Battery pack phase change thermal management method and system based on MPC hierarchical regulation

Abstract

The application discloses a battery pack phase change thermal management method and system based on MPC hierarchical regulation, and relates to the technical field of battery thermal management. The battery pack phase change thermal management method and system based on MPC hierarchical regulation comprises the following steps: S1, collecting battery phase change thermal management data, and performing preprocessing on the battery phase change thermal management data; S2, constructing a thermal risk assessment data set, assessing the thermal risk level of the battery, and generating a thermal risk grade label; S3, extracting the thermal risk grade label, setting a corresponding regulation mode, assessing the regulation parameter values of the condensate circulating pump speed, the gas export rate and the fire-retardant liquid injection rate, and generating a target control quantity sequence; and S4, constructing an MPC regulation prediction model, setting control constraints and state boundary limits, predicting future state responses, performing smoothing correction and issuing executable control instructions. The problems of steam aggregation leading to a decrease in cooling efficiency and further inducing thermal runaway under high-rate charging and discharging conditions are solved.

Description

Technical Field

[0001] This invention relates to the field of battery thermal management technology, specifically to a battery pack phase change thermal management method and system based on MPC hierarchical control. Background Technology

[0002] Lithium-ion batteries face complex operating conditions in applications such as electric transportation and energy storage power stations, including high-rate discharge, large ambient temperature differences, and concentrated heat loads, which can easily lead to localized overheating or thermal runaway risks. Phase change materials (PCMs), due to their excellent latent heat absorption and temperature-controlled release capabilities, are widely used in battery thermal management structures, becoming an important pathway to enhance thermal safety. To improve the cooling efficiency and control performance of PCM systems, various modeling optimization and simulation parameter tuning schemes have been proposed in existing technologies.

[0003] For example, the invention with announcement number CN111475963B relates to an optimization method for a phase change thermal management system of a lithium-ion battery. The steps are as follows: obtaining the external design parameters of the battery and module, the electrochemical parameters and thermophysical parameters of the electrode materials, and the thermophysical parameters of the phase change material; establishing a one-dimensional electrochemical-three-dimensional thermal coupling model; comparing the voltage curves and temperature curves obtained from experiments and simulations at different discharge rates to verify the correctness of the model; using the model to analyze the optimal value of a single variable under different discharge rates while satisfying the thermal management target of the battery pack; taking the maximum battery volume to the total module volume as the optimization objective function, the optimal combination of single variables as the initial value for optimization, and the thermal management target as a constraint, performing multi-parameter optimization to solve the variable values.

[0004] For example, the invention with announcement number CN120337331B provides a battery cooling method based on phase change materials, including: step S1, identifying multiple equivalent thermal nodes on a lithium-ion battery structure with phase change material layers covering both ends of the cell, and modeling a three-dimensional thermal network model by combining air temperature nodes and cooling medium temperature nodes; step S2, establishing heat conduction paths between nodes in the three-dimensional thermal network model to simulate battery heat dissipation; step S3, collecting temperature and material data during the battery heat dissipation simulation process; step S4, obtaining multiple phase change material thickness optimization schemes as individuals to form a population; step S5, extracting a preset number of individuals from the population, and using a pre-constructed weighted objective function to continuously perform population iterative optimization on each individual based on the temperature and material data to obtain the optimal individual, and then using the optimal individual to optimize the thickness of the phase change material layer.

[0005] However, the limitations of existing two-phase phase change thermal management systems are becoming increasingly apparent in practical applications: when the battery is operating at low rates, although basic heat dissipation can be achieved through phase change, the system lacks dynamic adaptability and cannot flexibly adjust the heat dissipation intensity according to heat changes, which can easily lead to energy waste or untimely heat dissipation; when the battery enters a high-rate charge and discharge state, the heat increases sharply, and the pressure inside the system is prone to rise due to vapor accumulation, which causes the phase change boiling point to rise accordingly, and the heat dissipation efficiency is greatly reduced, making it difficult to match the heat dissipation requirements under high heat density; more importantly, when facing the risk of thermal runaway, existing systems mostly rely on a single cooling method and lack a progressive safety control mechanism, making it difficult to effectively curb the spread of thermal runaway and posing serious hidden dangers to equipment safety and personnel safety.

[0006] Therefore, in order to address the above problems, there is an urgent need for a battery pack phase change thermal management method and system based on MPC hierarchical control. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a battery pack phase change thermal management method and system based on MPC hierarchical control, which solves the problem of vapor accumulation leading to decreased cooling efficiency and subsequent thermal runaway under high-rate charge and discharge conditions.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a battery phase change thermal management method and system based on MPC hierarchical control, comprising: S1, collecting battery phase change thermal management data, and performing time alignment, abnormal sample removal, smoothing and noise reduction, and standardization on the battery phase change thermal management data to obtain preprocessed battery phase change thermal management data; S2, constructing a thermal risk assessment dataset based on the preprocessed battery phase change thermal management data, assessing the thermal risk level of the battery under the current operating conditions, generating thermal risk level labels, and constructing a thermal risk state dataset; S3, extracting thermal risk level labels, setting corresponding control modes, and evaluating the control parameter values ​​of condensate circulation pump speed, gas exhaust rate, and flame retardant injection rate based on the thermal risk state dataset to generate a target control quantity sequence; S4, constructing an MPC control prediction model based on the target control quantity and battery phase change thermal management data, setting control constraints and state boundary limits, predicting future state responses, performing smoothing corrections and issuing executable control commands, and combining the risk state change trend within a continuous assessment period to achieve control mode rollback and strategy closed-loop update.

[0011] Furthermore, battery phase change thermal management data is collected, and time alignment, outlier removal, smoothing, noise reduction, and standardization are performed on the preprocessed battery phase change thermal management data. The specific steps to obtain preprocessed battery phase change thermal management data are as follows: Real-time acquisition of battery phase change thermal management data, including cell quality, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, and gas extraction rate; and, based on the battery surface temperature, battery current, and voltage data, a thermal model parameter inversion algorithm is used to obtain… The battery's specific heat capacity is used to obtain the battery's internal resistance through Ohm's law and transient voltage response analysis. Time alignment of the battery phase change thermal management data is performed using timestamp interpolation to unify the data time base under different sampling frequencies. Continuity identification of the battery phase change thermal management data is performed using a first-order difference detection method to eliminate discontinuous samples caused by data loss and duplication. The battery phase change thermal management data is smoothed using a moving average method to reduce the impact of short-term fluctuations on data stability. Finally, the battery phase change thermal management data is standardized using a max-min normalization method to unify the numerical scale of the battery phase change thermal management data.

[0012] Furthermore, based on the preprocessed battery phase change thermal management data, a thermal risk assessment dataset is constructed. The specific steps for assessing the battery's thermal risk level under current operating conditions are as follows: A continuous assessment period is divided according to a fixed time window. Within each assessment period, battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, and condensate flow rate are extracted to construct a thermal risk assessment dataset. Based on the thermal risk assessment dataset, the instantaneous thermal risk level of the battery under current operating conditions is assessed: The equivalent heating term is obtained by multiplying the square of the battery current by the battery internal resistance and then dividing by the product of the battery specific heat capacity and the battery mass. The absolute value of the temperature difference between the condensate inlet temperature and the condensate outlet temperature is taken to obtain the cooling temperature difference term. The battery surface temperature, the equivalent heating term, and the cooling temperature difference term are added together to form the thermal response term. The boiling point change correction term is obtained by adding one to the vapor chamber pressure, taking the natural logarithm, and then adding one again. The cooling regulation term is obtained by adding the condensate flow rate to the boiling point change correction term. The thermal risk assessment value is obtained by dividing the thermal response term by the cooling regulation term.

[0013] Further, the specific steps for generating thermal risk level labels and constructing a thermal risk status dataset are as follows: Real-time comparison of battery thermal risk assessment values ​​and thermal risk thresholds to classify battery thermal risk levels: When the battery thermal risk assessment value is less than or equal to the Level 1 thermal risk threshold, it is marked as Level 1 risk; when the battery thermal risk assessment value is greater than the Level 1 thermal risk threshold but less than the Level 2 thermal risk threshold, it is marked as Level 2 risk; when the battery thermal risk assessment value is greater than or equal to the Level 2 thermal risk threshold, it is marked as Level 3 risk; The battery thermal risk labels, battery thermal risk assessment values, and corresponding battery phase change thermal management data are structurally bound to output the thermal risk status dataset.

[0014] Further, the specific steps for extracting the thermal risk level label and setting the corresponding control mode are as follows: Extract the risk level label for the current assessment period and set the corresponding control mode: Level 1 risk triggers heat dissipation mode, which only performs condensate circulation control; Level 2 risk triggers enhanced mode, which performs condensate circulation and air extraction control; Level 3 risk triggers emergency mode, which simultaneously performs condensate circulation, air extraction and flame retardant injection control; After determining the corresponding control mode, proceed to the control command execution stage.

[0015] Furthermore, the specific steps for evaluating the control parameter values ​​of condensate circulation pump speed, gas extraction rate, and flame retardant injection rate based on the thermal risk state dataset are as follows: Add one to the vapor chamber pressure and take its natural logarithm to obtain the vapor correction term; add one to the vapor correction term and multiply by the battery thermal risk assessment value to obtain the risk correction term; divide the risk correction term by the sum of the condensate flow rate and the minimum term to obtain the condensation control assessment value; divide the battery surface temperature by the sum of the battery surface temperature, condensate level, and the minimum term, and multiply the resulting ratio by the exhaust correlation coefficient to obtain the gas extraction control assessment value; add the flame retardant injection amount and the gas extraction rate, divide by the sum of the battery thermal risk assessment value and a constant, and multiply the result by the flame retardant correlation coefficient to obtain the flame retardant control assessment value.

[0016] Furthermore, the specific steps for generating the target control quantity sequence are as follows: the condensation regulation evaluation value, the gas extraction regulation evaluation value, and the flame retardant regulation evaluation value are respectively mapped to the target control quantities for adjusting the condensate circulation pump speed, the gas exhaust rate, and the flame retardant injection rate, forming a target control quantity sequence, and the target control quantities are mapped and bound to the corresponding execution objects.

[0017] Furthermore, based on the target control quantity and battery phase change thermal management data, an MPC regulation and prediction model is constructed, control constraints and state boundary limits are set, future state responses are predicted, smooth corrections are performed, and executable control commands are issued. The specific steps are as follows: The target control quantity is extracted as the control input; the battery surface temperature, vapor chamber pressure, and condensate level are extracted as state variables to construct the MPC regulation and prediction model; the maximum and minimum control values ​​and single-cycle adjustment amplitudes of the condensate circulation pump speed, gas extraction rate, and flame retardant injection rate are set to construct the control input value constraints; the safe operating ranges of the battery surface temperature, vapor chamber pressure, and condensate level are set to construct the boundary constraints of the state variables; the target control quantity within the current evaluation period is input into the MPC regulation and prediction model to predict the state response sequence within future evaluation periods; if any predicted state response value touches the boundary conditions, the target control quantity is smoothly corrected within the control increment range, and a control command sequence is output; the control command sequence is data-bound with the condensate circulation pump, the extraction unit, and the flame retardant injection component, and issued to the execution end for real-time scheduling of the control commands.

[0018] Furthermore, based on the trend of risk status changes within a continuous assessment period, the specific steps for implementing the control mode rollback and strategy closed-loop update are as follows: After the control command is executed, the battery surface temperature, vapor chamber pressure, and condensate level are continuously collected and dynamically monitored in conjunction with the current risk level change trend; if the state variables remain stably within the safe range within a continuous fixed assessment period and the risk level continues to decrease, the control mode rollback mechanism is triggered, and the third-level and second-level control are exited in sequence, smoothly transitioning to the heat dissipation mode.

[0019] The second aspect of this invention provides a battery phase change thermal management system based on MPC hierarchical control, comprising: a thermal management data preprocessing module, a battery thermal risk assessment module, a hierarchical control parameter generation module, and a predictive control execution optimization module. The thermal management data preprocessing module collects battery phase change thermal management data and performs time alignment, outlier removal, smoothing, noise reduction, and standardization on the data to obtain preprocessed battery phase change thermal management data. The battery thermal risk assessment module constructs a thermal risk assessment dataset based on the preprocessed battery phase change thermal management data to assess the thermal risk level of the battery under current operating conditions. The system generates thermal risk level labels and constructs a thermal risk state dataset. A graded control parameter generation module extracts thermal risk level labels, sets corresponding control modes, and evaluates the control parameter values ​​for condensate circulation pump speed, gas extraction rate, and flame retardant injection rate based on the thermal risk state dataset, generating a target control quantity sequence. A predictive control execution optimization module constructs an MPC control prediction model based on the target control quantity and battery phase change thermal management data, sets control constraints and state boundary limits, predicts future state responses, performs smooth corrections, and issues executable control commands. Combined with the risk state change trend within a continuous evaluation period, it achieves control mode rollback and strategy closed-loop update.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) The battery pack phase change thermal management method and system based on MPC hierarchical control constructs a multi-level control framework covering heat dissipation mode, enhancement mode and emergency mode. Combined with the real-time identification results of battery thermal risk level within the assessment period, the control target and control path are dynamically set to ensure that the thermal management strategy can adaptively switch under different operating conditions to match the actual thermal load changes. This mechanism avoids the problem of lag in response of traditional thermal management systems under sudden high temperature conditions, and significantly improves the timeliness of control and the thermal steady-state maintenance capability of the system.

[0023] (2) The battery pack phase change thermal management method and system based on MPC hierarchical control reduces the pressure inside the cavity by pumping air, restoring effective heat exchange under phase change conditions; at the same time, when the risk increases, flame retardant liquid is injected to absorb heat and form thermal isolation, thus delaying the occurrence of thermal runaway to the maximum extent. It essentially solves the problem of cooling decay of traditional phase change systems under high pressure and high heat environment, thereby realizing active response and hierarchical suppression of battery thermal risks under high rate charge and discharge conditions, and significantly enhancing the system's adaptability to extreme conditions and safety protection capabilities.

[0024] (3) The battery pack phase change thermal management method and system based on MPC hierarchical control achieves forward-looking identification and rapid response to thermal runaway trends by setting state boundaries and control increment constraints. During the evaluation cycle, the risk change trend of the continuous cycle is also used as a feedback signal to realize the closed-loop retreat of the control mode, and gradually transition from the emergency mode to the basic heat dissipation mode, thereby forming a hierarchical, dynamic and adaptive thermal control system.

[0025] (4) This battery pack phase change thermal management method and system based on MPC hierarchical control maps the condensation control evaluation value, the gas extraction control evaluation value, and the flame retardant control evaluation value to the target control quantities of the condensate circulation pump speed, the gas exhaust rate, and the flame retardant injection rate, respectively. Based on the mapping relationship between the evaluation values ​​and the execution quantities, a structured parameter binding mechanism is constructed. This not only ensures that the physical meaning of the control parameters corresponds one-to-one with the operation behavior of the execution object, but also improves the interpretability of the control targets and the controllability of the system scheduling. Attached Figure Description

[0026] Figure 1 This is a flowchart of a battery pack phase change thermal management method based on MPC hierarchical control.

[0027] Figure 2 This is a structural diagram of a battery pack phase change thermal management system based on MPC hierarchical control.

[0028] Figure 3 A diagram for battery thermal risk classification assessment and control threshold determination;

[0029] Figure 4 This is a schematic diagram of the phase change thermal management system for a graded control type battery pack.

[0030] In the diagram, 1. Battery pack; 2. Radiator; 3. Cooling coil; 4. Sealed enclosure; 5. Vapor chamber pressure sensor; 6. Electrically controlled valve; 7. Condensate circulation pump; 8. Gas storage tank; 9. Cooling tower; 10. Flame retardant liquid storage tank; 11. Battery temperature sensor. Detailed Implementation

[0031] 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.

[0032] Please see Figures 1-4This invention provides a technical solution: a battery pack phase change thermal management method and system based on MPC hierarchical control, comprising: S1, collecting battery phase change thermal management data, and performing time alignment, abnormal sample removal, smoothing and noise reduction, and standardization processing on the battery phase change thermal management data to obtain preprocessed battery phase change thermal management data; S2, based on the preprocessed battery phase change thermal management data, constructing a thermal risk assessment dataset, assessing the thermal risk level of the battery under the current operating conditions, generating thermal risk level labels, and constructing a thermal risk state dataset; S3, extracting thermal risk level labels, setting corresponding control modes, and evaluating the control parameter values ​​of condensate circulation pump speed, gas exhaust rate, and flame retardant injection rate based on the thermal risk state dataset to generate a target control quantity sequence; S4, constructing an MPC control prediction model based on the target control quantity and battery phase change thermal management data, setting control constraints and state boundary limits, predicting future state responses, performing smoothing corrections and issuing executable control commands, and combining the risk state change trend within a continuous assessment period to achieve control mode rollback and strategy closed-loop update.

[0033] Specifically, the process involves collecting battery phase change thermal management data and performing time alignment, outlier removal, smoothing and noise reduction, and standardization on the preprocessed data. The specific steps for obtaining the preprocessed battery phase change thermal management data are as follows: Real-time acquisition of battery phase change thermal management data, including cell mass, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, and gas extraction rate. Cell mass is obtained through fixed-point mass sampling of the cell assembly using a weighing device; battery surface temperature is obtained through a battery temperature sensor. Device 11 achieves real-time sensing; the steam chamber pressure is obtained by detecting changes in gas pressure within the evaporation chamber using the steam chamber pressure sensor 5; the amount of flame retardant liquid injected is obtained by real-time statistical analysis of the actual injected volume using the injection metering unit; the condensate flow rate is obtained by a turbine flow meter installed in the condensate pipeline; the condensate level is obtained by collecting condensate height information from a float displacement sensor; the condensate inlet temperature and condensate outlet temperature are obtained and synchronously recorded by temperature sensors at the inlet and outlet, respectively; the battery voltage and battery current are continuously measured by a voltage sampler and a shunt current acquisition circuit, respectively; the gas extraction rate is calculated by a mass flow meter in the vacuum pumping circuit to obtain the exhaust flow rate per unit time. Furthermore, based on the battery surface temperature, battery current, and battery voltage data, the battery specific heat capacity is obtained through a thermal model parameter inversion algorithm. This process uses model correction results to dynamically and adaptively adjust the inversion accuracy; simultaneously, the battery internal resistance is obtained through Ohm's law and transient voltage response analysis, and a sudden change point detection mechanism is introduced to improve the sensitivity of resistance value change identification.Subsequently, time alignment processing was performed on the battery phase change thermal management data using timestamp interpolation to unify the time references for cell quality, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, and gas extraction rate at different sampling frequencies, ensuring consistency of time-series characteristics. A first-order differential detection method was used to identify the continuity of the battery phase change thermal management data, eliminating discontinuous data segments caused by missing or duplicate samples while preserving the integrity of the time axis. Finally, a moving average method was used to smooth the battery phase change thermal management data, reducing the impact of short-term data fluctuations on the overall trend. To mitigate interference in potential judgments and enhance signal stability, the battery phase change thermal management data is standardized using the minimax normalization method. Normalized interval boundaries are established for cell quality, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, gas extraction rate, and the inverted battery specific heat capacity and internal resistance. All variables involved in subsequent calculations are converted into dimensionless data. Subsequent thermal risk assessment, control parameter calculation, and MPC predictive control all utilize the corresponding dimensionless data, unifying the numerical scale and improving the feature adaptability and numerical stability of the model input.

[0034] In this implementation plan, by systematically collecting and standardizing data on cell quality, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, and gas extraction rate, the plan achieves temporal consistency, anomaly removal, signal smoothing, and numerical standardization of multi-source battery phase change thermal management data. This significantly improves the integrity, continuity, stability, and consistency of the original data and model input, ensuring the accuracy of the battery specific heat capacity and internal resistance calculation results. It also enhances the real-time response and stable control capabilities of the thermal model in subsequent thermal risk assessment and control decisions.

[0035] Specifically, based on the preprocessed battery phase change thermal management data, a thermal risk assessment dataset is constructed. The specific steps for assessing the thermal risk level of the battery under the current operating conditions are as follows: A continuous assessment period is divided according to a fixed time window. Within each assessment period, battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, and condensate flow rate are extracted to construct a complete, continuous, and consistent thermal risk assessment dataset. A stable time-series data structure is constructed to ensure consistency in the time dimension and physical quantity accuracy of each input quantity within each period. Based on the thermal risk assessment dataset, the instantaneous thermal risk level of the battery under the current operating conditions is assessed: the battery current is squared point-by-point and multiplied by the battery internal resistance to obtain the instantaneous heating power based on the actual current-carrying characteristics. Based on this, the product of the battery's specific heat capacity and its mass is used to obtain the temperature rise trend per unit heat capacity mass, forming the battery's equivalent heat generation term. The condensate inlet temperature is subtracted from the condensate outlet temperature, and the resulting temperature difference is subjected to an absolute value operation to extract the instantaneous cooling capacity reduction index, resulting in a cooling temperature difference term. The battery surface temperature is then sequentially superimposed with the battery's equivalent heat generation term and the cooling temperature difference term to form a comprehensive thermal response term characterizing the balance between heat input and cooling. The vapor chamber pressure is incremented by one and its natural logarithm is taken, then further incremented by one to avoid the influence of the zero logarithm, generating an exponential term used to correct boiling point changes, serving as a boiling point change correction term. The condensate flow rate is summed with the boiling point change correction term to form a cooling regulation term reflecting the coupling between cooling intensity and the system's physical state. Finally, the thermal response term is divided by the cooling regulation term to obtain a quantifiable, gradable, and physically meaningful battery thermal risk assessment value.

[0036] The specific formula for calculating the battery thermal risk assessment value is as follows:

[0037] ;

[0038] In the formula, H represents the battery thermal risk assessment value. Indicates the surface temperature of the battery. Indicates battery current. Indicates the battery's internal resistance. Indicates the battery's specific heat capacity. Indicates battery quality. Indicates the condensate inlet temperature. Indicates the condensate outlet temperature. Indicates the steam chamber pressure. This indicates the condensate flow rate.

[0039] In this embodiment, Table 1 is a data table of battery thermal risk assessment values, showing the dimensionless key thermal management data and the final calculated thermal risk assessment values ​​based on the aforementioned preprocessed data over five assessment cycles. Specifically, this includes: battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, condensate flow rate, and thermal risk assessment value. Specifically, in assessment cycle 1, the battery surface temperature is 48.3°C, the battery current is 110 kJ / m³, the battery internal resistance is 0.0023 kJ / m³, the specific heat capacity is 1000 kJ / m³, the mass is 0.5 kJ / m³, the condensate inlet temperature is 34.5°C, the condensate outlet temperature is 32.5°C, the vapor chamber pressure is 103.8°C, the condensate flow rate is 1.8 kJ / m³, and the final calculated thermal risk assessment value is 36.44. In assessment period 2, the battery surface temperature was 42.5°C, the battery current was 80 ohms, the battery internal resistance was 0.0025 N / A, the specific heat capacity was 1000 N / A, the mass was 0.5 N / A, the condensate inlet temperature was 35.0°C, the condensate outlet temperature was 32.0°C, the vapor chamber pressure was 101.3 L / 1000, the condensate flow rate was 1.5 L / 1000, and the thermal risk assessment value was 25.82. In assessment period 3, the battery surface temperature was 55.0°C, the battery current was 150 ohms, the battery internal resistance was 0.0021 N / A, the specific heat capacity was 1000 N / A, the mass was 0.5 N / A, the condensate inlet temperature was 34.0°C, the condensate outlet temperature was 33.0°C, the vapor chamber pressure was 106.2 L / 1000, the condensate flow rate was 2.0 L / 1000, and the thermal risk assessment value was 46.25. In assessment period 4, the battery surface temperature was 45.1°C, the battery current was 95 kW, the battery internal resistance was 0.0024 ohms, the specific heat capacity was 1000 N·m, the mass was 0.5 N·m, the condensate inlet temperature was 34.8°C, the condensate outlet temperature was 32.2°C, the vapor chamber pressure was 102.5 ohms, the condensate flow rate was 1.7 ohms, and the thermal risk assessment value was 31.36. In assessment period 5, the battery surface temperature was 51.2°C, the battery current was 130 ohms, the battery internal resistance was 0.0022 ohms, the specific heat capacity was 1000 N·m, the mass was 0.5 N·m, the condensate inlet temperature was 34.2°C, the condensate outlet temperature was 32.8°C, the vapor chamber pressure was 105.0 ohms, the condensate flow rate was 1.9 ohms, and the thermal risk assessment value was 41.31.

[0040] Table 1 Battery Thermal Risk Assessment Data Table

[0041]

[0042] like Figure 3The figure shows the battery thermal risk assessment values ​​and thermal risk levels for five assessment cycles. Different colors and graphic markers are used to clearly distinguish the thermal risk level corresponding to each assessment cycle. Specifically: green dots represent Level 1 risk, orange squares represent Level 2 risk, and red triangles represent Level 3 risk. In addition, the figure also shows the Level 1 risk threshold line (green dashed line) and the Level 2 risk threshold line (orange dashed line), providing an intuitive reference benchmark for determining the thermal risk level for each cycle, assisting in quickly identifying abnormal thermal states and formulating response measures. As can be seen from the figure, cycles 3 and 5 are in the Level 3 risk range, indicating that the battery faces a significant thermal safety threat; cycles 1 and 4 are in the Level 2 risk range; and cycle 2 is at the relatively safe Level 1 risk level. Figure 3 This intuitively reflects the sensitivity and discrimination ability of the thermal risk assessment method proposed in this invention under complex thermal conditions, providing decision support for the subsequent realization of graded response control and dynamic switching of regulation modes.

[0043] In this implementation plan, a thermal risk assessment dataset based on battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, and condensate flow rate is constructed to accurately characterize the thermal risk level of the battery under current operating conditions. The step-by-step construction process of the proposed battery equivalent heating term, cooling temperature difference term, thermal response term, boiling point change correction term, and cooling regulation term enhances the assessment model's analytical and expressive capabilities for multi-source heterogeneous temperature control parameters. This improves the accuracy and physical interpretability of the thermal risk assessment values ​​in characterizing the intrinsic correlation between instantaneous thermal shock, heat dissipation capacity, and risk level, laying a data foundation and decision support for the generation of graded control parameters and subsequent optimization of MPC control strategies.

[0044] Specifically, the steps for generating thermal risk level labels and constructing a thermal risk status dataset are as follows: Based on the battery thermal risk assessment value calculated within each assessment period, the battery thermal risk assessment value is compared in real time with the Level 1 and Level 2 thermal risk thresholds. Classification is performed according to a fixed risk level classification standard: when the battery thermal risk assessment value is less than or equal to the Level 1 thermal risk threshold, the current assessment period is marked as Level 1 risk; when the battery thermal risk assessment value is greater than the Level 1 thermal risk threshold but less than the Level 2 thermal risk threshold, the current assessment period is marked as Level 2 risk; when the battery thermal risk assessment value is greater than or equal to the Level 2 thermal risk threshold, the current assessment period is marked as Level 3 risk. After completing the risk level determination, the corresponding battery thermal risk label, battery thermal risk assessment value, battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, and condensate flow rate are bound to fields and uniformly numbered, and integrated into a structured thermal risk status dataset according to the time series, for subsequent control logic execution and control quantity prediction.

[0045] In this implementation plan, a complete risk level determination mechanism is constructed by comparing the battery thermal risk assessment value with the thermal risk threshold on a cycle-by-cycle basis. This achieves unified binding of battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure, and condensate flow rate under risk labels. This ensures that the battery thermal risk level classification standard is clear, the label generation process is continuous, and the data structure is well-organized. It significantly improves the accuracy of thermal state identification and the response efficiency of control decisions, and provides highly consistent and real-time risk data support for subsequent control parameter calculation and control mode selection.

[0046] Specifically, the steps for extracting the thermal risk level label and setting the corresponding control mode are as follows: Extract the risk level label for the current assessment period, and set a control mode that matches the risk level based on the battery thermal risk assessment value and the judgment result of the thermal risk threshold; When the risk level is determined to be Level 1, the heat dissipation mode is triggered, and only the condensate circulation pump 7 is controlled to adjust the condensate flow rate, relying on the cooling temperature difference between the condensate inlet temperature and the condensate outlet temperature to perform basic heat dissipation on the battery surface temperature; When the risk level is determined to be Level 2, the enhancement mode is triggered, and while maintaining the condensate flow rate control, the gas extraction unit is linked to adjust the gas extraction rate, actively reducing the vapor chamber pressure and enhancing the boiling point regulation capability; When the risk level is determined to be Level 3, the emergency mode is triggered, and while continuously controlling the condensate circulation pump 7 and the gas extraction unit, the flame retardant injection device is started, and thermal isolation and fire suppression control are performed according to the amount of flame retardant injected; After completing the matching setting of the control mode, the next stage of execution control command link is immediately entered, realizing the binding of the target control quantity of the corresponding execution object and the dynamic command issuance.

[0047] In this implementation plan, the risk level label of the current assessment period is extracted, the thermal risk level is determined based on the battery thermal risk assessment value, and the control strategies for condensate flow rate, gas extraction rate and flame retardant injection amount corresponding to the first-level risk, second-level risk and third-level risk are precisely set. This effectively realizes the hierarchical control linkage of condensate circulation pump 7, gas extraction unit and flame retardant injection device. While ensuring that the condensate inlet temperature and condensate outlet temperature form a cooling temperature difference, the vapor chamber pressure is reduced and a thermal isolation barrier is built, thereby improving the pertinence and timeliness of thermal management response and enhancing the safety control capability under different thermal risk levels.

[0048] Specifically, the steps for evaluating the control parameters of condensate circulation pump speed, gas exhaust rate, and flame retardant injection rate based on the thermal risk state dataset are as follows: The vapor chamber pressure is incremented by one and the natural logarithm is taken, then combined with the effect of vapor pressure on boiling point to form a vapor correction term; the vapor correction term is incremented by one and multiplied by the battery thermal risk assessment value to form a risk correction term, used to quantify the current degree of thermal imbalance; the risk correction term is divided by the sum of the condensate flow rate and the minimum term to obtain the condensation control assessment value, reflecting the intensity of condensate circulation control; where the minimum term is a minimal positive constant used in the calculation to prevent division by zero errors; the battery surface temperature is divided by the sum of the battery surface temperature, condensate level, and the minimum term to form the extraction control ratio, which is then multiplied by the exhaust correlation coefficient to obtain the extraction control assessment value, used to assess the gas exhaust rate of the condensate space; where... The exhaust correlation coefficient is obtained based on a multivariate regression model constructed using historical battery thermal risk state data. Input variables include condensate level, battery surface temperature, condensate flow rate, and vapor chamber pressure. Parameter weights are determined using the least squares fitting method to obtain the exhaust correlation coefficient, which ranges from 0 to 1. The flame retardant injection volume and gas extraction rate are added together, and the result is divided by the sum of the battery thermal risk assessment value and a constant. This result is then multiplied by the flame retardant correlation coefficient to obtain the flame retardant control assessment value, which is used to adjust the flame retardant injection rate. A sample set is constructed based on the flame retardant injection volume, battery thermal risk assessment value, and gas extraction rate. Feature contribution is extracted using a gradient boosting regression tree-based control sensitivity analysis method to obtain the flame retardant correlation coefficient, which ranges from 0 to 1 and is used to characterize the matching degree between flame retardant control intensity and response characteristics.

[0049] The specific formula for calculating the condensation control assessment value is as follows:

[0050] ;

[0051] In the formula, This represents the condensation control assessment value. This indicates the battery thermal risk assessment value. Indicates the steam chamber pressure. Indicates the condensate flow rate. Indicates a minus term.

[0052] The specific formula for calculating the air extraction control assessment value is as follows:

[0053] ;

[0054] In the formula, This indicates the evaluation value for air extraction control. Indicates the surface temperature of the battery. Indicates the height of the condensate liquid level. Represents the exhaust correlation coefficient. Indicates a minus term.

[0055] The specific formula for calculating the flame retardant control assessment value is as follows:

[0056] ;

[0057] In the formula, This indicates the flame retardant control assessment value. This represents the battery thermal risk assessment value, and Q represents the amount of flame retardant injected. Indicates the gas extraction rate. This represents the flame retardant correlation coefficient.

[0058] In this implementation plan, by introducing steam correction terms, risk correction terms, condensation control assessment values, extraction control assessment values, and flame retardant control assessment values, and combining battery thermal risk assessment values, vapor chamber pressure, condensate flow rate, battery surface temperature, condensate level, flame retardant injection volume, and gas extraction rate, a method for calculating sub-item control parameters for condensate circulation pump speed, gas extraction rate, and flame retardant injection rate is constructed. Based on this, a minima are set to avoid division-by-zero anomalies. At the same time, the sensitivity of the assessment model is adjusted by the exhaust correlation coefficient and the flame retardant correlation coefficient to achieve the stability, responsiveness, and adjustability of the control parameter calculation, thereby improving the accurate perception and differentiated control capability of thermal risk status.

[0059] Specifically, the steps for generating the target control quantity sequence are as follows: The condensation control evaluation value, the extraction control evaluation value, and the flame retardant control evaluation value are converted numerically using a mapping function to clearly identify the target control quantities for adjusting the condensate circulation pump speed, gas extraction rate, and flame retardant injection rate. Based on the conversion results, a target control quantity sequence is constructed, containing three items: the target value of the condensate circulation pump speed, the target value of the gas extraction rate, and the target value of the flame retardant injection rate. The target value of the condensate circulation pump speed is mapped and bound to the condensate circulation pump 7, the target value of the gas extraction rate is mapped and bound to the extraction unit, and the target value of the flame retardant injection rate is mapped and bound to the flame retardant injection component, ensuring that the target control quantities can be directly used for quantitative adjustment operations in the control execution process.

[0060] In this implementation scheme, by establishing a precise correspondence between the target values ​​of the condensate circulation pump speed, the gas extraction rate, and the flame retardant injection rate in the control logic, it is ensured that the condensation control evaluation value, the gas extraction control evaluation value, and the flame retardant control evaluation value can be stably and efficiently mapped into executable quantitative adjustment commands. These commands are then bound one-to-one with the parameters of the condensate circulation pump 7, the gas extraction unit, and the flame retardant injection component. This improves the real-time performance, accuracy, and executability of the target control quantities, providing a reliable foundation for the closed-loop scheduling and dynamic adjustment of subsequent control commands, and enhancing the command accuracy and control response capability throughout the entire battery thermal management process.

[0061] Specifically, the steps for constructing an MPC regulation and prediction model based on target control quantities and battery phase change thermal management data, setting control constraints and state boundary limits, predicting future state responses, performing smooth corrections, and issuing executable control commands are as follows: The target values ​​for condensate circulation pump speed, gas exhaust rate, and flame retardant injection rate are extracted as control input variables. Battery surface temperature, vapor chamber pressure, and condensate level are used as state variables to jointly construct the MPC regulation and prediction model. The MPC regulation and prediction model adopts a state-space modeling structure, integrating control input delay constraints and state evolution equations, enabling dynamic prediction and strategy feedforward adjustment of multi-cycle control behavior. The maximum and minimum control input values ​​and the maximum adjustment amplitude per cycle are set using equipment specifications, historical control behavior data, and thermal safety tolerance requirements. Multiple rounds of simulation verification are performed offline to ensure that the control quantity does not exceed the actuator's capability range. The safe operating boundaries of the state variables are set using thermal safety threshold data provided by the battery manufacturer, vapor chamber sealing level specifications, and condensate level sensor calibration data. All boundary limits are verified for safety using a simulation platform before system deployment. Subsequently, the target control quantity generated within the current evaluation period is used as the control input and fed into the MPC regulation and prediction model. Combined with the current state variable values, forward prediction is performed to obtain the predicted sequences for battery surface temperature, vapor chamber pressure, and condensate level height for multiple consecutive evaluation periods. During the prediction process, if the predicted value of any state variable exceeds the boundary limit at any time step, the MPC's built-in multi-round rolling optimization algorithm is immediately activated. Based on the control quantity adjustment range setting, iterative smooth correction is performed on the control input, and a full-cycle state prediction is performed again in real time after each correction until all state response sequences fall completely within the boundary limit range, ensuring stable convergence of the prediction path within a controllable range. Finally, the smoothed and optimized condensate circulation pump speed control command, gas extraction rate control command, and flame retardant injection rate control command are output. These are mapped and bound at the data structure level to the condensate circulation pump 7, the extraction unit, and the flame retardant injection component, respectively, to achieve safe scheduling and stable execution of the control commands.

[0062] In this implementation scheme, an MPC control prediction model is constructed and state variables such as battery surface temperature, vapor chamber pressure, and condensate level are integrated. Target values ​​for condensate circulation pump speed, gas extraction rate, and flame retardant injection rate are used as control inputs to achieve forward-looking prediction of state evolution trends over multiple consecutive evaluation periods. Control input constraints are set using equipment specifications, historical control behavior data, and thermal safety tolerance requirements. Boundary limits for state variables are set based on thermal safety thresholds provided by the battery manufacturer, vapor chamber sealing level specifications, and condensate level sensor calibration data to ensure prediction accuracy and control safety. When any predicted value of a state variable reaches a boundary limit, a multi-round smoothing correction mechanism is triggered to dynamically adjust the control input and repeat state prediction until all state response sequences converge within a safe range. Finally, the corrected control commands are bound to the condensate circulation pump 7, the extraction unit, and the flame retardant injection assembly via data structure to ensure the real-time performance, robustness, and closed-loop stability of the thermal control response.

[0063] Specifically, based on the risk state change trend within a continuous assessment cycle, the specific steps for implementing control mode rollback and strategy closed-loop update are as follows: After the control command is executed, the battery surface temperature, vapor chamber pressure, and condensate level are continuously collected and used as state variables for real-time dynamic monitoring; simultaneously, the thermal risk level label generated in the current assessment cycle is extracted, and the risk level label of the previous cycle is recorded to construct a risk state change sequence within the continuous assessment cycle; the sliding window analysis method is used to determine whether the state variables are always within the safe operating range within the current monitoring cycle; if within a fixed number of continuous assessment cycles, the battery surface temperature... If the temperature remains below the thermal safety limit, the steam chamber pressure remains stable within the required sealing level, the condensate level remains within the effective sensing threshold range, and the thermal risk level label shows a monotonically decreasing trend, then the current thermal risk level is determined to have been stabilized and alleviated, meeting the conditions for the control mode rollback. The control mode rollback mechanism is triggered, and the execution instructions included in the current control mode are revoked level by level. The flame retardant liquid injection control corresponding to the third-level risk and the gas exhaust control corresponding to the second-level risk are exited in sequence, leaving only the condensate circulation control. This achieves a smooth rollback from the emergency mode to the heat dissipation mode and provides a real-time input basis for subsequent risk status closed-loop updates.

[0064] In this implementation scheme, a step-by-step back control mechanism for the control mode is achieved by combining the changing trends of state variables such as battery surface temperature, vapor chamber pressure, and condensate level height during continuous evaluation cycles, as well as the dynamic change trajectory of thermal risk level labels. This mechanism effectively enhances the adaptability and sensitivity of the cooling control strategy. Under the premise of ensuring that the battery's thermal safety operation boundary is not breached, redundant control operations are promptly removed, unnecessary cooling resource consumption is reduced, the execution efficiency of condensate circulation pump speed control commands, gas exhaust rate control commands, and flame retardant injection rate control commands is improved, and the dynamic stability of battery phase change thermal management data is maintained, thereby achieving closed-loop coordination between battery thermal risk state identification and control strategy updates.

[0065] like Figure 2 As shown, the second aspect of this invention provides a battery phase change thermal management system based on MPC hierarchical control, comprising: a thermal management data preprocessing module, a battery thermal risk assessment module, a hierarchical control parameter generation module, and a predictive control execution optimization module, wherein: the thermal management data preprocessing module is used to collect battery phase change thermal management data and perform time alignment, outlier removal, smoothing, noise reduction, and standardization processing on the battery phase change thermal management data to obtain preprocessed battery phase change thermal management data; the battery thermal risk assessment module is used to construct a thermal risk assessment dataset based on the preprocessed battery phase change thermal management data to assess the thermal risk level of the battery under the current operating conditions. The system generates thermal risk level labels and constructs a thermal risk state dataset. A graded control parameter generation module extracts thermal risk level labels, sets corresponding control modes, and evaluates the control parameter values ​​for condensate circulation pump speed, gas extraction rate, and flame retardant injection rate based on the thermal risk state dataset, generating a target control quantity sequence. A predictive control execution optimization module constructs an MPC control prediction model based on the target control quantity and battery phase change thermal management data, sets control constraints and state boundary limits, predicts future state responses, performs smooth corrections, and issues executable control commands. Combined with the risk state change trend within a continuous evaluation period, it achieves control mode rollback and strategy closed-loop update.

[0066] like Figure 4The diagram illustrates the overall structure of a battery pack phase change thermal management system with hierarchical control capabilities. The system uses battery pack 1 as the core heat source, encapsulated within a sealed enclosure 4, forming a phase change thermal control space. The sealed enclosure 4 uses an airtight structure to define the dynamic adjustment boundaries of internal pressure and condensate level, ensuring the stability and controllability of thermal management operations. Battery surface temperature is acquired in real-time via battery temperature sensor 11, and pressure data from the top vapor chamber is collected using vapor chamber pressure sensor 5; both serve as key inputs for constructing a thermal risk assessment dataset. The hot vapor generated during the phase change process is conducted to an external cooling circuit via a cooling coil 3 located at the top of the vapor chamber, releasing latent heat and condensing into liquid condensate that falls back to the liquid surface, achieving phase change heat recovery. The cooling coil 3 is connected to a radiator 2 located outside the sealed enclosure 4, enhancing cooling flux through the external heat dissipation structure and effectively supporting heat exchange requirements under high-rate discharge and extreme thermal disturbances. To cope with different thermal conditions, the system is equipped with an extraction unit, including a condensate circulation pump 7, an electrically controlled valve 6, and a gas storage tank 8. The gas extraction rate is adjusted according to the steam chamber pressure, coordinating the control of internal pressure and boiling point stability. An external cooling tower 9 absorbs heat transferred from the coils, completing a closed-loop cooling process for the condensate. A flame retardant storage tank 10 and the corresponding electrically controlled valve 6 constitute a flame retardant injection assembly, dynamically injecting flame retardant according to the battery's thermal risk level, enhancing emergency response to battery thermal runaway. All sensor data and control quantities are centrally processed through a controller, achieving graded joint adjustment of the condensate circulation pump speed, gas extraction rate, and flame retardant injection rate. This constructs a closed-loop thermal management control system covering low-rate steady-state heat dissipation, high-rate enhanced cooling, and emergency response to extreme conditions, supporting the implementation of the MPC-based graded control strategy thermal management method proposed in this invention.

[0067] This implementation scheme achieves full-process coordination of battery phase change thermal management data, from raw acquisition to thermal risk identification, target control quantity generation, and closed-loop control command execution, by constructing a clearly defined functional structure including a thermal management data preprocessing module, a battery thermal risk assessment module, a graded regulation parameter generation module, and a predictive control execution optimization module. The modules work together around battery phase change thermal management data. The battery thermal risk assessment module outputs thermal risk level labels and thermal risk state datasets, driving the graded regulation parameter generation module to generate target values ​​for condensate circulation pump speed, gas extraction rate, and flame retardant injection rate. Furthermore, the predictive control execution optimization module constructs an MPC regulation prediction model, performing multiple rounds of state response prediction and smooth correction of control commands. This ensures the control strategy remains stable and dynamically adapts to risk level changes during multi-cycle assessments, ultimately achieving an organic unity of battery thermal risk assessment, cooling regulation strategy generation, and control command execution, thus improving the intelligence, adaptability, and safety level of battery thermal management.

[0068] 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.

[0069] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A battery pack phase change thermal management method based on MPC hierarchical control, characterized in that, Includes the following steps: S1: Collect battery phase change thermal management data, and perform time alignment, outlier sample removal, smoothing and noise reduction and standardization on the battery phase change thermal management data to obtain preprocessed battery phase change thermal management data; S2. Based on the preprocessed battery phase change thermal management data, construct a thermal risk assessment dataset, assess the thermal risk level of the battery under the current operating conditions, generate thermal risk level labels, and construct a thermal risk status dataset. S3, extract the thermal risk level label, set the corresponding control mode, and evaluate the control parameter values ​​of condensate circulation pump speed, gas exhaust rate and flame retardant injection rate based on the thermal risk state dataset, and generate the target control quantity sequence. S4 constructs an MPC regulation and prediction model based on target control variables and battery phase change thermal management data, sets control constraints and state boundary limits, predicts future state responses, performs smooth corrections and issues executable control commands, and combines the risk state change trend within a continuous evaluation period to achieve regulation mode rollback and strategy closed-loop update.

2. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for collecting battery phase change thermal management data and performing time alignment, outlier removal, smoothing, noise reduction, and standardization on the preprocessed battery phase change thermal management data are as follows: Real-time acquisition of battery phase change thermal management data, including cell mass, battery surface temperature, vapor chamber pressure, flame retardant injection volume, condensate flow rate, condensate level, condensate inlet temperature, condensate outlet temperature, battery voltage, battery current, and gas extraction rate; and based on battery surface temperature, battery current, and voltage data, obtaining the battery specific heat capacity through a thermal model parameter inversion algorithm, and obtaining the battery internal resistance through Ohm's law and transient voltage response analysis. The battery phase change thermal management data is time-aligned using timestamp interpolation to unify the data time base under different sampling frequencies; the continuity of the battery phase change thermal management data is identified using a first-order difference detection method to remove discontinuous samples caused by data loss and duplication; the battery phase change thermal management data is smoothed using a moving average method to reduce the impact of short-term fluctuations on data stability; and the battery phase change thermal management data is standardized using a max-min normalization method to unify the numerical scale of the battery phase change thermal management data.

3. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for constructing a thermal risk assessment dataset based on the preprocessed battery phase change thermal management data and assessing the battery's thermal risk level under current operating conditions are as follows: The continuous evaluation period is divided according to a fixed time window. Within each evaluation period, the battery surface temperature, battery current, battery internal resistance, battery specific heat capacity, battery mass, condensate inlet temperature, condensate outlet temperature, vapor chamber pressure and condensate flow rate are extracted to construct a thermal risk assessment dataset. Based on the thermal risk assessment dataset, the instantaneous thermal risk level of the battery under current operating conditions is evaluated as follows: The equivalent heating term is obtained by squared battery current, multiplied by battery internal resistance, and then divided by the product of battery specific heat capacity and battery mass; the absolute value of the temperature difference between condensate inlet temperature and condensate outlet temperature is obtained as the cooling temperature difference term; the thermal response term is obtained by summing the battery surface temperature, the equivalent heating term, and the cooling temperature difference term in sequence; the boiling point change correction term is obtained by adding one to the vapor chamber pressure, taking the natural logarithm, and then adding one again; the cooling regulation term is obtained by adding the condensate flow rate to the boiling point change correction term; the thermal regulation term is obtained by dividing the thermal response term by the cooling regulation term.

4. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for generating thermal risk level labels and constructing a thermal risk status dataset are as follows: Real-time comparison of battery thermal risk assessment value and thermal risk threshold to classify battery thermal risk level: when the battery thermal risk assessment value is less than or equal to the first-level thermal risk threshold, it is marked as first-level risk; When the battery thermal risk assessment value is greater than the Level 1 thermal risk threshold but less than the Level 2 thermal risk threshold, it is marked as Level 2 risk; When the battery thermal risk assessment value is greater than or equal to the level 2 thermal risk threshold, it is marked as level 3 risk; The battery thermal risk label, battery thermal risk assessment value, and corresponding battery phase change thermal management data are structured and bound together to output a thermal risk status dataset.

5. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for extracting the thermal risk level label and setting the corresponding control mode are as follows: Extract the risk level label for the current assessment period and set the corresponding control mode: Level 1 risk triggers heat dissipation mode, which only performs condensate circulation control; Level 2 risk triggers enhanced mode, which performs condensate circulation and air extraction control. The Level 3 risk triggers an emergency mode, simultaneously controlling condensate circulation, evacuation, and flame retardant injection. After determining the corresponding control mode, the process proceeds to execute control commands.

6. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for evaluating the control parameter values ​​of condensate circulation pump speed, gas removal rate, and flame retardant injection rate based on the thermal risk state dataset are as follows: Add one to the steam chamber pressure and take the natural logarithm to obtain the steam correction term; add one to the steam correction term and multiply it by the battery thermal risk assessment value to obtain the risk correction term; divide the risk correction term by the sum of the condensate flow rate and the minimum term to obtain the condensation control assessment value. Divide the battery surface temperature by the sum of the battery surface temperature, the condensate level, and the minimum term, and multiply the resulting ratio by the exhaust correlation coefficient to obtain the exhaust control evaluation value. The flame retardant liquid injection volume is added to the gas extraction rate, divided by the sum of the battery thermal risk assessment value and a constant, and the result is multiplied by the flame retardant correlation coefficient to obtain the flame retardant control assessment value.

7. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for generating the target control quantity sequence are as follows: The evaluation values ​​of condensation regulation, gas extraction regulation, and flame retardant regulation are respectively mapped to the target control quantities for adjusting the condensate circulation pump speed, gas exhaust rate, and flame retardant injection rate, forming a sequence of target control quantities. The target control quantities are then mapped and bound to the corresponding execution objects.

8. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for constructing the MPC regulation and prediction model based on the target control quantity and battery phase change thermal management data, setting control constraints and state boundary limits, predicting future state responses, performing smooth corrections, and issuing executable control commands are as follows: The target control variable is extracted as the control input, and the battery surface temperature, vapor chamber pressure and condensate level are extracted as state variables to construct an MPC regulation and prediction model. The maximum and minimum control values ​​and single-cycle adjustment ranges of the condensate circulation pump speed, gas discharge rate, and flame retardant injection rate are set to construct the value constraints of the control inputs; the safe operating ranges of the battery surface temperature, vapor chamber pressure, and condensate level are set to construct the boundary constraints of the state variables. The target control quantity within the current evaluation period is input into the MPC control prediction model to predict the state response sequence within future evaluation periods. If any predicted state response value touches the boundary conditions, the target control quantity is smoothly corrected within the control increment range, and a control command sequence is output. The control command sequence is data-bound with the condensate circulation pump (7), the air extraction unit and the flame retardant injection component, and sent to the execution end to execute the real-time scheduling of the control commands.

9. The battery pack phase change thermal management method based on MPC hierarchical control according to claim 1, characterized in that: The specific steps for combining the risk status change trend within a continuous assessment period to achieve the rollback of the control mode and the closed-loop update of the strategy are as follows: After the control command is executed, the battery surface temperature, vapor chamber pressure and condensate level are continuously collected and dynamically monitored in conjunction with the current risk level change trend. If the state variables remain stable within the safe range within a continuous fixed assessment period and the risk level continues to decrease, the control mode rollback mechanism is triggered, and the third-level and second-level control are exited in sequence, smoothly transitioning to the basic heat dissipation mode.

10. A battery pack phase change thermal management system based on MPC hierarchical control, characterized in that: include: The module includes a thermal management data preprocessing module, a battery thermal risk assessment module, a graded control parameter generation module, and a predictive control execution optimization module, among which: The thermal management data preprocessing module is used to collect battery phase change thermal management data and perform time alignment, abnormal sample removal, smoothing and noise reduction and standardization on the battery phase change thermal management data to obtain preprocessed battery phase change thermal management data. The battery thermal risk assessment module is used to construct a thermal risk assessment dataset based on preprocessed battery phase change thermal management data, assess the thermal risk level of the battery under the current operating conditions, generate thermal risk level labels, and construct a thermal risk status dataset. The graded control parameter generation module is used to extract thermal risk level labels, set corresponding control modes, and evaluate the control parameter values ​​of condensate circulation pump speed, gas exhaust rate and flame retardant injection rate based on the thermal risk state dataset, and generate a target control quantity sequence. The predictive control execution optimization module is used to construct an MPC regulation prediction model based on the target control quantity and battery phase change thermal management data, set control constraints and state boundary limits, predict future state responses, perform smooth corrections and issue executable control commands, and combine the risk state change trend within the continuous evaluation period to realize the control mode rollback and strategy closed-loop update.

Images