A thermal management method for a liquid-cooled battery pack energy storage system

Abstract

The application relates to the technical field of battery thermal management, and discloses a kind of independent liquid-cooled battery pack energy storage system thermal management method with warehouse.The method collects the real-time thermal parameter data of system, and identifies the current thermal state of system accordingly;Based on the current thermal state, the corresponding thermal management model is set, which provides the basic framework for thermal management.Meanwhile, the temperature distribution data of the system is monitored in real time, and the thermal adjustment rule is formulated according to the data.Finally, the system is dynamically adjusted according to the thermal adjustment rule set.The synergistic effect of real-time data acquisition, state identification, model setting, temperature monitoring and dynamic adjustment effectively optimizes the thermal distribution of the system, improves the efficiency and reliability of thermal management, and ensures the stable operation of the independent liquid-cooled battery pack energy storage system.The dynamic adjustment mechanism ensures that the thermal management strategy can be optimized in real time with the change of the system thermal state, and improves the stability and adaptability of the system operation.

Description

Technical Field

[0001] This invention relates to the field of battery thermal management technology, specifically to a thermal management method for an energy storage system with an independent liquid-cooled battery pack. Background Technology

[0002] In the field of energy storage, independent liquid-cooled battery pack energy storage systems have been widely used in new energy power plants, industrial and commercial energy storage, and emergency power supply scenarios due to their high-efficiency energy storage and conversion capabilities. The stable operation of such systems is highly dependent on proper thermal management, because battery packs generate a large amount of heat during charge-discharge cycles, long-term storage, and ambient temperature fluctuations. If the heat accumulates or is unevenly distributed, it can easily cause abnormal system operation.

[0003] In existing technologies, thermal management of independent liquid-cooled battery pack energy storage systems often employs control methods with preset fixed parameters. For example, some thermal management schemes set a single cooling or heating threshold based on the rated operating temperature range of the battery pack, initiating the corresponding thermal regulation action when the overall system temperature reaches the threshold. This approach does not consider the dynamic thermal characteristics of the system at different operating stages. For instance, the heat generation rate of the battery pack during high-rate charging and discharging is much higher than during low-rate operation, making it difficult for fixed thresholds to adapt to varying heat generation conditions.

[0004] Traditional thermal management methods lack real-time identification of system thermal status. Most solutions rely on periodic inspections or offline data analysis to determine thermal status, failing to capture subtle changes in thermal parameters during system operation. When problems such as poor local contact or obstructed coolant flow occur in the battery pack, localized heat accumulation may intensify rapidly, but existing methods struggle to detect this in time. Measures are often only taken when the overall temperature becomes abnormal, by which time battery performance has already been negatively impacted.

[0005] In terms of temperature distribution monitoring, existing technologies often employ a limited array of temperature sensors, monitoring only the temperature data on the battery pack surface or coolant inlet and outlet. This makes it difficult to comprehensively reflect the temperature field distribution of the entire energy storage system. In independent liquid-cooled battery pack energy storage systems, differences in battery module density and coolant piping routing can lead to significant temperature variations in different areas, making it difficult to fully capture localized overheating or overcooling. This incomplete temperature monitoring results in a lack of precise targeting in thermal management adjustments, easily leading to a "one-size-fits-all" approach that results in energy waste or persistent localized thermal problems.

[0006] Existing thermal management adjustment rules are mostly based on experience and do not dynamically correlate with real-time temperature distribution data. For example, adjustment parameters such as coolant flow rate and radiator fan speed are preset with fixed gradients, and are adjusted according to these fixed gradients regardless of the actual temperature distribution of the system. When the system experiences both localized high and low temperatures, these fixed rules cannot achieve differentiated adjustment, resulting in insufficient thermal management in some areas and over-adjustment in others, affecting the overall operating efficiency and stability of the system. Summary of the Invention

[0007] The purpose of this invention is to provide a thermal management method for an energy storage system with an independent liquid-cooled battery pack, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides a thermal management method for an energy storage system with an independent liquid-cooled battery pack, the method comprising:

[0009] Collect real-time thermal parameter data of the independent liquid-cooled battery pack energy storage system, and identify the current thermal state of the independent liquid-cooled battery pack energy storage system based on the real-time thermal parameter data;

[0010] The thermal management model of the independent liquid-cooled battery pack energy storage system is set based on the current thermal state.

[0011] Real-time monitoring of temperature distribution data of the independent liquid-cooled battery pack energy storage system;

[0012] The thermal adjustment rules for the independent liquid-cooled battery pack energy storage system are set based on the temperature distribution data.

[0013] The thermal management of the independent liquid-cooled battery pack energy storage system is dynamically adjusted according to the thermal adjustment rules.

[0014] Preferably, the real-time thermal parameter data acquisition of the independent liquid-cooled battery pack energy storage system includes: acquiring battery temperature data and coolant flow rate data of the independent liquid-cooled battery pack energy storage system through multiple types of sensors;

[0015] A temperature gradient sequence is generated based on the battery temperature data;

[0016] A flow distribution curve is generated based on the coolant flow data;

[0017] The real-time thermal parameter data is formed by fusing the temperature gradient sequence and the flow distribution curve.

[0018] Preferably, setting the thermal management model of the independent liquid-cooled battery pack energy storage system based on the current thermal state includes: determining the thermal diffusion characteristics of the independent liquid-cooled battery pack energy storage system based on the temperature gradient sequence;

[0019] A first thermal control model is constructed based on the aforementioned thermal diffusion characteristics;

[0020] The cooling efficiency characteristics of the independent liquid-cooled battery pack energy storage system are determined based on the flow distribution curve.

[0021] A second thermal control model is constructed based on the aforementioned cooling efficiency characteristics;

[0022] The thermal management model is generated by combining the first thermal control model and the second thermal control model.

[0023] Preferably, the real-time monitoring of temperature distribution data of the independent liquid-cooled battery pack energy storage system includes: continuously collecting real-time temperature values ​​of multiple battery pack units of the independent liquid-cooled battery pack energy storage system;

[0024] A temperature deviation trend line is generated based on the real-time temperature values ​​of the multiple battery pack units;

[0025] Based on the temperature deviation trend line, the thermal uniformity fluctuation data of the independent liquid-cooled battery pack energy storage system is identified.

[0026] The thermal uniformity fluctuation data is used as the temperature distribution data.

[0027] Preferably, setting the thermal adjustment rules for the independent liquid-cooled battery pack energy storage system based on the temperature distribution data includes: analyzing the ambient temperature change trend of the independent liquid-cooled battery pack energy storage system;

[0028] Potential thermal risk factors are identified by combining the thermal uniformity fluctuation data and the ambient temperature change trend.

[0029] Based on the aforementioned potential thermal risk factors, the risk response methods for the independent liquid-cooled battery pack energy storage system are set;

[0030] The heat adjustment rules are generated based on the aforementioned risk response methods.

[0031] Preferably, the analysis of the ambient temperature change trend of the independent liquid-cooled battery pack energy storage system includes: classifying the ambient temperature change trend to obtain classified environmental data;

[0032] Extract the key environmental features from the classified environmental data;

[0033] Historical environmental data are collected based on the aforementioned key environmental characteristics;

[0034] Based on the historical environmental data, analyze the environmental temperature fluctuation patterns and cooling demand patterns;

[0035] The ambient temperature change trend is generated by combining the ambient temperature fluctuation pattern and the cooling demand pattern.

[0036] Preferably, the step of identifying potential thermal risk factors by combining the thermal uniformity fluctuation data and the ambient temperature change trend includes: identifying the normal thermal parameter range of the independent liquid-cooled battery pack energy storage system based on the thermal uniformity fluctuation data;

[0037] The current thermal behavior deviation is generated by comparing the real-time temperature values ​​of the multiple battery pack units with the normal thermal parameter range;

[0038] An abnormal thermal threshold is determined based on the current thermal behavior deviation and the trend of ambient temperature change.

[0039] The potential thermal risk factors are identified based on the abnormal thermal threshold.

[0040] Preferably, the risk response method for setting the independent liquid-cooled battery pack energy storage system according to the potential thermal risk factors includes: locating the abnormal battery pack unit corresponding to the potential thermal risk factors;

[0041] Analyze the thermal predicament type of the abnormal battery pack unit;

[0042] The thermal strain strategy for the abnormal battery pack unit is set based on the thermal predicament type.

[0043] The risk response method is generated based on the thermal strain strategy.

[0044] Preferably, the step of dynamically adjusting the thermal management of the independent liquid-cooled battery pack energy storage system according to the thermal adjustment rules includes: identifying thermal management triggering conditions based on the thermal adjustment rules;

[0045] A coolant flow rate adjustment command is generated based on the aforementioned thermal management triggering conditions;

[0046] The coolant flow rate adjustment command is executed to adjust the coolant distribution in stages;

[0047] During the adjustment process, the temperature feedback data and the predicted values ​​of the thermal management model are verified in real time.

[0048] The thermal adjustment rules are iteratively updated based on the verification results.

[0049] Preferably, the method further includes: collecting the final temperature data of the independent liquid-cooled battery pack energy storage system;

[0050] The final temperature data is compared with the expected temperature range of the thermal management model to generate a model error signal;

[0051] Adjust the parameters of the thermal management model based on the model error signal;

[0052] The adjusted parameters are updated to the historical parameter library for subsequent thermal management.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] This thermal management method acquires real-time thermal parameter data from an independent liquid-cooled battery pack energy storage system, enabling timely capture of changes in the system's thermal characteristics at different operating stages. Unlike traditional thermal management methods that rely on fixed parameters, this real-time sensing mechanism allows for more accurate identification of the system's thermal state, avoiding untimely thermal management issues caused by lagging thermal parameters or preset deviations. When the system is under high-load charging and discharging conditions, thermal parameters change rapidly; real-time acquisition can detect these changes immediately, providing a basis for subsequent thermal management adjustments.

[0055] This approach sets up a thermal management model based on the current thermal state, allowing model parameters to dynamically adapt to the actual thermal conditions of the system. Traditional thermal management models often have a fixed architecture, making it difficult to cope with the differences in thermal demands under different ambient temperatures and operating loads. In this method, the model settings are directly linked to the real-time thermal state. When the system's thermal state changes due to changes in the external environment or internal operating conditions, the thermal management model can adjust accordingly, thereby more accurately matching the system's current thermal management needs and improving the adaptability of thermal management.

[0056] Real-time monitoring of temperature distribution data overcomes the limitations of traditional methods that only monitor localized temperatures, enabling a comprehensive understanding of the temperature conditions in all areas of the battery pack. The internal structure of an independent liquid-cooled battery pack energy storage system is complex, with varying heat dissipation conditions for battery modules in different locations, making localized temperature imbalances common. Comprehensive temperature distribution monitoring clearly identifies areas of excessively high or low temperatures, revealing the heat transfer paths and distribution patterns within the system. This provides a basis for subsequent targeted adjustments, preventing system problems caused by undetected localized temperature anomalies.

[0057] Setting thermal adjustment rules based on temperature distribution data makes thermal management adjustments more targeted and precise. Traditional adjustment rules often rely on overall temperature thresholds for coarse adjustments, failing to account for temperature differences in different regions. This method, however, formulates rules based on specific temperature distributions. For higher-temperature areas, coolant flow can be increased or localized heat dissipation enhanced; for lower-temperature areas, heat dissipation intensity can be appropriately reduced to avoid energy waste. This differentiated adjustment approach enables balanced temperature control within the system, reducing the impact of localized thermal stress on battery performance.

[0058] Dynamically adjusting the system using thermal adjustment rules enables more timely and flexible thermal management responses. During system operation, the thermal state and temperature distribution are constantly changing, and fixed adjustment methods are insufficient to adapt to these dynamic changes. A dynamic adjustment mechanism can continuously optimize the adjustment strategy based on real-time monitoring of thermal parameters and temperature distribution changes. When the system heat load increases, it promptly increases the heat dissipation capacity; when the heat load decreases, it appropriately reduces the heat dissipation intensity, ensuring the system remains in a reasonable thermal environment. This guarantees that the battery pack operates under suitable temperature conditions, reduces system fluctuations caused by improper thermal management, and extends the system's stable operating cycle. Attached Figure Description

[0059] Figure 1 This is a schematic diagram illustrating the working principle of the thermal management method for an energy storage system with an independent liquid-cooled battery pack as described in this invention.

[0060] Figure 2 A flowchart for real-time thermal parameter data acquisition;

[0061] Figure 3 Flowchart for setting up a thermal management model;

[0062] Figure 4 This is a flowchart for analyzing the trend of environmental temperature changes. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Please see Figure 1 This invention provides a thermal management method for an independent liquid-cooled battery pack energy storage system. This method achieves effective thermal management of the independent liquid-cooled battery pack energy storage system through real-time data acquisition, status identification, model setting, monitoring and analysis, and dynamic adjustment. Specific implementation methods are as follows:

[0065] This approach involves collecting real-time thermal parameter data from an independent liquid-cooled battery pack energy storage system, identifying the system's current thermal state based on this data, setting a thermal management model for the system based on the current thermal state, monitoring the system's temperature distribution data in real time, setting thermal adjustment rules for the system based on the temperature distribution data, and dynamically adjusting the system's thermal management according to the thermal adjustment rules. This solution ensures that the system maintains thermal stability during operation and avoids overheating or uneven temperature issues.

[0066] Example 1: See Figure 2The initial step in this thermal management method is to collect real-time thermal parameter data from an independent liquid-cooled battery pack energy storage system. Its implementation involves multi-layered sensing technologies, data generation mechanisms, and data fusion processing. The system deploys a temperature sensor network at multiple key locations. This network consists of numerous high-precision digital temperature sensor nodes, with sensor contacts located on the outer surface of each battery pack unit, key nodes between internal cells, and the positive and negative electrode connection areas. These sensors perform periodic temperature measurements according to a preset synchronous sequence, with a sampling frequency set to once per second to ensure the capture of transient temperature changes that may occur during battery charging, discharging, or resting. The temperature sensors transmit the collected raw temperature values ​​to the central data processing unit via standardized digital communication interfaces (such as I2C or SPI). Upon receiving this raw data, the data processing unit performs preprocessing operations, including eliminating signal noise caused by environmental electromagnetic interference, smoothing short-term fluctuations using a sliding window averaging filter algorithm, and marking and removing obviously abnormal measurement values ​​that exceed the physically possible range. This ultimately forms a valid battery temperature dataset that can be used for subsequent analysis.

[0067] Based on the preprocessed battery temperature data, the system dynamically generates a temperature gradient sequence. This sequence is an important indicator characterizing the spatial distribution and changing trend of heat within the system. The generation process involves spatial relationship calculations of synchronous temperature data collected from multiple battery pack units. Specifically, the data processing unit first identifies pairs or groups of battery units that are spatially adjacent or have significant heat conduction paths, based on the physical layout and topological connections (series or parallel structures) of the battery pack units. For each such spatially related unit group, the temperature difference at the same sampling point is calculated. The system maintains a continuous difference calculation process, and over time, these time-sequentially arranged discrete unit temperature differences constitute the basic temperature difference sequence. To enhance the sequence's sensitivity to thermal change trends and reduce the interference of instantaneous disturbances, the system applies post-processing to this basic sequence. Post-processing includes applying a moving average algorithm over a certain time span to calculate the average value sequence of differences within a selected time window, while simultaneously performing dispersion analysis on the sequence data points and smoothing interpolation correction for abnormal difference points that significantly deviate from neighboring points. The processed sequence can clearly reflect the rate and direction of heat transfer between different regions within the system.

[0068] At the coolant flow monitoring level, the system installs flow sensors at key locations in the cooling circulation loop. High-precision flow meters are installed on the main coolant inlet and outlet pipes, as well as at the inlet of the coolant branches supplying each independent battery pack unit. These flow meters can measure the volumetric or mass flow rate of the coolant flowing through the pipes in real time and output the measured values ​​at a constant frequency (e.g., several times per second). The raw sampled values ​​of the coolant flow data are also transmitted to the central data processing unit. Based on the timestamps of the flow meter measurements, the system constructs a raw data point set showing the change of coolant flow over time. To form a flow distribution curve characterizing the dynamic cooling intensity of the entire system and its local areas, the data processing module performs trajectory processing on the raw flow data. This processing, based on the time axis coordinate, connects the received discrete flow data points into continuous line segments, visually displaying the flow velocity trajectory in the current and short-term past. For the multi-level flow control structure that may exist in the system (such as the total flow of the main loop, the flow of regional branches, and the flow of a single pack inlet), the system can simultaneously generate and maintain multiple corresponding flow distribution curves, reflecting the coolant distribution status at different levels.

[0069] The fusion of temperature gradient sequences and flow distribution curves to form the final real-time thermal parameter data is a process of unifying the heat distribution state and cooling response state in the time dimension. The data processing unit stores this fused integrated data in a structured table in the system database. Each record includes a strictly synchronized timestamp, an associated spatial location identifier (the unit group corresponding to the temperature gradient), the current characteristic values ​​of the gradient sequence (such as moving average, sequence slope), and key feature points of the relevant coolant flow distribution curve (such as the current flow rate, the rate of change of flow rate in the previous time window, and the associated branch identifier). The entire acquisition process is managed by a dedicated task scheduler in the central controller. This scheduler ensures that sensor data reading, preprocessing, sequence / curve generation, data fusion, and storage are completed within a specified hard real-time period, meeting the timeliness requirements of the thermal management system. The database storage design supports efficient historical data backtracking and real-time data stream reading, providing complete, accurate, and consistent basic data support for subsequent identification of the current thermal state.

[0070] Example 2: See Figure 3In setting up the thermal management model for an independent liquid-cooled battery pack energy storage system, the system first utilizes a generated temperature gradient sequence to deeply analyze thermal diffusion characteristics. The temperature gradient sequence serves as the input dataset, with its sequence values ​​representing the trend of temperature difference changes between adjacent battery pack cells at different times. The system calculates thermal diffusion characteristic parameters by analyzing the temporal evolution characteristics of this sequence. Specifically, the sequence differencing method is used to calculate the change in gradient values ​​at consecutive time points, reflecting the thermal diffusion rate. Simultaneously, the statistical characteristics of the sequence within a certain time window are calculated, such as the mean and variance of the gradient change amplitude. These statistics reveal the uniformity of heat transfer within the system and the overall conduction efficiency. By establishing a correlation function between the gradient sequence values ​​and previous historical temperature data, the response delay characteristics of thermal diffusion to temperature changes are further identified. These parameters together constitute a complete quantitative description of the system's thermal diffusion characteristics.

[0071] Thermal diffusivity tensor The components are quantified through the statistical characteristics of the temperature gradient sequence. The specific calculation steps are as follows: take the temperature gradient sequence data of 10 consecutive sampling periods (sampling frequency 1 second / time), and calculate the average temperature difference between adjacent battery pack units in each period. with standard deviation If the battery pack is in series topology, The main diagonal component takes the value of (unit: For non-main diagonal components, the value is 0.3 times that of the main diagonal components; for parallel topologies, the value of the main diagonal components is [value missing]. The non-principal diagonal component is taken as 0.5 times the principal diagonal component; when the battery pack type is ternary lithium battery, Multiply by a correction factor of 1.1 for the whole; if it is a lithium iron phosphate battery, multiply by a correction factor of 0.9.

[0072] Based on the identified thermal diffusion characteristics, a first thermal control model is constructed. The core objective of this model is to describe and predict the spatial propagation behavior of heat between battery pack cells. Its mathematical expression employs a partial differential equation framework to simulate the variation of the temperature field with time and spatial location. The equations are constructed based on the principle of thermal energy conservation and utilize the previously extracted thermal diffusion characteristic parameters to parameterize the model coefficients. The specific equations are as follows:

[0073]

[0074] in: Representing the temperature field of the battery system, it is determined by spatial location and time. The function; Represents a time variable; It is a coefficient tensor directly related to thermal diffusion characteristics. Its components are determined by the thermal diffusion rate and directional characteristics obtained from gradient sequence analysis, reflecting the combined effect of material thermal conduction and inter-unit contact thermal resistance. It is the Laplace operator, representing the thermal diffusion process in space; It is a coupling coefficient that incorporates the influence intensity of the heat source term into the model; This represents a function representing the heat source / heat sink term equivalent to the action of the coolant, and this function depends on the location of the coolant flow path. and time Its functional form is influenced by the cooling efficiency characteristics and serves as the input to the second thermal control model. Before the second thermal control model outputs the cooling efficiency characteristics, the heat source term... Assigning values ​​according to temporary rules: Initial stage uses ambient temperature. With the battery's rated operating temperature The difference, i.e. ,in Proportional coefficient (ternary lithium) Lithium iron phosphate When the flow distribution curve first outputs stable data (fluctuation ≤ 5% for 5 consecutive sampling periods), the cooling heat dissipation is immediately calculated using the second thermal control model. replace ,Finish The initial update.

[0075] This model is discretized during the solution process and mapped onto the geometric layout mesh of the battery pack unit. The discretization of the first thermal control model adopts the finite volume method, and the specific mapping rule is as follows: each battery pack unit is divided into 20 mesh nodes according to the physical dimensions of length × width × height = 10cm × 10cm × 20cm (1 node every 1cm along the height direction); the heat capacity parameters of the mesh nodes are... According to battery material density (ternary lithium) Lithium iron phosphate ), specific heat capacity (ternary lithium) Lithium iron phosphate ) and node volume Calculation, i.e. ;

[0076] 3) Thermal resistance of adjacent grid nodes Based on the thermal conductivity of battery materials (ternary lithium) Lithium iron phosphate ) and node spacing Calculation, i.e. ( (The contact area between nodes is taken as 1 cm²). The physical properties of the mesh nodes (such as heat capacity and thermal resistance of adjacent nodes) are initialized by the mechanical and material structural parameters of the system. The initial state parameters of the model (such as the initial temperature distribution) are loaded from the current temperature sensor data. The model solver uses the finite volume or finite difference method to perform numerical solutions and obtain the predicted spatiotemporal evolution of temperature.

[0077] Simultaneously, the system determines cooling efficiency characteristics based on the flow distribution curve. The flow distribution curve, with time on the horizontal axis and coolant velocity on the vertical axis, depicts the dynamic history of coolant delivery intensity. The analysis of cooling efficiency characteristics focuses on the curve's changing features: the instantaneous velocity value characterizes the cooling intensity potential at that time; the change in velocity within a time window (calculated through the curve slope) reflects the agility of the cooling response; and the average velocity within a specific operating range (such as the high-rate discharge period) indicates the average cooling capacity level under that condition. The curve's fluctuation pattern and trend inflection points are used to determine the dynamic response capability and potential control inertia of the cooling system. These characteristic values ​​are quantified and extracted to form a set of efficiency characteristic parameters describing the system's cooling performance. Cooling efficiency characteristics are expressed as heat dissipation power per unit flow rate. Quantification, the calculation formula is as follows ,in: The outlet temperature of the coolant is (°C). The coolant inlet temperature (°C) is collected by the pipeline temperature sensor. The coolant volumetric flow rate (L / min) is collected by a flow sensor. The specific heat capacity of the coolant (water-ethylene glycol solution (50:50)). mineral oil .when At this time, it is determined to be in a high-efficiency cooling state; This is the normal state; This is an inefficient state.

[0078] Based on the extracted cooling efficiency characteristic parameters, a second thermal control model is constructed. This model focuses on describing the heat exchange effect of the coolant flow and its dynamic adjustment process. The model adopts a coupled equation framework of fluid dynamics and heat transfer, with key equations including the Navier-Stokes equations (or simplified pipe flow equations) describing the flow characteristics of the coolant in the channel network, and the heat transfer equations describing the convective heat exchange between the coolant and the battery pack housing / cell. Key coefficients in the model equations, such as the channel fluid resistance coefficient and the convective heat transfer coefficient, are calibrated by the cooling efficiency characteristic parameters. For example, the heat transfer coefficient is related to the average flow velocity and the response speed; the flow resistance parameter is related to the flow fluctuation mode and the characteristics of the control mechanism. The model establishes a computational grid according to the channel topology (such as parallel and series branches) and sets boundary conditions (such as pump pressure and valve opening). During simulation execution, the model receives current inputs such as the coolant inlet temperature and valve control signals, and calculates the temperature rise of the coolant at different locations and the resulting heat dissipation distribution of the battery cells.

[0079] In the second thermal control model, the flow resistance coefficient convective heat transfer coefficient The correlation rule with the flow distribution curve is: flow resistance coefficient When the fluctuation range of the flow distribution curve hour, Take a fixed value (water-ethylene glycol solution) mineral oil ;when hour, according to Adjustment( (Base value); convective heat transfer coefficient When the flow distribution curve shows an upward trend (for 3 consecutive sampling periods) When increasing), according to Adjustment( Based on the value, (Rated flow rate); when it shows a downward trend, according to Adjustment, and The minimum value is not lower than .

[0080] The boundary conditions for the second thermal control model are set according to the coolant type: initial pump pressure value. Water-glycol solution system mineral oil system Initial value of valve opening When the system is in a static state, When in a charging / discharging state, Coolant inlet temperature : Default is based on ambient temperature Setting, if the ambient temperature , Increased through the preheating module .

[0081] After the first and second thermal control models are constructed, the system performs model fusion to generate a unified thermal management model. The fusion process is iteratively coupled across discrete time steps. Within each time step, the first thermal control model provides the currently predicted temperature field distribution. .Will The results are passed to the second thermal control model as boundary conditions for calculating the heat exchange between the coolant and the battery cells. Based on the current coolant flow rate and control input, the second thermal control model calculates the updated heat exchange distribution field. The updated version The heat source term is then substituted into the first thermal control model, driving the model to proceed to the next time step for prediction. This process iterates repeatedly, forming a closed loop. The fusion model ultimately outputs a comprehensive prediction of the system's future state (temperature field, cooling demand, hotspot risk) and has the function of optimizing the calculation of the optimal coolant flow control command. After initial deployment, the coefficients of the thermal management model are periodically calibrated and updated based on newly generated temperature gradient sequences and flow distribution curve data to correct model errors and enhance prediction accuracy.

[0082] Iteration time steps of the first and second thermal control models Fluctuation frequency of the temperature deviation trend line Dynamic adjustment: Calculate the number of fluctuations in the temperature deviation trend line within 1 minute (the number of intervals between adjacent peaks / troughs), and obtain... ;like , ;like , ;like , .

[0083] The convergence criterion for iterative coupling is a dual-threshold criterion: ① Temperature field deviation threshold: Within three consecutive iteration steps, the predicted temperature field output by the first model... Temperature field calculation values ​​fed back from the second model Maximum deviation ② Cooling power deviation threshold: Within three consecutive iteration steps, the cooling power output by the second model Cooling power required by the first model deviation ③ When both of the above thresholds are met, the model is determined to be converged and the iteration is stopped; otherwise, the coupled calculation continues until the maximum number of iterations is reached (default 100 times; if it still does not converge, the model calibration mechanism is triggered).

[0084] When the following abnormal operating conditions occur, the iterative coupling should be paused and emergency adjustments should be performed first: If the temperature of a battery pack unit rises by ≥3℃ within 10 seconds, the iteration should be paused immediately, the opening of the coolant branch valve corresponding to that unit should be increased by 30%, and the iteration should be restarted after the temperature increase rate is ≤0.5℃ / minute; If the coolant flow rate drops by ≥20% within 5 seconds, the iteration should be paused, the pump pressure should be increased by 10%, and the cooling branches of two non-critical battery packs should be closed (prioritizing the unit with the lowest temperature). The iteration should be restarted after the flow rate recovers to more than 90% of the rated value.

[0085] In the real-time temperature distribution data monitoring stage, the system continuously and intensively collects the current temperature readings of each battery pack unit. This relies on a multi-point temperature sensor network deployed on the surface of each individual battery pack unit. Sensor data is uploaded to the central processing unit at a high frequency (milliseconds). Upon receiving the real-time temperature data stream, the data processing unit immediately performs spatial statistical analysis: calculating the average temperature of all units at a given sampling time point, and then calculating the deviation of each unit's current temperature value from the average value. The system maintains a set of deviation data continuously calculated for all units over a period of time. In the time dimension, the deviation data is filtered to remove measurement noise. The system further utilizes time series analysis technology to track the trajectory of temperature deviation changes of each unit over time. By aggregating the deviation sequence data of all units, a temperature deviation trend line characterizing the overall temperature non-uniformity of the system is generated. This trend line is a comprehensive reflection of the spatial discreteness of temperature values ​​(quantified by standard deviation) and the temporal drift (described by the slope and curvature of the trend line). The system extracts key statistical features from the temperature deviation trend line: peak fluctuation amplitude, duration of rise / fall, frequency of main fluctuations, etc., which are comprehensively expressed as thermal uniformity fluctuation data. This data provides a quantitative indicator for subsequent analysis of the system's stability changes in the temperature dimension.

[0086] Example 3: See Figure 4Based on temperature distribution data, the thermal adjustment rules for the independent liquid-cooled battery pack energy storage system are established, starting with a detailed analysis of the ambient temperature change trend. The system is equipped with an ambient temperature sensor that records the surrounding air temperature at fixed time intervals, forming a continuous raw ambient temperature data stream. This data stream is preprocessed by the data processing module, including removing transient interference noise and filling in missing values ​​caused by brief communication failures. The processed ambient temperature sequence is then classified to obtain categorized environmental data. Classification is based on preset temperature range division rules and a dynamic threshold for the rate of temperature change: temperature values ​​are mapped to discrete temperature levels (e.g., low temperature, normal temperature, and high temperature). Simultaneously, the absolute value of the temperature difference between adjacent sampling points is compared with a set threshold to determine the change state at each time point: basically stable, fluctuating upwards, or fluctuating downwards. Finally, the ambient temperature state at each moment is jointly labeled by its temperature level and change state, forming a categorized environmental data sequence with multi-attribute tags.

[0087] Extracting key environmental features from categorized environmental data is a crucial step in trend analysis. Feature extraction focuses on indicators that characterize the long-term properties and dynamic changes of environmental temperature. Core extracted features include: the mean and extreme values ​​of environmental temperature within a selected observation window and their frequency; extreme values ​​of environmental temperature change over a continuous 24-hour period (i.e., diurnal temperature range); and whether environmental temperature exhibits a periodic pattern. The system accumulates and identifies various periodic pattern libraries through historical data analysis, such as seasonal patterns (e.g., high-temperature and high-fluctuation patterns in summer, and low-temperature and stable patterns in winter) and weather patterns (e.g., sunny-day radiation warming patterns, and cloudy-rainy and cooling and stable patterns). For the current categorized environmental data sequence, its statistical characteristics are calculated and compared with pattern library templates to determine the most matching historical environmental pattern type. This matching result and its corresponding statistical feature values ​​together constitute the key environmental feature set.

[0088] Based on the extracted key environmental features, the system retrieves and collects relevant historical environmental data. This historical data originates from environmental temperature records and accompanying system operation parameter logs stored during the system's long-term operation. The search criteria are indexed by key environmental features: for identified specific historical environmental patterns, the system retrieves the complete historical time-series dataset for that pattern; simultaneously, it focuses on specific numerical ranges defined by indicators such as average temperature and diurnal temperature range within the current environmental features, retrieving data from all historical periods falling within that range. The collected historical environmental datasets form the foundation for environmental trend analysis.

[0089] The analysis of environmental temperature fluctuation patterns and cooling demand patterns is based on historical environmental data. Environmental temperature fluctuation pattern analysis utilizes time series modeling techniques to detect regular temperature fluctuations in the collected historical dataset. The analysis process includes: decomposing the historical environmental temperature series into long-term trend terms, periodic terms, and short-term fluctuation terms on a time axis; identifying recurring temperature change inflection point patterns (such as rapid heating followed by slow cooling, and step-like heating patterns) and their duration and amplitude characteristics; and statistically analyzing the frequency and intensity distribution of various fluctuation patterns under different historical environmental modes. Cooling demand pattern analysis focuses on the actual thermal management response records of the system within the corresponding time period of the historical environmental data. These records document the actual control operations performed by the system at that time to maintain the target temperature, such as coolant flow rate and valve opening. By correlating environmental temperature change data with cooling control command data from the same period, a mapping relationship is established between environmental temperature indicators (such as instantaneous temperature values, heating rates over the past N minutes) and cooling intensity demand levels (such as maintaining flow rate, slight increase, and significant increase). The distribution of required cooling intensity levels and average response delay time under different environmental modes is quantitatively analyzed. The analysis results of the two modes serve as characteristic labels describing the environmental requirements for system thermal management.

[0090] Combining ambient temperature fluctuation patterns and cooling demand patterns, the system ultimately generates an ambient temperature change trend within a future prediction time window. This trend is expressed as a multi-dimensional prediction vector, whose dimensions include: a sequence of point estimates (or expected values) of ambient temperature at future time steps within the prediction window; descriptions of major fluctuation patterns (such as anticipated rapid warming events or periods of sustained high temperatures) and their predicted intensity; and a sequence of major cooling demand levels inferred from this trend, along with their predicted change time points. This trend information is not a single temperature curve but rather integrates the most probable temperature change path driven by the environment and the corresponding expected cooling responses. The trend generation calculation can be modeled as follows:

[0091]

[0092] in: Represents the time index With spatial / environmental pattern index The overall trend of ambient temperature change; It is a weighting factor for environmental fluctuation patterns, and its value is determined by the intensity of the currently identified environmental pattern; It is the characteristic function output of the historical environmental fluctuation pattern, and its input is... It is the currently identified pattern type and its quantization parameters; It is a weighting factor for the cooling demand pattern, which depends on the degree of deviation between the current cooling demand level and the historical average level; It is the output of the characteristic function of historical cooling demand patterns, and the input is... These are parameters for predicting environmental fluctuations. This indicates the current operating load status of the energy storage system.

[0093] The system then combines thermal uniformity fluctuation data with the generated ambient temperature change trend. A comprehensive assessment is conducted to identify potential thermal risk factors. Thermal uniformity fluctuation data, presented as a time series, reflects the dispersion of temperature differences between different battery pack units within the system. The identification process first uses historical safe operating data to statistically summarize the typical normal range of thermal uniformity (i.e., the expected normal thermal parameter range) under specific operating conditions (such as different environmental modes and different charge / discharge powers). During real-time operation, the system calculates the relative position of the thermal uniformity value measured at the current time point to this preset normal range, specifically checking whether it exceeds the upper limit boundary value of this range (representing excessively dispersed temperature distribution and the risk of local hot spots). If a thermal uniformity exceeding the limit is observed continuously or cumulatively for a certain threshold time, the event is marked as a sign of temperature dispersion risk. Subsequently, the system will... The predicted trend of ambient temperature change is superimposed into this risk symptom analysis: for example, if the current thermal uniformity is within the boundary safety zone but The predicted rapid rise in ambient temperature or entry into a high-temperature mode in the short term significantly amplifies the risk of thermal runaway. Conversely, if the current thermal uniformity fluctuates abnormally but... The predicted environmental conditions will stabilize and cool down, potentially mitigating the risks. The system quantifies the current thermal behavior deviation (i.e., the degree to which observed thermal uniformity deviates from the normal range) and its trend. The coupling values ​​of the intensities of various adverse changes are dynamically used to determine a comprehensive, real-time abnormal thermal threshold that takes into account future environmental drivers. Any coupling state that reaches or exceeds this abnormal thermal threshold is identified as a potential thermal risk factor. The description of the risk factor includes its type (e.g., the risk of potential hotspot expansion driven by high-temperature environment, the stress risk of excessive temperature difference during cold start, etc.), location information (located by the abnormal thermal uniformity area), potential time window of impact, and estimated severity level.

[0094] Based on the specific attributes of identified potential thermal risk factors (such as type, location, impact time window, and severity level), the system sets corresponding risk response methods. These methods essentially establish a set of logical rules or parameter adjustment instructions to counteract specific risk factors. For example, for a high-temperature environment-driven potential hotspot risk identified in a specific battery pack area, responses might include: increasing the baseline opening of the coolant branch valve leading to that area, increasing the target value of the coolant flow rate in that area, setting a maximum allowable flow rate limit (to prevent excessive local pressure), and shortening the monitoring-feedback response cycle of relevant flow control. For identified cold-start large temperature difference risks, responses might include instructing a small flow rate of coolant to pre-circulate during the initial startup phase to preheat the cooling system, limiting the heating rate, and applying gradient temperature rise instructions. Each risk response method is configured with detailed encoding of its triggering conditions (i.e., the type and level of a specific risk factor) and the sequence of control actions to be executed. Finally, based on the set of risk response methods set for all currently identified risk factors, the system abstracts and encodes complete thermal adjustment rules. These rules are stored in the control database in the form of computer-executable decision tables or state transition logic.

[0095] Example 4: When identifying potential thermal risk factors, the normal thermal parameter range of the independent liquid-cooled battery pack energy storage system is first determined based on thermal uniformity fluctuation data. The system maintains a historical database storing thermal uniformity data records from past operations. This data is categorized and archived according to different environmental conditions and operating states (such as charging, discharging, and resting). The normal thermal parameter range is defined through statistical analysis of this historical data: the system calculates the statistical distribution of thermal uniformity data points under similar operating conditions and extracts the value range corresponding to safe operation. Specifically, this process includes setting lower and upper thresholds. The lower threshold indicates the minimum allowable temperature difference (to avoid overcooling), and the upper threshold indicates the maximum allowable temperature difference (to prevent hotspot formation). These thresholds are stored numerically in the system configuration file and are periodically updated based on new data to ensure adaptability. For example, in the current environment, with the system operating at a base temperature of 25°C, the upper threshold of the normal thermal parameter range is 5.0°C (representing the maximum allowable temperature difference between cells), and the lower threshold is 0.5°C; the range exists in tabular form in the control unit's memory and can be retrieved for comparison at any time.

[0096] The system monitors the real-time temperature values ​​of multiple battery pack units and compares them with normal thermal parameter ranges to generate the current thermal behavior deviation. This step involves reading the sampled temperature values ​​from the sensors of each battery pack unit (sampling period on the order of milliseconds) and calculating the real-time temperature difference between units (i.e., the maximum temperature difference between units minus the minimum temperature difference). The system maintains a unit location index table, where each unit has a unique identifier (such as a physical location code) for easy location. The deviation calculation process includes comparing the real-time temperature difference between the unit and the aforementioned upper and lower thresholds. The deviation value is defined as the difference between the real-time temperature difference and the upper threshold (a positive value indicates a risk of exceeding the limit), or the absolute difference between the real-time temperature difference and the lower threshold (a negative value indicates a deviation). The results are processed in structured data form, including the magnitude and direction of the deviation (e.g., a positive deviation indicates an excessive temperature difference leading to a high risk of overheating, while a negative deviation indicates an insufficient temperature difference leading to inadequate heat dissipation). After deviation calculation, a status flag is marked to indicate whether the deviation exceeds the normal range boundary. The table below shows data from a specific example scenario: The system has five battery pack units (labeled P1 to P5). At time T, real-time temperature values ​​are collected from sensors. The upper limit of the normal range is set to 5.0°C (representing the maximum allowable temperature difference), and the lower limit is set to 0.5°C (the minimum allowable temperature difference). Deviation and risk status are calculated. The table is stored in the system cache for subsequent analysis. The data is automatically generated and verified by the control processor; see Tables 1 and 2.

[0097] Table 1: Battery Pack Cell Temperature Status Table

[0098] Location identifier Real-time temperature (°C) Upper limit of normal range (°C) Lower limit of normal range (°C) P1 26.0 5.0 0.5 P2 28.5 5.0 0.5 P3 30.0 5.0 0.5 P4 27.0 5.0 0.5 P5 31.0 5.0 0.5

[0099] Table 2: Thermal Safety Range Checklist

[0100] Statistical items Value (°C) Upper limit of normal range (°C) Lower limit of normal range (°C) Calculate the deviation (°C). Deviation direction Is it outside the normal range? inter-unit temperature difference 4.0 5.0 0.5 -1.0 negative deviation no

[0101] Note: Location identifiers "P1" to "P5" represent five battery pack units. The real-time temperature value is the absolute temperature of the unit. The real-time temperature difference is calculated by subtracting the lowest temperature (26.0°C) from the highest temperature (31.0°C for P5), resulting in 5.0°C. The calculated deviation is the real-time temperature difference (5.0°C) minus the upper limit of the normal range (5.0°C), which equals 0°C (no deviation). However, in actual storage, a deviation value of 0°C is marked as a neutral state. During comparison, it was found that the current temperature difference equals the upper limit value, and the system marked it as a potential risk, but it was not exceeded.

[0102] Based on the current thermal behavior deviation and the ambient temperature change trend, the system determines the abnormal thermal threshold. The ambient temperature change trend is obtained from external sensors, predicting the temperature change path in the future time window (e.g., rising from the current 25°C to 28°C). The system analyzes the magnitude and direction of the deviation data: in the example scenario, a deviation value of 0°C indicates that the temperature difference is close to the upper limit threshold; simultaneously, the trend predicts an ambient temperature rise of 2°C, which is expected to increase the heat load by 20%. The abnormal thermal threshold is dynamically calculated by the control algorithm: the algorithm combines deviation data (e.g., deviation magnitude of 0°C) and trend parameters (e.g., heating rate, expected load) to output a comprehensive safety boundary value. The calculation process takes into account the deviation magnitude, direction, and environmental trend intensity coefficient, and outputs an adjustment threshold (e.g., set to 5.5°C). If the real-time temperature difference exceeds this adjustment threshold, the system marks an abnormal state. The algorithm runs on the local computing core and uses a real-time feedback mechanism to update the threshold to prevent false positives. In the current example, the abnormal thermal threshold is set to 5.2°C (considering the buffer of the ambient temperature rise), but since the real-time temperature difference is 5.0°C, which is not exceeded, the system remains in a warning state.

[0103] Potential thermal risk factors are identified based on abnormal heat thresholds. The identification process scans and compares the real-time temperature difference of all units with an adjusted threshold (5.2°C). If the threshold is exceeded, a risk type code is assigned (e.g., type code "Overheat" represents hotspot risk). In the example, the temperature difference is not exceeded, but it is close to the threshold. The system checks the details of the deviation distribution: through the location identifier, it is found that the temperature of unit P5 (31.0°C) is significantly higher than others, while the environmental trend shows a warming trend, leading to the predicted risk of local hotspot generation. The risk factor description includes type label, location, and potential impact level (e.g., high risk), and is stored as a risk dataset. The system scans the entire dataset to locate the specific unit: the location index identifies P5 as an abnormal battery pack unit, and the temperature data is directly mapped to its physical location (e.g., the upper right unit in the system layout) to complete the location.

[0104] Analyze the thermal distress type of the abnormal battery pack cells. Thermal distress type identification is based on temperature values ​​and cell status data: extract historical temperature curves and coolant-related parameters (such as flow rate logs) from cell P5. In the example, P5's temperature is consistently higher than surrounding cells, with a stable upward trend (e.g., rising from 28°C to 31°C within 2 hours). The coolant flow rate is normal, but heat exchange efficiency is low, leading to a comprehensive assessment of the thermal distress type "heat accumulation." The type code is output by the decision logic, based on a preset rule base for matching (e.g., temperature difference expansion + normal flow rate corresponds to heat accumulation). System processing includes detailed type descriptions and root cause inferences, avoiding direct intervention.

[0105] Thermal strain strategies are set for abnormal battery pack units based on thermal distress types. Strategy settings access a predefined strategy library: for the "thermal accumulation" type, strategies include increasing coolant flow and prioritizing cooling resource allocation. The system instantiates this as a sequence of instructions: an independent liquid-cooled control valve is assigned to unit P5, the flow rate is increased by 20%, and a temperature monitoring feedback loop is initiated. The strategy is output as control parameters, such as target flow rate values ​​and adjustment step sizes. Risk response methods are generated based on the thermal strain strategy, encoded as an executable instruction set, including a risk identifier, corresponding strategy parameters, and execution conditions (such as trigger time points). In the final stage, the system integrates the methods into the thermal adjustment rule library for dynamic execution. All operations are scheduled by the central controller, with processing time controlled within milliseconds to meet real-time requirements.

[0106] Example 5: When dynamically adjusting the thermal management of an independent liquid-cooled battery pack energy storage system according to thermal adjustment rules, the system parses the thermal adjustment rule entries stored in the rule base. The rule entries exist in a structured data format, with each rule containing a clear trigger condition definition and a corresponding response action sequence. Trigger conditions are typically expressed as logical judgment expressions, with variables derived from real-time monitored temperature data, environmental data, and system status data. The system runs a continuous condition matching process, which periodically substitutes the currently collected data into the trigger condition expressions in the rules for logical evaluation. When the calculation result of the trigger condition expression for a rule entry is true, the system determines that the thermal management trigger condition corresponding to that rule is met. For example, a rule might set the trigger condition as "the maximum temperature difference between battery pack units continuously exceeds a set threshold of 5.2 degrees Celsius for a duration greater than 30 seconds." When real-time data meets this condition, the trigger condition is identified and marked as active.

[0107] Upon identifying the activated thermal management trigger condition, the system immediately generates a specific coolant flow adjustment command based on the preset response action sequence in the rule entry. The command generation process strictly follows the action logic defined in the rule. The command content typically includes the target coolant flow value, the identifier of the target battery pack unit or area to be adjusted, the rate limit for flow adjustment, and the time window expected to reach the target flow. The command is encapsulated in a standardized control command format, containing information such as the command type code, target object address code, target flow value, and execution timestamp. The system maintains a command queue, and newly generated coolant flow adjustment commands are added to the tail of the queue to await execution scheduling.

[0108] When executing coolant flow adjustment commands, the system employs a phased adjustment strategy to control coolant distribution. Phased adjustment aims to balance adjustment speed with system stability, avoiding hydraulic shocks or temperature fluctuations caused by sudden flow changes. The adjustment process is typically divided into three main phases: initial adjustment, steady-state maintenance, and fine-tuning. In the initial adjustment phase, the system quickly adjusts the coolant flow to a preset proportion (e.g., 70%-80% of the target flow value) by controlling the proportional control valves or variable frequency pumps in the corresponding branches. This phase is rapid and aims to initially respond to thermal management needs. In the steady-state maintenance phase, the system maintains the adjusted flow level for a period while closely monitoring the temperature response data in the target area. This phase allows the system to observe temperature change trends under relatively stable flow conditions and evaluate the effectiveness of the initial adjustment. Finally, in the fine-tuning phase, the system finely adjusts the coolant flow based on the deviation between real-time temperature feedback and the target temperature, gradually approaching and ultimately stabilizing near the target flow value. The entire phased adjustment process is a closed-loop control system composed of flow control actuators (valves, pumps) and flow sensor feedback.

[0109] Throughout the entire coolant flow adjustment process, the system simultaneously performs real-time verification of temperature feedback data. The core of this verification lies in comparing the actual collected battery pack unit temperature data with the predicted temperature values ​​generated by the thermal management model under the same adjustment command input. After each flow adjustment action, the system records the future temperature change curve predicted by the model based on the current state and the adjustment action. Simultaneously, the system continuously collects readings from actual temperature sensors, forming a measured temperature change trajectory. The verification process aligns and compares the measured trajectory data points with the model's predicted curve at the same time points, calculating the difference sequence between the two. This difference sequence reflects the degree of agreement between the model prediction and the actual response.

[0110] Based on the difference sequence (i.e., the deviation between predicted and measured values) obtained from the above verification process, the system iteratively updates the thermal adjustment rules. The update logic focuses on correcting potential parameter biases or logical deficiencies in the rules. For example, if verification reveals that the model-predicted temperature drop rate is much faster than the measured drop rate, it indicates that the initial flow adjustment range set in the rule may be insufficient, and the system will increase the corresponding flow adjustment ratio parameter in the rule according to a preset algorithm. Conversely, if the predicted cooling is excessive while the measured cooling is insufficient, the flow adjustment range may be reduced. The update process may also involve adjusting the threshold setting of the rule triggering conditions, or modifying the time length and flow ratio parameters of each stage in the phased adjustment. The updated thermal adjustment rule parameters are written back to the rule base, replacing the original relevant parameter items, for use in subsequent thermal management decision-making cycles.

[0111] After completing a round of dynamic thermal management adjustments and operating stably for a period of time, the system collects the final temperature data of the energy storage system. The final temperature data refers to the set of steady-state temperature values ​​read from the temperature sensors of all battery pack units after the adjustment actions have been completed and the temperature has stabilized. These data represent the actual temperature distribution results after this thermal management intervention.

[0112] The system compares the collected final temperature data with the expected temperature range predicted by the thermal management model before or during adjustment. The expected temperature range is the interval within which the temperature should fall, calculated by the model based on the initial state and the adjusted actions. The comparison process calculates the difference between the measured final temperature value of each battery pack unit and the upper and lower limits of the expected temperature range corresponding to that unit. If the measured value exceeds the expected range, an exceedance event record is generated. The system summarizes the exceedance event information of all units, calculates the overall deviation statistics (such as the proportion of units exceeding the range, the average exceedance magnitude), and finally forms the model error signal. This signal is a structured data object containing an error type identifier (such as overall deviation, local hotspot deviation), a quantitative value of the error magnitude, and the location information of the battery pack units involved.

[0113] Based on the generated model error signal, the system initiates the parameter adjustment program for the thermal management model. The adjustment program accesses the model parameter library to locate the model coefficients associated with the error signal. For example, if the error signal indicates that the model generally underestimates the temperature drop, the coefficients related to heat capacity or heat exchange efficiency in the model may be adjusted. The adjustment algorithm calculates new parameter values ​​based on the amplitude and direction of the error signal, according to a preset coefficient update rule (such as a simplified application of the gradient descent principle). The calculated new parameter values ​​are written to the model parameter library, replacing the original corresponding parameter entries.

[0114] The system updates the adjusted thermal management model parameters to the historical parameter database. The historical parameter database is a versioned parameter storage system that not only saves the latest effective parameter set but also records historical versions of parameter adjustments. Each time an update occurs, the system generates a unique version identifier for the new parameter set (usually containing a timestamp and update reason code) and stores this identifier associated with the adjusted parameter set. Simultaneously, the system records a summary of the model error signal that triggered the parameter update, along with the corresponding thermal management event identifier (such as which adjustment triggered the verification). Maintaining the historical parameter database allows the system to trace the parameter evolution process and roll back to previous parameter versions when necessary. This historical parameter data provides a reference for subsequent thermal management model operation and optimization.

[0115] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0116] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A thermal management method for an independent liquid-cooled battery pack energy storage system, characterized in that, The thermal management method comprises the following steps: collecting real-time thermal parameter data of the independent liquid-cooled battery pack energy storage system, identifying a current thermal state of the independent liquid-cooled battery pack energy storage system according to the real-time thermal parameter data; setting a thermal management model of the independent liquid-cooled battery pack energy storage system based on the current thermal state; real-time monitoring temperature distribution data of the independent liquid-cooled battery pack energy storage system; setting a thermal adjustment rule of the independent liquid-cooled battery pack energy storage system based on the temperature distribution data; performing dynamic thermal management adjustment on the independent liquid-cooled battery pack energy storage system according to the thermal adjustment rule; The real-time monitoring of the temperature distribution data of the independent liquid-cooled battery pack energy storage system comprises the following steps: continuously collecting real-time temperature values of a plurality of battery pack units of the independent liquid-cooled battery pack energy storage system; generating a temperature deviation trend line according to the real-time temperature values of the plurality of battery pack units; identifying thermal uniformity fluctuation data of the independent liquid-cooled battery pack energy storage system based on the temperature deviation trend line; The thermal uniformity fluctuation data is used as the temperature distribution data; The setting of the thermal adjustment rule of the independent liquid-cooled battery pack energy storage system based on the temperature distribution data comprises the following steps: analyzing an environmental temperature change trend of the independent liquid-cooled battery pack energy storage system; combining the thermal uniformity fluctuation data and the environmental temperature change trend to identify potential thermal risk factors; setting a risk response mode of the independent liquid-cooled battery pack energy storage system according to the potential thermal risk factors; generating the thermal adjustment rule based on the risk response mode.

2. A thermal management method for a standalone liquid-cooled battery pack energy storage system as claimed in claim 1, wherein, The collection of the real-time thermal parameter data of the independent liquid-cooled battery pack energy storage system comprises the following steps: acquiring battery temperature data and cooling liquid flow data of the independent liquid-cooled battery pack energy storage system through a plurality of types of sensors; generating a temperature gradient sequence according to the battery temperature data; generating a flow distribution curve according to the cooling liquid flow data; fusing the temperature gradient sequence and the flow distribution curve to form the real-time thermal parameter data.

3. A thermal management method for a standalone liquid-cooled battery pack energy storage system as claimed in claim 2, wherein, The setting of the thermal management model of the independent liquid-cooled battery pack energy storage system based on the current thermal state comprises the following steps: determining thermal diffusion characteristics of the independent liquid-cooled battery pack energy storage system based on the temperature gradient sequence; constructing a first thermal control model based on the thermal diffusion characteristics; determining cooling efficiency characteristics of the independent liquid-cooled battery pack energy storage system based on the flow distribution curve; constructing a second thermal control model based on the cooling efficiency characteristics; combining the first thermal control model and the second thermal control model to generate the thermal management model.

4. A thermal management method for a stand-alone liquid-cooled battery pack energy storage system as claimed in claim 1, wherein, The analysis of the environmental temperature change trend of the independent liquid-cooled battery pack energy storage system comprises the following steps: classifying the environmental temperature change trend to obtain classified environmental data; extracting key environmental features of the classified environmental data; collecting historical environmental data based on the key environmental features; analyzing an environmental temperature fluctuation mode and a cooling demand mode according to the historical environmental data; combining the environmental temperature fluctuation mode and the cooling demand mode to generate the environmental temperature change trend.

5. A thermal management method for a stand-alone liquid-cooled battery pack energy storage system as claimed in claim 4, wherein, The combining the thermal uniformity fluctuation data and the ambient temperature change trend to identify a potential thermal risk factor comprises: identifying a normal thermal parameter range of the independent liquid-cooled battery pack warehouse energy storage system based on the thermal uniformity fluctuation data; comparing real-time temperature values of the plurality of battery pack units with the normal thermal parameter range to generate a current thermal behavior deviation; determining an abnormal thermal threshold based on the current thermal behavior deviation and the ambient temperature change trend; identifying the potential thermal risk factor according to the abnormal thermal threshold.

6. A thermal management method for a stand-alone liquid-cooled battery pack energy storage system as claimed in claim 5, wherein, The setting a risk response mode of the independent liquid-cooled battery pack warehouse energy storage system according to the potential thermal risk factor comprises: locating an abnormal battery pack unit corresponding to the potential thermal risk factor; analyzing a thermal dilemma type of the abnormal battery pack unit; setting a thermal strain strategy for the abnormal battery pack unit based on the thermal dilemma type; generating the risk response mode according to the thermal strain strategy.

7. A thermal management method for a stand-alone liquid-cooled battery pack energy storage system as claimed in claim 6, wherein, The dynamically adjusting thermal management of the independent liquid-cooled battery pack warehouse energy storage system according to the thermal adjustment rule comprises: identifying a thermal management trigger condition based on the thermal adjustment rule; generating a cooling liquid flow adjustment instruction according to the thermal management trigger condition; adjusting the cooling liquid distribution in stages by executing the cooling liquid flow adjustment instruction; verifying temperature feedback data in real time during the adjustment process and the predicted value of the thermal management model; iteratively updating the thermal adjustment rule based on the verification result.

8. A thermal management method for a stand-alone liquid-cooled battery pack energy storage system as claimed in claim 7, wherein, The method further comprises: collecting final temperature data of the independent liquid-cooled battery pack warehouse energy storage system; comparing the final temperature data with an expected temperature range of the thermal management model to generate a model error signal; adjusting parameters of the thermal management model according to the model error signal; updating the adjusted parameters to a historical parameter library for subsequent thermal management.

Images