Model prediction-based power lithium battery thermal management control method and system

Abstract

The application provides a model prediction-based power lithium battery thermal management control method and system, which comprises the following steps: obtaining demand power information of an electric vehicle power device, electrical, thermal state and cooling loop parameters of a power lithium battery pack, establishing a current path and cooling branch corresponding table and forming a current thermal management decision cycle data package; generating an orthogonal cooling topology coding sequence, collecting cooling loop and battery module responses, and forming a cooling topology coding observation matrix; establishing an electro-thermal fluid coupling prediction model, identifying space heat generation parameters and cooling heat exchange parameters, generating an internal heat source image, a cooling reachable capability matrix and a heat generation center; determining a to-be-migrated heat generation center, a target migration area, a candidate current path and a candidate cooling topology, and obtaining a target joint control instruction set by using an ant colony algorithm; and executing the target joint control instruction set and verifying an actual heat generation center by local verification coding.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for power lithium batteries, and in particular to a model-based predictive thermal management control method and system for power lithium batteries. Background Technology

[0002] Lithium-ion batteries used in electric vehicles need to withstand large and continuously changing currents during acceleration, hill climbing, fast charging, and energy recovery. Due to factors such as module internal resistance, contact resistance at connection points, state of charge, and temperature differences, heat may concentrate in areas near the current inlet, outlet, or localized areas within the module, creating localized high temperatures. If this heat cannot be dissipated in time, it can easily lead to increased module temperature differences, accelerated performance degradation, and increased thermal safety risks.

[0003] Existing thermal management methods for power lithium batteries typically adjust the cooling pump speed, cooling branch valve opening, or cooling power based on the module surface temperature, coolant inlet and outlet temperatures, or the average temperature of the battery pack. Due to the limited number of temperature measurement locations and the fact that the common return temperature incorporates the thermal responses of multiple cooling branches, existing methods struggle to accurately determine the battery module, connection location, and localized area corresponding to heat generation. They also cannot determine the actual cooling capacity of each cooling branch for different heat-generating areas, and therefore may only respond after a localized hot spot has formed.

[0004] Furthermore, existing methods typically treat the current transmission path as a fixed connection, primarily reducing temperature by increasing cooling flow. When a high-heat-generating area is located in a less cooling area, simply increasing the cooling pump speed not only increases energy consumption but is also limited by common pipeline pressure, branch flow distribution, and heat exchange structure, making it difficult to achieve targeted cooling. Existing technologies also lack control methods to shift the heat-generating location towards a region with stronger cooling capacity by adjusting the current inlet, current outlet, and conductive path.

[0005] Therefore, this invention proposes a model-predictive-based thermal management control method and system for power lithium batteries. The information disclosed in the background section is only for enhancing understanding of the background of this disclosure and may therefore contain prior art information that is not common knowledge to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by providing a model-based predictive thermal management control method and system for power lithium batteries, thereby solving the technical problems mentioned in the background section.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The model-based prediction-based thermal management control method for power lithium batteries includes the following steps:

[0009] S1. Obtain the power demand information of the electric vehicle power unit and the electrical, thermal status and cooling circuit parameters of the power lithium battery pack, determine the available current path and the coverage area of ​​the cooling branch, establish a correspondence table between the current path and the cooling branch, and form the current thermal management decision cycle data package.

[0010] S2. Generate an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, execute each coding state and collect return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response to form a cooling topology coding observation matrix.

[0011] S3. Based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, establish an electrothermal fluid coupling prediction model, identify spatial heat generation parameters and cooling heat transfer parameters, and generate internal heat source images, cooling reachability matrix, heat generation center and heat generation credibility.

[0012] S4. Based on the internal heat source image, cooling reachability matrix, heat generation center and electrothermal fluid coupling prediction model, determine the heat generation center to be relocated, the target migration area, candidate current paths and candidate cooling topologies, and use the ant colony algorithm to search for a joint control scheme that satisfies electrical, thermal and hydraulic constraints to form a target joint control instruction set.

[0013] S5. Execute the target joint control instruction set and collect joint control response data. Invert the actual heat generation center through local verification coding, calculate the heat generation center migration deviation, update the electrothermal fluid coupling prediction model accordingly, and perform rolling control or safe rollback.

[0014] S1 specifically includes: acquiring the power demand information of the electric vehicle's power unit; collecting the voltage, current, surface temperature, state of charge, connection status, and cooling circuit parameters of the battery module; determining the contact resistance and allowable current based on the voltage drop change before and after a current step, thus forming an initial operating parameter set; screening available current paths based on the electrical connection diagram and liquid cooling circuit diagram; determining the battery modules and thermal state areas covered by the cooling branches; establishing a correspondence table between current paths and cooling branches; performing time alignment and validity verification on the initial operating parameter set; estimating the cooling branch flow rate based on valve opening and common pipeline pressure difference; and associating valid data, data validity flags, and the correspondence table to form the current thermal management decision cycle data package.

[0015] S2 specifically includes: generating cooling topology coding states that meet the requirements of flow rate, pressure, and fault isolation based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch; screening states with full column rank and column correlation not exceeding the threshold to form an orthogonal cooling topology coding sequence; executing each coding state in an open-then-close manner; and collecting inlet and return liquid temperatures, total coolant flow rate, common pipeline pressure difference, module surface temperature, and actual output power after the valve opening reaches the target range to form an original response sequence; performing benchmark correction, time unification, and identifiability checks on the original response sequence; and arranging the coding vector with each response sequence to form a cooling topology coding observation matrix.

[0016] S3 specifically includes: dividing the battery module into thermal state regions based on the thermal management decision cycle data package and the cooling topology coding observation matrix; establishing an electrothermal-fluid coupling prediction model with module current, estimated flow rate of cooling branches, common inlet temperature, and cooling topology coding state as inputs, and regional temperature and common return temperature as outputs; using the observation matrix to iteratively identify the regional heat generation correction coefficient, equivalent thermal resistance, heat transfer coefficient correction, return propagation delay, and common return mixing correction coefficient to obtain spatial heat generation parameters, cooling heat transfer parameters, and identification validity indicators; when identification is valid, generating an internal heat source image, cooling reachability matrix, heat generation center, heat generation confidence level, and model version number based on the parameters.

[0017] S4 specifically includes: determining the heat generation center to be migrated and the target migration area based on the internal heat source image, cooling reachability matrix, heat generation center, electrothermal fluid coupling prediction model, and available current path, forming candidate current paths and candidate cooling topologies; combining candidate current paths and candidate cooling topologies into candidate joint control schemes, using the electrothermal fluid coupling prediction model to predict temperature, common pipeline pressure difference, power tracking error, and effective heat generation center migration amount, deleting schemes that do not meet the constraints, and using the ant colony algorithm to search according to the comprehensive evaluation value; determining the target current path and target cooling topology based on the search results, forming a target joint control instruction set containing switching sequence, valve opening, pump speed, predicted migration results, and safety backoff information.

[0018] S5 specifically includes: verifying and executing the target joint control instruction set, switching the target current path using a controlled overlapping method, executing the target cooling topology in a first-on, then-off manner, collecting electrical, thermal, hydraulic, and execution feedback data, forming joint control response data and execution validity flags; executing local verification coding when the execution validity flag is valid, inputting the joint control response data into the electrothermal fluid coupling prediction model, inverting the actual heat generation center and actual heat generation intensity, calculating the heat generation center migration deviation and determining whether reverse migration is possible; updating relevant model parameters in a bounded manner based on the heat generation center migration deviation, verifying the updated prediction error, generating a new model version number, and executing rolling control or safety rollback based on the verification results.

[0019] A model-based predictive thermal management control system for power lithium batteries includes:

[0020] The module includes a data acquisition and mapping module, an encoding and observation module, a model identification module, a joint optimization module, and an execution and update module.

[0021] The data acquisition and mapping module is used to acquire the power demand information of the electric vehicle power unit and the electrical, thermal status and cooling circuit parameters of the power lithium battery pack, determine the available current path and the coverage area of ​​the cooling branch, establish a correspondence table between the current path and the cooling branch, and form the current thermal management decision cycle data package.

[0022] The coding observation module is used to generate and execute an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, and to collect the return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response to form a cooling topology coding observation matrix.

[0023] The model identification module is used to establish an electrothermal fluid coupling prediction model based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, identify spatial heat generation parameters and cooling heat transfer parameters, and generate internal heat source images, cooling reachability matrix, heat generation center and heat generation confidence;

[0024] The joint optimization module is used to determine the heat generation center to be relocated, the target migration area, candidate current paths and candidate cooling topologies based on the internal heat source image, cooling reachability matrix, heat generation center and electrothermal coupling prediction model. It uses ant colony algorithm to search for joint control schemes that meet electrical, thermal and hydraulic constraints, and forms a target joint control instruction set.

[0025] The execution update module is used to execute the target joint control instruction set and collect joint control response data. It inverts the actual heat generation center through local verification coding, calculates the heat generation center migration deviation, updates the electrothermal fluid coupling prediction model based on the heat generation center migration deviation, and executes rolling control or safety rollback.

[0026] The beneficial effects of this invention are as follows:

[0027] This invention establishes a correspondence between current paths and cooling branches, and performs orthogonal cooling topology coding. This allows for the differentiation of heat generation contributions from different battery modules and thermal states based on limited temperature, flow, and pressure responses, improving the accuracy of identifying localized heat generation locations. By using an electrothermal coupling prediction model to identify spatial heat generation and cooling heat transfer parameters, it generates internal heat source images and heat generation centers. This enables the detection of localized heat concentrations before a significant increase in module surface temperature, reducing the risk of persistent localized hotspots.

[0028] This invention constructs a cooling accessibility matrix, which quantifies the cooling effect of different cooling branches on various thermal states, avoiding the need to uniformly increase cooling flow rate based solely on average temperature and improving the targeted allocation of cooling resources. By screening available current paths and adjusting current inlet and outlet, the heat generation center is shifted to areas with higher cooling accessibility, overcoming the limitation of passive cooling from the cooling side and improving the heat extraction capacity of local high heat generation areas.

[0029] This invention utilizes an ant colony algorithm to jointly optimize current paths and cooling topologies. While meeting constraints related to power requirements, current carrying capacity, temperature, and pipeline pressure, it also considers hotspot migration, module temperature differences, cooling pump power, and switching frequency, thereby improving the overall performance of thermal management control. By inverting the actual heat generation center through local verification coding and updating the electrothermal coupling prediction model based on the heat generation center migration deviation, the model can adapt to changes in contact resistance, valve deviations, and heat transfer performance degradation, achieving closed-loop rolling optimization of thermal management for power lithium batteries. Attached Figure Description

[0030] Figure 1 This is a flowchart of the model prediction-based thermal management control method for power lithium batteries according to the present invention.

[0031] Figure 2 This is a structural diagram of the power lithium battery thermal management control system based on model prediction according to the present invention. Detailed Implementation

[0032] 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.

[0033] Example 1: As Figure 1 As shown, this embodiment provides a model-predictive-based thermal management control method for power lithium batteries, including the following steps:

[0034] S1. Obtain the power demand information of the electric vehicle power unit and the electrical, thermal status and cooling circuit parameters of the power lithium battery pack, determine the available current path and the coverage area of ​​the cooling branch, establish a correspondence table between the current path and the cooling branch, and form the current thermal management decision cycle data package.

[0035] S2. Generate an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, execute each coding state and collect return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response to form a cooling topology coding observation matrix.

[0036] S3. Based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, establish an electrothermal fluid coupling prediction model, identify spatial heat generation parameters and cooling heat transfer parameters, and generate internal heat source images, cooling reachability matrix, heat generation center and heat generation credibility.

[0037] S4. Based on the internal heat source image, cooling reachability matrix, heat generation center and electrothermal fluid coupling prediction model, determine the heat generation center to be relocated, the target migration area, candidate current paths and candidate cooling topologies, and use the ant colony algorithm to search for a joint control scheme that satisfies electrical, thermal and hydraulic constraints to form a target joint control instruction set.

[0038] S5. Execute the target joint control instruction set and collect joint control response data. Invert the actual heat generation center through local verification coding, calculate the heat generation center migration deviation, update the electrothermal fluid coupling prediction model accordingly, and perform rolling control or safe rollback.

[0039] S1 specifically includes the following sub-steps:

[0040] S110. Obtain power demand information of the electric vehicle power unit and initial operating data of the power lithium battery pack. The vehicle controller in the electric vehicle power unit sends the current power demand, the predicted power demand sequence, and the allowable power fluctuation range through the vehicle communication bus; the predicted power demand sequence includes the prediction start time, prediction time domain, time interval, and power demand at each prediction time.

[0041] The battery management system acquires the module voltage, module current, surface temperature, and state of charge of each battery module. The first three items are collected by the voltage sampling circuit, the current sensor, and the temperature sensor located in the vicinity of the positive current inlet, the vicinity of the negative current outlet, and the center of the module, respectively. The state of charge is provided based on the current integration result and the open-circuit voltage correction result.

[0042] The connection status of each current inlet and outlet is obtained through the drive feedback signal of the current path switching matrix, and the voltage drop on both sides of the connection position is collected. When the battery current undergoes a step change not lower than the identification current threshold (the identification current threshold is preferably in the range of 20A to 50A), and the switching device is stably turned on and the sampled data is valid, the contact resistance is calculated according to the following formula:

[0043]

[0044] in, The contact resistance at the connection point corresponding to the r-th current inlet or outlet is expressed in Ω. This indicates the change in voltage drop at the connection point before and after a change in current, expressed in volts (V). This indicates the corresponding change in current, expressed in amperes (A).

[0045] Take the median value after obtaining three consecutive results. For example, if the battery current changes from 80A to 120A, and the voltage drop across a certain current inlet increases by 6mV, then the contact resistance of that current inlet is 0.15mΩ.

[0046] Based on the rated current-carrying capacity of the connecting conductor, the allowable conduction current of the switching device, the thermal safety current corresponding to the contact resistance, and the allowable charge / discharge current of the battery module, the minimum value is taken as the allowable current in the corresponding direction. The common inlet temperature, common return temperature, total coolant flow rate, common inlet pressure, common return pressure, and the valve opening of each cooling branch are acquired through corresponding sensors, and these, along with the sampling timestamps, form the initial operating parameter set. The data sampling period refers to the time interval between two adjacent samples; the thermal management decision period refers to the time range for regenerating the target joint control command set.

[0047] S120. Establish a table corresponding to current paths and cooling branches. Read the hardware connection configuration pre-written into the controller's non-volatile memory based on the electrical connection diagram and liquid cooling circuit diagram. As a specific implementation, the current path switching matrix used to change the connection state of the current inlet and current outlet is composed of a high-current bidirectional solid-state relay array or a high-current MOSFET switch array with an anti-arc buffer circuit, to meet the requirement of achieving microsecond to millisecond-level contactless safe switching of the power battery under a rated current of hundreds of amperes.

[0048] The path contact resistance of the v-th current path is determined by the following formula:

[0049]

[0050] in, Let v represent the total equivalent resistance of the v-th current path. , and These represent the contact resistance at the current inlet, the contact resistance at the current outlet, and the on-resistance of the switching device in the path, respectively, all in Ω.

[0051] A safe current path refers to a preset current path that has been confirmed by factory electrical safety tests, meets the above conditions, and allows a current flow that is not lower than the minimum power sustaining current. For example, battery module M1 has current inlets IN1 and IN2 and current outlets OUT1 and OUT2; when the switching device corresponding to IN2 fails, and the allowable current flow of the path from IN1 to OUT2 is lower than the module's required current, only the path from IN1 to OUT1 is determined as a usable current path.

[0052] Based on the actual connection relationships, determine the battery module and the area adjacent to the positive current inlet, the area adjacent to the negative current outlet, or the central area of ​​the module covered by each cooling branch. Associate the available current paths with the battery modules and the cooling branches with the covered areas, forming a current path to cooling branch correspondence table. The current path to cooling branch correspondence table includes the numbers of the battery module, current path, current inlet, current outlet, cooling branch, and cooling branch valves, as well as the permissible current direction, permissible current flow, path contact resistance, covered area, and version number.

[0053] S130. Perform time alignment and validity verification on the initial operating parameter set to form the current thermal management decision cycle data packet. The start time of the thermal management decision cycle is taken as the unified sampling time; continuously changing data is linearly interpolated using two valid sampled values ​​before and after this time, and the connection status and valve status are taken from the most recent valid feedback value before this time.

[0054] Based on the cooling branch flow coefficient, the effective flow area corresponding to the valve opening, the difference between the common inlet pressure and the common return pressure, and the coolant density, calculate the estimated flow rate of the j-th cooling branch:

[0055]

[0056] in, This represents the estimated flow rate of the j-th cooling branch, in m³ / s. 3 / s; This represents the flow coefficient of the j-th cooling branch; Indicates the valve opening degree as Effective circulation area at that time, in m 2 ; and These represent the common inlet pressure and the common return pressure, respectively, in Pa. This indicates the density of the coolant, expressed in kg / m³. 3 The above parameters were obtained from the factory flow rate calibration. Defined as the pressure difference in the common pipeline.

[0057] Verify timestamps, sensor ranges, data change rates, switch feedback, and flow consistency. Safe cooling status refers to the preset valve opening and cooling pump speed combination, confirmed by factory thermoelectric joint testing, ensuring that the cooling flow rate of each battery module is not lower than the minimum cooling flow rate, and that the common inlet pressure, common return pressure, and common pipeline pressure differential are within allowable ranges.

[0058] The verified data, data validity flags, current path and cooling branch correspondence table, and their version numbers are associated to form the current thermal management decision cycle data packet. If the critical data is invalid, the safe current path and safe cooling state are maintained, and cooling topology coding is not performed; otherwise, the current thermal management decision cycle data packet and the current path and cooling branch correspondence table are used as inputs to S210 and S310.

[0059] S2 specifically includes the following sub-steps:

[0060] S210. Generate an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the current path-cooling branch correspondence table. Read the data validity flag, surface temperature of each battery module, estimated flow rate of each cooling branch, common inlet pressure, common return pressure, common pipeline pressure difference, total coolant flow rate, and allowable power fluctuation range of the electric vehicle's power unit. If critical data is invalid, the surface temperature of any battery module reaches the upper limit of the coding start temperature, the pressure parameter exceeds the allowable range, or the cooling pump fails, the cooling topology coding is not executed, and a safe cooling state is maintained.

[0061] The cooling topology coding state refers to the cooling circuit state determined by the target opening degree of each cooling branch valve and the target speed of the cooling pump within a single coding duration. Based on the current path-cooling branch correspondence table, the combinations of coded opening degree and minimum safe opening degree for each cooling branch valve are enumerated. The coded opening degree is used to identify flow rate changes, and the minimum safe opening degree is used to maintain the minimum cooling flow rate for the corresponding battery module. Combinations that cannot guarantee the minimum cooling flow rate, form a closed high-pressure liquid section, have excessive pressure parameters, or contain faulty branches are deleted.

[0062] The l-th candidate cooling topology encoding state is represented as an encoding vector. Where M represents the number of battery modules; when the cooling branch corresponding to the m-th battery module is at the coded opening degree, When at the minimum safe opening degree, The correlation between any two battery module corresponding to the code column is calculated according to the following formula:

[0063]

[0064] Where L represents the number of encoded states. This indicates the correlation between the corresponding coded columns of the m-th battery module and the n-th battery module. and These represent the average values ​​of the corresponding coded columns. When the variance of any coded column is 0, the coded matrix is ​​directly determined to be unidentifiable.

[0065] The coding states that achieve full column rank in the coding matrix and whose inter-column correlation does not exceed the correlation threshold are selected and arranged in ascending order of the number of valve switching between adjacent coding states to form an orthogonal cooling topology coding sequence. The correlation threshold is determined by the factory thermoelectric joint calibration. For example, for four battery modules, six coding vectors can be used: 1100, 1010, 1001, 0110, 0101, and 0011. Each battery module is in the coding open state three times, which can distinguish the independent contribution of each battery module to the common return fluid temperature.

[0066] The temperature warning threshold is lower than the upper limit of the coding start temperature, the upper limit of the coding start temperature is lower than the upper limit of the coding stop temperature, the upper limit of the coding stop temperature is not higher than the safety fallback temperature, and the safety fallback temperature is lower than the maximum allowable operating temperature of the battery module. For example, if the maximum allowable operating temperature of a certain model of battery module is known to be 60℃, then the thresholds can be set as follows: temperature warning threshold 45℃, upper limit of coding start temperature 50℃, upper limit of coding stop temperature 53℃, and safety fallback temperature 55℃. The specific thresholds are ultimately determined through factory thermal runner calibration.

[0067] S220. Execute the orthogonal cooling topology coding sequence and collect the raw response data for each coding state. The coding observation window refers to the time range within which the cooling topology coding state is executed and the hydrothermal response is collected. Adjacent coding states are switched using a first-open-then-close method: first, adjust the cooling branch valves that require increased flow in the next coding state to the coded opening degree; after the common inlet pressure, common return pressure, and common pipeline pressure difference are within the allowable range, then adjust the cooling branch valves that require decreased flow to the minimum safe opening degree, and adjust the target speed of the cooling pump according to the effective flow area of ​​all valves.

[0068] The start time of the coded response is determined when the feedback opening of each cooling branch valve falls within the target opening tolerance range. In each coded state, the following data are collected: common inlet temperature, common return temperature, total coolant flow rate, common inlet pressure, common return pressure, actual valve opening, estimated flow rate of each cooling branch, surface temperature of each battery module, and actual output power of the electric vehicle's power unit, forming the original response sequence. If the actual output power exceeds the allowable power fluctuation range, the current coded state is marked as invalid.

[0069] The rate of change of adjacent sampled values ​​of total coolant flow, common pipeline pressure difference, and common return fluid temperature were used as stability criteria. Hydraulic response was considered stable when all three rates of change did not exceed their corresponding stability thresholds within N consecutive data sampling periods, where N represents the number of consecutive data sampling periods used for stability determination. Subsequently, thermal response data was collected, and the observation period was the time required for the common return fluid temperature to reach 90% of its final step change.

[0070] If the module surface temperature reaches the upper limit of the coding stop temperature, the estimated flow rate of any cooling branch is lower than the minimum cooling flow rate, the common inlet pressure, common return pressure or common pipeline pressure difference exceeds the allowable range, the valve feedback is abnormal or the key data fails, then the safe cooling state is restored and the corresponding original response sequence is discarded.

[0071] S230. Perform benchmark correction, time unification, and identifiability checks on the effective raw response sequence to form a cooling topology coding observation matrix. Taking the start time of the coded response as time 0 and the previous effective sample value as the benchmark, convert the common return temperature, total coolant flow rate, common pipeline pressure difference, and surface temperature of each battery module into a relative change sequence.

[0072] The common return liquid temperature rise response is calculated according to the following formula:

[0073]

[0074] in, This represents the common return liquid temperature rise response at the k-th uniform sampling time under the l-th coded state; and These represent the common return liquid temperature and the common inlet liquid temperature at the corresponding time, respectively, both in °C. and These represent the common return liquid temperature and the common inlet liquid temperature at time 0 (i.e., the start time of the coding response) in the l-th coding state, respectively, both in °C.

[0075] All valid response sequences are resampled to the same sampling interval and observation duration; continuous measurement data are interpolated linearly, valve status is maintained by previous values, and coded states with a missing sampling point ratio exceeding the missing ratio threshold are deemed invalid.

[0076] The encoded vector, actual output power, common inlet temperature, common return temperature rise, total coolant flow rate, common pipeline pressure difference, estimated flow rate of cooling branches, and battery module surface temperature response sequence are arranged in a fixed field order to form a cooling topology encoding observation matrix, and associated with the current thermal management decision cycle number, the version number of the current path and cooling branch correspondence table, and the data validity flag.

[0077] If the coding matrix is ​​not full column rank, the number of valid coding states is less than the number of battery modules, the correlation of coding columns exceeds the correlation threshold, or the common return liquid temperature rise response does not exceed the temperature sensor noise threshold, then return to S210 to regenerate the orthogonal cooling topology coding sequence; otherwise, use the cooling topology coding observation matrix as the input to S310 and S320.

[0078] S3 specifically includes the following sub-steps:

[0079] S310. Establish an electrothermal coupling prediction model based on the current thermal management decision cycle data package, cooling topology coding observation matrix, and factory calibration data. The factory calibration data is obtained from current step tests, cooling branch flow step tests, and thermal balance tests of the same model of power lithium battery pack, including battery module geometry, temperature measurement location coordinates, equivalent heat capacity, equivalent thermal resistance, the correspondence between cooling branch flow and heat transfer coefficient, and common return liquid propagation delay. Among them, the equivalent heat capacity can be obtained by physical calculation based on the specific heat capacity and mass of the cell material constituting the battery module; the initial value of the equivalent thermal resistance is calculated based on the thickness and thermal conductivity of the heat-conducting medium between the thermal state regions using a one-dimensional heat conduction formula, and is finally corrected in the thermal balance test.

[0080] Each battery module is divided into a region adjacent to the positive electrode current inlet, a module center region, and a region adjacent to the negative electrode current outlet. Each region serves as a thermal state region, representing the temperature, heat generation power, equivalent heat capacity, and heat exchange capacity within its corresponding spatial range. Based on the current path-cooling branch correspondence table, each thermal state region is associated with its corresponding current inlet, current outlet, temperature sensor, and cooling branch.

[0081] The module's equivalent internal resistance is obtained from the calibration table based on the module's state of charge and temperature, and is corrected according to the changes in module voltage and current when the current step condition is met. The contact heat generation power at the r-th current inlet or outlet connection point at the k-th sampling time is calculated using the following formula:

[0082]

[0083] in, This indicates the heat generation power from contact, expressed in W. Indicates the current at the connection point, in amperes (A). This represents the contact resistance, measured in Ω. The heat generated at the contact point is assigned to the region adjacent to the corresponding current inlet or outlet. The heat generated within the module is determined based on the module current and the module's equivalent internal resistance, and is distributed according to the initial heat generation allocation coefficient.

[0084] The initial heat generation distribution coefficient refers to three coefficients that are set in advance based on the volume or mass ratio of the region near the positive current inlet, the central region of the module, and the region near the negative current outlet, and the sum of the three coefficients is 1. In one specific embodiment, the initial heat generation distribution coefficients of the region near the positive current inlet, the central region of the module, and the region near the negative current outlet are preferably set to 0.3, 0.4, and 0.3, respectively.

[0085] The temperature of each thermal state region is updated according to the following formula:

[0086]

[0087] in, This represents the equivalent heat capacity of the z-th thermal state region, in J / K. and These represent the temperatures of the z-th thermal state region at the k-th and (k+1)-th calculation times, respectively, in °C. This represents the time interval for model calculations, expressed in seconds (s). This indicates the regional heat production capacity, expressed in W. This represents the equivalent thermal resistance between this region and the g-th adjacent thermal state region, expressed in K / W. This represents the temperature of the g-th adjacent thermal state region at the k-th calculation time. This represents the regional heat transfer coefficient, with units of W / (m²). 2 ·K); This represents the effective heat exchange area, in m². 2 ; This indicates the common inlet temperature. The heat transfer coefficient is obtained from the factory calibration relationship based on the estimated flow rate of the corresponding cooling branch.

[0088] Assume the number of cooling branches is The predicted outlet temperatures of each cooling branch are weighted by flow rate to form the predicted common return temperature.

[0089]

[0090] in, This indicates the predicted common return liquid temperature. and Let represent the estimated flow rate and predicted outlet temperature of the j-th cooling branch, respectively.

[0091] This results in an electrothermal fluid coupling prediction model that takes module current, estimated flow rate of cooling branch, common inlet temperature and cooling topology encoding state as inputs, and thermal state region temperature, cooling branch outlet temperature and common return temperature as outputs.

[0092] S320. The spatial heat generation and cooling heat transfer parameters of the electrothermal coupling prediction model are identified using the cooling topology coding observation matrix. The heat generation correction coefficient of the hot state region, the equivalent thermal resistance between hot state regions, the heat transfer coefficient correction between the hot state region and the cooling branch, the cooling branch return liquid propagation delay, and the common return liquid mixing correction coefficient are determined as the parameters to be identified. The equivalent heat capacity, effective heat transfer area, and temperature measurement location coordinates are taken from the factory calibration values. The initial values ​​and boundaries of the parameters to be identified are determined by the factory calibration data and device tolerances.

[0093] For each valid coded state, the actual output power, common inlet temperature, estimated flow rate of the cooling branch, and coded vector are input into the electrothermal fluid coupling prediction model to obtain the theoretical common return temperature rise response and the theoretical surface temperature response. The weighted square error between the theoretical response and the measured response is used as the identification error, and the weights of each item are determined and normalized according to the reciprocal of the variance of the corresponding sensor measurement error.

[0094] The parameters to be identified are adjusted using an iterative approach, and the theoretical response of all valid encoded states is recalculated after each iteration. The process stops when the identification error is lower than the identification error threshold, or when the error decrease over three consecutive iterations is lower than the convergence threshold; if the maximum number of iterations is reached and the conditions are still not met, the parameter identification is deemed to have failed.

[0095] After parameter identification is completed, a sensitivity check is performed. If two parameters have approximately the same effect on all encoded responses, the parameter with lower sensitivity is fixed, or the process returns to S210 to add a cooling topology encoding state that can distinguish the corresponding thermal state region. The output includes spatial heat generation parameters, cooling heat transfer parameters, and a valid identification flag.

[0096] S330: Read the valid identification flag; if parameter identification fails, return to S210 to supplement the cooling topology code. If the parameter identification is valid, generate an internal heat source image, a cooling reachability matrix, and a heat generation center based on the space heat generation parameters and cooling heat transfer parameters.

[0097] The internal heat source image is a dataset arranged according to the spatial location of each thermal state region, recording the region's heat generation power, predicted temperature, spatial coordinates, and heat generation reliability. Its smallest spatial resolution unit is the thermal state region. The heat generation center of the corresponding region is obtained by weighting the heat generation power of each thermal state region. When the proportion of the heat generation power of a certain thermal state region to the total heat generation power of the module exceeds the concentration ratio threshold (the concentration ratio threshold is preferably between 60% and 80%), the center coordinates of that thermal state region are directly used.

[0098] For example, when the heat generation power of the three thermal states is 12W, 3W and 2W respectively, the heat generation center is located in the region near the positive current inlet.

[0099] The cooling capacity of the j-th cooling branch to the z-th thermal state region is calculated according to the following formula:

[0100]

[0101] in, This indicates the cooling capability of the j-th cooling branch to the z-th thermal state region; This indicates the amount of heat removed from the z-th thermal state region after the flow rate of the cooling branch increases, in W. This represents the flow rate increment of the j-th cooling branch, in meters. 3 / s. Based on the results of the maximum cooling test and the minimum safe flow test, the cooling reachability is normalized to 0-1, and a cooling reachability matrix is ​​formed with each thermal state region as the row and each candidate cooling branch as the column.

[0102] The coding completeness, fitting accuracy, parameter sensitivity, data effectiveness, and parameter boundary margin are normalized to 0-1, and the minimum value is taken as the heat generation confidence level. If the heat generation confidence level is lower than the confidence threshold, return to S210 for supplementary observations; if the heat generation confidence level meets the requirements, use the internal heat source image, cooling reachability matrix, heat generation center, heat generation confidence level, electrothermal fluid coupling prediction model with completed parameter identification, and model version number as inputs to S410.

[0103] S4 specifically includes the following sub-steps:

[0104] S410. Based on the internal heat source image, cooling reachability matrix, heat generation center, heat generation confidence, electrothermal fluid coupling prediction model, predicted demand power sequence, available current path and hardware connection relationship, determine the heat generation center to be relocated, the target migration area, the candidate current path set and the candidate cooling topology set.

[0105] Heat-producing centers with a heat production reliability not lower than the reliability threshold, and meeting at least one of the following risk conditions: regional heat production power exceeding the heat production power threshold, predicted temperature exceeding the temperature warning threshold, cooling achievable capacity below the achievable capacity threshold, or heat production power growth rate exceeding the heat production growth threshold, are identified as heat-producing centers to be relocated. The thresholds are determined by thermal safety tests, normal heat production statistics, minimum safe flow tests, and the 95th percentile of historical data, respectively. When multiple heat-producing centers to be relocated exist, they are ranked according to the degree of predicted temperature exceeding the limit, the severity of heat production anomalies, and the degree of insufficient cooling achievable capacity, with the center ranked first being prioritized for relocation.

[0106] The target migration region is defined as the thermal state region within the same battery module where the current density distribution can be altered by switching current paths, the cooling capacity is higher than the original thermal state region, and the predicted temperature is lower than the temperature warning threshold. Paths that allow insufficient current, excessive path contact resistance, or short circuits and abnormal connections after switching are deleted, forming a set of candidate current paths. The spatial heat generation distribution of each candidate current path is recalculated using an electrothermal-fluid coupling prediction model, and candidate current paths that cause the heat generation center to move towards the target migration region are retained.

[0107] Based on the cooling reachability matrix, candidate cooling topologies are generated, consisting of the target valve opening degree of each cooling branch, the target speed of the cooling pump, and the branch connectivity. Candidate cooling topologies that have any battery module flow rate below the minimum cooling flow rate, common inlet pressure, common return pressure, or common pipeline differential pressure exceeding limits, cooling pump power exceeding limits, or forming a closed high-pressure liquid section are deleted. These candidate cooling topologies are used for continuous thermal management control and differ from the cooling topology coding status used for parameter identification in S210.

[0108] S420. The ant colony algorithm is used to jointly search for candidate current paths and candidate cooling topologies. Each candidate current path is set as a current path node, and each candidate cooling topology is set as a cooling topology node. Each artificial ant selects one current path node and one cooling topology node in turn to form a candidate joint control scheme.

[0109] The candidate joint control scheme, current module temperature, state of charge, predicted demand power sequence, common inlet temperature and current model parameters are input into the electrothermal fluid coupling prediction model to obtain the predicted heat generation center location, predicted heat generation intensity, predicted highest area temperature, predicted maximum module temperature difference, predicted common pipeline pressure difference, predicted cooling pump power, predicted power tracking error and expected number of switching operations.

[0110] The displacement component of the predicted heat generation center along the target migration direction is defined as the effective heat generation center migration amount; when the predicted heat generation center moves in the opposite direction, the effective heat generation center migration amount is 0. Hard constraint screening is first performed. If a candidate joint control scheme cannot meet any of the following conditions: electric vehicle power unit required power, module voltage, module current, path allowable current, safe backoff temperature, pressure parameter upper limit, or hardware switching rules, the candidate joint control scheme is directly deleted.

[0111] Each evaluation index is normalized according to the device's allowable range and the results of the factory joint control test, and the comprehensive evaluation value is calculated according to the following formula:

[0112]

[0113] in, The value represents the comprehensive evaluation value of the s-th candidate joint control scheme; the underlined parameters represent, in order, the normalized highest regional temperature, the maximum temperature difference of the module, the pressure difference of the common pipeline, the cooling pump power, the number of switching operations, the power tracking error, and the effective heat generation center migration. to This represents the corresponding evaluation weight, and the sum of the weights is 1. The smaller the overall evaluation value, the better the corresponding candidate joint control scheme.

[0114] All allowed nodes are assigned the same positive initial pheromone, and the heuristic information is the reciprocal of the sum of the comprehensive evaluation value and a preset positive number. The probability that an artificial ant will choose node v from node u is determined by the following formula:

[0115]

[0116] in, This represents the probability of node selection. This indicates the pheromone concentration connected to the corresponding node. Indicates heuristic information, and These represent the pheromone influence coefficient and the heuristic information influence coefficient, respectively. Let w represent the set of optional nodes corresponding to node u, and w represent the set of nodes. Optional node numbers in the table.

[0117] After each iteration, the pheromone concentration is reduced according to the volatility coefficient, and the pheromone concentration of candidate joint control schemes with smaller comprehensive evaluation values ​​is increased by the reciprocal of their comprehensive evaluation values; the pheromone concentration is limited between a preset minimum and maximum value. The number of artificial ants, algorithm coefficients, and maximum number of iterations were determined by offline experiments.

[0118] In a typical implementation scenario, the number of artificial ants is 10 to 30, the pheromone influence coefficient α ranges from 1.0 to 2.0, the heuristic pheromone influence coefficient β ranges from 2.0 to 5.0, the pheromone evaporation coefficient is set to 0.1 to 0.3, and the maximum number of iterations is set to 50 to 100.

[0119] For example, candidate path IN2 to OUT1 can move the heat generation center 42mm toward the target migration area, and its allowed current is 160A, which is higher than the predicted demand current of 140A, so it can be included in the search; although candidate path IN2 to OUT2 can move 55mm, its allowed current is only 120A, so it should be deleted in the hard constraint screening stage.

[0120] S430. Form a target joint control instruction set based on the ant colony algorithm iteration results. Stop the search when the maximum number of iterations is reached, the change in the optimal comprehensive evaluation value after three consecutive iterations is less than the convergence threshold, or the remaining computation time is insufficient to complete the next round of iterations; select the scheme with the smallest comprehensive evaluation value from the candidate joint control schemes that meet the hard constraints, and determine the target current path and target cooling topology.

[0121] If the difference in the comprehensive evaluation value among multiple candidate joint control schemes is less than the equivalent threshold, then the scheme with fewer switching times, smaller difference from the current cooling topology, and lower predicted maximum temperature in the region will be selected in order of priority.

[0122] The target joint control instruction set is a set of control data that completes one joint thermal management action. It includes the current thermal management decision cycle number, model version number, target current path number, target current inlet, target current outlet, current path switching sequence and switching time, target valve opening degree of each cooling branch, target speed of cooling pump, valve opening and closing sequence, instruction validity duration, predicted heat generation center location, predicted heat generation intensity, predicted effective heat generation center migration amount, predicted highest area temperature, predicted common pipeline pressure difference, safe backoff current path, and safe cooling status.

[0123] If none of the candidate joint control schemes meet the hard constraints, or if the ant colony algorithm fails to obtain a valid scheme within the specified computation time, a "no feasible scheme" flag is set, a safe current path and safe cooling state are maintained, and a request to reduce the permissible output power is sent to the electric vehicle power unit. The target joint control command set, the "no feasible scheme" flag, and the reduced permissible output power are used as inputs to S510.

[0124] S5 specifically includes the following sub-steps:

[0125] S510: Verify and execute the target joint control instruction set to generate joint control response data. Read the "no feasible solution" flag; if this flag is valid, maintain the safe current path and safe cooling state, and send the reduced permissible output power to the electric vehicle power unit. If this flag is invalid, check the thermal management decision cycle number, model version number, instruction validity duration, target current path, target cooling topology, and current power demand; if any one of these is not satisfied, maintain the safe state and re-execute S410.

[0126] After successful verification, a controlled overlap method is used to switch the current path: first, the module current is reduced to a safe switching current, then the target current path is turned on; after confirming that the switch feedback, path voltage, and path current are normal, the current is transferred to the target current path and the original current path is disconnected within a preset overlap time. If the hardware does not allow two current paths to be turned on simultaneously, a first-off-then-on switching is performed. The safe switching current, overlap time, and confirmation time are determined by the rated parameters of the switching device and the factory switching test. To prevent device breakdown or severe arcing caused by high current switching, a safe reference range is provided: the safe switching current is usually set to no more than 5A, the overlap time is controlled between 10ms and 50ms, and the confirmation time is usually no less than 20ms.

[0127] For example, when the current path IN1 to OUT1 switches to IN2 to OUT1, the module current is first reduced from 140A to 3A, the corresponding switch for IN2 is turned on, and after confirming current establishment, the current transfer is completed within 100ms and the corresponding switch for IN1 is turned off. Simultaneously, the target cooling topology is executed according to the first-on, then-off rule, and the cooling pump speed is adjusted.

[0128] Power point tracking error is calculated using the following formula:

[0129]

[0130] in, This represents the power tracking error at the k-th sampling time. and These represent the actual output power of the lithium-ion battery pack and the power required by the electric vehicle's power unit, respectively, both in watts (W). When the power tracking error exceeds the allowable power fluctuation range, the current borne by each battery module is first adjusted; if this is still insufficient, a request to reduce the allowable output power is sent.

[0131] Record module voltage, module current, path current, switch feedback status, module surface temperature, common inlet temperature, common return temperature, common inlet pressure, common return pressure, total coolant flow rate, estimated flow rate of cooling branches, actual valve opening, actual cooling pump speed, and power tracking error to form joint control response data. When the target current path and target cooling topology are executed correctly and the data is valid, a valid joint control execution flag is generated.

[0132] When the deviation between the actual highest area temperature or the common pipeline pressure difference and the corresponding predicted value exceeds the execution deviation threshold, the current joint control is terminated, and a safe backoff current path and safe cooling state are executed.

[0133] S520. When the joint control execution validity flag is valid, perform local verification coding and invert the actual heat generation center migration result. Select coding states from the orthogonal cooling topology coding sequence that cause different flow changes between the thermal state region where the heat generation center to be migrated is located and the target migration region, and the temperature response sensitivity difference between the two thermal state regions is not lower than the sensitivity threshold, to form a local verification coding sequence; if a single coding state cannot distinguish between the two thermal state regions, select more than two coding states.

[0134] During the local verification coding, the target current path remains unchanged, the flow rate change of each cooling branch is limited to 5%-15% of the current flow rate, and the cooling flow rate of each battery module is guaranteed to be no less than the minimum cooling flow rate. The common inlet pressure, common return pressure, common pipeline pressure difference and module temperature do not exceed the allowable range.

[0135] The joint control response data, local verification coding sequence, common inlet temperature, and actual output power are input into the electrothermal fluid coupled prediction model. The target current path and the confirmed cooling heat transfer parameters are fixed. Within the physical boundary, the heat generation power of each thermal state region is adjusted to minimize the weighted error between the theoretical response and the measured response. When the weighted error exceeds the verification error threshold (preferably ranging from 5% to 10%), the verification is marked as invalid, and S210 is executed in the next thermal management decision cycle.

[0136] Let the coordinates of the heat-generating center to be relocated before execution be... The coordinates of the predicted heat production center of S430 are: The actual coordinates of the inverted heat production center are: The deviation of the heat generation center migration is calculated according to the following formula:

[0137]

[0138] in, This represents the Euclidean distance between the actual migration vector and the predicted migration vector, in mm. The actual migration distance, actual migration direction, projection of the actual migration vector onto the target migration direction, and the deviation between the actual and predicted heat generation intensity are calculated simultaneously. When the projection is less than 0, it is determined that the heat generation center has migrated in the opposite direction.

[0139] For example, if the predicted heat production center location is 42mm and the actual location is 35mm, the heat production center migration deviation is 7mm; if the actual location is -8mm, it is determined that the heat production center has migrated in the opposite direction, and a safe rollback is performed. The output includes the actual heat production center location, actual heat production intensity, actual migration vector, heat production center migration deviation, and a valid verification flag.

[0140] S530. Based on the valid verification results, the electrothermal fluid coupling prediction model is updated in a bounded manner, and rolling control or safe rollback is performed. Only the target current path contact resistance correction coefficient, the heat distribution coefficient of the target current inlet and the target current outlet adjacent area, the equivalent thermal resistance between the thermal state area where the heat generation center to be migrated is located and the target migration area, the heat transfer coefficient correction of the target cooling branch, the return liquid propagation delay, and the common return liquid mixing correction coefficient are updated; the equivalent heat capacity, spatial coordinates, and hardware connection relationships are not updated.

[0141] A sensitivity matrix is ​​constructed based on the response sensitivity of the parameters to be updated to the common return liquid temperature rise and module surface temperature. The parameter correction is calculated using the least squares method with physical boundary constraints. If the sensitivity matrix does not have a full column rank, the parameter update is not performed. The i-th parameter is updated according to the following formula:

[0142]

[0143] in, and These represent the model parameters before and after the update, respectively. This represents the parameter correction amount obtained using the least squares method with physical boundary constraints. Indicates the update gain. and These represent the lower and upper physical limits of the parameters, respectively. This means that the update results will be restricted to the corresponding physical boundaries.

[0144] The update result is saved and a new model version number is generated only if the updated prediction error of the common return liquid temperature rise decreases, the prediction error of the surface temperature of each battery module does not increase, and the prediction error of other valid coding states does not exceed the allowable range; otherwise, the model parameters before the update are restored.

[0145] If the following occurs: the heat generation center migrates in the opposite direction, the migration distance is lower than the minimum effective migration distance, the actual heat generation intensity continues to increase, the module surface temperature reaches the safe retreat temperature, the common inlet pressure, common return pressure or common pipeline pressure difference exceeds the allowable range, or the model parameters fail to update continuously, the safe retreat current path and safe cooling state are executed, and a request to reduce the allowable output power is sent to the electric vehicle power unit.

[0146] The actual location of the heat generation center, the actual heat generation intensity, the joint control response data, the model version number, and the safety rollback results are written to the historical state cache. After entering the next heat management decision cycle, S110-S130 are executed first to reacquire and update the current heat management decision cycle data package; S210 is executed when supplementary heat source observations are needed, and S410 is executed when the target joint control scheme needs to be redefined.

[0147] When the predicted temperature of all thermal states is below the temperature warning threshold, the heat generation power is below the heat generation power threshold, the cooling reachability is not below the reachability threshold, the power tracking error is within the allowable power fluctuation range, and the common inlet pressure, common return pressure, and common pipeline pressure difference are all within the allowable range for three consecutive thermal management decision cycles, the heat generation center migration control will end and the system will switch to conventional thermal management control.

[0148] Example 2: Figure 2 As shown, this embodiment provides a model-predictive-based thermal management control system for power lithium batteries, including:

[0149] The module includes a data acquisition and mapping module, an encoding and observation module, a model identification module, a joint optimization module, and an execution and update module.

[0150] The data acquisition and mapping module is used to acquire the power demand information of the electric vehicle power unit and the electrical, thermal status and cooling circuit parameters of the power lithium battery pack, determine the available current path and the coverage area of ​​the cooling branch, establish a correspondence table between the current path and the cooling branch, and form the current thermal management decision cycle data package.

[0151] The coding observation module is used to generate and execute an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, and to collect the return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response to form a cooling topology coding observation matrix.

[0152] The model identification module is used to establish an electrothermal fluid coupling prediction model based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, identify spatial heat generation parameters and cooling heat transfer parameters, and generate internal heat source images, cooling reachability matrix, heat generation center and heat generation confidence;

[0153] The joint optimization module is used to determine the heat generation center to be relocated, the target migration area, candidate current paths and candidate cooling topologies based on the internal heat source image, cooling reachability matrix, heat generation center and electrothermal coupling prediction model. It uses ant colony algorithm to search for joint control schemes that meet electrical, thermal and hydraulic constraints, and forms a target joint control instruction set.

[0154] The execution update module is used to execute the target joint control instruction set and collect joint control response data. It inverts the actual heat generation center through local verification coding, calculates the heat generation center migration deviation, updates the electrothermal fluid coupling prediction model based on the heat generation center migration deviation, and executes rolling control or safety rollback.

[0155] Those skilled in the art will understand that the preset parameters, weights, and thresholds involved in the above formulas can be conventionally set, calibrated, and adjusted according to the physical constraints of the actual power battery system and specific application scenarios.

[0156] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0157] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0158] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A model-based predictive thermal management control method for power lithium batteries, characterized in that, Includes the following steps: S1. Obtain the power demand information of the electric vehicle power unit and the electrical, thermal status and cooling circuit parameters of the power lithium battery pack, determine the available current path and the coverage area of ​​the cooling branch, establish a correspondence table between the current path and the cooling branch, and form the current thermal management decision cycle data package. S2. Generate an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, execute each coding state and collect return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response to form a cooling topology coding observation matrix. S3. Based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, establish an electrothermal fluid coupling prediction model, identify spatial heat generation parameters and cooling heat transfer parameters, and generate internal heat source images, cooling reachability matrix, heat generation center and heat generation credibility. S4. Based on the internal heat source image, cooling reachability matrix, heat generation center and electrothermal fluid coupling prediction model, determine the heat generation center to be relocated, the target migration area, candidate current paths and candidate cooling topologies, and use the ant colony algorithm to search for a joint control scheme that satisfies electrical, thermal and hydraulic constraints to form a target joint control instruction set. S5. Execute the target joint control instruction set and collect joint control response data. Invert the actual heat generation center through local verification coding, calculate the heat generation center migration deviation, update the electrothermal fluid coupling prediction model accordingly, and perform rolling control or safe rollback.

2. The model-predictive-based thermal management control method for power lithium batteries according to claim 1, characterized in that, S1 specifically includes: The system obtains the power demand information of the electric vehicle's power unit, collects the voltage, current, surface temperature, state of charge, connection status and cooling circuit parameters of the battery module, and determines the contact resistance and allowable current based on the voltage drop change before and after the current step, thus forming an initial set of operating parameters. Based on the electrical connection diagram and liquid cooling circuit diagram, available current paths are selected, the battery modules and thermal state areas covered by the cooling branches are determined, and a table of correspondence between current paths and cooling branches is established.

3. The model-predictive-based thermal management control method for power lithium batteries according to claim 2, characterized in that, Also includes: The initial set of operating parameters is time-aligned and validated. The flow rate of the cooling branch is estimated based on the valve opening and the pressure difference of the common pipeline. The valid data, data validity flags and corresponding tables are associated to form the current thermal management decision cycle data package.

4. The model-predictive-based thermal management control method for power lithium batteries according to claim 1, characterized in that, S2 specifically includes: Based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, generate cooling topology coding states that meet the requirements of flow, pressure and fault isolation, filter the states with full column rank of the coding matrix and column correlation degree not higher than the threshold, and form an orthogonal cooling topology coding sequence. Each coded state is executed in the first-open-then-close manner. After the valve opening reaches the target range, the inlet and outlet liquid temperatures, total coolant flow rate, common pipeline pressure difference, module surface temperature and actual output power are collected to form the original response sequence. The original response sequences are subjected to benchmark correction, time unification, and identifiability checks. The coding vectors are then arranged with each response sequence to form a cooling topology coding observation matrix.

5. The model-predictive-based thermal management control method for power lithium batteries according to claim 1, characterized in that, S3 specifically includes: Based on the thermal management decision cycle data package and the cooling topology coding observation matrix, the battery module is divided into thermal state regions. An electrothermal-fluid coupling prediction model is established with module current, estimated flow rate of cooling branch, common inlet temperature and cooling topology coding state as inputs, and regional temperature and common return temperature as outputs. The observation matrix is ​​used to iteratively identify the regional heat generation correction coefficient, equivalent thermal resistance, heat transfer coefficient correction, return liquid propagation delay, and common return liquid mixing correction coefficient, so as to obtain the space heat generation parameters, cooling heat transfer parameters, and identification effective indicators.

6. The model-predictive-based thermal management control method for power lithium batteries according to claim 5, characterized in that, Also includes: When the identification is valid, an internal heat source image, a cooling reachability matrix, a heat generation center, a heat generation confidence level, and a model version number are generated based on the parameters.

7. The model-predictive-based thermal management control method for power lithium batteries according to claim 1, characterized in that, S4 specifically includes: Based on the internal heat source image, cooling reachability matrix, heat generation center, electrothermal fluid coupling prediction model and available current path, determine the heat generation center to be relocated and the target relocation area, and form candidate current paths and candidate cooling topologies. Candidate current paths and candidate cooling topologies are combined into candidate joint control schemes. The electrothermal fluid coupling prediction model is used to predict temperature, common pipeline pressure difference, power tracking error and effective heat generation center migration. Schemes that do not meet the constraints are deleted. Ant colony algorithm is used to search based on comprehensive evaluation value. Based on the search results, the target current path and target cooling topology are determined, and a target joint control instruction set containing switching sequence, valve opening degree, pump speed, predicted migration results and safety backoff information is formed.

8. The model-predictive-based thermal management control method for power lithium batteries according to claim 1, characterized in that, S5 specifically includes: Verify and execute the target joint control instruction set, switch the target current path in a controlled overlapping manner, and execute the target cooling topology in a first-on, then-off manner. Collect electrical, thermal, hydraulic and execution feedback data to form joint control response data and execution validity flags. When the valid flag is valid, local verification coding is performed. The joint control response data is input into the electrothermal fluid coupling prediction model to invert the actual heat generation center and the actual heat generation intensity, calculate the heat generation center migration deviation, and determine whether the migration is reversed.

9. The model-predictive-based thermal management control method for power lithium batteries according to claim 8, characterized in that, Also includes: The relevant model parameters are updated in a bounded manner based on the migration deviation of the heat production center. The updated prediction error is verified, a new model version number is generated, and rolling control or safe rollback is performed based on the verification results.

10. A model-predictive-based thermal management control system for power lithium batteries, employing the model-predictive-based thermal management control method for power lithium batteries as described in any one of claims 1 to 9, characterized in that, include: The module includes a data acquisition and mapping module, an encoding and observation module, a model identification module, a joint optimization module, and an execution and update module. The data acquisition and mapping module is used to acquire the power demand information of the electric vehicle's power unit, as well as the electrical, thermal status, and cooling circuit parameters of the power lithium battery pack. The coding observation module is used to generate and execute an orthogonal cooling topology coding sequence based on the current thermal management decision cycle data packet and the correspondence table between current path and cooling branch, and to collect the return liquid temperature rise, coolant flow rate, common pipeline pressure difference and module surface temperature response. The model identification module is used to establish an electrothermal fluid coupling prediction model based on the cooling topology coding observation matrix and the current thermal management decision cycle data packet, and to identify the space heat generation parameters and cooling heat transfer parameters. The joint optimization module is used to determine the heat generation center to be relocated, the target migration area, candidate current paths, and candidate cooling topologies based on the internal heat source image, cooling reachability matrix, heat generation center, and electrothermal fluid coupling prediction model. The execution update module is used to execute the target joint control instruction set and collect joint control response data. It then uses local verification coding to invert the actual heat generation center and calculate the heat generation center migration deviation.

Images