example, if the preset deviation threshold is ±2°C, then cell 8 with a deviation of +2.5°C is marked as an outlier. Identifying outliers helps identify localized areas of overheating or undercooling. If cell 8 deviates by more than 2°C multiple times within an hour, an outlier deviation distribution is generated, recording the number and time of occurrence. This distribution provides a visual representation of the thermal behavior of the problematic cell.

[0038] In one embodiment, the frequency and amplitude of temperature fluctuations are calculated using the distribution of abnormal deviations. For example, if unit 8 experiences 10 deviations exceeding 2°C within an hour, the frequency is 10 times per hour, and the amplitude is the absolute value range of the deviations, such as 2°C to 3°C. This fluctuation characteristic quantifies the dynamic nature of temperature changes and provides a basis for thermal management.

[0039] A spatial distribution model is constructed based on the fluctuation characteristics to describe the temperature distribution within the battery pack. Assuming a two-dimensional battery pack, with cell 8 located in the center, the model shows that its temperature is higher than that of surrounding cells, forming a localized hotspot. Preferably, the spatial distribution model is presented as a visual heat map, clearly demonstrating the location and spread of high-temperature areas. Outlier detection provides timely warnings of potential failures, fluctuation characteristic analysis optimizes heat dissipation strategies, and the spatial distribution model provides a reference for battery pack design. These technologies collectively improve battery pack safety and service life. Identified hotspots guide cooling system optimization to reduce the risk of overheating.

[0040] It's important to note that this data-driven analysis method is adaptable to different battery pack sizes, making it highly versatile. Fluctuation characteristics can also be combined with historical data to predict future temperature trends. For example, if the frequency of fluctuations in cell 8 gradually increases, it may indicate insufficient heat dissipation, requiring proactive intervention. This predictive capability further enhances system reliability. Preferably, the spatial distribution model can also be dynamically updated to adapt to changes in the battery pack's operating status, ensuring accurate long-term monitoring.

[0041] S102. According to the temperature fluctuation distribution characteristics, an adaptive PID algorithm is used to adjust the cooling system parameters, and the proportional, integral and differential gains are dynamically updated based on the deviation value to determine the cooling power output information.

[0042] Based on the temperature fluctuation distribution characteristics, the deviation from the target temperature is calculated and a deviation sequence is determined. An adaptive PID algorithm is used to dynamically update the proportional gain, integral gain, and differential gain based on the deviation sequence to generate a gain parameter set. This gain parameter set is used to adjust the cooling system operating parameters and calculate the cooling power output. If the cooling power output exceeds the preset threshold, a secondary optimization of the gain parameter set is performed and the cooling power output is recalculated.

[0043] Acquiring temperature fluctuation distribution characteristics is fundamental to battery pack thermal management. This temperature fluctuation distribution reflects the dynamic characteristics of the temperature changes of each cell within the battery pack over time. For example, assuming a battery pack contains 10 cells, and a sensor collects temperature data once per second, a temperature sequence consisting of multiple time points is generated over time. Based on this sequence, statistical methods can be used to extract fluctuation characteristics, such as periodic changes in each cell temperature or instantaneous mutation points. These characteristics provide the data foundation for subsequent deviation calculations.

[0044] Specifically, when calculating the deviation value sequence from the target temperature, the target temperature is usually the optimal operating temperature preset in the battery pack design, such as 30 degrees Celsius. In one embodiment, assume that the temperature of a certain unit at a certain moment is 32 degrees Celsius, and the deviation from the target temperature is 2 degrees Celsius. By performing similar calculations on the temperatures of all units and time nodes, a deviation value sequence can be generated. The deviation sequence not only reflects the temperature difference at a single point in time, but also reveals the temperature change trend through the time dimension, such as whether a certain unit is continuously high or periodically deviates from the target.

[0045] The core of the adaptive PID algorithm is to dynamically adjust control parameters based on the deviation sequence. Proportional gain influences response speed, integral gain eliminates steady-state errors, and differential gain suppresses rapid changes. For example, if the deviation sequence indicates that a unit's temperature is consistently high, the algorithm might increase the proportional gain to accelerate the cooling response while appropriately adjusting the integral gain to prevent long-term deviation accumulation. For example, if the deviation suddenly increases from 2 degrees Celsius to 5 degrees Celsius, the differential gain will increase rapidly to suppress the rapid temperature rise. This dynamic adjustment ensures precise and real-time control. For example, when adjusting cooling system operating parameters, the gain parameter set directly affects the cooling power output. The cooling system achieves temperature control by adjusting fan speed or liquid cooling pump flow. Assuming the gain parameter set increases the fan speed from 1000 rpm to 1500 rpm, the cooling power output will increase accordingly. The power output calculation should comprehensively consider system hardware characteristics, such as the fan's power curve or the heat exchange efficiency of the liquid cooling system. This approach enables the cooling system to respond quickly to temperature changes.

[0046] In one embodiment, if the cooling power output value exceeds a preset threshold range, for example, exceeding a maximum power of 500 watts or falling below a minimum power of 100 watts, a secondary optimization of the gain parameter set is required. This optimization process may be achieved by reducing the proportional gain or re-evaluating the weight of the deviation sequence.

[0047] For example, if the power output reaches 600 watts in a certain calculation, the system will automatically roll back the gain parameters and recalculate the output value to 450 watts. This secondary optimization design improves the stability and adaptability of the system and avoids overload or inefficient operation. Specifically, each link of the above method is closely connected, from temperature fluctuation feature extraction to deviation calculation, to adaptive control and power optimization, forming a complete thermal management closed loop. For example, the generation of the deviation sequence depends on the accuracy of the fluctuation characteristics, while the dynamic adjustment of the PID algorithm is centered on the deviation sequence, and the optimization of the operating parameters of the cooling system further guarantees the overall efficiency. This multi-link collaborative approach not only improves the response speed of the battery pack thermal management, but also enhances the robustness of the system, providing reliable protection for the safe operation of the battery.

[0048] S103 , extracting the power allocation weight from the cooling power output information, and combining it with the temperature gradient prediction model to calculate the temperature change trend of each battery cell in the next period to obtain the predicted temperature field distribution.

[0049] Power allocation weights are derived from the cooling power output data, and statistical analysis methods are used to determine the cooling power allocation ratio for each battery cell. The power allocation weights are combined with the temperature gradient model to obtain the heat content of each battery cell and generate an initial temperature distribution. If the initial temperature distribution exceeds a preset threshold, a linear regression algorithm is used to adjust the temperature gradient model to obtain an optimized temperature distribution. Based on the optimized temperature distribution, the temperature change trend of each battery cell in the next time period is obtained to generate time series data. Using the time series data, an interpolation method is used to determine the temperature value of each battery cell in the next time period. Spatial correlations are extracted from the temperature values to generate a temperature field distribution. If there are outliers in the temperature field distribution, smoothing filtering is performed to obtain the final temperature field distribution data.

[0050] For example, when deriving power allocation weights from cooling power output data, statistical analysis can be used to determine the power requirements of each battery cell. For example, assuming a battery pack contains 10 cells, power output data is collected over a one-hour period, revealing that some cells have higher power requirements due to their proximity to heat sources. Statistical analysis can use a weighted average method, combined with historical data, to calculate the cooling power allocation ratio for each cell. The key is to dynamically adjust weights based on actual power fluctuations to ensure a reasonable cooling power distribution.

[0051] The power allocation weights are combined with a temperature gradient model, which captures the heat input to generate an initial temperature distribution. The temperature gradient model can be constructed based on the spatial location and thermal conductivity of the battery cells. For example, cells near the center have poorer heat dissipation and higher heat input. Assuming a cell's heat dissipation power weight is 0.3, and the model estimates its heat input to be 200W, the initial temperature distribution shows a cell temperature of 45°C. If the temperature in any area of the distribution exceeds the 50°C threshold, further optimization is required.

[0052] Specifically, when adjusting the temperature gradient model, a linear regression algorithm can be used to analyze the relationship between temperature and heat dissipation power. For example, by collecting multiple sets of power and temperature data, regression analysis reveals that for every 100W increase in heat dissipation power, the temperature decreases by approximately 5°C. Based on this information, model parameters can be adjusted to make the temperature distribution more realistic. For example, after optimization, the temperature of a unit dropped from 48°C to 43°C, making the overall distribution more uniform.

[0053] Preferably, time series data is generated based on the optimized temperature distribution to predict temperature trends over the next period. For example, analyzing the temperature data for the previous six hours reveals a temperature change of approximately 2°C per hour for a cell, and the predicted temperature is 45°C one hour later. Trend fitting can be used to clearly demonstrate the dynamic characteristics of each cell in time series data.

[0054] In one embodiment, an interpolation method is used to determine the temperature value of the next time period. For example, the temperature of a unit at 0.5 hours and 1 hour is 38°C and 40°C respectively, and the temperature at 0.75 hours is estimated to be approximately 39.5°C through linear interpolation. This method is suitable for scenarios with fewer data points and can effectively fill time gaps. It is understandable that when extracting spatial correlation from temperature values to generate temperature field distribution, it can be achieved through spatial autocorrelation analysis. For example, the analysis found that the temperature changes of units near the edge are highly correlated with the adjacent units, and the generated temperature field shows the heat concentration area. If the temperature in a certain area is abnormally higher than the surrounding area, such as a point reaching 55°C, it can be regarded as an abnormal point.

[0055] For example, when processing outliers, smoothing filtering can effectively eliminate noise. For example, using a moving average filter, the temperature at the outlier point is replaced by the average of the four surrounding points. Assuming the surrounding temperature is around 40°C, the filtered temperature at the outlier point is adjusted to 42°C. The resulting temperature field distribution data is smoother, reflecting the true heat distribution characteristics.

[0056] It's important to note that each step in the above method focuses on thermal management of the battery cooling system. For example, dynamic adjustment of power allocation weights optimizes energy utilization, while interpolation and filtering improve the accuracy of temperature prediction. These operations are closely linked to ensure a logically rigorous and efficient process for generating the temperature field distribution.

[0057] S104 : For the predicted temperature field distribution, if there is an area where the temperature gradient exceeds a preset threshold, the thermal conductivity of the variable thermal interface material is adjusted to generate a heat diffusion path optimization solution.

[0058] According to the temperature field distribution data obtained above, the temperature value of each area in the space is determined. If the temperature gradient of an area exceeds the preset threshold, the high gradient area is divided according to the temperature field distribution to obtain a set of abnormal areas. For the set of abnormal areas, the corresponding list of variable thermal conductive materials is obtained from the interface material database to determine the thermal conductivity range of each material. According to the temperature field distribution of the high gradient area, the finite element analysis algorithm is used to calculate the thermal conductivity adjustment value of each material to obtain an optimized thermal conductivity set. Through the optimized thermal conductivity set, a heat diffusion path is generated to determine the heat flow direction from the high gradient area to the low gradient area. If the heat flow direction of the heat diffusion path covers all abnormal areas, the path generation algorithm is used to update the heat diffusion path to obtain the final optimization solution.

[0059] For example, extracting temperature values for each region from temperature field distribution data can be achieved through a grid partitioning method. Assume that a battery module is divided into multiple grid cells, each of which records the temperature value at a specific location. For example, the temperature values of grid points within a certain region are 45°C, 47°C, and 50°C, respectively. By comparing these values with a preset threshold, such as 48°C, areas with abnormal temperatures can be quickly located. This method relies on spatial sampling and can accurately capture the details of the temperature distribution. If the temperature gradient in a region exceeds the threshold, such as the temperature difference between adjacent grid points exceeds 5°C / cm, a high-gradient region needs to be identified. High-gradient regions generally indicate concentrated heat, which may affect battery performance. By analyzing the temperature field distribution, high-gradient regions can be classified as a collection of abnormal regions. For example, in a module, due to uneven heat flow near the heat dissipation channel, a collection of abnormal regions with large temperature differences is formed, including grid points A and B. Specifically, when obtaining a list of variable thermal conductive materials from the interface material database, suitable thermal conductive materials can be selected based on their material properties.

[0060] For example, the database includes silicon-based thermal pads, phase change materials, and graphene composites, with thermal conductivities ranging from 2W / m·K to 10W / m·K. Given the high temperatures in unusually high areas, graphene composites with higher thermal conductivity are preferred. This screening method ensures that the material is perfectly matched to the specific scenario.

[0061] For example, when calculating thermal conductivity adjustments using finite element analysis, we can simulate the heat transfer process in abnormal areas. For example, if the initial thermal conductivity of a high-gradient area is 5 W / m·K, analyzing the heat flux distribution reveals that it needs to be increased to 7 W / m·K to achieve temperature equilibrium. This adjustment, based on the characteristics of the regional temperature field, can effectively optimize heat distribution.

[0062] In one embodiment, when generating a heat diffusion path, the transfer of heat from a high-temperature area to a low-temperature area can be simulated based on the optimized set of thermal conductivity coefficients. For example, a heat flow path is formed from the abnormal area A to the heat dissipation channel, covering the area where the temperature drops from 50°C to 40°C. This path generation relies on the law of heat conduction to ensure efficient heat conduction. Preferably, if the heat flow direction covers all abnormal areas, it can be further optimized by a path generation algorithm. For example, if the initial path only covers area A and does not cover area B, the path direction can be adjusted to add an auxiliary path from area B to the radiator. This optimization method ensures comprehensive heat conduction in abnormal areas through multi-path collaboration.

[0063] It's important to note that the above approach, through layered analysis and dynamic adjustments, can accurately address abnormal temperature conditions. Each step is logically structured around the thermal management requirements of the battery module. For example, from temperature extraction to path optimization, a complete thermal management solution is formed. This solution ensures balanced heat distribution through multi-faceted collaboration.

[0064] S105. Through the heat diffusion path optimization solution, the phase change material layer and the graphene heat diffusion layer are driven to work together to build a dynamic heat channel network and balance the temperature distribution characteristics of the battery pack.

[0065] Based on a heat diffusion path optimization scheme, the heat capacity characteristics of the phase change material layer are utilized to assign phase change material activation points to high-heat areas, generating a heat absorption distribution. Leveraging the high thermal conductivity of the graphene thermal layer, a dynamic heat channel network is constructed. Heat flow direction is extracted from the heat absorption distribution to determine the heat conduction path and efficiency. Based on the heat conduction path, the heat channel network topology is adjusted, and the interaction frequency between the phase change material layer and the graphene layer is optimized to achieve a balanced heat flow distribution characteristic for the battery pack temperature.

[0066] For example, for the heat diffusion path optimization solution, when the heat capacity characteristics of the phase change material layer are adopted, the temperature of the high-heat area can be adjusted by utilizing the characteristics of the phase change material to absorb or release a large amount of heat at a specific temperature. Phase change materials are usually composed of organic or inorganic materials, such as paraffin or salt compounds, which can absorb heat near the phase change temperature point without significantly increasing the temperature. In the battery pack thermal management scenario, suppose that the temperature of a module reaches 55°C when the battery pack is running, exceeding the safety threshold of 50°C. One possible implementation method is to arrange a paraffin-based phase change material layer near the module, and set its phase change temperature to 48°C. When the temperature approaches 48°C, the material begins to melt, absorb excess heat, and keep the module temperature stable.

[0067] Preferably, the allocation of phase change material activation points is based on the temperature field distribution data, with priority given to covering the areas with the highest temperature. For example, through real-time monitoring, the top and middle of the battery pack are identified as high-heat areas, and phase change material layers with thicknesses of 5 mm and 3 mm are respectively configured to form a heat absorption distribution.

[0068] Specifically, the high thermal conductivity of the graphene thermal layer can be used to construct a dynamic heat channel network. Graphene has an extremely high thermal conductivity of approximately 2000W / m·K, which can quickly transfer heat from hot areas to cold areas.

[0069] For example, a 0.2mm thick graphene film was placed in the high-heat area at the top of the battery pack and connected to the heat sink. Heat absorption distribution showed that heat was concentrated in the top module, and the graphene layer conducted heat to the heat sink in a predetermined direction, forming a heat conduction path.

[0070] It should be noted that the efficiency of the heat conduction path depends on the continuity and contact quality of the graphene layer. In one embodiment, laser cutting technology is used to ensure a smooth contact surface between the graphene layer and the battery module, reducing thermal resistance by approximately 20%, making the heat flow direction more controllable, and significantly improving the rate of heat conduction from the top module to the bottom heat sink.

[0071] In one possible implementation, when adjusting the topology of the heat channel network, the distribution of the graphene layer can be dynamically optimized based on the heat conduction path. For example, if the initial topology is a grid structure, and if heat flow is detected to be concentrated in the center of the battery pack, the topology can be adjusted to a radial topology. The density of graphene channels in this central region is increased, and the channel width is increased from 0.5mm to 0.8mm, thereby improving heat conduction capacity. It is understood that topological adjustments must balance material cost and processing difficulty. The preferred approach is to add local branch channels in high-heat areas rather than a complete reconstruction. For example, the interaction frequency between the phase change material layer and the graphene layer is optimized to balance heat absorption and conduction efficiency. If the temperature in the central region of the battery pack is still relatively high after 30 minutes of operation, the interaction frequency can be adjusted by controlling the activation cycle of the phase change material layer. In one embodiment, the phase change material is set to enter a heat-absorbing state every 5 minutes, while the graphene layer continues to conduct heat, thereby achieving a more uniform heat flow distribution. Preferably, monitoring data shows that the overall temperature fluctuation range of the battery pack is reduced from ±5°C to ±2°C, and the heat flow distribution characteristics are more balanced. It's understandable that this approach, through the synergistic effect of phase change materials and graphene, precisely regulates heat in high-heat areas. Dynamic adjustments to heat conduction paths and topology further enhance thermal management efficiency. This approach is highly applicable to battery pack thermal management and can effectively address heat distribution issues under complex operating conditions.

[0072] S106. Obtain heat conduction efficiency data of the dynamic heat channel network, combine it with the battery operating status collected by the real-time operating condition sensor, input it into the operating condition recognition algorithm, and match the optimal cooling strategy in the typical operating condition library.

[0073] The system obtains thermal conductivity efficiency data from the dynamic heat channel network and battery operating status data collected by real-time sensors, performs data fusion processing, and obtains fused operating data. This fused operating data is then input into an operating condition recognition algorithm, and feature extraction is performed to obtain operating condition characteristics. If the operating condition characteristics match the preset characteristics in the typical operating condition library, the corresponding cooling strategy is retrieved from the typical operating condition library to determine a preliminary cooling strategy. Based on this preliminary cooling strategy and the dynamic efficiency data of the heat channel network, an optimization algorithm is used to adjust the strategy parameters to obtain the optimized cooling strategy.

[0074] For example, when obtaining the heat conduction efficiency data of a dynamic heat channel network, the heat transfer situation can be monitored in real time by heat flux sensors placed at key nodes of the heat channel. The sensor collects data once per second and records the heat conduction rate of different areas of the heat channel. For example, if the heat flux density at a node is 500W / m 2 , the other node is 300W / m 2 These data reflect the dynamic efficiency differences of the hot channel network and provide a basis for subsequent optimization.

[0075] Real-time sensors collect battery operating status data, including temperature and voltage sensors, to monitor battery pack health. Multi-point data collection provides a comprehensive picture of operating status, providing a foundation for data fusion. The collection frequency can be set to once a minute to balance real-time performance and system load.

[0076] In one possible implementation, data fusion involves integrating heat transfer efficiency data with battery operating status data. Preferably, heat flux data can be correlated with battery temperature data using a weighted average method to generate fused operating data. For example, if a region has high heat flux and elevated temperature, the fused data may indicate that this region requires priority cooling. The fusion process must consider the temporal synchronization of the data to ensure accurate correlation between heat transfer efficiency and battery status.

[0077] It is understandable that the feature extraction process of the operating condition identification algorithm needs to mine key information from the fused data. For example, by analyzing the fused data, the distribution characteristics of the high-temperature area of the battery are extracted, such as a certain area with a continuous high temperature for more than 10 minutes, or a temperature gradient in a certain area exceeding 5°C / cm. These features reflect the abnormal points of the operating conditions. The algorithm can use a clustering method to compare the extracted features with the typical operating condition library to quickly locate the current operating condition. In one embodiment, if the operating condition characteristics match the typical operating condition library, for example, if a "high temperature and heavy load condition" is identified, the corresponding cooling strategy is retrieved from the library, such as increasing the coolant flow to 2L / min and prioritizing it to the high-temperature area. The preliminary cooling strategy needs to be combined with the real-time nature of the operating condition to ensure timely response. The strategies in the library are based on the accumulation of historical data and can quickly adapt to common operating conditions.

[0078] For example, when adjusting the cooling strategy parameters, the efficiency data of the dynamic heat channel network can be combined with the optimization algorithm to further refine it. Assume that the efficiency of a certain heat channel is low and the heat flux density is only 200W / m 2 By adjusting the coolant flow rate or heat channel topology, heat removal in this area can be prioritized. The optimization algorithm, based on the principle of gradient descent, iteratively adjusts parameters until the heat flow distribution is more even. This adjusted strategy can more accurately adapt to the current operating conditions.

[0079] It's important to note that the optimized cooling strategy requires real-time verification of its suitability. For example, battery temperature data can be collected again to see if the high-temperature range has dropped to a safe level, such as from 50°C to 40°C. This verification process can be combined with heat channel efficiency data to ensure improved heat transfer efficiency after the strategy adjustment. This entire process forms a closed loop, from data collection to strategy optimization, ensuring stable battery pack operation.

[0080] S107 , extracting hierarchical control parameters from the matching cooling strategy, decomposing them into strategy layer instructions and execution layer actions using a hierarchical control architecture, and generating an energy allocation optimization plan.

[0081] The hierarchical control parameters are obtained from the cooling strategy, and a preset extraction algorithm is used to determine the parameter set. If the parameter set contains policy-level instructions, the instructions are decomposed to generate a first instruction sequence. Based on the first instruction sequence, execution-level actions are obtained, and it is determined whether the actions meet the preset execution conditions. If the actions meet the execution conditions, a first action sequence is generated, and the priority of the action sequence is determined. Using the first action sequence and the priority, an initial energy allocation plan is generated. Based on the initial energy allocation plan, the input parameters of the optimization algorithm are obtained, and it is determined whether the optimization conditions are met. If the optimization conditions are met, an optimized energy allocation plan is generated, and the output parameters of the final plan are determined.

[0082] In one possible implementation, extracting hierarchical control parameters from a cooling policy involves a structured decomposition of the policy. A cooling policy typically contains multiple layers of control logic, such as a policy layer that determines overall cooling targets and an execution layer that handles specific actions. The extraction algorithm can be rule-based or machine learning-based, analyzing key fields in the policy, such as time, temperature, and power, to generate a set of parameters.

[0083] For example, if a battery cooling strategy specifies that high-power cooling be initiated when the temperature exceeds 45 degrees Celsius, the algorithm can extract "temperature threshold 45 degrees Celsius" and "high-power mode" as a parameter set. This approach ensures comprehensive parameter coverage and provides a foundation for subsequent instruction decomposition.

[0084] For example, if the parameter set includes a policy-level instruction, such as "prioritize lowering the core area temperature," the instruction needs to be decomposed to generate a first instruction sequence. The decomposition process can be based on a predefined instruction template to split the instruction into specific subtasks.

[0085] For example, reducing the core temperature can be decomposed into "starting the core fan" and "increasing the coolant flow rate."

[0086] In one embodiment, assuming the core temperature is 50 degrees Celsius, the system can generate an instruction sequence: first start the fan to 80% power, then increase the coolant flow rate to 2 liters / minute. This decomposition method ensures that the instruction sequence is clearly executable.

[0087] Specifically, after obtaining the execution layer action based on the first instruction sequence, it is necessary to further determine whether the action meets the execution conditions. The execution conditions may include hardware status, energy consumption limit, etc.

[0088] For example, fan operation requires confirmation that its current speed is lower than the target value, and coolant flow rate adjustment requires checking the available power of the pump.

[0089] In one embodiment, if the fan's current speed is 50% and the instruction requires 80%, the action condition is met; if the pump power is insufficient, the flow rate instruction needs to be adjusted. This judgment process ensures the feasibility of the action.

[0090] Preferably, after the first action sequence is generated, the priority of the action sequence needs to be determined.

[0091] For example, a core area with excessive temperature may take precedence over an edge area, so fan action takes precedence over coolant adjustments.

[0092] In one possible implementation, the system assigns weights based on temperature distribution, with each 1°C increase in core zone temperature increasing the priority by 10%. For example, if the core zone temperature is 50°C and the edge zone is 40°C, the fan priority can be set to 80, and the coolant priority to 60. This priority distribution improves responsiveness in critical areas.

[0093] It is understandable that when generating the initial energy allocation plan through the first action sequence and priority, energy consumption balance needs to be comprehensively considered.

[0094] For example, the initial energy allocation plan is used as an input parameter of the optimization algorithm, and it is necessary to determine whether the optimization conditions are met, such as whether the total energy consumption does not exceed the threshold or the temperature drop rate meets the standard.

[0095] In one embodiment, if the solution causes the core area temperature to drop at a rate of 2 degrees Celsius per minute and the total power is less than 900 watts, the condition is met. Otherwise, the fan power or flow rate needs to be adjusted. This judgment process improves the adaptability of the solution. In one possible implementation, when generating an optimized energy allocation solution, the parameters can be adjusted through iterative optimization. For example, in the initial solution, the fan power is too high, resulting in excessive energy consumption. The optimization algorithm can reduce the fan power to 500 watts and increase the coolant flow rate to 2.5 liters per minute, ultimately maintaining the temperature drop rate at 1.5 degrees Celsius per minute. This optimization method takes into account both energy efficiency and cooling requirements. It should be noted that the output parameters of the final solution must clearly specify specific execution values, such as a fan speed of 7000 rpm and a coolant flow rate of 2.2 liters per minute. These parameters provide clear guidance for the execution layer to ensure efficient operation of the cooling system.

[0096] S108. According to the energy allocation optimization plan, the operating parameters of the actuators in the cooling system are adjusted, the real-time operating status of the thermal management system is updated, and a new temperature fluctuation distribution characteristic is obtained.

[0097] Actuator parameters are collected through the cooling system and adjusted using a preset energy distribution model to obtain an optimized parameter set. Adjustment instructions are obtained from the optimized parameter set, and the operating status of the thermal management system is updated to obtain real-time status data. If the real-time status data exceeds the preset threshold, the temperature fluctuation is calculated using a distribution analysis algorithm to obtain a fluctuation feature set. Distribution patterns are extracted from the fluctuation feature set, and a clustering algorithm is used to divide the fluctuation intervals to obtain classified fluctuation interval data. Abnormal intervals are obtained from the classified fluctuation interval data, and the actuator parameters are adjusted to obtain an updated parameter configuration. Energy distribution is optimized using the updated parameter configuration, and system operation is adjusted using a feedback control algorithm to obtain a stable temperature distribution feature. Key indicators are extracted from the stable temperature distribution feature, and the thermal management system status monitoring is updated to obtain the final operating status data.

[0098] For example, when collecting actuator parameters through a cooling system, data can be obtained from devices such as the heat exchanger, fan speed, and pump flow rate. Suppose the heat exchanger's temperature sensor indicates a current value of 45 degrees Celsius and the fan speed is 1200 rpm. These parameters reflect the system's real-time operating status. The data collection process typically relies on a sensor network to ensure data accuracy. Preferably, the data undergoes preliminary filtering to remove outliers to improve the reliability of subsequent analysis. When adjusting parameters using a pre-set energy allocation model, energy can be reallocated using a linear mapping method based on the collected data. For example, the fan speed might be adjusted from 1200 rpm to 1500 rpm to address rising temperatures. The adjusted parameter set will contain a more optimal operating configuration, such as a 10% increase in pump flow rate. The core of this approach is to ensure more efficient energy allocation through model prediction. Specifically, when adjustment instructions are obtained from the optimized parameter set, specific commands can be generated for the thermal management system. For example, the command might require increasing the coolant flow rate to 2 liters per minute while reducing the power of the secondary fan. These instructions are directly applied to the controller to ensure that the system operates according to the optimized configuration. It should be noted that instruction generation needs to take into account the response time of the device to avoid frequent adjustments that may cause system instability.

[0099] In one embodiment, after updating the operating status of the thermal management system, real-time status data may show that the temperature of a core component is 40 degrees Celsius, slightly lower than expected. When the temperature exceeds a preset threshold (e.g., 42 degrees Celsius), further analysis is triggered. For example, temperature fluctuations may manifest as periodic increases. Using a distribution analysis algorithm, it can be identified that the fluctuations are concentrated in the range of 38 to 42 degrees Celsius. This analysis relies on statistical methods to extract the frequency and amplitude of the fluctuations. For example, after extracting a set of fluctuation features, it may be found that the fluctuations are mainly caused by changes in the external ambient temperature. The clustering algorithm divides the fluctuation intervals into high-frequency and low-frequency categories. The classified interval data clearly demonstrates the distribution pattern of abnormal fluctuations. For example, high-frequency fluctuations may be concentrated in scenarios where the temperature rises rapidly within a short period of time. This division facilitates subsequent precise adjustments. It is understandable that after obtaining the abnormal interval, the fan speed may be specifically reduced to 1000 revolutions per minute when adjusting the actuator parameters to reduce unnecessary energy consumption. The updated parameter configuration will rebalance the system's operating load. For example, the coolant flow rate may be fine-tuned to 1.8 liters per minute. This adjustment is based on the characteristics of the abnormal interval and ensures more stable system operation. Preferably, a proportional control mechanism can be introduced when optimizing energy distribution through a feedback control algorithm. For example, when the temperature deviates from the target value, the pump power is gradually adjusted to avoid sudden changes. The core of this method is dynamic response to ensure that the temperature distribution tends to be stable. The stable temperature distribution characteristics may be manifested as the temperature of the core components maintained at around 39 degrees Celsius, with a fluctuation amplitude of less than 1 degree Celsius. In one embodiment, when extracting key indicators from the stable temperature distribution characteristics, attention can be paid to temperature uniformity and energy efficiency. For example, the uniformity index may show that the temperature difference of all components is less than 2 degrees Celsius. These indicators are directly used to update status monitoring. For example, the monitoring system may record the daily temperature change trend to facilitate long-term optimization. The core of this approach is to improve the predictability of the system through data-driven decision-making.

[0100] S109: If there are still areas where the deviation exceeds the threshold in the new temperature fluctuation distribution characteristics, return to the adaptive PID algorithm to recalculate the cooling parameters, and cyclically optimize until the temperature fluctuation meets the preset range.

[0101] If there are areas in the temperature fluctuation distribution where the deviation exceeds the threshold, the fluctuation data of the exceeding areas are extracted from the distribution characteristics to obtain a set of regional deviations. Through the set of regional deviations, the adaptive PID algorithm is called to calculate the adjustment values of the cooling parameters to obtain a set of parameter increments. According to the set of parameter increments, the cooling parameters are adjusted to generate a new parameter configuration to obtain an updated cooling scheme. Using the updated cooling scheme, the temperature fluctuation distribution is recalculated to obtain a new distribution feature. If there are still areas in the new distribution feature where the deviation exceeds the threshold, the fluctuation data are extracted from the new distribution feature to obtain an updated set of deviations. Through the updated set of deviations, a loop optimization is performed to iteratively adjust the cooling parameters to obtain an optimized parameter configuration. According to the optimized parameter configuration, it is verified whether the temperature fluctuation meets the preset range to obtain the final fluctuation distribution.

[0102] In one possible implementation, deviation analysis of temperature fluctuation distribution can begin with data acquisition. The cooling system uses sensors to acquire real-time temperature data from heat exchangers and pipelines, generating a distribution map. Identifying deviation areas relies on visual analysis of the distribution map, with the system automatically marking areas exceeding the standard for subsequent processing. This approach uses intuitive data presentation to quickly pinpoint problem areas.

[0103] Specifically, extracting regional deviation sets further organizes data from areas exceeding the standard. For example, if the temperature points in the exceeding area are 45°C, 46°C, and 44°C, the system will classify these points into a deviation set and record their locations and timestamps. This aggregation process facilitates subsequent algorithm calls because it consolidates scattered data points into an actionable whole.

[0104] Preferably, the application of an adaptive PID algorithm is key to adjusting cooling parameters. In one embodiment, the algorithm analyzes the duration and magnitude of the deviations described above. For example, if the temperature exceeds a threshold for three consecutive minutes, the algorithm might recommend increasing the coolant flow rate by 10%. A parameter increment set is generated, containing the flow adjustment value and the pump speed change value. This adaptive adjustment is based on learning from historical operating data and can dynamically adapt to different operating conditions. In one possible implementation, once a new parameter configuration is generated, the cooling system immediately applies it. For example, the coolant flow rate may be adjusted from 100 liters / hour to 110 liters / hour, and the pump speed may be increased from 1200 rpm to 1300 rpm. It should be noted that the generation of the new configuration takes into account the overall balance of the system to avoid instability in other areas caused by a single parameter adjustment. For example, when recalculating the temperature fluctuation distribution, the system simulates the heat transfer process under the new configuration. For example, after the adjustment, the heat exchanger temperature distribution shows that the temperature in the exceeding range has dropped to 41 degrees Celsius, but a small area still exceeds the threshold. In this case, the system will extract a new deviation set, such as the point sets at 41 degrees Celsius and 42 degrees Celsius. It can be understood that this iterative process ensures that the deviations gradually converge.

[0105] Specifically, loop optimization relies on multiple iterative adjustments. In the second iteration, the system further increases the coolant flow rate to 115 liters / hour while also fine-tuning the fan speed. Preferably, a maximum number of iterations, such as five, is set to avoid infinite loops. After several adjustments, the number of deviation points is significantly reduced, and the temperature distribution becomes more uniform.

[0106] In one possible implementation, the final step is verifying that temperature fluctuations meet a preset range. For example, the adjusted distribution shows that all zone temperatures are between 38 and 40 degrees Celsius, meeting the threshold requirement. It should be noted that this verification not only focuses on the numerical values but also checks the stability of the distribution, such as whether the frequency of fluctuations is too high. This comprehensive check ensures the long-term reliability of the system. For example, the generation of the final fluctuation distribution results in a detailed operation report. For example, the report records a comparison of the temperature curves before and after the adjustment, as well as the changes in the deviation points for each iteration. This data record provides a reference for subsequent optimization and demonstrates the system's efficiency in dynamic adjustment.

[0107] On the other hand, Figure 2 As shown, the present invention provides a lithium-ion power battery thermal management online control system, which mainly includes:

[0108] The temperature acquisition and deviation calculation module is used to collect battery pack temperature data in real time through sensors, calculate the deviation between each cell temperature value and the average temperature based on the preset temperature acquisition frequency, and obtain the temperature fluctuation distribution characteristics;

[0109] Adaptive PID control module, which uses an adaptive PID algorithm to adjust cooling system parameters based on temperature fluctuation distribution characteristics, dynamically updates proportional, integral and differential gains based on deviation values, and determines cooling power output information;

[0110] The temperature prediction module is used to extract the power allocation weight from the cooling power output information, and combine it with the temperature gradient prediction model to calculate the temperature change trend of each battery cell in the next period to obtain the predicted temperature field distribution;

[0111] The thermal diffusion optimization module is used to predict the temperature field distribution. If there are areas where the temperature gradient exceeds a preset threshold, the thermal conductivity of the variable thermal interface material is adjusted to generate a thermal diffusion path optimization plan;

[0112] Dynamic thermal channel construction module, which is used to drive the phase change material layer and graphene thermal diffusion layer to work together through thermal diffusion path optimization scheme to build a dynamic thermal channel network and balance the temperature distribution characteristics of the battery pack;

[0113] The operating condition identification and strategy matching module is used to obtain the thermal conduction efficiency data of the dynamic heat channel network, combine it with the battery operating status collected by the real-time operating condition sensor, input it into the operating condition identification algorithm, and match it with the optimal cooling strategy in the typical operating condition library;

[0114] The hierarchical control and energy allocation module is used to extract hierarchical control parameters from the matching cooling strategy, decompose them into strategy-level instructions and execution-level actions using a hierarchical control architecture, and generate an energy allocation optimization plan;

[0115] The actuator parameter adjustment module is used to adjust the operating parameters of the actuators in the cooling system according to the energy allocation optimization plan, update the real-time operating status of the thermal management system, and obtain new temperature fluctuation distribution characteristics;

[0116] The loop optimization module is used to return to the adaptive PID algorithm to recalculate the cooling parameters if there are still areas where the deviation exceeds the threshold in the new temperature fluctuation distribution characteristics, and to loop optimization until the temperature fluctuation meets the preset range.

[0117] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A lithium-ion power battery thermal management online control method, characterized in that: include: Acquiring battery pack temperature data of a lithium-ion power battery, performing statistical analysis on the battery pack temperature data to obtain temperature fluctuation distribution characteristics; obtaining cooling power output information through a feedback control method based on the temperature fluctuation distribution characteristics; obtaining a power allocation weight based on the cooling power output information, and performing a temperature field prediction simulation based on the power allocation weight to obtain a predicted temperature field distribution; The thermal conductivity of the variable thermal interface material in the cooling system is adjusted according to the predicted temperature field distribution to obtain a heat diffusion path optimization scheme. Based on the heat diffusion path optimization scheme, a dynamic heat channel network is constructed. The dynamic heat channel network is combined with the battery operating status to match and obtain a cooling strategy. Hierarchical control parameters are extracted from the cooling strategy, and an energy distribution optimization scheme is obtained through the hierarchical control parameters. Based on the energy distribution optimization scheme, the cooling system of the lithium-ion power battery is controlled to achieve thermal management regulation.

2. The method according to claim 1, characterized in that The process of acquiring the temperature fluctuation distribution characteristics includes: The battery pack temperature data is divided into units, and the temperature mean of the divided units is statistically calculated. The deviation value is calculated based on the temperature and mean of each unit, and the deviation value is judged. Based on the judgment result, an abnormal deviation distribution is obtained. The fluctuation characteristics of each unit are calculated based on the abnormal deviation distribution. Based on the fluctuation characteristics, a temperature spatial distribution model of each unit is constructed to obtain the temperature distribution characteristics.

3. The method according to claim 1, characterized in that The process of obtaining the cooling power output information includes: Deviation calculation is performed based on the temperature fluctuation distribution characteristics and the control target value to obtain a deviation value sequence. The cooling power output value is obtained by controlling the calculation through the adaptive PID algorithm based on the deviation value sequence, wherein the gain parameter of the adaptive PID algorithm is updated according to the deviation value sequence.

4. The method according to claim 1, wherein The process of obtaining the predicted temperature field distribution includes: According to the temperature fluctuation distribution characteristics, the cooling power output information is weighted to obtain the power distribution weight; according to the temperature distribution characteristics, a temperature gradient model is constructed, and the power allocated according to the power distribution weight is combined with the temperature gradient model simulation to obtain an optimized temperature distribution. According to the optimized temperature analysis and prediction, the temperature change trend at subsequent moments is obtained, and according to the temperature change trend, the temperature field distribution at the next moment is obtained. The temperature field distribution at the next moment is smoothed and filtered to obtain a predicted temperature field distribution.

5. The method according to claim 1, wherein The process of obtaining the heat diffusion path optimization solution includes: Based on the predicted temperature field distribution, the predicted temperature value of each unit in the space is obtained, and the predicted temperature values are judged and divided to obtain a set of abnormal areas. Based on the temperature control of the abnormal area set, the corresponding variable thermal conductivity material is obtained. The thermal conductivity adjustment value of each material is calculated by using the finite element analysis algorithm to obtain an optimized thermal conductivity set. Based on the optimized thermal conductivity set, the heat diffusion path is obtained. The heat diffusion path is optimized and updated using the path generation algorithm to obtain the final heat diffusion path optimization solution.

6. The method according to claim 5, characterized in that The acquisition process of the dynamic heat channel network includes: According to the heat diffusion path optimization scheme, material activation points are assigned to the abnormal area set to obtain the heat absorption distribution. Based on the heat absorption distribution and the thermal conductivity of the variable thermal conductivity material, a dynamic heat channel network is constructed. The heat flow direction is extracted from the heat absorption distribution, and the heat conduction path and efficiency of the dynamic heat channel network are obtained based on the heat flow direction.

7. The method according to claim 1, characterized in that The cooling strategy acquisition process includes: The thermal conductivity efficiency data of the dynamic heat channel network and the battery operating status are fused to obtain fused operating data, the fused operating data are identified by a working condition recognition algorithm to obtain operating condition characteristics, the operating condition characteristics are matched with typical working conditions to obtain a corresponding initial cooling strategy; wherein the working condition recognition algorithm is a clustering method, and the typical working condition is a typical working condition under pre-stored historical data, corresponding to a cooling strategy; the preliminary cooling strategy is optimized according to the thermal conductivity efficiency data of the dynamic heat channel network to obtain a cooling strategy.

8. The method according to claim 1, characterized in that The process of obtaining the energy allocation optimization solution includes: Obtain the hierarchical control parameters in the cooling strategy, perform feature extraction on the hierarchical control parameters, strategy layer instructions, decompose the strategy layer instructions, obtain execution layer actions, judge the execution layer actions, obtain the action sequence, sort the action sequence, obtain the priority of the action sequence, obtain the initial energy allocation plan based on the action sequence and priority, optimize the initial energy allocation plan, and obtain the energy allocation optimization plan.

9. The method according to claim 1, characterized in that The process of controlling the cooling system of the lithium-ion power battery includes: According to the energy distribution optimization plan, equipment parameters of different areas of the cooling system are adjusted to obtain adjustment parameters, adjustment instructions are obtained according to the adjustment parameters, and the cooling system controlled by the thermal management system is controlled through the adjustment instructions.

10. A lithium-ion power battery thermal management online control system, characterized in that: Used to execute the method according to any one of claims 1 to 9.

Images