A thermal management system and control method for a liquid-cooled energy storage tank

Abstract

The application belongs to the technical field of process control, and particularly relates to a heat management system and control method for a liquid-cooled energy storage cabinet. The system is connected with battery cluster charging and discharging current, voltage, state of charge, module temperature, liquid inlet temperature, liquid return temperature, circulating pump speed and cooling unit output, generates future heat generation sequence based on charging and discharging power change and module temperature change, establishes cooling liquid transmission lag parameters according to liquid inlet temperature, liquid return temperature and historical control response, generates heat arrears according to future heat generation sequence, cooling liquid heat carrying prediction value and module equivalent heat capacity, and then generates long-time domain cooling capacity preset instruction and short-time domain circulating flow correction instruction by a double-time domain control unit. The application integrates heat generation advance amount, cooling transmission lag amount and module heat inertia amount into the same process state, so that liquid inlet temperature setting, cooling unit output and circulating pump speed are coordinately adjusted based on the same source state.

Description

Technical Field

[0001] This invention belongs to the field of process control technology, specifically relating to a thermal management system and control method for liquid-cooled energy storage cabinets. Background Technology

[0002] Thermal management of liquid-cooled energy storage cabinets is typically accomplished collaboratively by the battery management system (BMS), thermal management controller, and liquid-cooled actuators. The BMS provides the thermal management controller with data on battery cluster charging / discharging current, voltage, state of charge, and individual module temperatures. The liquid-cooled actuators provide inlet and outlet liquid temperatures, circulation pump speed, and cooling unit output. Conventional process control schemes often employ a closed-loop system using temperature thresholds, temperature difference thresholds, and fixed hysteresis. When the highest module temperature exceeds the start-up threshold, the cooling unit output is increased or the inlet liquid temperature setpoint is decreased; when the temperature difference within the cabinet exceeds the allowable range, the circulation pump speed is increased or a higher flow rate is maintained; when the temperature falls below the hysteresis threshold, the cooling unit output is reduced or the circulation pump speed is reduced. Some schemes incorporate average temperature, inlet / outlet liquid temperature difference, and ambient temperature within the cabinet as compensation parameters to reduce false triggering caused by individual temperature points. They may also set an upper limit on the output change rate to prevent frequent start-stop operations of the actuators. The control object, acquisition object, and execution object of the above scheme are all matched with the liquid-cooled energy storage cabinet, and the implementation process is clear. It can maintain a basic temperature closed loop when the charging and discharging is stable and the heat load changes slowly. However, its control basis is still the temperature result that has already been formed. The cooling action is often triggered only after the heat is conducted through the battery cell, module structure and liquid cooling plate. The starting point of process control lags behind the heat generation process.

[0003] When the energy storage cabinet is under high-rate charging and discharging or frequent power switching, battery heat generation does not occur synchronously with the module temperature. Instead, it is related to the square term of the current, voltage deviation, state of charge range, and current temperature. After a sudden increase in charging and discharging power, heat first accumulates inside the cell. This heat is then transferred to the temperature sampling location via the module materials, thermal interface, and liquid cooling plate. After the cooling unit output changes, the inlet liquid temperature needs to go through circulation pipeline transmission, liquid cooling plate heat exchange, and module heat diffusion processes before the target module temperature can be changed. Even if the existing temperature closed-loop system increases the sampling frequency or lowers the cooling start threshold, it still adjusts the cooling output only after the temperature change has been sensed by the sampling point. It cannot eliminate the time difference between heat generation preceding temperature response and cooling effect lagging behind the execution command. The module temperature curve may remain flat for a short period of time, but the internal heat storage has already increased; when the inlet liquid temperature curve has already decreased, the target module temperature may not have shown a corresponding drop yet. If a low inlet temperature or a high circulation pump speed is maintained for an extended period to cover this time difference, the low heat load module will continuously lose heat, causing a deviation between the inlet temperature setting and the actual heat demand. The circulation pump output will also form a uniform adjustment between the high heat load area and the low heat load area, making it difficult to reflect the degree of heat accumulation in different control cycles.

[0004] Existing improved control schemes incorporate predictive control, inlet / outlet liquid temperature difference assessment, or pump speed feedback correction. Predictive control typically estimates future temperature rise based on historical temperature, charging / discharging power, and ambient temperature, and provides a preset output for the cooling unit accordingly. Inlet / outlet liquid temperature difference assessment is usually used to estimate the heat carried away by the coolant and serves as the basis for whether to increase cooling output. Pump speed feedback correction is usually used to determine whether the circulating flow rate change has reached the set state. Although the above processes contain more process information than simple threshold control, heat load prediction, coolant transfer lag, and execution constraints are often processed as independent steps. The prediction results are only used to estimate cooling demand, lag compensation is only used to correct response time, and output change rate limits are only used to suppress fluctuations at the execution end. Due to the lack of intermediate state quantities that can simultaneously characterize future heat generation, the heat that the coolant can carry away, and the module's heat storage margin, the controller has difficulty determining how much heat gap remains to be offset by the cooling system within the current control cycle, and it is also difficult to synchronously transmit this heat gap to the cooling unit and circulating pump. If long-cycle cooling capacity presets and short-cycle flow rate corrections are generated based on different state quantities, they may exhibit opposite adjustment trends under the same operating conditions. When the predicted temperature rise begins to fall but the residual heat inside the module has not yet been released, the execution quantity may decrease ahead of the predicted value, while the temperature change continues to develop along the direction of thermal inertia.

[0005] Therefore, the main technical problem in the thermal management process of existing liquid-cooled energy storage cabinets lies in the fact that control methods based on temperature thresholds or separate predictive compensation cannot unify the advance amount of heat generation during charging and discharging, the lag amount of coolant transfer, and the thermal inertia of the battery module into the same process state. This results in a lack of common basis for presetting cooling capacity and correcting circulation flow. During the rapid rise in heat load, the output of the cooling unit and the inlet temperature settings cannot be established in advance according to the future heat gap, and the correction of the circulation pump speed cannot be coordinated with the time when the cooling effect reaches the module. During the decline in heat load, the control output is prone to regress before the residual heat of the module is completely released. The essence of this problem is not insufficient data from a single sensor, nor is it insufficient cooling capacity at the actuator end, but rather that the three dynamic processes of heat generation, cooling transfer, and module heat storage are not evaluated in a unified manner, resulting in a lack of coordination basis for inlet temperature, cooling unit output, and circulation pump speed within the same control cycle. Since the same heat load change presents different time sequences on the electrical side, liquid side, and temperature side, relying solely on data from any one side will cause deviations in the control state. The resulting control deviation is continuous; temperature peak overshoot will occur before the cooling effect arrives; the temperature difference inside the cabinet will continue to increase under uniform flow regulation; and the execution quantity will be repeatedly adjusted near the threshold. Summary of the Invention

[0006] The purpose of this invention is to provide a thermal management system and control method for liquid-cooled energy storage cabinets, which can solve the problems in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A thermal management system for a liquid-cooled energy storage cabinet includes: an operation data access unit, an electrothermal-fluid coupling prediction unit, a cooling hysteresis identification unit, a heat deficit generation unit, and a dual-time-domain control unit. The operation data access unit is used to access battery cluster charging / discharging current, voltage, state of charge, module temperature, inlet temperature, return temperature, circulating pump speed, and cooling unit output. The electrothermal-fluid coupling prediction unit generates a future heat generation sequence based on changes in charging / discharging power and module temperature. The cooling hysteresis identification unit establishes coolant transport hysteresis parameters based on inlet temperature, return temperature, and historical control responses. The heat deficit generation unit generates a heat deficit based on the future heat generation sequence, the predicted value of heat that the coolant can carry away, and the equivalent heat capacity of the module. The dual-time-domain control unit generates a long-time-domain cooling capacity preset command and a short-time-domain circulating flow correction command based on the heat deficit.

[0008] Preferably, the electrothermal-current coupling prediction unit includes a condition segmentation processing unit, an equivalent heat generation mapping unit, and a thermal inertia recursion unit; the condition segmentation processing unit divides the heat load into stages according to the direction of charging and discharging current, the rate of power change, and the state of charge interval; the equivalent heat generation mapping unit maps the current square term, voltage deviation term, and temperature rise slope term in each heat load stage to the stage heat generation; the thermal inertia recursion unit recursively integrates the stage heat generation with the module's equivalent heat capacity, historical residual heat, and current temperature gradient to generate a future heat generation sequence corresponding to the future control cycle.

[0009] Preferably, the cooling hysteresis identification unit includes a response timing extraction unit, a path hysteresis estimation unit, and a hysteresis parameter update unit; the response timing extraction unit extracts the control quantity change time, liquid temperature response time, and module temperature response time from the timing data of the cooling unit output, circulating pump speed, inlet liquid temperature, return liquid temperature, and module temperature; the path hysteresis estimation unit generates transmission hysteresis parameters based on the time difference between the control quantity change time and the liquid temperature response time, and between the liquid temperature response time and the module temperature response time; the hysteresis parameter update unit updates the transmission hysteresis parameters in groups according to different circulating pump speed ranges.

[0010] Preferably, the dual-time-domain control unit includes a long-time-domain control unit, a short-time-domain control unit, and a shared state coordination unit. The long-time-domain control unit takes the future heat generation sequence, transmission lag parameters, and heat deficit as inputs to generate the liquid inlet temperature setpoint and the cooling unit output preset value. The short-time-domain control unit takes the real-time module temperature, the predicted module temperature, and the rate of change of the heat deficit as inputs to generate the circulating pump speed correction value. The shared state coordination unit synchronously writes the heat deficit value into the control state cache of the long-time-domain control unit and the short-time-domain control unit, so that the cooling capacity preset command and the circulating flow correction command are generated based on the same thermal state value.

[0011] Preferably, the thermal inertia recursion unit includes a residual heat storage subunit, a heat capacity correction subunit, and a sequence smoothing subunit; the residual heat storage subunit generates historical residual heat based on the difference between the future heat generation sequence in the previous control cycle and the predicted value of heat that can be carried away by the coolant; the heat capacity correction subunit corrects the equivalent heat capacity of the module online based on the module temperature change rate, the state of charge range, and the historical residual heat; the sequence smoothing subunit performs time-series recursion on the stage heat generation according to the corrected equivalent heat capacity of the module to form a future heat generation sequence with a thermal inertia continuation term.

[0012] Preferably, the hysteresis parameter update unit includes a response sample screening subunit, a hysteresis consistency verification subunit, and a partition parameter writing subunit; the response sample screening subunit selects historical control segments where the output of the cooling unit or the speed of the circulating pump changes abruptly and the rate of change of charging and discharging power is lower than a preset threshold; the hysteresis consistency verification subunit eliminates control segments that do not meet the cooling transmission sequence based on the sequential relationship between the liquid inlet temperature decrease sequence, the liquid return temperature change sequence, and the module temperature change sequence; the partition parameter writing subunit writes the transmission hysteresis parameters corresponding to the remaining control segments into the corresponding circulating pump speed range.

[0013] Preferably, the long-time domain control unit includes a constraint construction subunit and a preset solution subunit; the constraint construction subunit constructs long-time domain control constraints based on the future highest module temperature, the temperature difference between modules inside the cabinet, the cooling unit output change rate, and the liquid inlet temperature change rate; the preset solution subunit, within the long-time domain control constraints, uses the remaining value of the heat deficit at the end of the prediction window as the state variable to solve for the liquid inlet temperature setpoint and the cooling unit output preset value; after the liquid inlet temperature setpoint and the cooling unit output preset value are determined, the short-time domain control unit generates a circulating pump speed correction value based on the real-time temperature residual.

[0014] Preferably, the heat deficit generation unit includes a deficit initial value calculation unit, a deficit rolling update unit, and a deficit limiting unit; the deficit initial value calculation unit generates a prediction window initial deficit value based on the difference between the future heat generation sequence with thermal inertia continuation term and the current predicted value of heat that can be carried away by the coolant; the deficit rolling update unit accumulates the heat generation increment, cooling zone heat, and historical residual heat in each future control cycle according to the online corrected equivalent heat capacity of the module to form a heat deficit sequence; the deficit limiting unit applies boundary constraints to the heat deficit sequence based on the heat storage margin corresponding to the module's highest allowable temperature.

[0015] Preferably, the dual time-domain control unit further includes a rollback coordination unit. When the charging and discharging power decreases and the heat deficit sequence still has a positive residual value, the rollback coordination unit freezes the decrease range of the preset output of the cooling unit and keeps the circulation pump speed correction amount in a non-zero compensation range that matches the real-time temperature residual. After the heat deficit sequence decreases to the rollback boundary determined by the equivalent heat capacity of the module, the rollback coordination unit releases the decrease step size of the preset output of the cooling unit and the circulation pump speed correction amount according to the coolant transfer hysteresis parameter.

[0016] A thermal management process control method for a liquid-cooled energy storage cabinet includes: receiving data on battery cluster charging / discharging current, voltage, state of charge, module temperature, inlet liquid temperature, return liquid temperature, circulating pump speed, and cooling unit output; generating a future heat generation sequence based on changes in charging / discharging power and module temperature; establishing coolant transport hysteresis parameters based on inlet liquid temperature, return liquid temperature, and historical control responses; generating a heat deficit based on the future heat generation sequence, a predicted value of heat that the coolant can carry away, and the equivalent heat capacity of the module; generating a long-term cooling capacity preset instruction based on the heat deficit, and generating a short-term circulating flow correction instruction based on the heat deficit; and outputting the long-term cooling capacity preset instruction and the short-term circulating flow correction instruction to the thermal management execution terminal of the liquid-cooled energy storage cabinet.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention integrates battery cluster charging / discharging current, voltage, state of charge, module temperature, inlet temperature, return temperature, circulating pump speed, and cooling unit output into the same process control chain. It predicts and generates a future heat generation sequence through electrothermal-fluid coupling, and identifies coolant transport lag parameters through cooling lag identification. Finally, it calculates the heat deficit using the future heat generation sequence, the predicted heat that the coolant can carry away, and the module's equivalent heat capacity. The heat deficit expresses the difference between future heat generation, cooling transport, and module heat storage as a single state quantity. Long-term control generates a cooling capacity preset command based on this, while short-term control generates a circulating flow correction command. The inlet temperature setting, cooling unit output, and circulating pump speed are no longer triggered by independent thresholds. When the heat load increases, the future heat generation sequence provides a lead time for heat generation, the transmission lag parameter provides a time reference for the cooling effect to reach the module, and the module's equivalent heat capacity provides constraints on heat absorption and continuation. These three types of information are collectively included in the heat deficit, enabling the cooling capacity to be preset before the temperature sampling value rises significantly, and the circulation flow rate to be synchronously corrected around the same heat gap. When the heat load decreases but residual heat still exists, the remaining state of the heat deficit limits the premature retreat of cooling output, allowing the cooling effect to cover the residual heat release process. Long-time domain control and short-time domain control no longer follow the temperature threshold and pump speed feedback separately, but instead form a control state transmission through the same heat deficit. The set-up time of the cooling unit output, the magnitude of the decrease in the inlet liquid temperature setting, and the duration of the circulation pump speed correction have the same judgment basis. The cooling preset and flow compensation in each control cycle are all based on the heat gap that has not yet been offset by the cooling system, and a correspondence is formed between the actuator actions and the electric side heat generation, liquid side transmission, and module heat storage. As a result, heat generation is advanced, cooling is delayed, and thermal inertia are incorporated into the same process control judgment. The formation chain of temperature peak overshoot is moved forward, and the expansion of temperature difference in the cabinet is jointly suppressed by cooling capacity and circulation flow driven by a unified state quantity.

[0018] 2. This invention utilizes segmented processing of operating conditions, equivalent heat generation mapping, and thermal inertia recursion to ensure that the future heat generation sequence includes information on changes in charging and discharging power, historical residual heat of the module, and corrections to equivalent heat capacity. Response timing extraction, path lag estimation, and lag parameter updates allow the coolant transport lag parameters to be updated with the circulating pump speed range and historical response segments. Rolling updates of heat deficit and amplitude limiting ensure that the generation of control quantities is constrained by the heat storage margin corresponding to the module's highest allowable temperature. Long-time domain control solves for preset values ​​within constraints of the future highest module temperature, the temperature difference between modules within the cabinet, the cooling unit output change rate, and the inlet liquid temperature change rate. Short-time domain control corrects the circulating pump speed based on the real-time temperature residual after the preset values ​​are determined. Backoff coordination limits the cooling output decrease step size when charging and discharging power decreases and there is still a heat deficit. This process ensures that cooling capacity presets, inlet liquid temperature adjustments, circulating flow corrections, and backoff control are continuously updated around the same thermal state, reducing output jumps caused by repeated triggering of a single threshold. Because the inlet temperature setting, cooling unit output preset, and circulating pump speed correction are all constrained by heat deficit, transmission lag parameters, and temperature residuals, the probability of low heat load areas being overcooled is reduced, and the circulating pump speed no longer changes independently from the cooling unit output preset. Response sample screening and lag consistency verification eliminate historical segments that do not meet the cooling transmission sequence, ensuring that lag parameter updates are consistent with the actual timing of the liquid cooling process; deficit limiting controls the output according to the heat storage margin constraint, ensuring that changes in control quantities do not deviate from the module's allowable heat capacity boundary. Rollback coordination releases the cooling unit output and circulating pump speed decrease step size according to the remaining value of the heat deficit, avoiding immediate cancellation of cooling capacity after power reduction. The module temperature difference, maximum temperature, and rate of change of executed quantities within the cabinet are incorporated into the same solution process, transforming thermal management process control from single-point temperature triggering to multi-thermal state collaborative constraints. Attached Figure Description

[0019] Figure 1 This is an overall flowchart of the liquid-cooled energy storage cabinet thermal management system of the present invention; Figure 2 This is a flowchart illustrating the generation of future heat generation sequences and the recursive calculation of thermal inertia in this invention. Figure 3 This is a flowchart of the coolant transport hysteresis parameter identification and partition update process of the present invention. Figure 4 This is a flowchart of the dual-time-domain process control and rollback coordination driven by the heat deficit of the present invention. Detailed Implementation

[0020] refer to Figure 1In one embodiment, the thermal management system for the liquid-cooled energy storage cabinet is deployed within the thermal management controller of the energy storage cabinet and establishes data interaction relationships with the battery management side data interface, the liquid cooling execution side data interface, and the control output interface. The thermal management system includes an operating data access unit, an electrothermal flow coupling prediction unit, a cooling hysteresis identification unit, a heat deficit generation unit, and a dual-time domain control unit. The operating data access unit receives battery cluster charging and discharging current, voltage, state of charge, module temperature, inlet liquid temperature, return liquid temperature, circulating pump speed, and cooling unit output, and caches and aligns the above data according to a unified sampling time scale. The electrothermal flow coupling prediction unit generates a future heat generation sequence based on changes in charging and discharging power and module temperature. The cooling hysteresis identification unit establishes coolant transfer hysteresis parameters based on inlet liquid temperature, return liquid temperature, and historical control responses. The heat deficit generation unit generates a sequence based on future heat generation... The heat deficit is generated by the quantity sequence, the predicted value of heat that the coolant can carry away, and the equivalent heat capacity of the module. The dual time-domain control unit generates a long-term cooling capacity preset command and a short-term circulation flow correction command based on the heat deficit. The long-term cooling capacity preset command is used to form the inlet temperature setpoint and the cooling unit output preset value. The short-term circulation flow correction command is used to form the circulation pump speed correction value. This allows the heat generation advance, coolant transmission lag, and module thermal inertia to be evaluated under the same process state, and enables the cooling unit output, inlet temperature setpoint, and circulation pump speed correction to generate linkage control based on the same heat deficit. This embodiment uses a unified heat deficit as a shared state for long-term and short-term control, which can avoid the lag response of temperature threshold triggering to the temperature rise that has already occurred, and enables the thermal management control process to perform closed-loop adjustment around the heat gap that has not yet been offset by the cooling system.

[0021] In this embodiment, the operation data access unit performs time unification, anomaly removal, and calculable quantity conversion on the accessed data. Specifically, the battery cluster charging and discharging current, voltage, and state of charge generate an electrical-side operation vector according to the current control cycle of the energy storage cabinet; the module temperature generates a temperature matrix according to the module number and sampling time; and the liquid inlet temperature, return temperature, circulating pump speed, and cooling unit output generate a liquid-side operation vector according to the liquid-cooled side sampling time scale. When the sampling times of data from different sources are inconsistent, the operation data access unit uses the end time of the control cycle as the alignment reference, employs adjacent sampled values ​​to maintain the electrical-side data, linear interpolation for the liquid-side temperature, and a combination of the most recent effective value and the rate of change extrapolation for the module temperature for synchronization processing. When any temperature point experiences an isolated jump exceeding the temperature change trend of its adjacent modules within an adjacent control cycle, the operation data access unit does not directly use this jump value as the control basis, but marks it as a value to be verified and retains the original value for subsequent spatial consistency judgment. The electrothermal-fluid coupling prediction unit receives the data... After aligning the data, the initial form of the future heat generation sequence is determined by the direction and magnitude of the change in charging and discharging power. The module temperature change rate is used as the feedback quantity to determine whether heat generation has been conducted to the sampling point. The cooling hysteresis identification unit extracts the transmission time sequence from the changes in the output of the cooling unit, the change in the speed of the circulating pump, the response of the inlet liquid temperature, and the response of the return liquid temperature. The heat deficit generation unit subtracts the future heat generation sequence from the predicted value of the heat that the coolant can carry away by the coolant according to the same prediction window, and converts the difference into a heat deficit that can be used for control solution by combining the equivalent heat capacity of the module. The dual time domain control unit controls the output of the cooling unit and the setting of the inlet liquid temperature in the long time domain, and controls the correction of the speed of the circulating pump in the short time domain. This provides a unified calculation basis for the cooling preset at the beginning of the high heat load and the residual heat coverage after the heat load decreases. The advantage of this embodiment is that there is a continuous data link between data access, prediction, hysteresis identification, deficit calculation and dual time domain control, avoiding the problem of inconsistent control direction caused by the use of different state quantities in each control link.

[0022] Specifically, when generating the future heat generation sequence, the electrothermal-fluid coupling prediction unit does not directly equate the current module temperature with the future heat load. Instead, it uses the change in charging and discharging power as the heat source lead time, the temperature change rate as the result after heat is conducted to the sampling location, and the state of charge interval as an indirect representation of the battery's internal resistance change. This information is then synthesized into the heat generation for multiple future control cycles. The predicted value of the heat that the coolant can carry away is determined by the inlet temperature, return temperature, the flow rate corresponding to the circulation pump speed, and the coolant temperature difference. The module's equivalent heat capacity is used to represent the degree of hysteresis in the module's temperature response after absorbing heat within the control cycle. The heat deficit numerically represents the uncompensated portion between the future heat generation and the heat that the coolant can carry away within the prediction window, and is limited to the module's allowable... Within the heat storage margin, if the charging and discharging power increases in the current cycle but the module temperature has not yet increased significantly, the future heat generation sequence will increase first, and the heat deficit will increase accordingly. The long-term control unit generates a cooling capacity preset command in advance based on the increased heat deficit, and the short-term control unit generates a circulation flow correction command based on the rate of change of the same heat deficit. If the charging and discharging power decreases but the rate of change of the module temperature remains positive, the heat deficit will not immediately return to zero. The long-term control unit and the short-term control unit still retain necessary control outputs to prevent the cooling effect from being canceled before the module's residual heat is released. The advantage of this embodiment is that it transforms the advance change of electric side heat generation, the transmission delay of liquid side cooling, and the thermal inertia response of temperature side into the same control state, so that process control no longer depends on the hysteresis triggering of a single temperature threshold.

[0023] The outstanding calorie debt can be calculated as follows, whereby... Indicates the first The heat deficit per predictive control cycle, in physical terms, is the difference in heat that has not yet been carried away by the coolant within the current prediction window and still needs to be covered by thermal management control. Indicates the first The future heat generation for each predictive control cycle Indicates the first The predicted amount of heat that the coolant can remove during each predictive control cycle is estimated. Indicates the first The module's equivalent heat capacity for each predictive control cycle. Indicates the first The equivalent temperature rise caused by residual heat within each predictive control cycle Indicates when Take when greater than 0 ,when If less than or equal to 0, take 0; if within a certain period Take 1200, Take 900, If we take 150, then The calculation result is 450, indicating that there is still a heat difference of 450 in this cycle that will be included in the subsequent control solution.

[0024]

[0025] refer to Figure 2 In one embodiment, the electrothermal-current coupling prediction unit includes a condition segmentation processing unit, an equivalent heat generation mapping unit, and a thermal inertia recursion unit. The condition segmentation processing unit divides the heat load into stages according to the direction of charging and discharging current, the rate of power change, and the state of charge interval. When the direction of charging and discharging current is charging and the rate of power change is continuously positive, the corresponding control cycle is assigned to the charging load increase stage. When the direction of charging and discharging current is discharging and the rate of power change is continuously positive, the corresponding control cycle is assigned to the discharging load increase stage. When the rate of power change is close to 0 and the module temperature change rate remains stable, the corresponding control cycle is assigned to the steady-state heat load stage. When the rate of power change is negative and the module temperature change rate is still positive, the corresponding control cycle is assigned to the waste heat release stage. The equivalent heat generation mapping unit maps the current square term, voltage deviation term, and temperature rise slope term in each heat load stage to the stage heat generation. The current square term is used to characterize the main trend of ohmic heat change in the battery cell, the voltage deviation term is used to characterize the energy shift between the current charge / discharge state and the reference state, the temperature rise slope term is used to characterize the thermal response intensity that has been conducted to the temperature sampling point, and the thermal inertia recursive unit recursively integrates the stage heat generation with the module's equivalent heat capacity, historical residual heat and the current temperature gradient to generate a future heat generation sequence corresponding to the future control cycle. The future heat generation sequence is not a single-point heat generation estimate, but a data sequence corresponding to multiple control cycles in the prediction window. Each sequence value includes the current electric side heat source, the residual heat of the previous cycle and the trend of the module temperature field change. The advantage of this embodiment is that the electric heat prediction result simultaneously reflects the changes in charge / discharge conditions and the continuation of module thermal inertia, so that the subsequent heat deficit calculation can obtain the heat generation advance amount before the temperature rises significantly.

[0026] Specifically, the operating condition segmentation processing unit calculates the charge / discharge power change rate for each control cycle and forms an operating condition identifier based on the current direction. When the operating condition identifier changes between adjacent control cycles, the operating condition segmentation processing unit divides the prediction window into multiple heat load stages and establishes an independent heat generation mapping coefficient for each stage to avoid confusing the thermal behavior of the load increase stage, steady state stage, and waste heat release stage into the same linear process. The equivalent heat generation mapping unit calculates the current square term, voltage deviation term, and temperature rise slope term within each heat load stage and selects the corresponding weight combination according to the stage type. The voltage deviation term can be obtained from the difference between the current voltage and the reference voltage in the same state of charge interval. The temperature rise slope term can be obtained from the difference between the current module temperature and the previous... The difference in module temperature during the control cycle is divided by the control cycle length. After obtaining the stage heat generation, the thermal inertia recursive unit recursively calculates the historical residual heat according to the equivalent heat capacity of the module. At the same time, the current temperature gradient is added to the sequence calculation as a representation of the heat already diffused inside the module. When a module is in a high-temperature zone and the temperature gradient between adjacent modules is large, the thermal inertia recursive unit will keep the subsequent terms of the corresponding heat generation sequence at a high level. When the power decreases but the temperature rise slope is still positive, the future heat generation sequence will not immediately drop to a low-load state with the power. The advantage of this embodiment is that it performs staged, source-specific, and inertial recursive processing on the future heat generation, so that the predicted value can reflect the time continuity of the thermal process of the liquid-cooled energy storage cabinet.

[0027] The equivalent heat production mapping can be implemented as follows, where Indicates the first The module in the first The future heat generation for each predictive control cycle Indicates the heat load stage The weights of the squared current terms are determined. Indicates the first The battery cluster charge / discharge current corresponding to each predictive control cycle Indicates the heat load stage Determined weights for voltage deviation terms, This indicates the deviation of the current voltage from the reference voltage within the same state of charge range. Indicates the heat load stage The weights of the determined temperature rise slope term, Indicates the first Temperature rise slope of each module Indicates the first The residual heat generated by each module in the previous control cycle, if during the load increase phase... Take 0.8, Take 400, Take 0.5, Take 20, Take 30, Take 0.4, If we take 60, then The value is 402.

[0028]

[0029] refer to Figure 3 In one embodiment, the cooling hysteresis identification unit includes a response timing extraction unit, a path hysteresis estimation unit, and a hysteresis parameter update unit. The response timing extraction unit extracts the control quantity change time, liquid temperature response time, and module temperature response time from the timing data of the cooling unit output, circulating pump speed, inlet liquid temperature, return liquid temperature, and module temperature. The control quantity change time is determined by the starting point of a continuous change in the cooling unit output or circulating pump speed within adjacent control cycles. The liquid temperature response time is determined by the point where the inlet liquid temperature or return liquid temperature shows a continuous change in the same direction as the control quantity change. The module temperature response time is determined by the point where the target module temperature rise slope begins to decrease or the temperature curve turns... Once the inflection point is determined, the path lag estimation unit generates transmission lag parameters based on the time difference between the control quantity change time and the liquid temperature response time, and between the liquid temperature response time and the module temperature response time. The lag parameter update unit updates the transmission lag parameters in groups according to different circulating pump speed ranges, so that the same cooling control action has different lag descriptions under different flow rate states, avoiding the use of the same coolant transmission delay under low pump speed and high pump speed states. The advantage of this embodiment is that the cooling lag parameters are derived from the actual response time sequence between the control action and the liquid temperature and module temperature, which can provide a time reference consistent with the liquid cooling process for long-term cooling preset and short-term flow correction.

[0030] Specifically, the response timing extraction unit sets a sliding analysis window in the historical control data and searches for step change segments in the output of the cooling unit or the speed of the circulating pump within the window. When the output of the cooling unit increases and the inlet liquid temperature decreases within a continuous control cycle, and the return liquid temperature changes accordingly in subsequent control cycles, the response timing extraction unit records the time of the output change, the time of the inlet liquid temperature response, and the time of the return liquid temperature response. When the rate of change of the return liquid temperature and the slope of the module temperature rise change correspondingly after the speed of the circulating pump increases, the response timing extraction unit records the time of the pump speed change and the time of the module temperature response. The path lag estimation unit records the time difference between the change in the control quantity and the liquid temperature response as the execution lag to the liquid side, and records the liquid temperature response to the module temperature. The time difference of the response is recorded as the liquid-to-module lag, and the two lags are combined into a coolant transmission lag parameter. The lag parameter update unit maintains multiple sets of lag parameters according to the interval of the circulating pump speed. When a new response sample passes the timing consistency check, it is merged into the sample set of the corresponding pump speed interval and the old parameters are replaced by a weighted update method. If the new sample differs too much from the existing sample and cannot meet the cooling transmission sequence of first changing the inlet temperature, then changing the return temperature, and then changing the module temperature, it will not participate in the parameter update. The advantage of this embodiment is that the lag identification is not a fixed delay assumption, but a partitioning process that updates with the operating data, so that the cooling control can determine the arrival time of the cooling effect in combination with the actual operating state of the circulating pump.

[0031] The transmission lag parameter can be calculated as follows, where Indicates the first The coolant transport lag parameter used in each control cycle Indicates the moment when the control variable changes. Indicates the liquid temperature response time. Indicates the module's temperature response time. Indicates the current circulating pump speed range The determined partition correction coefficients are calculated as follows: if the control quantity change time is 10, the liquid temperature response time is 13, the module temperature response time is 18, and the current circulating pump speed range corresponds to... If we take 0.9, then The equivalent lag is 7.2 control cycles.

[0032]

[0033] In one embodiment, the dual-time-domain control unit includes a long-time-domain control unit, a short-time-domain control unit, and a shared state coordination unit. The long-time-domain control unit takes the future heat generation sequence, transmission lag parameters, and heat deficit as inputs to generate an inlet temperature setpoint and a cooling unit output preset value. The short-time-domain control unit takes the real-time module temperature, the predicted module temperature, and the rate of change of the heat deficit as inputs to generate a circulating pump speed correction value. The shared state coordination unit synchronously writes the heat deficit value into the control state caches of the long-time-domain control unit and the short-time-domain control unit, so that the cooling capacity preset command and the circulating flow correction command are generated based on the same thermal state value. The prediction window of the long-time-domain control unit covers multiple control... The cycle is suitable for handling slow changes in the output and inlet liquid temperature settings of the cooling unit. The correction cycle of the short-time domain control unit is consistent with the real-time temperature feedback cycle, which is suitable for handling the rapid correction of the local heat exchange response by the circulating pump speed. The shared state coordination unit locks the heat deficit and its rate of change before the end of each control cycle and releases it to the long-time domain control unit and the short-time domain control unit at the beginning of the next cycle. This avoids the long-time domain control based on the previous state and the short-time domain control based on the current state, which would cause control asynchrony. The advantage of this embodiment is that the cooling capacity and circulation flow rate are no longer determined by two independent closed loops, but form a continuous control relationship of preset, corrected, and rolled back under the same heat deficit constraint.

[0034] Specifically, the long-time domain control unit determines the cooling capacity that needs to be established in advance based on the heat deficit sequence within the prediction window. Within the allowable range of the cooling unit output change rate and the inlet liquid temperature change rate, it calculates the preset output of the cooling unit and the setpoint of the inlet liquid temperature. If the transmission lag parameter is large, the long-time domain control unit will move the start period of the preset cooling capacity forward, so that the time for the coolant to reach the target heat exchange position matches the period of the predicted peak heat deficit. The short-time domain control unit uses the difference between the real-time module temperature and the predicted module temperature as the temperature residual, and uses the heat deficit change rate as the compensation direction. When the real-time module temperature is higher than the predicted module temperature and the heat deficit is still increasing... During the extension, the short-time domain control unit generates a positive circulation pump speed correction. When the real-time module temperature is lower than the predicted module temperature and the heat deficit is close to the rollback boundary, the short-time domain control unit gradually releases the circulation pump speed correction. The shared state coordination unit performs consistency verification on the input states of the long-time domain control unit and the short-time domain control unit. If the timestamps of the heat deficit read by the two are different, a new short-time domain correction is generated with a delay, and the correction of the previous cycle is used for transition, so as to avoid abrupt changes in the control output when the time alignment is abnormal. The advantage of this embodiment is that the long-time domain preset control covers the cooling effect lag, and the short-time domain flow correction covers the real-time temperature residual. The two maintain the same direction of adjustment through the shared state quantity.

[0035] Table 1 shows the data correspondence between the operation data access unit, prediction unit, hysteresis identification unit, heat deficit generation unit, and dual time-domain control unit, which is used to illustrate the source, processing method, and output attribution of each data object in the process control chain.

[0036] Table 1. Correspondence between Data Objects in the Process Control Link

[0037] In a preferred embodiment, the thermal inertia recursion unit includes a residual heat storage subunit, a heat capacity correction subunit, and a sequence smoothing subunit. The residual heat storage subunit generates historical residual heat based on the difference between the future heat generation sequence and the predicted value of heat that can be carried away by the coolant in the previous control cycle. The heat capacity correction subunit corrects the equivalent heat capacity of the module online based on the module temperature change rate, the state of charge interval, and the historical residual heat. The sequence smoothing subunit recursively calculates the stage heat generation according to the corrected equivalent heat capacity of the module, forming a future heat generation sequence with a thermal inertia continuation term. At the end of each control cycle, the residual heat storage subunit calculates the heat generation difference that has not been carried away by the coolant and sets the remaining heat generation amount accordingly. The heat generation difference is stored in the residual heat cache according to the module number. The heat capacity correction sub-unit reads the residual heat cache and the module temperature change rate. When the temperature rise slope corresponding to the same heat generation difference is small, it is determined that the equivalent heat capacity of the module should be increased. When the temperature rise slope corresponding to the same heat generation difference is large, it is determined that the equivalent heat capacity of the module should be decreased. The sequence smoothing sub-unit recursively extrapolates the stage heat generation within the prediction window according to the heat capacity correction result, so as to avoid the future heat generation sequence from jumping inconsistent with the module thermal inertia when switching operating conditions. The advantage of this embodiment is that through residual heat storage, online heat capacity correction and sequence smoothing extrapolation, the heat generation prediction can cover the waste heat release process and the module thermal response hysteresis.

[0038] Specifically, the residual heat storage sub-unit does not simply copy the heat deficit from the previous cycle to the next cycle. Instead, it calculates for each module the cumulative value of the future heat generation sequence within the previous prediction window and the cumulative value of the predicted heat that the coolant can carry away. The positive difference is taken as historical residual heat. If the cumulative heat generation value in the previous control cycle is greater than the cumulative heat value of the cooling zone, then the corresponding module has residual heat. If the cumulative heat value of the cooling zone is greater than or equal to the cumulative heat generation value, then the residual heat of the corresponding module is treated as 0. The heat capacity correction sub-unit compares the historical residual heat with the module temperature change rate. If the residual heat is high but the temperature change rate is low, it indicates that the module... The absorption and release of heat is slow, and the corrected equivalent heat capacity of the module increases. If the residual heat is low and the temperature change rate is high, it indicates that the module temperature responds quickly to thermal disturbances. The corrected equivalent heat capacity of the module decreases. The sequence smoothing sub-unit distributes the heat generated in the current stage to the adjacent control cycles within the prediction window according to the corrected equivalent heat capacity of the module, and adds the historical residual heat to the subsequent sequence terms in a decaying manner to avoid the future heat generation sequence from dropping to a low value when the power decreases instantaneously. The advantage of this embodiment is that the rate of change of the future heat generation sequence is constrained by the module heat capacity, which enables long-term control to still cover the heat that has not yet been released after the heat load decreases.

[0039] The online correction of the module's equivalent heat capacity can be achieved as follows, wherein... Indicates the first The module in the first The equivalent heat capacity used in each control cycle This represents the equivalent heat capacity of the previous control cycle. Indicates the heat capacity update coefficient. Indicates residual heat from history. Indicates the current temperature rise slope. This indicates the prevention of small positive numbers with a denominator of 0. Take 1000, Take 0.2, Take 80, Take 0.5, If we take 0.1, the correction increment is 26.67. It is 1026.67.

[0040]

[0041] Furthermore, the hysteresis parameter update unit includes a response sample screening subunit, a hysteresis consistency verification subunit, and a partition parameter writing subunit. The response sample screening subunit selects historical control segments where the output of the cooling unit or the speed of the circulating pump experiences a step change and the rate of change of charging / discharging power is below a preset threshold. The hysteresis consistency verification subunit eliminates control segments that do not meet the cooling transmission sequence based on the sequential relationship between the inlet temperature decrease time sequence, the return temperature change time sequence, and the module temperature change time sequence. The partition parameter writing subunit writes the transmission hysteresis parameters corresponding to the remaining control segments into the corresponding circulating pump speed range. The response sample screening subunit selects segments with small changes in charging / discharging power. To reduce the interference of sudden changes in the electric heat source on the cooling response timing extraction, the hysteresis consistency verification sub-unit requires the liquid cooling process to meet the timing relationship that the change in control quantity precedes the liquid inlet or return response, and the liquid temperature response precedes the module temperature response. If the module temperature changes in the opposite direction before the liquid temperature after the change in control quantity, the corresponding segment is not used for hysteresis parameter update. The partition parameters are written into the sub-unit to maintain the transmission hysteresis parameter table according to the circulation pump speed range, and the most recently verified response sample and its calculation result are retained in each range. The advantage of this embodiment is that the hysteresis parameter update process excludes heat source disturbances and timing abnormal segments, making the coolant transmission hysteresis parameters more suitable for process control solutions.

[0042] Specifically, the response sample screening sub-unit starts with a step change in the output of the cooling unit or the speed of the circulating pump in the historical control data, and checks the rate of change of charging and discharging power over several control cycles. If the rate of change exceeds a threshold, it is considered that the change of the heat source has a strong impact on the subsequent temperature response, and the segment is not considered a cooling lag sample. If the rate of change is below the threshold, the change sequence of the inlet liquid temperature, return liquid temperature, and module temperature is read forward. The lag consistency verification sub-unit sorts the time of drop in inlet liquid temperature, change in return liquid temperature, and change in module temperature, and determines whether they conform to the transmission sequence of the cooling medium from the cooling end to the module side. The process involves checking whether the direction of the inlet liquid temperature change corresponds to the direction of the cooler unit output change, whether the return liquid temperature change corresponds to the circulating heat exchange state, and whether the module temperature change occurs after the liquid temperature change. The partition parameters are written into the sub-section, which stores the samples that have passed the above screening and verification according to the circulating pump speed range. The transmission lag parameters within the same range are then updated with weights, with newer samples receiving higher weights and older samples that have not been updated for a long time gradually having their weights reduced. The advantage of this embodiment is that the lag parameters can be updated according to the actual operating conditions, but the update source is constrained by the physical timing, thus avoiding erroneous samples from altering the controller's judgment on the arrival time of the cooling effect.

[0043] In one embodiment, the long-time domain control unit includes a constraint construction subunit and a preset solution subunit. The constraint construction subunit constructs long-time domain control constraints based on the future maximum module temperature, the temperature difference between modules inside the cabinet, the rate of change of the cooling unit output, and the rate of change of the inlet liquid temperature. The preset solution subunit, within the long-time domain control constraints, uses the remaining value of the heat deficit at the end of the prediction window as a state variable to solve for the inlet liquid temperature setpoint and the preset output of the cooling unit. After the inlet liquid temperature setpoint and the preset output of the cooling unit are determined, the short-time domain control unit generates a circulating pump speed correction based on the real-time temperature residual. The constraint construction subunit uses the future maximum module temperature and the temperature difference between modules inside the cabinet as thermal safety constraints to adjust the cooling... The unit output change rate and the inlet liquid temperature change rate are used as execution stability constraints to ensure that the preset solution does not only pursue a rapid reduction in the heat deficit while ignoring the continuity of the execution. The preset solution sub-unit takes the remaining heat deficit at the end of the prediction window as the control state and prioritizes solving the inlet liquid temperature setpoint and the preset output of the cooling unit that can bring the remaining heat deficit into the allowable boundary. After the long-time domain result is determined, the short-time domain control unit only makes local corrections to the circulating pump speed to avoid the short-time domain control from changing the cooling capacity that has been determined by the long-time domain control. The advantage of this embodiment is that the long-time domain preset and the short-time domain correction have a sequential relationship and boundary constraints, which can reduce the oscillation of the control output when the heat load changes rapidly.

[0044] Specifically, the constraint construction sub-unit calculates the temperature trajectory of each module within the prediction window based on the future heat generation sequence and transmission lag parameters in each control cycle, and extracts the future maximum module temperature and the temperature difference between modules in the cabinet. If the future maximum module temperature approaches the allowable boundary or the temperature difference between modules in the cabinet continues to widen, the preset solution sub-unit prioritizes increasing the preset output of the cooling unit or decreasing the inlet temperature setpoint. If the rate of change of the executed quantity reaches the boundary, the preset solution sub-unit retains the unresolved heat deficit and passes it to subsequent control cycles for further processing. The short-time domain control unit reads the real-time temperature residual. If the real-time module temperature is higher than the predicted module temperature, it indicates that the real-time temperature is higher than the predicted module temperature. The actual thermal response is faster than predicted. The short-time domain control unit adds a circulation pump speed correction based on the cooling capacity determined in the long-time domain. If the real-time module temperature is lower than the predicted module temperature, it indicates that the cooling effect or heat release is stronger than predicted. The short-time domain control unit reduces the circulation pump speed correction, but does not directly modify the long-time domain liquid inlet temperature setpoint and the cooling unit output preset value. The shared state coordination unit transmits the heat deficit, remaining heat deficit, and state timestamp between the two control units. The advantage of this embodiment is that the control division of labor is clear. The long-time domain handles the establishment of cooling capacity and liquid temperature setting, while the short-time domain handles real-time residual compensation. The two are kept consistent through shared state.

[0045] refer to Figure 4In a preferred embodiment, the heat deficit generation unit includes a deficit initial value calculation unit, a deficit rolling update unit, and a deficit limiting unit. The deficit initial value calculation unit generates a prediction window initial deficit value based on the difference between the future heat generation sequence with thermal inertia continuation term and the current predicted value of heat that can be carried away by the coolant. The deficit rolling update unit accumulates the heat generation increment, cooling zone heat, and historical residual heat in each future control cycle according to the online corrected equivalent heat capacity of the module to form a heat deficit sequence. The deficit limiting unit applies boundary constraints to the heat deficit sequence based on the heat storage margin corresponding to the module's highest allowable temperature. The deficit initial value calculation unit calculates the deficit values ​​that have been generated in the current control cycle but have not yet been... The heat difference offset by the cooling system serves as the starting point of the prediction window. The deficit rolling update unit adds the future heat generation amount and deducts the predicted value of heat that can be taken away by the coolant in each cycle within the prediction window. At the same time, the historical residual heat is pushed to the subsequent cycle according to the heat release rate corresponding to the equivalent heat capacity of the module. The deficit limiting unit prevents the heat deficit amount from being used as a normal state quantity in the solution after it exceeds the module's allowable heat storage margin. When the heat deficit amount reaches the limiting boundary, the dual time domain control unit generates a stronger cooling capacity preset and circulation flow correction according to the boundary state. The advantage of this embodiment is that the heat deficit amount sequence can reflect the accumulation process of future heat gaps and is also constrained by the module's thermal safety boundary.

[0046] Specifically, the initial debt calculation unit reads the first term of the future heat generation sequence after thermal inertia recursion and the current predicted value of heat that can be carried away by the coolant. If the first term of the future heat generation is greater than the predicted value of the coolant heat, the difference forms the initial debt value. If the predicted value of the coolant heat is greater than or equal to the first term of the future heat generation, the initial debt value is 0. The debt rolling update unit recursively calculates the heat debt from the previous period, the future heat generation in the current period, the predicted value of heat that can be carried away by the coolant in the current period, and the residual heat release item in the current period, forming a heat debt sequence. If in a certain week... During the period, the charging and discharging power decreases, but the residual heat release term remains positive. The heat deficit sequence will remain positive and will not immediately drop to 0. The deficit limiting unit calculates the heat storage margin based on the difference between the module's maximum allowable temperature and the current module temperature, as well as the module's equivalent thermal capacity after online correction. When any value in the heat deficit sequence exceeds the heat storage margin, the value is limited to the heat storage margin and the boundary state is transmitted to the long-term domain control unit. The advantage of this embodiment is that the heat deficit calculation covers the initial gap, rolling accumulation, and safety boundary, enabling the controller to identify the formation, continuation, and limitation state of the heat gap within the same prediction window.

[0047] The rolling update of the heat deficit sequence can be achieved as follows, where Indicates the first The module within the prediction window The amount of heat deficit per control cycle. This indicates the amount of heat deficit from the previous forecast period. Indicates the future heat generation. This indicates the predicted value of heat that the coolant can remove. Indicates the historical residual heat release coefficient. This represents the historical residual heat from the previous forecast period. This represents the heat storage margin calculated from the module's maximum allowable temperature. This indicates that the smaller value is taken. If the heat deficit in the previous forecast period was 300, and the future heat generation is 500, the predicted heat value for the cooling zone is 420. If the value is 60 and the thermal storage capacity is 600, then the result before the update is 440, and after the limit is applied... It is 440.

[0048]

[0049] Furthermore, the dual-time-domain control unit also includes a rollback coordination unit. When the charging / discharging power decreases but the heat deficit sequence still has a positive residual value, the rollback coordination unit freezes the decrease range of the preset output of the cooling unit and keeps the circulating pump speed correction amount in a non-zero compensation range that matches the real-time temperature residual. After the heat deficit sequence decreases to the rollback boundary determined by the module's equivalent heat capacity, the rollback coordination unit releases the decrease step size of the preset output of the cooling unit and the circulating pump speed correction amount according to the coolant transfer hysteresis parameter. The rollback boundary is jointly determined by the module's equivalent heat capacity, the difference between the current temperature and the target temperature, and the residual heat release rate. When the charging and discharging power decreases, if the cooling capacity is withdrawn solely based on the power change, the heat that has not yet been released inside the module will continue to be conducted to the temperature sampling point. The rollback coordination unit determines whether the residual heat still needs to be covered by reading the positive residual value of the heat deficit sequence, and determines the retention range of the circulation pump speed correction by using the real-time temperature residual. When the heat deficit sequence is lower than the rollback boundary and the temperature residual no longer increases, the rollback coordination unit gradually releases the decreasing step size according to the transmission hysteresis parameter. The advantage of this embodiment is that the control output rollback is no longer triggered solely by the power decrease or the temperature being lower than the threshold, but is jointly limited by the heat deficit remaining state, the module's thermal capacity, and the cooling transmission delay.

[0050] Specifically, in each control cycle, the rollback coordination unit compares the current charge / discharge power change rate, the end value of the heat deficit sequence, and the real-time temperature residual. When the charge / discharge power change rate is negative and the end value of the heat deficit sequence is greater than the rollback boundary, the rollback coordination unit maintains the decrease in the preset output of the cooling unit at 0 or limits it to within the decrease range of the previous cycle. Simultaneously, it requires the short-time domain control unit to continue outputting a non-zero circulating pump speed correction. The non-zero compensation range is determined by the sign and magnitude of the real-time temperature residual. If the real-time temperature residual is positive, the circulating pump speed correction is maintained at... In the positive compensation range, if the real-time temperature residual is close to 0 but the heat deficit is still positive, the circulating pump speed correction will be rolled back in small steps. When the heat deficit sequence drops below the rollback boundary, the rollback coordination unit will not immediately cancel all cooling capacity, but will allocate the decreasing step size to several subsequent control cycles according to the coolant transmission hysteresis parameter, so that the cooling effect already transmitted in the liquid cooling path matches the module's residual heat release process. The advantage of this embodiment is that the rollback control covers the thermal inertia stage after the power decreases, reducing the secondary temperature rise caused by the premature cancellation of cooling output.

[0051] Table 2 shows the corresponding processing methods for heat deficit, transmission lag parameters, and dual time-domain control outputs in different control stages, which are used to illustrate the continuous use of the same thermal state quantity by the controller in the load increase, steady state, and rollback stages.

[0052] Table 2. Staged Dual-Time Domain Control Processing Rules

[0053] In one embodiment, the thermal management process control method for the liquid-cooled energy storage cabinet corresponds to the above-described system operation process. The method includes accessing the battery cluster's charging and discharging current, voltage, state of charge, module temperature, inlet temperature, return temperature, circulating pump speed, and cooling unit output. It generates a future heat generation sequence based on changes in charging and discharging power and module temperature. It establishes coolant transport hysteresis parameters based on the inlet temperature, return temperature, and historical control responses. It generates a heat deficit based on the future heat generation sequence, the predicted heat that the coolant can carry away, and the module's equivalent heat capacity. It generates a long-term cooling capacity preset command based on the heat deficit and a short-term circulating flow correction command based on the heat deficit. The long-term cooling capacity preset command and the short-term circulation flow correction command are output to the thermal management execution terminal of the liquid-cooled energy storage cabinet. The method performs time-scale alignment and operating condition division after data access. When generating the future heat generation sequence, thermal inertia recursion is introduced. When establishing the coolant transmission lag parameter, response timing extraction and partition update are introduced. When generating the heat deficit, initial deficit calculation, rolling update and amplitude limiting processing are introduced. When outputting control commands, long-term preset, short-term correction and rollback coordination are introduced. The advantage of this embodiment is that the method flow and system structure form a corresponding relationship, and prediction, lag compensation, deficit evaluation and dual-time-domain execution control in the thermal management of the liquid-cooled energy storage cabinet can be completed in a software process control manner.

[0054] Specifically, the method reads the operating data from the battery side and the liquid cooling side at the beginning of each control cycle and writes the data into a unified time axis. Then, it calculates the charge / discharge power change rate, module temperature rise slope, inlet / outlet liquid temperature difference, and the timing of the execution quantity change. Next, it determines the heat load stage based on the current direction, power change rate, and state of charge interval. A future heat generation sequence is generated from the current square term, voltage deviation term, temperature rise slope term, and historical residual heat. Simultaneously, it reads cooling response segments that meet the sample conditions from historical control data. Based on the timing of control quantity change, liquid temperature response, and module temperature response, it generates coolant transport hysteresis parameters. Finally, it calculates the heat deficit sequence based on the future heat generation sequence, the predicted value of heat that the coolant can carry away, and the equivalent heat capacity of the module. After the heat deficit sequence is generated, it is synchronously written into the long-term control state cache and the short-term control state cache. The long-term control generates the preset output of the cooling unit and the set value of the liquid inlet temperature based on the heat deficit sequence, the future highest temperature, the temperature difference of the modules in the cabinet, and the constraint of the change rate of the execution quantity. After the long-term control result is determined, the short-term control generates the circulation pump speed correction quantity based on the residual between the real-time module temperature and the predicted module temperature. If the residual heat release stage is entered, the backoff coordination releases the control output decrease step size based on the remaining value of the heat deficit, the equivalent heat capacity of the module, and the coolant transfer hysteresis parameter. The advantage of this embodiment is that each method process revolves around the same heat deficit, so that a continuous process control closed loop is formed between heat generation prediction, cooling hysteresis, and execution output.

[0055] The predicted heat value of the cooling zone can be calculated as follows, where Indicates the first The module in the first The predicted heat removal value of the coolant corresponding to each predictive control cycle. This represents the flow rate quantity derived from the circulating pump speed. Indicates the first The heat transfer distribution coefficient of the cooling path of each module Indicates the first The return liquid temperature for each predictive control cycle. Indicates the first The inlet temperature for each predictive control cycle. This indicates the coolant transport hysteresis parameter. This represents the hysteresis decay coefficient. Take 50, Take 0.6, Take 8, Take 0.05, If we take 4, then It is 196.49.

[0056]

[0057] In this embodiment, before the long-term cooling capacity preset command and the short-term circulation flow correction command are output to the thermal management execution terminal, a command consistency check is performed. Specifically, if the direction of the cooling unit output preset amount generated by the long-term control unit is increasing, while the direction of the circulation pump speed correction amount generated by the short-term control unit based on the real-time temperature residual is decreasing, the shared state coordination unit checks whether the heat deficit timestamp and the real-time temperature residual timestamp are consistent. If the timestamps are inconsistent, the update of the short-term circulation flow correction command is temporarily suspended and the correction amount of the previous control cycle is used. If the timestamps are consistent and the heat deficit is not, the update of the short-term circulation flow correction command is temporarily suspended and the correction amount of the previous control cycle is used. If the heat deficit is still positive, the short-time domain control unit recalculates the circulating pump speed correction amount according to the rate of change of the heat deficit. If the heat deficit is already below the backoff boundary, the backoff coordination unit takes over the control output release process. The long-time domain cooling capacity preset command and the short-time domain circulating flow correction command both include the current control cycle number, the heat deficit index, and the hysteresis parameter index. The thermal management execution end executes the corresponding output according to the control cycle number in the command. The advantage of this embodiment is that the control command undergoes a same-source state verification before output, which reduces the opposite execution command caused by the inconsistency between the predicted state and the real-time feedback state.

[0058] In a preferred embodiment, when the liquid-cooled energy storage cabinet contains multiple battery module temperature sampling points, the electrothermal-fluid coupling prediction unit generates a future heat generation sequence according to the module dimension, and the heat deficit generation unit generates a heat deficit sequence according to the module dimension, forming a cabinet-level heat deficit within the dual-time-domain control unit. The cabinet-level heat deficit can be determined by the maximum value, average value, and temperature difference diffusion trend of the heat deficit of each module. The long-time-domain control unit uses the cabinet-level heat deficit to solve for the preset output of the cooling unit and the liquid inlet temperature setpoint, while the short-time-domain control unit uses the module-level heat deficit... The heat deficit and real-time temperature residual generate the circulating pump speed correction. If the heat deficit of a certain module is higher than that of other modules, but the average heat deficit of the cabinet is lower, the long-term control unit avoids directly and significantly increasing the overall cooling output of the cabinet. The short-term control unit prioritizes adjusting the circulating flow correction to enhance the heat exchange process. If the heat deficit of multiple modules increases simultaneously, the long-term control unit increases the cabinet-level cooling capacity and lowers the inlet temperature setpoint. The advantage of this embodiment is that the cabinet-level cooling capacity and the module-level flow correction have a hierarchical control basis, which can reduce the overcooling of the entire cabinet caused by a single point temperature rise.

[0059] Specifically, the formation process of cabinet-level heat deficit includes reading module heat deficit, reading temperature residual, and reading temperature difference trend. The temperature difference trend is obtained from the rate of change of the difference between the highest and lowest module temperatures. When the heat deficit corresponding to the highest module continues to increase and the temperature difference trend is expanding, the cabinet-level heat deficit is adjusted by adding a diffusion correction term to the maximum module heat deficit. When the heat deficits of multiple modules are close and the temperature difference trend is stable, the cabinet-level heat deficit approaches the average of the heat deficits of each module. Under low heat load conditions... When the temperature residual of the module is negative and the heat deficit of the high heat load module is positive, the long-term control unit limits the liquid inlet temperature setpoint from continuing to decrease, and the short-term control unit covers the heat deficit of the high heat load module by correcting the circulation pump speed. The processing method does not change the hardware structure of the liquid-cooled energy storage cabinet, and only achieves cabinet-level and module-level control coordination through the allocation of state variables in the process control. The advantage of this embodiment is that the thermal imbalance state between modules can enter the same heat deficit evaluation process, reducing the tendency of low-temperature areas to be overcooled caused by uniform cooling.

[0060] Furthermore, the storage of historical control response data adopts a circular caching method, and an index is established according to the control cycle number, operating condition stage, circulating pump speed range, cooling unit output change direction, and response timing label. When the cooling hysteresis identification unit needs to update the transmission hysteresis parameters, it does not directly search from all historical data, but reads candidate segments according to the current operating condition stage and the current circulating pump speed range, and then filters them by the response sample screening sub-unit and the hysteresis consistency verification sub-unit. The response timing label includes the control quantity change time label, the liquid temperature response time label, and the module temperature response time label. If a historical segment has been marked as timing abnormal during caching, it will not participate in the new hysteresis parameter estimation. If a historical segment is supplemented by new temperature response data in subsequent operation, its response timing label is recalculated and added to the candidate sample set. The partition parameter writing sub-unit saves the average hysteresis parameter, sample number, and most recent update time for each circulating pump speed range. The advantage of this embodiment is that historical response data can be reused and partition updates are maintained, reducing the dependence of hysteresis parameter identification on a single response segment.

[0061] In this embodiment, the future heat generation sequence, coolant transfer lag parameters, heat deficit sequence, inlet temperature setpoint, cooling unit output preset, and circulating pump speed correction are all written into the control record. The control record is stored according to the control cycle and used for the thermal inertia recursion, residual heat storage, and lag parameter update of the next cycle. If the execution end fails to complete the corresponding action according to the output instruction within a certain control cycle, an execution deviation label is written into the control record. The control cycle with the execution deviation label is not used for cooling lag parameter update, but is still used for heat deficit rolling update, because the execution deviation will affect the actual cooling zone heat. If there is a value to be verified in the temperature sampling within a certain control cycle, the prediction unit retains the original value and the reconstructed value of that temperature point. The reconstructed value is used when predicting electrothermal-fluid coupling, and the original value is used when checking spatial consistency, to avoid single-point anomalies directly changing the future heat generation sequence. The advantage of this embodiment is that the control record serves both the prediction of the next cycle and the identification of historical responses, and the data closed loop can maintain the continuity of the control state.

[0062] In one embodiment, the thermal management system can be implemented by a processor executing a program stored in a storage medium. The running data access unit, the electrothermal flow coupling prediction unit, the cooling hysteresis identification unit, the heat deficit generation unit, and the dual time-domain control unit can all be implemented by software functional units formed by program instructions. The program instructions of the running data access unit are used to receive and cache the operating data of the battery side and the liquid cooling side. The program instructions of the electrothermal flow coupling prediction unit are used to perform operating condition segmentation, equivalent heat generation mapping, and thermal inertia recursion. The program instructions of the cooling hysteresis identification unit are used to perform response timing extraction, path hysteresis estimation, and hysteresis parameter partitioning update. The program instructions of the heat deficit generation unit are used to perform deficit initial value calculation, deficit rolling update, and deficit limiting. The program instructions of the dual time-domain control unit are used to perform long time-domain preset solution, short time-domain flow correction, and rollback coordination. The program implementation is not limited to the processor model and storage medium type, only requiring that it can complete data reading, state calculation, and instruction output according to the control cycle. The advantage of this embodiment is that the system can be implemented through software process control, which facilitates the deployment of corresponding data processing and control logic on the existing liquid-cooled energy storage cabinet control platform.

[0063] In this embodiment, all control quantities are generated around the heat deficit. If the future heat generation sequence increases but the predicted value of heat that the coolant can carry away has not increased, the heat deficit increases and triggers long-term domain control to establish cooling capacity in advance. If the coolant transmission lag parameter is large, the preset time of cooling capacity moves to the front of the prediction window. If the real-time temperature residual is higher than the prediction residual, the short-term domain control increases the circulation pump speed correction. If the heat deficit drops below the backoff boundary, the backoff coordination releases the cooling output decrease step size. This forms a closed-loop process from data access to prediction, from prediction to deficit, from deficit to control, and from control to historical update. The closed-loop process does not introduce hardware structures unrelated to the liquid-cooled energy storage cabinet, nor does it use general threshold control as the main control logic. Instead, it completes process control through the same state expression of electric side heat generation, liquid side heat carrying, and module heat storage. The advantage of this embodiment is that it can provide an implementable software control scheme for the main problems of heat generation advance, cooling lag, and thermal inertia coexisting in the thermal management of liquid-cooled energy storage cabinets.

Claims

1. A thermal management system for a liquid-cooled energy storage cabinet, characterized in that, include: The system includes a data access unit, an electrothermal flow coupling prediction unit, a cooling hysteresis identification unit, a heat deficit generation unit, and a dual time-domain control unit. The operation data access unit is used to access the battery cluster charging and discharging current, voltage, state of charge, module temperature, inlet temperature, return temperature, circulation pump speed, and cooling unit output. The electrothermal-fluid coupling prediction unit generates a future heat generation sequence based on changes in charge / discharge power and module temperature. The cooling hysteresis identification unit establishes coolant transport hysteresis parameters based on inlet temperature, return temperature and historical control response. The heat deficit generation unit generates heat deficit based on the future heat generation sequence, the predicted value of heat that can be carried away by the coolant, and the equivalent heat capacity of the module. The dual-time-domain control unit generates a long-time-domain cooling capacity preset command and a short-time-domain circulation flow correction command based on the heat deficit.

2. The thermal management system for a liquid-cooled energy storage cabinet according to claim 1, characterized in that, The electrothermal-fluid coupling prediction unit includes a working condition segmentation processing unit, an equivalent heat generation mapping unit, and a thermal inertia recursion unit. The operating condition segmentation processing unit divides the heat load into stages according to the direction of charging and discharging current, the rate of power change, and the state of charge interval. The equivalent heat generation mapping unit maps the current square term, voltage deviation term, and temperature rise slope term in each heat load stage to the stage heat generation. The thermal inertia recursive unit recursively integrates the stage heat generation with the module's equivalent heat capacity, historical residual heat, and current temperature gradient to generate a future heat generation sequence corresponding to future control cycles.

3. The thermal management system for a liquid-cooled energy storage cabinet according to claim 1, characterized in that, The cooling hysteresis identification unit includes a response timing extraction unit, a path hysteresis estimation unit, and a hysteresis parameter update unit; The response timing extraction unit extracts the control quantity change time, liquid temperature response time, and module temperature response time from the timing data of the cooling unit output, circulating pump speed, inlet liquid temperature, return liquid temperature, and module temperature. The path lag estimation unit generates transmission lag parameters based on the time difference between the control quantity change time and the liquid temperature response time, and between the liquid temperature response time and the module temperature response time. The hysteresis parameter update unit updates the transmission hysteresis parameters in groups according to different circulating pump speed ranges.

4. The thermal management system for a liquid-cooled energy storage cabinet according to claim 1, characterized in that, The dual time-domain control unit includes a long time-domain control unit, a short time-domain control unit, and a shared state coordination unit; The long-time domain control unit takes the future heat generation sequence, transmission lag parameters and heat deficit as inputs to generate the liquid inlet temperature setpoint and the cooling unit output preset value; The short-time domain control unit takes the real-time module temperature, the predicted module temperature, and the rate of change of the heat deficit as inputs to generate a circulating pump speed correction amount; The shared state coordination unit synchronously writes the heat deficit into the control state cache of the long-term control unit and the short-term control unit, so that the cooling capacity preset instruction and the circulation flow correction instruction are generated based on the same thermal state quantity.

5. The thermal management system for a liquid-cooled energy storage cabinet according to claim 2, characterized in that, The thermal inertia recursion section includes a residual heat storage subsection, a heat capacity correction subsection, and a sequence smoothing subsection; The residual heat storage sub-unit generates historical residual heat based on the difference between the future heat generation sequence in the previous control cycle and the predicted value of heat that can be carried away by the coolant. The heat capacity correction sub-unit corrects the equivalent heat capacity of the module online based on the module temperature change rate, state of charge range, and historical residual heat. The smoothing sub-unit of the sequence recursively extrapolates the stage heat generation according to the modified equivalent heat capacity of the module, forming a future heat generation sequence with a thermal inertia continuation term.

6. The thermal management system for a liquid-cooled energy storage cabinet according to claim 3, characterized in that, The hysteresis parameter update unit includes a response sample screening subunit, a hysteresis consistency verification subunit, and a partition parameter writing subunit; The response sample screening sub-section selects historical control segments in which the output of the cooling unit or the speed of the circulating pump changes abruptly and the rate of change of the charging and discharging power is lower than a preset threshold. The hysteresis consistency verification sub-unit eliminates control segments that do not meet the cooling transmission sequence based on the sequential relationship between the liquid inlet temperature decrease sequence, the liquid return temperature change sequence, and the module temperature change sequence. The partition parameter writing sub-section writes the transmission lag parameters corresponding to the remaining control segments into the corresponding circulating pump speed range.

7. The thermal management system for a liquid-cooled energy storage cabinet according to claim 4, characterized in that, The long-time domain control unit includes a constraint construction subunit and a preset solution subunit; The constraint construction sub-unit constructs long-term time-domain control constraints based on the future highest temperature of the module, the temperature difference of the module inside the cabinet, the output change rate of the cooling unit, and the change rate of the liquid inlet temperature. The preset solution sub-unit, within the long-term control constraints, uses the remaining value of the heat deficit at the end of the prediction window as the state variable to solve for the inlet temperature setpoint and the preset output of the cooling unit. After the inlet temperature setpoint and the cooling unit output preset value are determined, the short-time domain control unit generates a circulating pump speed correction amount based on the real-time temperature residual.

8. The thermal management system for a liquid-cooled energy storage cabinet according to claim 5, characterized in that, The heat debt generation unit includes a debt initial value calculation unit, a debt rolling update unit, and a debt limit unit; The initial debt calculation unit generates the initial debt value of the prediction window based on the difference between the future heat generation sequence with thermal inertia continuation term and the current predicted value of heat that can be carried away by the coolant. The debt rolling update unit accumulates the heat generation increment, cooling zone heat and historical residual heat in each future control cycle into a heat debt sequence according to the online corrected equivalent heat capacity of the module. The debt limiting unit applies boundary constraints to the heat debt sequence based on the heat storage margin corresponding to the module's highest allowable temperature.

9. The thermal management system for a liquid-cooled energy storage cabinet according to claim 7, characterized in that, The dual time-domain control unit also includes a rollback coordination unit. When the charging and discharging power decreases and the heat deficit sequence still has a positive residual value, the rollback coordination unit freezes the decrease of the preset amount of the cooling unit output and keeps the circulating pump speed correction amount in a non-zero compensation range that matches the real-time temperature residual. After the heat deficit sequence decreases to the fallback boundary determined by the module's equivalent heat capacity, the fallback coordination unit releases the decrease step size of the cooling unit output preset amount and the circulating pump speed correction amount according to the coolant transmission hysteresis parameter.

10. A thermal management process control method for a liquid-cooled energy storage cabinet, characterized in that, include: Connect to the battery cluster charging and discharging current, voltage, state of charge, module temperature, inlet temperature, return temperature, circulation pump speed, and cooling unit output. A future heat generation sequence is generated based on changes in charging and discharging power and module temperature. Establish coolant transport hysteresis parameters based on inlet temperature, return temperature, and historical control response; The heat deficit is generated based on the future heat generation sequence, the predicted heat that the coolant can carry away, and the equivalent heat capacity of the module. A long-term cooling capacity preset instruction is generated based on the heat deficit, and a short-term circulation flow correction instruction is generated based on the heat deficit. The long-term cooling capacity preset command and the short-term circulation flow correction command are output to the thermal management execution terminal of the liquid-cooled energy storage cabinet.

Images