Energy storage battery temperature management system and method based on liquid cooling technology

Abstract

The invention relates to the technical field of battery temperature management, and discloses an energy storage battery temperature management system and method based on a liquid cooling technology. The method comprises the following steps: carrying out design parameter population initialization on the battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations; performing discharge process heat dissipation simulation and three-dimensional temperature distribution construction on the target energy storage battery based on the plurality of first liquid cooling system design parameter combinations to obtain three-dimensional temperature distribution data; constructing a first relation vector between the number of channels and the temperature, and constructing a second relation vector between the flow of the cooling liquid and the temperature; and inputting the first relation vector and the second relation vector into a multi-task deep learning model for optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management precision of the battery discharge process.

Description

Technical Field

[0001] The present application relates to the technical field of battery temperature management, and in particular to a temperature management system and method for energy storage batteries based on liquid cooling technology. Background Art

[0002] The rapid development of energy storage battery technology has become an indispensable part of the modern energy system. However, with the increase in battery energy density and the complexity of application scenarios, the thermal management problem of batteries has become increasingly prominent. Poor thermal management will not only lead to a decline in battery performance, but may also cause safety risks, such as thermal runaway or even fire caused by overheating.

[0003] Traditional air cooling technology is difficult to meet the needs of current high-performance battery systems due to its insufficient cooling efficiency under high-load conditions. Liquid cooling technology has become a powerful alternative due to its excellent heat conduction capability and better temperature uniformity. Although the liquid cooling system can effectively manage the temperature of the battery, the system design is complex and requires precise control of the flow rate and channel layout of the coolant to adapt to the thermal load characteristics of different battery packs. This requires the system to not only be highly adaptable, but also to be able to dynamically adjust according to the real-time thermal state of the battery in actual operation, which places higher requirements on the control strategy and design of the system. Summary of the invention

[0004] The present application provides a temperature management system and method for an energy storage battery based on liquid cooling technology, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management accuracy of the battery discharge process.

[0005] In a first aspect, the present application provides a temperature management method for an energy storage battery based on liquid cooling technology, and the temperature management method for an energy storage battery based on liquid cooling technology comprises:

[0006] Initializing a design parameter population for a battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate;

[0007] Based on the multiple first liquid cooling system design parameter combinations, respectively, a discharge process heat dissipation simulation and a three-dimensional temperature distribution construction are performed on the target energy storage battery to obtain three-dimensional temperature distribution data for each first liquid cooling system design parameter combination;

[0008] Constructing a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and constructing a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data;

[0009] The first relationship vector and the second relationship vector are input into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

[0010] A second aspect of the present application provides a temperature management system for an energy storage battery based on liquid cooling technology, and the temperature management system for an energy storage battery based on liquid cooling technology comprises:

[0011] An initialization module is used to initialize the design parameter population of the battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate;

[0012] A simulation module, used to perform heat dissipation simulation and three-dimensional temperature distribution construction on the target energy storage battery during discharge based on the multiple first liquid cooling system design parameter combinations, and obtain three-dimensional temperature distribution data of each first liquid cooling system design parameter combination;

[0013] A construction module, configured to construct a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and to construct a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data;

[0014] A prediction module is used to input the first relationship vector and the second relationship vector into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

[0015] The third aspect of the present application provides a computer device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the computer device executes the above-mentioned energy storage battery temperature management method based on liquid cooling technology.

[0016] A fourth aspect of the present application provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned energy storage battery temperature management method based on liquid cooling technology.

[0017] In the technical solution provided by the present application, by using liquid as a cooling medium, the liquid cooling system can transfer and disperse heat more effectively than the traditional air cooling system. The battery can cool down more quickly when running at high load, thereby improving the performance and life of the battery. By designing optimized flow channels and flow control, temperature uniformity between battery cells and within a single battery is achieved. Uniform temperature distribution is crucial to the long-term stability and safety of the battery, and can avoid structural damage or performance degradation caused by local overheating and large temperature gradients. A multi-task deep learning model is used to predict and optimize the design parameters, coolant flow rate and number of channels of the liquid cooling system. Through intelligent adaptive adjustment, the cooling strategy can be dynamically adjusted according to the actual working conditions and environmental changes of the battery to ensure that the battery system can maintain the best working state in various environments. By precisely controlling the parameters of the liquid cooling system, unnecessary energy consumption can be effectively reduced because the system can avoid overcooling and uneven cooling. The effectiveness of temperature management directly affects the safe operation of the battery. The liquid cooling system can respond quickly to changes in battery temperature and perform effective control, thereby reducing the safety risks caused by battery overheating, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management accuracy of the battery discharge process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a schematic diagram of an embodiment of a method for managing the temperature of an energy storage battery based on liquid cooling technology in an embodiment of the present application;

[0019] Figure 2 This is a schematic diagram of an embodiment of a temperature management system for an energy storage battery based on liquid cooling technology in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The embodiments of the present application provide a temperature management system and method for an energy storage battery based on liquid cooling technology, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management accuracy of the battery discharge process.

[0021] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the energy storage battery temperature management method based on liquid cooling technology includes:

[0023] Step 101: Initialize a design parameter population for a battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate;

[0024] It is understandable that the execution subject of the present application can be a temperature management system for energy storage batteries based on liquid cooling technology, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0025] Specifically, the channel number threshold and flow parameter threshold of the battery liquid cooling system are defined. These thresholds are upper or lower limits in the design process to ensure the feasibility and efficiency of the design parameters in practical applications. In order to generate multiple initial liquid cooling system design parameter combinations, a particle swarm optimization algorithm (PSO) is used, which is a population-based optimization tool that updates the particle position by simulating the hunting behavior of a flock of birds to effectively search for the optimal solution. In the particle swarm optimization algorithm, a group of particles are randomly initialized, each particle represents a possible design parameter combination, that is, a first random value of the number of channels and a second random value of the coolant flow rate. Random numbers are generated within the defined channel number threshold and flow parameter threshold to form an initial set of design parameter candidates. The randomly generated parameter values ​​are combined to form multiple candidate liquid cooling system design parameter combinations. The target fitness data of each candidate design parameter combination is calculated based on the expected cooling effect, energy efficiency, and cost-effectiveness of the design parameter combination, which can be obtained by simulation or analysis based on previous experimental data. The target fitness data is compared with the preset standard fitness data, and the performance of each design parameter combination is evaluated by calculating the difference data between the two. After obtaining the target difference data for each candidate design parameter combination, the average of these difference data is calculated. The average difference data provides a quantitative indicator that reflects the degree of deviation of each design parameter combination relative to the standard fitness. Parameter combination selection is performed based on the average difference data to screen out those design parameter combinations with the smallest differences, which are closer to or better than the preset performance standards. Through the screening and optimization process, multiple first liquid cooling system design parameter combinations are obtained, including the optimal solution for the initial number of channels and the initial coolant flow rate.

[0026] Step 102: Based on a plurality of first liquid cooling system design parameter combinations, heat dissipation simulation and three-dimensional temperature distribution construction are performed on the target energy storage battery during discharge, and three-dimensional temperature distribution data of each first liquid cooling system design parameter combination is obtained;

[0027] Specifically, a heat dissipation simulation model is constructed for each combination of liquid cooling system design parameters. The model is based on the principle of computational fluid dynamics (CFD) and is used to simulate and analyze the path and speed of the coolant flowing inside the battery and its impact on the temperature of the battery pack. The model of each design parameter combination has a different number of channels and coolant flow rate. The simulation helps predict and optimize the performance of the liquid cooling system to ensure that it can effectively control the temperature rise of the battery during discharge. Based on preset discharge parameters, such as discharge rate and ambient temperature, each liquid cooling heat dissipation simulation model is subjected to actual discharge process heat dissipation test. Battery temperature data is collected through specialized software and hardware tools to record the temperature changes inside the battery pack at different discharge stages. These data provide a direct basis for the performance evaluation of the liquid cooling system and show the impact of the cooling system design parameters on the heat dissipation efficiency. The collected battery temperature data is subjected to a three-dimensional temperature distribution analysis. Each temperature data point inside the battery is converted into a three-dimensional spatial coordinate to understand the distribution of temperature inside the battery pack. Through three-dimensional mapping, the distribution of temperature in the entire battery pack is visualized. Based on the three-dimensional spatial coordinates of each temperature data point obtained, the three-dimensional temperature distribution field of the liquid cooling simulation model is mapped to obtain the heat dissipation performance of the liquid cooling system in space. It can also be used to adjust and optimize the design parameters to ensure that the battery can be maintained within the optimal operating temperature range in actual use.

[0028] Step 103: constructing a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and constructing a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data;

[0029] It should be noted that the three-dimensional temperature distribution data is feature analyzed to extract multiple temperature distribution features of the battery under each combination of the first liquid cooling system design parameters. These features may include the maximum, minimum, average and temperature gradient of the temperature, which together describe the thermal behavior pattern of the battery under a specific cooling system configuration. The mean of multiple temperature distribution features is calculated, and the volatility of the temperature data is analyzed based on the mean, and the temperature coefficient of variation is calculated. The temperature coefficient of variation is an important indicator for measuring the stability of the temperature distribution characteristics, which can reflect the consistency and fluctuation range of the battery temperature distribution under different design parameters. At the same time, for each design parameter, such as the initial number of channels and the initial coolant flow rate, the corresponding coefficient of variation is calculated, and the calculation results show the degree of change of each design parameter under different experimental or simulation conditions. The obtained quantitative coefficient of variation is combined with the temperature coefficient of variation to calculate the quantitative weight of the initial number of channels. At the same time, the flow weight of the initial coolant flow rate is calculated by combining the flow coefficient of variation and the temperature coefficient of variation. These weights reflect the relative importance of each design parameter in battery temperature management and provide a weighted benchmark for subsequent data processing. Using the calculated quantity weights and the first weight, the initial number of channels and multiple temperature distribution characteristics are vectorized and vector weighted analyzed. Through mathematical modeling and data processing technology, a first relationship vector reflecting the relationship between the number of channels and the battery temperature distribution is constructed. Similarly, the initial coolant flow and temperature distribution characteristics are processed using flow weights and second weights to generate a second relationship vector reflecting the relationship between the coolant flow and the battery temperature. These two relationship vectors not only provide a quantitative relationship between design parameters and battery performance, but can also guide future system design and optimization work to ensure that the battery is maintained within the optimal temperature range during operation, thereby improving the battery's efficiency and life.

[0030] The three-dimensional temperature distribution data is searched for feature points to identify multiple initial three-dimensional temperature feature points from a complex data set. These feature points are usually key locations in the temperature distribution, such as the local maximum temperature point, the minimum temperature point or other significant change points, which are crucial to understanding the overall temperature distribution pattern. Subsequently, the initial three-dimensional temperature feature points are analyzed, including the identification of neighboring points. By searching the surrounding area of ​​each feature point, the neighboring temperature feature point sets are determined. These sets contain other temperature points similar or related to the initial feature point, thereby capturing and reflecting the continuity and local changes of the battery temperature in space. Using the neighboring temperature feature point sets, a temperature feature point distribution map corresponding to each initial three-dimensional temperature feature point is constructed. The distribution map is a visual representation of the spatial pattern of the temperature distribution inside the battery, providing an intuitive way to analyze how the temperature diffuses and concentrates inside the battery. Each distribution map not only shows the spatial distribution of temperature, but also reveals how the cooling effect is affected by the design parameters of the liquid cooling system. The cluster center is calculated for each temperature feature point distribution map, and the temperature distribution center or key area is determined by mathematical methods. Through cluster analysis, the core areas where the temperature is concentrated in each distribution map are identified, which are usually the key to controlling or optimizing the temperature management strategy. Based on the cluster center, multiple target three-dimensional temperature feature points are determined, which represent the key temperature control points under each design parameter combination. Based on these target three-dimensional temperature feature points, multiple temperature distribution features of each first liquid cooling system design parameter combination are generated, which are used to evaluate and optimize the performance of the liquid cooling system.

[0031] Step 104: Input the first relationship vector and the second relationship vector into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

[0032] Specifically, a multi-task deep learning model is constructed, which can simultaneously process and optimize multiple outputs, such as the number of channels and coolant flow rate. The model includes two branches of convolutional neural networks (CNNs), namely the first convolutional neural network and the second convolutional neural network, which are responsible for extracting spatial features from the input relationship vector. These convolutional neural networks are followed by a shared hidden layer, which acts as a hub for information integration and merges the features of the two networks to provide a richer and more comprehensive feature expression for subsequent processing. Subsequently, two Transformer networks are introduced, namely the first Transformer network and the second Transformer network, which are specialized for the prediction of the number of channels and the coolant flow rate, respectively. Transformer networks have significant advantages in dealing with such problems because they can process sequence data and excel in capturing long-distance dependencies. By inputting the first relationship vector and the second relationship vector into this multi-task deep learning model, the model can comprehensively consider the complex relationship between the two parameters and the battery temperature management performance, and predict the optimal number of channels and the optimal coolant flow rate. The prediction process is carried out through multi-layer processing and feature extraction within the model to ensure that the output parameter combination can maximize the cooling efficiency and overall performance of the battery. After obtaining the optimal number of channels and optimal coolant flow predicted by the model, a second liquid cooling system design parameter combination for the battery liquid cooling system is generated based on these data. This new design parameter combination is verified and adjusted in actual applications to ensure its effectiveness and practicality in real-world environments. By combining the multi-task deep learning model of convolutional neural networks and Transformer networks, the accuracy and operational efficiency of liquid cooling system design can be significantly improved, thereby providing a more stable and efficient temperature management solution for energy storage batteries.

[0033] The first relationship vector is input into the first convolutional neural network for convolution feature operation to obtain the first convolution vector, and the features related to the number of channels are extracted. The second relationship vector is input into the second convolutional neural network for convolution feature operation to obtain the second convolution vector, and the features related to the coolant flow rate are extracted. The first and second convolution vectors represent the deep, nonlinear features of the input data. Subsequently, the first relationship vector and the second relationship vector are vector concatenated to obtain the concatenated relationship vector. This allows the model to take into account the correlation and interaction between the two vectors in subsequent processing, thereby more comprehensively understanding and predicting the design parameters. The concatenated relationship vector is input into the shared hidden layer for feature extraction, and the concatenated vector is converted into a target feature vector, which contains all the necessary information for accurate prediction in subsequent steps. The first convolution vector and the target feature vector are input into the first Transformer network, which is specially configured to predict the optimal number of channels. The Transformer network optimizes the processing of long-range dependencies through its attention mechanism, which helps to understand and predict the channel configuration of the liquid cooling system. Similarly, the second convolution vector and the target feature vector are input into the second Transformer network to predict the optimal coolant flow rate. The two networks output the predicted optimal number of channels and coolant flow rate, respectively, based on complex patterns and relationships learned from the data. The optimal number of channels and optimal coolant flow rate obtained from the two Transformer networks are integrated and output through the output layer. The output layer adjusts and optimizes the final output according to the weights and biases determined in the model training to ensure that the output results can maximize the performance of the battery liquid cooling system.

[0034] In the embodiment of the present application, by using liquid as a cooling medium, the liquid cooling system can transfer and disperse heat more effectively than the traditional air cooling system. The battery can cool down more quickly when running at high load, thereby improving the performance and life of the battery. By designing optimized flow channels and flow control, temperature uniformity between battery cells and within a single battery is achieved. Uniform temperature distribution is crucial to the long-term stability and safety of the battery, and can avoid structural damage or performance degradation caused by local overheating and large temperature gradients. A multi-task deep learning model is used to predict and optimize the design parameters, coolant flow rate and number of channels of the liquid cooling system. Through intelligent adaptive adjustment, the cooling strategy can be dynamically adjusted according to the actual working conditions and environmental changes of the battery to ensure that the battery system can maintain the best working state in various environments. By precisely controlling the parameters of the liquid cooling system, unnecessary energy consumption can be effectively reduced because the system can avoid overcooling and uneven cooling. The effectiveness of temperature management directly affects the safe operation of the battery. The liquid cooling system can respond quickly to changes in battery temperature and perform effective control, thereby reducing the safety risks caused by battery overheating, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management accuracy of the battery discharge process.

[0035] In a specific embodiment, the process of executing step 101 may specifically include the following steps:

[0036] (1) Define the channel number threshold and flow parameter threshold of the battery liquid cooling system;

[0037] (2) performing random initialization according to a channel number threshold and a flow parameter threshold by using a particle swarm optimization algorithm to obtain a plurality of first random values ​​of the channel number and a plurality of second random values ​​of the coolant flow rate;

[0038] (3) combining the plurality of first random values ​​and the plurality of second random values ​​to generate a plurality of candidate liquid cooling system design parameter combinations;

[0039] (4) calculating the target fitness data of each candidate liquid cooling system design parameter combination respectively, and obtaining the preset standard fitness data;

[0040] (5) calculating the difference data between the target fitness data and the standard fitness data to obtain the target difference data of each candidate liquid cooling system design parameter combination;

[0041] (6) Calculating the average difference data corresponding to the target difference data of each candidate liquid cooling system design parameter combination, and selecting parameter combinations for multiple candidate liquid cooling system design parameter combinations based on the average difference data to obtain multiple first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combination includes: an initial number of channels and an initial coolant flow rate.

[0042] Specifically, the key design parameters of the battery liquid cooling system, namely the thresholds of the number of channels and the coolant flow rate, are defined. These thresholds are key constraints in the entire design process, determining the upper and lower limits of the system design, ensuring that the liquid cooling system can meet the cooling requirements without causing unnecessary waste of resources due to over-design. After setting the thresholds, the particle swarm optimization algorithm (PSO) is used to randomly initialize the parameters. The particle swarm optimization algorithm is a swarm intelligence algorithm that simulates the hunting behavior of a flock of birds to find the optimal solution and achieves the global optimum through information sharing between individuals. The preliminary design parameters are randomly generated according to the threshold of the number of channels and the threshold of the coolant flow rate to obtain the first random value and the second random value. For example, if the threshold of the number of channels is set between 10 and 50, and the threshold of the coolant flow rate is set between 100 and 500 liters per hour, the particle swarm algorithm will generate a series of random combination values ​​within these ranges, which reflect different possible design solutions. Subsequently, these randomly generated numbers of channels and coolant flows are combined into multiple candidate liquid cooling system design parameter combinations. For example, a specific combination may be 30 channels and a flow rate of 250 liters / hour, while another combination may be 45 channels and a flow rate of 300 liters / hour, each of which represents a potential liquid cooling system configuration. The candidate liquid cooling system design parameter combinations are evaluated and their target fitness data are calculated. The target fitness data is calculated based on a series of predefined performance criteria, which may include the system's cooling efficiency, energy consumption, cost, and feasibility. The performance of each design parameter combination is compared with the preset standard fitness data to determine how effective each combination is. The standard fitness data usually comes from previous design experience or the expected performance of an ideal model. By comparing the difference between the target fitness data and the standard fitness data, the target difference data of each candidate liquid cooling system design parameter combination is obtained. This data reflects the degree of deviation of each combination from the ideal state, and the smaller the difference, the closer the design is to the ideal model. For example, if the fitness value of a design parameter combination differs from the standard value by only 5%, while the difference of another combination is 10%, the former is better. By calculating the average of the target difference data of all candidate design parameter combinations, they are ranked and screened. The average difference data provides a quantitative indicator to evaluate the overall performance of each design parameter combination. A lower average difference value indicates a better liquid cooling system design. Based on this data, multiple design parameter combinations are selected, which constitute the final first liquid cooling system design parameter combination, including the optimal number of channels and coolant flow rate.

[0043] In a specific embodiment, the process of executing step 102 may specifically include the following steps:

[0044] (1) constructing a heat dissipation simulation model for each target energy storage battery based on multiple first liquid cooling system design parameter combinations to obtain a liquid cooling heat dissipation simulation model for each first liquid cooling system design parameter combination;

[0045] (2) Based on the preset discharge parameter data, the liquid cooling heat dissipation simulation model is subjected to a heat dissipation test during the discharge process and battery temperature data is collected to obtain battery temperature data of each liquid cooling heat dissipation simulation model;

[0046] (3) performing a three-dimensional temperature distribution analysis on multiple temperature data points in the battery temperature data to obtain the three-dimensional spatial coordinates of each temperature data point;

[0047] (4) According to the three-dimensional spatial coordinates of each temperature data point, the three-dimensional temperature distribution field of the liquid cooling simulation model is mapped to obtain the three-dimensional temperature distribution data of each first liquid cooling system design parameter combination.

[0048] Specifically, a heat dissipation simulation model is constructed for the target energy storage battery according to multiple first liquid cooling system design parameter combinations. The design parameter combination generally includes factors such as the layout and quantity of the cooling channels, the type of coolant and its flow rate. Each parameter combination is optimized based on different cooling requirements and expected heat loads to simulate various thermal management scenarios that the battery may encounter in actual use. For example, if the battery pack is designed to work under high load conditions, a liquid cooling system with more cooling channels and a higher coolant flow rate may be required. In this case, a liquid cooling system design parameter combination may include 50 cooling channels and a coolant flow rate of 500 liters per hour. For this design parameter combination, a corresponding liquid cooling heat dissipation simulation model is constructed using computational fluid dynamics (CFD) software to simulate the flow of coolant in the battery module and its impact on the battery temperature. Based on the preset discharge parameter data, each liquid cooling heat dissipation simulation model is subjected to a discharge process heat dissipation test and battery temperature data collection. The discharge parameters include the discharge rate, discharge depth and ambient temperature of the battery, which directly affect the heat generation and heat dissipation requirements of the battery. By simulating the actual working conditions of the battery, the temperature data of the battery under various operating conditions are collected. For example, when simulating a high discharge rate scenario, a sharp rise in temperature may be observed in certain areas of the battery, which is critical for evaluating the effectiveness of the liquid cooling system. Subsequently, a three-dimensional temperature distribution analysis is performed on multiple temperature data points in the collected battery temperature data, which can reveal the detailed distribution of the temperature inside the battery. By converting the temperature data into three-dimensional spatial coordinates, the high and low temperatures are obtained, and the specific locations of these high and low temperature points inside the battery are obtained. Through data analysis software and algorithms, simple temperature readings are converted into temperature maps in three-dimensional space. According to the three-dimensional spatial coordinates of each temperature data point, a three-dimensional temperature distribution field mapping is performed for each liquid cooling simulation model. Combining the three-dimensional temperature data with the liquid cooling simulation model, a comprehensive three-dimensional temperature distribution map is obtained, which shows how the battery temperature is distributed throughout the battery module under a specific combination of liquid cooling system design parameters. For example, if it is found that a certain design parameter combination causes the temperature of a certain area of ​​the battery to be too high, the layout of the cooling channel can be further adjusted or the flow rate of the coolant can be changed to achieve a better heat dissipation effect.

[0049] In a specific embodiment, the process of executing step 103 may specifically include the following steps:

[0050] (1) performing feature analysis on the three-dimensional temperature distribution data to obtain a plurality of temperature distribution features of each first liquid cooling system design parameter combination;

[0051] (2) performing mean calculation on a plurality of temperature distribution features to obtain a feature mean, and calculating temperature variation coefficients of the plurality of temperature distribution features based on the feature mean;

[0052] (3) calculating the coefficient of variation of the initial number of channels to obtain the coefficient of variation of the number, and calculating the coefficient of variation of the initial coolant flow rate to obtain the coefficient of variation of the flow rate;

[0053] (4) calculating a quantity weight of the initial number of channels and a first weight of the temperature variation coefficient according to the quantity variation coefficient and the temperature variation coefficient, and calculating a flow weight of the initial coolant flow and a second weight of the temperature variation coefficient according to the flow variation coefficient and the temperature variation coefficient;

[0054] (5) performing vector conversion and vector weighted analysis on the initial number of channels and multiple temperature distribution features according to the number weight and the first weight, to obtain a first relationship vector between the number of channels and the temperature;

[0055] (6) According to the flow weight and the second weight, vector conversion and vector weighted analysis are performed on the initial coolant flow and multiple temperature distribution characteristics to obtain a second relationship vector between the coolant flow and the temperature.

[0056] Specifically, by performing feature analysis on the three-dimensional temperature distribution data, multiple temperature distribution features of each combination of liquid cooling system design parameters are identified, including the highest temperature point, the lowest temperature point, the average temperature, and the temperature gradient. By using data analysis tools and techniques, such as computational fluid dynamics (CFD) simulation and subsequent data processing software, key data is extracted from the simulation results. The temperature distribution features are averaged to obtain the average thermal behavior of the battery under different liquid cooling system designs. For example, if a combination of liquid cooling system design parameters produces multiple highest temperature points, calculating the average of these highest points helps the system evaluate the cooling efficiency of the design parameter combination. Based on the feature mean, the temperature coefficient of variation of multiple temperature distribution features is calculated. This coefficient is an important indicator for measuring the consistency of temperature distribution and can reflect the impact of design parameters on temperature control stability. Subsequently, the coefficient of variation of the initial number of channels and the coolant flow rate is calculated. The coefficient of variation refers to the degree of change of the parameter under different operating conditions, which can help the system understand the sensitivity of the number of channels and coolant flow rate to battery temperature control within a predetermined operating range. For example, if a design parameter combination shows a high coefficient of variation under high load operating conditions, this may mean that its performance is sensitive to operating conditions and may require further optimization to improve the robustness of the system. Calculate the weights of the coefficient of variation for the number of channels and coolant flow rate. The weights are calculated based on the degree of correlation between the coefficient of variation and the temperature coefficient of variation, quantifying the influence of each design parameter on the overall temperature management. For example, if the coefficient of variation of the number of channels is high, its quantity weight will also be correspondingly high, indicating that special attention should be paid to the adjustment of the number of channels during the design optimization process. By combining the quantity weight with the first weight (calculated based on the temperature coefficient of variation), the initial number of channels and multiple temperature distribution features are vectorized and vector weighted, and finally a first relationship vector describing the relationship between the number of channels and temperature is generated. Similarly, the initial coolant flow rate and temperature distribution features are processed using the flow weight and the second weight (also calculated based on the temperature coefficient of variation) to generate a second relationship vector describing the relationship between the coolant flow rate and temperature.

[0057] In a specific embodiment, the process of performing feature analysis on the three-dimensional temperature distribution data to obtain multiple temperature distribution features of each first liquid cooling system design parameter combination may specifically include the following steps:

[0058] (1) performing feature point search on the three-dimensional temperature distribution data to obtain a plurality of initial three-dimensional temperature feature points;

[0059] (2) performing neighboring point identification on a plurality of initial three-dimensional temperature feature points respectively to obtain a set of neighboring temperature feature points corresponding to each initial three-dimensional temperature feature point;

[0060] (3) constructing a temperature feature point distribution map corresponding to each initial three-dimensional temperature feature point based on the adjacent temperature feature point set;

[0061] (4) Calculating the cluster centers of the temperature feature point distribution maps respectively to obtain the target cluster center of each temperature feature point distribution map;

[0062] (5) Determine a corresponding plurality of target three-dimensional temperature feature points according to the target cluster center, and generate a plurality of temperature distribution features of each first liquid cooling system design parameter combination according to the plurality of target three-dimensional temperature feature points.

[0063] Specifically, a feature point search is performed on the three-dimensional temperature distribution data to identify meaningful temperature feature points in the three-dimensional temperature distribution data. These feature points usually include extreme temperature points, such as the highest temperature point and the lowest temperature point, as well as other thermodynamically significant change points, which may represent key areas where heat accumulates or dissipates. For example, if there is an area inside a battery cell that continuously shows a higher temperature than the surrounding environment, the highest temperature point in this area will be marked as an initial three-dimensional temperature feature point. This identification usually relies on advanced image processing or data analysis software, which can process data from temperature sensors or simulation software and automatically identify key hot spots. After determining the initial three-dimensional temperature feature point, neighboring point identification is performed. For each feature point, the surrounding area is searched to identify a set of points with a temperature close to the main feature point, forming a set of neighboring temperature feature points. The neighboring point set helps to understand the thermal behavior of the feature point and its impact on the surrounding environment. For example, the neighboring point set around a high-temperature feature point may show a temperature gradient that gradually decreases from the center to the outside, which helps to reveal how heat diffuses in the area. Based on each set of neighboring temperature feature points, a corresponding temperature feature point distribution map is constructed. The distribution graph graphically displays the temperature distribution of each feature point and its neighboring points, allowing the system to intuitively see the distribution pattern of temperature in three-dimensional space. The cluster center of the temperature feature point distribution graph is calculated to identify the temperature cluster center in each distribution graph. Cluster analysis techniques such as K-means or DBSCAN use algorithms to determine the center points of a group of points, which represent the main temperature cluster areas. Each cluster center reflects a key thermal management area that requires special attention in the design of the liquid cooling system. Based on the cluster center, the corresponding target three-dimensional temperature feature point is determined. The temperature distribution characteristics of each first liquid cooling system design parameter combination are generated or optimized based on the feature points. For example, if the analysis shows that a specific cluster center is located in a corner of the battery module and the area continues to show higher temperatures in all simulation scenarios, the system may decide to add more cooling channels or increase the flow rate of the coolant in the area to solve the overheating problem.

[0064] In a specific embodiment, the process of executing step 104 may specifically include the following steps:

[0065] (1) constructing a multi-task deep learning model, which includes a first convolutional neural network, a second convolutional neural network, a shared hidden layer, a first Transformer network, a second Transformer network, and an output layer;

[0066] (2) inputting the first relationship vector and the second relationship vector into a multi-task deep learning model to predict the optimal number of channels and the optimal coolant flow rate, and outputting the optimal number of channels and the optimal coolant flow rate;

[0067] (3) Generate a second liquid cooling system design parameter combination of the battery liquid cooling system based on the optimal number of channels and the optimal coolant flow rate.

[0068] Specifically, a multi-task deep learning model is constructed. The core architecture of the model includes two convolutional neural networks (CNNs) and two Transformer networks, which are connected to each other through a shared hidden layer. This structural design takes advantage of the ability of CNN in processing images and spatial data, and the advantages of Transformer in processing sequence data and capturing long-distance dependencies. The first convolutional neural network mainly processes features related to the number of channels, while the second convolutional neural network processes features related to coolant flow. These two features are extracted from their respective relationship vectors, which may include temperature distribution data, system efficiency, and historical operation data. After processing by the convolutional layer, the feature vector is passed to the shared hidden layer to merge the information of the two streams for comprehensive analysis. The shared layer helps to reduce the complexity of the model and enables the model to learn common features across tasks during training, thereby improving the accuracy of predictions for different tasks. The features processed by the shared hidden layer are input into two different Transformer networks. The first Transformer network is responsible for predicting the optimal number of channels, while the second Transformer network is responsible for predicting the optimal coolant flow. The Transformer network is used because its self-attention mechanism can effectively capture the complex relationship between input features and provide accurate prediction output. After obtaining the predicted values ​​of the optimal number of channels and the optimal coolant flow rate from the Transformer network, these data are passed to the output layer of the model. The output layer integrates the prediction results and outputs the final design parameter combination. For example, if the prediction results show that the optimal number of channels is 40 and the optimal coolant flow rate is 200 liters / hour under the current test conditions and target performance indicators, these data will be directly used to guide the actual design and manufacture of the liquid cooling system. Using multi-task deep learning models to predict design parameters can significantly improve the efficiency of the design and optimize system performance through accurate data-driven decision-making. For example, in the energy management system of electric vehicles or large-scale data centers, the design of the liquid cooling system can be quickly adjusted according to different usage scenarios and environmental conditions to achieve the best heat dissipation and energy efficiency.

[0069] In a specific embodiment, the execution step inputs the first relationship vector and the second relationship vector into the multi-task deep learning model to predict the optimal number of channels and the optimal coolant flow rate, and the process of outputting the optimal number of channels and the optimal coolant flow rate may specifically include the following steps:

[0070] (1) inputting the first relationship vector into the first convolutional neural network to perform convolution feature operation to obtain a first convolution vector;

[0071] (2) inputting the second relationship vector into the second convolutional neural network to perform convolution feature operation to obtain a second convolution vector;

[0072] (3) performing vector concatenation on the first relationship vector and the second relationship vector to obtain a concatenated relationship vector, and inputting the concatenated relationship vector into a shared hidden layer for feature extraction to obtain a target feature vector;

[0073] (4) inputting the first convolution vector and the target feature vector into the first Transformer network to predict the number of channels and obtain the optimal number of channels, and inputting the second convolution vector and the target feature vector into the second Transformer network to predict the coolant flow rate and obtain the optimal coolant flow rate;

[0074] (5) The optimal number of channels and the optimal coolant flow rate are integrated and output through the output layer.

[0075] Specifically, the first relationship vector and the second relationship vector represent data of different parameters and operating states related to the number of channels and coolant flow of the battery liquid cooling system, respectively. For example, the first relationship vector may contain information such as temperature gradient, heat flux density, and layout of the liquid cooling channel extracted from historical operation data, while the second relationship vector may include data such as physical properties, flow rate, and temperature of the coolant. These vectors are respectively input into two convolutional neural networks, namely the first convolutional neural network and the second convolutional neural network, which are designed to extract spatial features related to their respective tasks. Convolutional neural networks detect local features in the input data by applying multiple filters, each of which is able to identify specific patterns in the data, which is the key to achieving effective feature extraction. In this process, the first convolutional neural network may identify patterns related to channel layout and heat distribution, while the second network focuses on capturing patterns of coolant flow and temperature changes. Subsequently, the first and second convolutional vectors obtained from the two convolutional networks are spliced ​​together to form a comprehensive spliced ​​relationship vector. Combine all key information related to channel design and coolant characteristics. The spliced ​​vector is input into a shared hidden layer, and the features are synthesized and refined through nonlinear transformation to generate a target feature vector. The target feature vector is input into two different Transformer networks. The first Transformer network uses the target feature vector and the first convolution vector to predict the number of channels, while the second Transformer network uses the target feature vector and the second convolution vector to predict the coolant flow rate. Each network deeply learns the dependencies in the input data through the self-attention mechanism and generates prediction results for its respective tasks. The prediction results (optimal number of channels and optimal coolant flow rate) are input into the output layer of the model. At the output layer, these data are further integrated and output as a final design parameter combination, which directly guides the actual design and manufacture of the liquid cooling system.

[0076] The above describes the energy storage battery temperature management method based on liquid cooling technology in the embodiment of the present application. The following describes the energy storage battery temperature management system based on liquid cooling technology in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the energy storage battery temperature management system based on liquid cooling technology includes:

[0077] An initialization module 201 is used to initialize the design parameter population of the battery liquid cooling system and generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate;

[0078] A simulation module 202 is used to perform heat dissipation simulation and three-dimensional temperature distribution construction on the target energy storage battery during the discharge process based on multiple first liquid cooling system design parameter combinations, and obtain three-dimensional temperature distribution data of each first liquid cooling system design parameter combination;

[0079] A construction module 203 is used to construct a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and to construct a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data;

[0080] The prediction module 204 is used to input the first relationship vector and the second relationship vector into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

[0081] Through the synergy of the above components, the liquid cooling system can transfer and disperse heat more effectively than the traditional air cooling system by using liquid as the cooling medium. The battery can cool down more quickly when running at high load, thereby improving the performance and life of the battery. The temperature uniformity between battery cells and within a single battery is achieved by designing optimized flow channels and flow control. The uniform distribution of temperature is crucial for the long-term stability and safety of the battery, and can avoid structural damage or performance degradation caused by local overheating and large temperature gradients. A multi-task deep learning model is used to predict and optimize the design parameters, coolant flow rate and number of channels of the liquid cooling system. Through intelligent adaptive adjustment, the cooling strategy can be dynamically adjusted according to the actual working conditions and environmental changes of the battery to ensure that the battery system can maintain the best working state in various environments. By accurately controlling the parameters of the liquid cooling system, unnecessary energy consumption can be effectively reduced because the system can avoid overcooling and uneven cooling. The effectiveness of temperature management directly affects the safe operation of the battery. The liquid cooling system can respond quickly to battery temperature changes and effectively control them, thereby reducing the safety risks caused by battery overheating, thereby improving the liquid cooling heat dissipation efficiency of the energy storage battery and the thermal management accuracy of the battery discharge process.

[0082] The present application also provides a computer device, which includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the energy storage battery temperature management method based on liquid cooling technology in the above-mentioned embodiments.

[0083] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein instructions are stored in the computer, and when the instructions are executed on a computer, the computer executes the steps of the energy storage battery temperature management method based on liquid cooling technology.

[0084] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0085] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0086] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A temperature management method for energy storage batteries based on liquid cooling technology, characterized in that: The energy storage battery temperature management method based on liquid cooling technology includes: Initializing a design parameter population for a battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate; Based on the multiple first liquid cooling system design parameter combinations, respectively, a discharge process heat dissipation simulation and a three-dimensional temperature distribution construction are performed on the target energy storage battery to obtain three-dimensional temperature distribution data for each first liquid cooling system design parameter combination; Constructing a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and constructing a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data; The first relationship vector and the second relationship vector are input into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

2. The energy storage battery temperature management method based on liquid cooling technology according to claim 1 is characterized in that: The battery liquid cooling system is initialized with a design parameter population to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate, including: Define the channel quantity threshold and flow parameter threshold of the battery liquid cooling system; By using a particle swarm optimization algorithm, random initialization is performed according to the channel number threshold and the flow parameter threshold to obtain multiple first random values ​​of the channel number and multiple second random values ​​of the coolant flow rate; Combining the plurality of first random values ​​and the plurality of second random values ​​to generate a plurality of candidate liquid cooling system design parameter combinations; Calculate the target fitness data of each candidate liquid cooling system design parameter combination respectively, and obtain the preset standard fitness data; Calculating difference data between the target fitness data and the standard fitness data to obtain target difference data for each candidate liquid cooling system design parameter combination; Calculate the average difference data corresponding to the target difference data of each candidate liquid cooling system design parameter combination, and select parameter combinations for the multiple candidate liquid cooling system design parameter combinations based on the average difference data to obtain multiple first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combination includes: an initial number of channels and an initial coolant flow rate.

3. The energy storage battery temperature management method based on liquid cooling technology according to claim 1 is characterized in that: The method of performing heat dissipation simulation and three-dimensional temperature distribution construction on the target energy storage battery during the discharge process based on the multiple first liquid cooling system design parameter combinations to obtain three-dimensional temperature distribution data for each first liquid cooling system design parameter combination includes: Based on the multiple first liquid cooling system design parameter combinations, heat dissipation simulation models are constructed for the target energy storage batteries respectively to obtain a liquid cooling heat dissipation simulation model for each first liquid cooling system design parameter combination; Based on the preset discharge parameter data, the liquid cooling and heat dissipation simulation model is subjected to a discharge process heat dissipation test and battery temperature data collection to obtain battery temperature data of each liquid cooling and heat dissipation simulation model; Performing a three-dimensional temperature distribution analysis on a plurality of temperature data points in the battery temperature data to obtain a three-dimensional spatial coordinate of each temperature data point; According to the three-dimensional spatial coordinates of each temperature data point, the three-dimensional temperature distribution field mapping is performed on the liquid cooling heat dissipation simulation model to obtain the three-dimensional temperature distribution data of each first liquid cooling system design parameter combination.

4. The energy storage battery temperature management method based on liquid cooling technology according to claim 1 is characterized in that: The step of constructing a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and constructing a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data, comprises: Performing feature analysis on the three-dimensional temperature distribution data to obtain a plurality of temperature distribution features of each first liquid cooling system design parameter combination; Calculating the mean of the plurality of temperature distribution features to obtain a feature mean, and calculating the temperature variation coefficient of the plurality of temperature distribution features according to the feature mean; Calculating the coefficient of variation of the initial number of channels to obtain the coefficient of variation of the number, and calculating the coefficient of variation of the initial coolant flow rate to obtain the coefficient of variation of the flow rate; Calculating the quantity weight of the initial number of channels and the first weight of the temperature variation coefficient according to the quantity variation coefficient and the temperature variation coefficient, and calculating the flow weight of the initial coolant flow and the second weight of the temperature variation coefficient according to the flow variation coefficient and the temperature variation coefficient; Performing vector conversion and vector weighted analysis on the initial number of channels and the plurality of temperature distribution features according to the number weight and the first weight to obtain a first relationship vector between the number of channels and the temperature; The initial coolant flow rate and the multiple temperature distribution characteristics are vector-converted and vector-weighted analyzed according to the flow rate weight and the second weight to obtain a second relationship vector between the coolant flow rate and the temperature.

5. The energy storage battery temperature management method based on liquid cooling technology according to claim 4 is characterized in that: The characteristic analysis of the three-dimensional temperature distribution data to obtain a plurality of temperature distribution characteristics of each first liquid cooling system design parameter combination includes: Performing feature point search on the three-dimensional temperature distribution data to obtain a plurality of initial three-dimensional temperature feature points; Respectively identify neighboring points of the multiple initial three-dimensional temperature feature points to obtain a set of neighboring temperature feature points corresponding to each initial three-dimensional temperature feature point; Constructing a temperature feature point distribution map corresponding to each initial three-dimensional temperature feature point based on the adjacent temperature feature point set; Calculating cluster centers of the temperature characteristic point distribution maps respectively to obtain target cluster centers of each temperature characteristic point distribution map; A plurality of corresponding target three-dimensional temperature feature points are determined according to the target cluster center, and a plurality of temperature distribution features of each first liquid cooling system design parameter combination are generated according to the plurality of target three-dimensional temperature feature points.

6. The energy storage battery temperature management method based on liquid cooling technology according to claim 1, characterized in that: The step of inputting the first relationship vector and the second relationship vector into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination includes: Constructing a multi-task deep learning model, wherein the multi-task deep learning model includes a first convolutional neural network, a second convolutional neural network, a shared hidden layer, a first Transformer network, a second Transformer network, and an output layer; Inputting the first relationship vector and the second relationship vector into the multi-task deep learning model to predict the optimal number of channels and the optimal coolant flow rate, and outputting the optimal number of channels and the optimal coolant flow rate; A second liquid cooling system design parameter combination of the battery liquid cooling system is generated according to the optimal number of channels and the optimal coolant flow rate.

7. The energy storage battery temperature management method based on liquid cooling technology according to claim 6 is characterized in that: The step of inputting the first relationship vector and the second relationship vector into the multi-task deep learning model to perform optimal channel number prediction and optimal coolant flow prediction, and outputting the optimal channel number and the optimal coolant flow, comprises: Inputting the first relationship vector into the first convolutional neural network to perform convolution feature operation to obtain a first convolution vector; Inputting the second relationship vector into the second convolutional neural network to perform convolution feature operation to obtain a second convolution vector; Performing vector concatenation on the first relationship vector and the second relationship vector to obtain a concatenated relationship vector, and inputting the concatenated relationship vector into the shared hidden layer to perform feature extraction to obtain a target feature vector; Inputting the first convolution vector and the target feature vector into the first Transformer network to predict the number of channels to obtain an optimal number of channels, and inputting the second convolution vector and the target feature vector into the second Transformer network to predict the coolant flow rate to obtain an optimal coolant flow rate; The optimal number of channels and the optimal coolant flow rate are integrated and outputted through the output layer.

8. A temperature management system for energy storage batteries based on liquid cooling technology, characterized in that: The energy storage battery temperature management system based on liquid cooling technology includes: An initialization module is used to initialize the design parameter population of the battery liquid cooling system to generate a plurality of first liquid cooling system design parameter combinations, wherein the first liquid cooling system design parameter combinations include: an initial number of channels and an initial coolant flow rate; A simulation module, used to perform heat dissipation simulation and three-dimensional temperature distribution construction on the target energy storage battery during discharge based on the multiple first liquid cooling system design parameter combinations, and obtain three-dimensional temperature distribution data of each first liquid cooling system design parameter combination; A construction module, configured to construct a first relationship vector between the number of channels and the temperature according to the initial number of channels and the three-dimensional temperature distribution data, and to construct a second relationship vector between the coolant flow rate and the temperature according to the initial coolant flow rate and the three-dimensional temperature distribution data; A prediction module is used to input the first relationship vector and the second relationship vector into a preset multi-task deep learning model to perform optimal combination parameter prediction to obtain a second liquid cooling system design parameter combination.

9. A computer device, characterized in that: The computer device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the computer device to execute the energy storage battery temperature management method based on liquid cooling technology as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by the processor, the energy storage battery temperature management method based on liquid cooling technology as described in any one of claims 1 to 7 is implemented.

Images