IoT-based adaptive temperature control method for lithium battery packs

Abstract

This invention specifically relates to an adaptive temperature control method for lithium battery packs based on the Internet of Things (IoT), encompassing the fields of lithium battery thermal management and fluid control technology. The method includes: constructing an energy balance algebraic loop by combining viscous dissipative heat disturbance power, dynamic polarization, ohmic series resistance, and the maximum permissible transient temperature rise rate threshold; inversely deriving the maximum permissible physical limit current command and executing non-negative protection cutoff. This invention constructs a fluid momentum dimensionality reduction solution architecture including a slip layer and a high-viscosity cold core region, enabling relatively accurate analysis of the cold core morphology and viscous dissipative parasitic heat sources within the tube under extremely cold conditions, to a certain extent overcoming the feedback hysteresis limitation of traditional temperature sensors that rely on heat conduction; based on the fluid-electrochemical coupling energy balance algebraic loop, the current safety boundary can be inversely derived before the temperature rise trend becomes apparent, realizing a mode shift from passive temperature control to active thermal safety constraint, which helps improve the thermal safety margin in high-rate charge-discharge scenarios.

Description

Technical Field

[0001] This invention relates to the field of lithium battery thermal management and fluid control technology, and in particular to an adaptive temperature control method for lithium battery packs based on the Internet of Things. Background Technology

[0002] With the rapid development of new energy vehicle power and energy storage systems, the charging and discharging safety and cycle life of lithium battery packs have become core indicators of industry concern. The temperature control capability of the thermal management system is a key foundation for ensuring stable battery operation. Current mainstream liquid cooling thermal management solutions typically use cell temperature sampling values ​​as feedback, adjusting the water pump speed and coolant flow rate to achieve heat dissipation or heating regulation. Under normal operating conditions, this can meet basic temperature control requirements and has been widely used in various power and energy storage battery systems.

[0003] In extremely cold environments, the dynamic viscosity of liquid cooling media changes significantly with decreasing temperature. The flow pattern inside the pipe tends to exhibit a stratified characteristic of a high-viscosity region in the center and a slip layer near the wall. Conventional flow-pressure difference calculation models based on the assumption of uniform flow deviate from the actual fluid state. Existing liquid cooling control strategies are mostly designed for room temperature heat dissipation scenarios, with relatively limited attention paid to the changes in fluid rheological characteristics under low-temperature, high-viscosity conditions. There is still room for optimization in areas such as the parasitic heat generation from viscous dissipation and the suppression of pipeline impact during pump speed regulation.

[0004] The widespread adoption of IoT sensing technology and edge computing capabilities provides hardware support for multi-dimensional state perception and real-time online computation of battery thermal management systems. Deeply integrating fluid dynamics dimensionality reduction methods with battery electrochemical thermal safety boundaries allows for thermal risk prediction and flow-limiting intervention before the thermal inertia hysteresis of temperature sensors takes effect. Simultaneously, it enables smooth pump speed regulation and online flow monitoring under high viscosity conditions, further improving the adaptability and operational reliability of liquid-cooled thermal management systems under extreme conditions. Summary of the Invention

[0005] The purpose of this invention is to propose an adaptive temperature control method for lithium battery packs based on the Internet of Things in order to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] An IoT-based adaptive temperature control method for lithium battery packs includes:

[0008] The raw data of total charging and discharging current, pressure drop at the inlet and outlet of the liquid cooling pipeline network and total volumetric flow rate are obtained by the sensing nodes. After time stamp alignment and noise removal, a high-confidence state estimation array is output.

[0009] Based on the high-confidence state estimation array, the local hydrodynamic viscosity, dynamic polarization and ohmic combined series internal resistance are calculated. A fluid momentum dimensionality reduction solution framework including the slip layer and the high-viscosity cold core region is established, and the unknown cold core physical radius and radial velocity gradient are iteratively reconstructed.

[0010] A local dissipation rate density function is established based on the local fluid dynamic viscosity, the unknown cold core physical radius, and the radial velocity gradient. The viscous dissipation heat disturbance power generated by fluid shear friction is calculated by spatial volume integration.

[0011] An energy balance algebraic loop is constructed by combining the viscous dissipation heat disturbance power, dynamic polarization, ohmic series resistance, and the maximum permissible transient limit temperature rise rate threshold. The maximum permissible physical limit current command is derived in reverse and non-negative protection cutoff is executed.

[0012] The maximum permissible physical limit current command is mapped to the mechanical torque safety upper limit command for current limiting execution, and the maximum permissible angular acceleration is calculated based on the maximum transient shock wave pressure increment physical threshold to smoothly control the electronic water pump.

[0013] Preferably, before acquiring the original data of total charging and discharging current, inlet and outlet pressure drop of the liquid cooling pipeline network, and total volumetric flow rate, a forced silent calibration mechanism is performed. Specifically, the operation is as follows:

[0014] During the silent time window before the liquid-cooled main water pump receives the start command, the microprocessor forcibly reads the static residual value sequence of the differential pressure sensor to calculate the mean value, and writes the mean value into the non-volatile memory as the zero-point offset compensation amount. In the subsequent dynamic operation, all the original pressure drop data collected are unconditionally subtracted from the zero-point offset compensation amount to restore the true flow pressure drop.

[0015] Preferably, the timestamp alignment and noise stripping process employs a first-order inertial adaptive anti-aliasing filtering algorithm with a follow-up smoothing coefficient, which is calculated in real time based on the following exponential decay physical model: ;

[0016] in, This indicates that the current discrete time step generates the dynamic smoothing coefficient; This indicates the raw observation value containing environmental noise captured at the current moment; This indicates the historical value of the state estimate generated at the previous discrete time step; This represents the system configuration forgetting factor scalar; This indicates that the preset physical channel has a calibrated reference noise variance;

[0017] Within each calculation cycle, by evaluating the absolute value of the residual between the original observation value containing environmental noise captured at the current moment and the historical value of the state estimate generated at the previous discrete moment, a rigid barrier is established in the digital domain to block the pollution of the control system by glitch signals. Conversely, when the data variance returns to the range of stable physical variation, the follow-up smoothing coefficient is adaptively reduced and the weight of the current fresh observation data is increased again.

[0018] Preferably, the establishment of the fluid momentum dimensionality reduction solution architecture, which includes the slip layer and the high-viscosity cold core region, depends on the following discrete space partitioning rule:

[0019] In order to adapt to the limited computing power of the vehicle edge controller and avoid the watchdog timer of the control system being frequently reset and restarted due to computing power overload, the microprocessor forcibly discretizes the fluid space inside the pipe into a high-viscosity cold core region that extends from the center to the unknown cold core physical radius and a high-viscosity slip layer region that extends close to the pipe wall to the unknown cold core physical radius.

[0020] The fluid velocity gradient at any radial position within the high-viscosity cold core region is forcibly set to be absolutely zero. At the same time, the flow velocity in the high-viscosity slip layer region is set to decrease sharply from the translational velocity at the cold core interface to the absolute zero velocity at the pipe wall, satisfying the fluid dynamics no-slip physical boundary condition.

[0021] Preferably, the method further includes:

[0022] The microprocessor calls the internal Newton-Raphson iterative optimization algorithm to perform mass conservation residual convergence calculation. The microprocessor assigns an initial safe guess value close to the physical inner radius of the pipe to the unknown cold core physical radius and iterates. It uses the gradient of the partial derivative of the residual with respect to the unknown cold core physical radius to quickly adjust the next iteration value and sets a maximum number of iterations limit. If the residuals on both sides of the equation fail to converge to the allowable small error band within the specified computing cycle, the microprocessor triggers the exception handling logic and temporarily freezes the current state update to continue using the safe physical size of the previous cycle.

[0023] Preferably, the calculation of the viscous dissipation heat disturbance power generated by fluid shear friction needs to be initiated under the condition of satisfying the nonlinear order of magnitude reversal triggering discrimination matrix. The specific initiation condition is as follows:

[0024] Only when the system detects that the ambient reference temperature is lower than the threshold for abrupt change in the dynamic viscosity of the extremely cold fluid, and simultaneously detects that the real-time discharge rate of the battery is higher than the threshold for high-rate heat generation of the cell, forming a nonlinear intersection zone, can the enormous electrochemical heat only heat the extremely thin slip layer fluid while the central high-viscosity cold core remains as solid as a rock. This causes the liquid-cooled water pump to continuously output mechanical work, which is directly converted into internal heat energy of the system. Only then does the microprocessor activate the local dissipation rate density function to perform deep integration calculations to extract the values ​​of the parasitic heat source.

[0025] Preferably, the reverse derivation of the maximum permissible physical limit current command relies on the following energy boundary derivation formula:

[0026] ;

[0027] in, This indicates the maximum permissible physical limit current command; This indicates that the composite node contains an equivalent combined thermal mass; This indicates that the composite node material exhibits an equivalent overall specific heat capacity; This represents the threshold for the maximum permissible transient limiting heating rate; This represents the power of viscous dissipative heat generation disturbance; This represents an estimated value of the actual effective surface convection heat dissipation power. This represents the series resistance of dynamic polarization and Ohm's combined internal resistance;

[0028] By using the above energy boundary derivation formula, before any physical temperature sensor detects a surge in the actual temperature and before the battery's massive thermal inertia hysteresis takes effect, the inevitable trend of thermodynamic deterioration can be predicted in advance, and direct cross-domain intervention can be made at the underlying electrical safety boundary of the battery management system.

[0029] Preferably, the specific operation of performing non-negative protection cutoff is as follows:

[0030] Before executing the square root machine instruction, the derivation formula for the energy boundary is performed. When polarity verification is performed, if the high-viscosity laminar flow state inside the tube deteriorates rapidly, causing the viscous dissipation heat generation disturbance power to increase explosively, making the molecule part easily dragged down below zero by a huge subtraction, once it is determined that the numerator term inside the square root is less than or equal to zero, the square root operation is bypassed directly, and the maximum permissible physical limit current command is forcibly assigned to an absolute zero ampere value, and all Joule heating current inside the cell is immediately and unconditionally cut off.

[0031] Preferably, the calculation of the maximum permissible angular acceleration based on the physical threshold of the maximum transient shock wave pressure increment depends on the following physical constraint formula derived from momentum-force analysis:

[0032] ;

[0033] in, This indicates the maximum permissible angular acceleration that the electric water pump rotor is allowed to perform; This indicates the physical threshold value at which the piping component can withstand the maximum transient shock wave pressure increment. This indicates that the key flow channel cross-sectional area is required to generate the water jet effect; Represents the actual fluid density; This indicates the length of the pipe section with distinctive physical characteristics that are highly susceptible to water column impact. This represents the theoretical flow rate, rotational speed, and volumetric mapping constant of the liquid-cooled water pump.

[0034] Within each control step, the physical constraint formula is calculated and integrated to generate a nonlinear speed ramp control curve, which guides the massive fluid mass to accelerate smoothly and ensures the system safely crosses the hydrodynamic physical impact boundary caused by rheological abrupt changes, preventing the internal ceramic mechanical seal of the electronic water pump from being crushed and failing due to impact.

[0035] Preferably, the smooth control electronic water pump process includes a low-frequency micro-pulse flaw detection and follow-up mode. The water pump drive module sends a preset frequency low duty cycle electrical signal, causing the liquid-cooled pump impeller to generate tiny periodic pressure fluctuations at extremely low speeds, insufficient to cause any destructive viscous shear friction heat generation. The transient fluid impedance of the pipeline network is continuously detected and calculated online. At the same time, the microprocessor determines whether the flow thawing release condition is met based on the following composite Boolean logic discriminant:

[0036] ;

[0037] in, Indicates the trigger flag for the release of the flow thawing condition; Indicates the unknown physical radius of the cold nucleus; Indicates the physical inner radius of the pipe; This indicates that a minimum safe cold core ratio boundary has been pre-defined; This represents the power of viscous dissipative heat generation disturbance; This represents the tolerance reference power dissipation noise baseline.

[0038] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0039] 1. This invention constructs a fluid momentum dimensionality reduction solution architecture that includes a slip layer and a high-viscosity cold core region. This architecture can accurately analyze the morphology of the cold core and the viscous dissipation parasitic heat source inside the pipe under extremely cold conditions, which to some extent breaks through the feedback hysteresis limitation of traditional temperature sensors that rely on heat conduction. Based on the energy balance algebraic loop of fluid and electrochemical coupling, the current safety boundary can be reversed before the temperature rise trend appears, realizing the mode transformation from passive temperature control to active thermal safety constraint, which helps to improve the thermal safety margin in high-rate charge and discharge scenarios.

[0040] 2. This invention achieves smooth speed regulation of the electronic water pump through a nonlinear speed ramp curve, which can suppress water column impact and pipeline pressure shock during the acceleration process of high viscosity fluid, reduce the risk of damage to precision components such as the water pump mechanical seal, and help extend the overall service life of the liquid cooling system; combined with a low-frequency micro-pulse flaw detection mode, it can monitor the flow thawing process online and adaptively switch control strategies to achieve flow adaptive matching under all operating conditions from extreme cold to normal temperature. Attached Figure Description

[0041] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0042] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0043] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0044] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0045] Example 1

[0046] Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0047] In this embodiment, it includes:

[0048] Step 1: Multi-source data acquisition under extreme operating conditions

[0049] Before implementing the thermo-electric coordinated control strategy based on internal viscous dissipation heat generation feedforward compensation, a physical information sensing network that can maintain high confidence and high accuracy under extreme cold environmental temperatures, severe electromagnetic interference caused by complex vehicle driving conditions, and large hysteresis characteristics of fluid dynamics is constructed.

[0050] By using IoT sensor nodes widely distributed inside and outside the battery pack, the electrical, thermodynamic, and hydrodynamic parameters of the entire lithium battery pack and its thermal management liquid cooling network under complex nonlinear conditions are acquired asynchronously. These multi-source heterogeneous raw data need to be physically aligned in time and space and noise stripped within the edge computing nodes to provide a solid, distortion-free data foundation for subsequent calculations and microfluidic reconstruction.

[0051] At the level of specific IoT hardware topology deployment, the sensing layer needs to have an extremely high density of physical nodes distributed along the complex liquid cooling pipeline inside the lithium battery pack. For sensing temperature field data, the system deploys high-precision platinum resistance temperature sensors (e.g., PT1000 type, to obtain a wider linear response range) with extremely low temperature drift characteristics at the main fluid inlet of the liquid cooling pipeline, the branch manifold nodes of each parallel cooling branch, and the main fluid outlet.

[0052] Meanwhile, an extremely thin, flexible film thermocouple matrix is ​​laid in the heat-intensive areas inside the lithium battery module (specifically, the welding area of ​​the top tabs of the cell and the center of the large surface of the cell sidewall). To ensure that all surface-mount temperature probes can accurately and without delay reflect the transient thermal conduction state of the battery skin and tube wall in extremely cold environments (e.g., minus 30 degrees Celsius or even lower), the bottom of all contact-type sensing probes must be coated with a special carbon nanotube thermal grease with ultra-high thermal conductivity that will not undergo glass transition at extremely low temperatures.

[0053] On the outside of the back of the sensor, a layer of aerogel insulation pad with low thermal conductivity must also be covered. This is to effectively block the measurement heat loss disturbance caused by the convection of the surrounding air, and to ensure that every temperature fluctuation captured by the sensor comes purely from the heat transfer inside the cell or fluid.

[0054] To collect key fluid dynamic parameters, miniature piezoresistive differential pressure sensors and high-frequency dual-channel ultrasonic flow meters were deployed at the high-pressure output end of the liquid-cooled main pump and at both ends of the longest and most flow-resistance critical liquid-cooled branch in the system.

[0055] Under extremely cold conditions, the kinematic viscosity of commonly used cooling media (such as aqueous ethylene glycol solutions) can increase dramatically by two to three orders of magnitude, or even become semi-gel-like. Traditional mechanical impeller flow meters may experience mechanical stoppage, blade damage, or severe step loss due to the enormous fluid shear resistance, thus uploading false zero flow velocity data.

[0056] A non-contact ultrasonic time-of-flight method is employed to capture transient changes in minute and viscous flow velocities. Furthermore, a high-precision miniature piezoresistive differential pressure sensor is used to measure the macroscopic pressure drop distribution as fluid flows through a specific pipeline section in real time. This pressure drop data provides the most direct external physical evidence for subsequently determining whether a "high-viscosity cold core blockage" phenomenon exists within the pipeline.

[0057] For acquiring electrical status data, the system uses high-frequency Hall current sensors deployed on the copper busbar assembly and voltage sampling harnesses distributed on the terminals of each individual cell to collect real-time and synchronous total charging and discharging current of the lithium battery pack and terminal voltage data of each individual cell. In order to capture transient Joule heating spikes caused by rapid acceleration, energy recovery braking, or high-frequency pulse fast charging, the sampling frequency of electrical data acquisition is set at the microsecond level, which is significantly higher than the millisecond-level thermodynamic and fluid dynamic data acquisition frequency.

[0058] After acquiring the massive amounts of heterogeneous data from multiple sources, it is necessary to perform timestamp alignment and physical filtering of the asynchronous data within edge computing nodes with concentrated computing power (such as battery thermal management domain controllers with multi-core microprocessor architectures). In real IoT communication networks (such as automotive CAN-FD or industrial Ethernet), due to differences in the analog-to-digital conversion time of each sensor node, queuing delays caused by bus protocol arbitration, and strong electromagnetic interference generated by high-power inverters in the field, the timestamps of data packets arriving at the microprocessor's receive buffer exhibit a large amount of discrete random delay and out-of-order phenomena.

[0059] To prevent mathematical divergence in the control loop due to sudden jumps in observed data, a first-order inertial adaptive anti-aliasing filtering algorithm based on a physical state-space model is introduced.

[0060] Let the discrete time step be The original physical observation sequence collected for a specific location Establish discrete state equations that incorporate thermal inertia or fluid inertia characteristics.

[0061] Define the true smooth physical state Considering that physical changes in temperature or pressure are constrained by the heat capacity and thermal resistance of materials in a very short time, their state transition characteristics are constrained by physical laws.

[0062] The filtering algorithm includes a core anti-aliasing digital filter state update formula, the expression of which is:

[0063] ;

[0064] in, This represents the state estimate after physical evolution smoothing at the current moment; This represents the estimated historical value of the state at the previous discrete moment; This represents the raw observation value, including ambient noise, uploaded by the sensor bus at the current moment.

[0065] parameter It is a follow-up smoothing coefficient, and its physical value range is limited to between 0 and 1.

[0066] In each computation cycle, the microprocessor calculates the current filter coefficients in real time based on the following exponential decay physical model by evaluating the absolute value of the residual between the current observation and the historical prediction:

[0067] ;

[0068] in:

[0069] Defined as a scalar forgetting factor of the system, it is used to adjust the rate at which the system’s trust in historical data decays. Its value is obtained from the factory calibration based on the sensor’s own background noise bandwidth and is a dimensionless constant.

[0070] Defined as the calibration reference noise variance, it represents the standard deviation tolerance of the white noise inherent in the sensor signal under conditions of no strong external interference for a specific physical channel. Its unit is the same as that of the observed variable. Maintaining consistency. By executing this explicit mathematical formula, the microprocessor can achieve intelligent allocation at the physical response level in real time: when a strong, instantaneous electromagnetic interference causes a data spike, the residual is much larger than the baseline noise variance, and the exponential term rapidly approaches zero, thus ensuring... It rapidly approaches 1. This is equivalent to building a rigid barrier in the digital domain, effectively preventing glitch signals from contaminating the control system;

[0071] Conversely, when the data variance reverts to the stationary physical range, the exponential term increases. The value of is adaptively reduced, and the system re-increases the weight of the current fresh observation data to ensure a highly agile response to transient high-rate discharge conditions.

[0072] To address the zero-point drift issue that may occur in differential pressure sensors under extremely cold conditions due to changes in the shrinkage stress of the metal diaphragm, a forced silent calibration mechanism was designed.

[0073] During the quiet time window between each power-on wake-up of the entire vehicle system and before the liquid-cooled main water pump receives a start command, the microprocessor forcibly executes a hydrostatic reference calibration program. The system reads the static residual value sequence of the differential pressure sensors at each node at this time, takes the average value, and writes it into the controller's non-volatile memory as the dynamic zero-point offset compensation amount under the current temperature conditions.

[0074] During the dynamic operation of the liquid-cooled pump after its official start-up, all the raw real-time pressure drop data collected must be unconditionally subtracted from the zero-point offset compensation amount in order to accurately reconstruct the real flow pressure drop generated by the fluid overcoming viscous resistance.

[0075] Through multi-dimensional sensor redundancy deployment, time-scale alignment, explicit nonlinear physical filtering formula calculation, and zero-point drift compensation processing, this step outputs a state array with extremely high confidence, providing initial boundary conditions for solving subsequent complex partial differential equations.

[0076] Step 2: Reconstruction of Stratified Flow Characteristics within the Pipe

[0077] After obtaining a high-confidence multi-source data array that has been cleaned and physically verified, this step delves into the interior of the liquid cooling pipeline to solve a hidden microscopic physical problem that has long plagued traditional thermal management control. In extremely low-temperature conditions, when the battery pack instantly switches to a high-rate charge / discharge state, the huge heat flux density generated inside the battery cells will instantly break through the thermal grease and rapidly heat the thin layer of cooling medium tightly attached to the liquid cooling pipe wall.

[0078] Because the cooling medium in the center of the pipe is still at extremely low temperature and extremely high viscosity, strong non-isothermal rheological stratification will occur radially inside the pipe. The fluid at the pipe wall will suddenly decrease in viscosity after being heated, forming a highly fluid "slip layer";

[0079] The center of the pipeline remains a hard "cold core" occupied by a high-viscosity, even gel-like, medium. The core task of this step is to use limited external macroscopic measurement data, combined with simplified fluid dynamics control equations, to reverse and accurately reconstruct the geometric spatial characteristics and velocity profile distribution of this microscopic stratified flow state within the pipe in the edge controller.

[0080] First, it is necessary to establish a strongly nonlinear algebraic relationship between the dynamic viscosity of the cooling medium and the local temperature, and then adopt the Arrhenius modified physical equation, which better reflects the change of intermolecular micro-forces with temperature.

[0081] The microprocessor can calculate the local hydrodynamic viscosity at a specific physical node in real time and continuously by calling this equation.

[0082] The mathematical expression for the core viscosity physical equation is:

[0083] ;

[0084] in,

[0085] Indicates the local absolute temperature of the cooling medium The dynamic viscosity value at this temperature is expressed in Pascal-second (Pa·s). );

[0086] This represents the local absolute temperature of a fluid element at the current coordinate system, expressed in Kelvin (K). );

[0087] parameter , as well as These are the characteristic constants that describe the unique rheological properties of this type of cooling medium. The process of obtaining these three characteristic constants is as follows:

[0088] Before the system leaves the factory, it is necessary to conduct hundreds of rheological calibration tests in the laboratory stage using a high-precision rotary thermostat in an extremely wide low-temperature gradient tank. Then, the nonlinear curve is fitted using the least squares method, and finally permanently burned and solidified in the read-only memory area of ​​the vehicle controller.

[0089] This equation can describe the exponential explosive growth of dynamic viscosity of a fluid as it approaches the freezing point or enters a severely supercooled state with extreme smoothness and precision.

[0090] The microprocessor must calculate the transient thermal boundary conditions on the inner surface of the tube wall online. Based on the electrical terminal data collected in microseconds in step one, the system calculates the dynamic heat generation power of the current battery cell according to the equivalent circuit model.

[0091] Total heat generation The composition encompasses the reversible entropy change heat of the electrochemical reaction, as well as the irreversible Joule polarization heat caused by internal ion migration resistance. In simplified engineering calculations, the main heat generation can be expressed by the formula... Approximately obtained.

[0092] in The total charging and discharging current is collected synchronously in real time; This refers to the dynamic polarization and ohmic combined series internal resistance of the current battery cell.

[0093] Assuming that the dynamic heat generation power is transferred to the inner wall of the fluid pipe with almost no heat loss through the heat conduction network of special thermally conductive silicone grease and aluminum liquid cooling plate, and considering the equivalent contact thermal conductivity area between the liquid cooling plate and the pipe wall... Given the known thermal conductivity of the aluminum alloy material used in the tube wall, the microprocessor uses the one-dimensional transient Fourier thermal conductivity law to calculate and reconstruct the instantaneous contact temperature of the inner surface of the tube wall. .

[0094] The system enters the mathematical modeling stage of the microscopic flow pattern distribution characteristics along the pipe path. For ease of calculation, it is assumed that the internal microstructure of the cooling channel is equivalent to a straight cylindrical pipe with a uniform cross-section. A polar coordinate system is established inside the pipe, and radial coordinates are defined. (Set the center axis of the pipeline) Inner surface of the pipe wall ,in (where the pipe's physical inner radius is the radius of the pipe).

[0095] In the initial stage of extremely cold start-up, due to the objective existence of the large volume of the highly viscous cold core, the effective flow cross-sectional area of ​​the fluid is much smaller than the physical geometric cross-sectional area of ​​the pipe. Therefore, the system introduces the simplified partial differential form of the Navier-Stokes fluid momentum equation for steady-state unidirectional laminar flow within a circular pipe.

[0096] In this model, the pressure drop driving force gradient that maintains fluid motion along the pipeline axis is defined as... ,in This refers to the actual pressure difference between the pipeline inlet and outlet obtained after zero-point drift cleaning in step one. The length of the characteristic physical pipe segment.

[0097] Due to the large temperature gradient distribution in the radial direction, the dynamic viscosity of the fluid... It is no longer a spatial constant, but rather varies with radial physical position. drastically changing functions In a rapidly flowing fluid thin layer (i.e., a slip layer), the shear internal friction stress experienced by adjacent fluid elements is... The dynamic equilibrium relationship between the external pressure drop gradient and the local velocity gradient is governed by the following fundamental fluid dynamics differential equation:

[0098] ;

[0099] In the above second-order partial differential equation This indicates the radial distance from the central axis of the pipe. The axial flow velocity of the fluid particles at that location.

[0100] For the aforementioned fluid control partial differential equations, this invention considers the extremely limited computing power and insufficient memory space of the on-board edge microprocessor, and the need to output results within millisecond-level control cycles. If the three-dimensional mesh partitioning in traditional computational fluid dynamics is directly used for numerical solution, the watchdog timer of the control system will inevitably be frequently reset and restarted due to computing power overload.

[0101] Complex global fluid equations were mapped to edge hardware using computational power reduction, resulting in a "two-region non-isothermal approximate dimensionality reduction solution architecture" that can be efficiently executed by microprocessors. This architecture endows pure mathematical deduction with strong technical attributes for solving practical engineering computing power bottlenecks.

[0102] The dimensionality reduction solution architecture forcibly divides the continuous fluid space inside the pipe into two independent regions with distinct rheological characteristics:

[0103] The first region is defined as the high-viscosity cold core region centered on the circle, with its radial coverage extending from the center. Extending all the way to ( This refers to the unknown cold nucleus physical radius that needs to be calculated and reconstructed. In the logic processing of the edge controller, the extremely high viscosity of the fluid in this region due to the extreme low temperature is equivalently abstracted as a "plunger" similar to a rigid solid that has completely lost its interlayer shear slip capability.

[0104] This means that the microprocessor directly assigns the same velocity value to all particles within that region, directly setting the internal velocity gradient boundary conditions. This completely eliminates the need for complex partial differential equation integration within this vast region, freeing up enormous computing resources.

[0105] The second region is defined as the high-speed slip layer zone closely adhering to the pipe wall, and its radial range is defined from... Until the pipe wall ( (This refers to the physical inner radius of the pipe). Within this thin-layer region, the microprocessor calculates the pipe wall temperature. The local viscosity decrease curve was calculated based on the distribution of extremely low temperatures at the cold core interface. The flow velocity in this region was assumed to decrease sharply from the translational velocity at the cold core interface to absolute zero velocity at the pipe wall, satisfying the no-slip physical boundary conditions of fluid mechanics.

[0106] In order to ultimately solve for that crucial but unmeasurable unknown geometric variable in reverse engineering, the cold core radius... The microprocessor simultaneously established the total mass flow conservation constraint equations.

[0107] The total volumetric flow rate of the current pipeline network collected by the ultrasonic flow meter in step one. In physics, it must be equal to the sum of the flow rate in the central cold core plunger region and the integral flow rate in the edge slip layer region.

[0108] To further reduce computational overhead, the microprocessor simplifies the dynamic viscosity of the fluid within the slip layer to a spatially averaged effective viscosity. This value is obtained by arithmetic weighting of the extremely low viscosity corresponding to the highest temperature of the tube wall and the high viscosity corresponding to the extremely low temperature of the cold core boundary.

[0109] The microprocessor integrates the velocity profile integral mathematical laws of laminar flow in a circular tube. Within each control interruption cycle, it invokes an internally optimized Newton-Raphson iterative optimization algorithm to solve the following nonlinear implicit algebraic mass conservation equations:

[0110] ;

[0111] Where parameters Defined as the uniform translational axial velocity of the central cold core entity region.

[0112] In the microprocessor's code execution logic, an initial safe guess value is first assigned to the unknown physical radius of the cold core. (That is, the initial state is assumed to be extremely blocked to ensure that the iteration process does not cause mathematical singularities due to excessively small guesses), and then the residuals on both sides of the mass conservation equation are calculated. The system sets the absolute convergence tolerance for iteration to be... (This value is less than the background measurement noise of the ultrasonic flow meter and will not introduce additional systematic errors.)

[0113] By utilizing residual pairs The directional gradient of the partial derivative is used to quickly adjust the value for the next iteration until the residual falls within the convergence tolerance band.

[0114] The system has a maximum limit on the number of iterations. If the residual fails to converge to the allowable small error band within the specified computing cycle, the microprocessor will trigger the exception handling logic, temporarily freeze the current state update, and continue to use the safe physical size of the previous cycle.

[0115] Through this efficient dimensionality reduction iteration, the edge system can accurately and continuously reconstruct the spatial geometry of the "high-viscosity cold core" inside the tube and the microscopic physical thickness of the slip layer in the current transient state within milliseconds.

[0116] This reconstructed computing mechanism, which delves into the microscopic mechanisms of fluid media, effectively avoids the drawbacks of traditional thermal management systems that rely on a single external surface temperature sensor for blind PID regulation of pump speed, leading to control failures.

[0117] When the edge system undergoes iterative calculations, the proportion of the huge cold core radius is obtained (for example, by determining...). When the critical state is reached, the physical reality clearly shows that the entire thermal management system has irreversibly fallen into a severe "strong non-isothermal slip stratification" catastrophic boundary.

[0118] At this point, most of the cooling medium is frozen in the center to form a plunger, losing its physical regulation ability for heat absorption and exchange. Only a negligible thin layer of fluid slides ineffectively and rapidly off the pipe wall surface.

[0119] Step 3: Calculation of the viscous dissipative heat generation boundary

[0120] After successfully reconstructing the spatial physical dimensions of the "high-viscosity cold core" inside the cooling pipe and the geometric thickness of the edge slip layer through the edge hardware computing power dimensionality reduction architecture in step two, this step aims to accurately quantify and remove a fatal physical quantity that has long been ignored as a minor error in traditional battery thermal management systems: the viscous dissipation heat power generated by the internal friction of the fluid inside the pipe.

[0121] Under normal ambient temperatures, the dynamic viscosity of fluids is low, and most of the mechanical energy applied to the fluid by the pump is converted into effective kinetic energy to propel the coolant through the pipe network and carry away heat; the frictional heat generation is negligible. Under extreme conditions of extreme cold and high-rate battery discharge, a nonlinear order-of-magnitude flip-triggered discrimination matrix is ​​introduced within the microprocessor.

[0122] This discrimination matrix clearly defines the physical triggering boundary conditions for the transition of viscous dissipative heat generation from a "negligible amount" to a "dominant destructive heat source." The system monitors the ambient reference temperature in real time during each control cycle. and the real-time discharge rate of the battery. The microprocessor internally sets a threshold for abrupt changes in viscosity at extremely cold temperatures (e.g., At this temperature, the viscosity of a specific alcohol-water solution increases exponentially (and the cell's high-rate heat generation threshold (e.g., ...)). ).

[0123] The system only activates its deep computing engine for this step when the system state variables simultaneously break through the nonlinear intersection region formed by the intersection of these two multidimensional thresholds. Under this extreme intersection condition, the enormous electrochemical heat can only heat the extremely thin slip layer fluid, while the central highly viscous cold core remains as solid as a rock. At this point, a terrifying velocity gradient is generated inside the slip layer. The enormous mechanical work continuously output by the liquid-cooled water pump cannot effectively drive the cold core; instead, due to the intense shearing and internal friction between fluid molecules within the slip layer, it is directly and efficiently converted into internal thermal energy of the system.

[0124] This parasitic heat, generated by the irreversible transformation of mechanical energy, adheres closely to the battery wall, essentially creating a new high-power heat source unexpectedly within the battery system, which already requires urgent heat dissipation.

[0125] To calculate this internal nonlinear thermal disturbance, the system microprocessor invokes the energy dissipation theory from fluid mechanics to establish a mathematical-physical model of the dissipation function based on the high coupling between the local radial velocity gradient and the real-time dynamic viscosity. In the cylindrical coordinate system of the incompressible viscous fluid, since the flow inside the pipe has been physically defined in step two as a unidirectional laminar flow with strong radial stratification, the axial translational velocity of the fluid only changes drastically with the radial physical coordinate.

[0126] Therefore, the local volume dissipation rate density (i.e., the heat generation power per unit physical volume per unit time) describing the irreversible conversion of mechanical kinetic energy into heat energy by a fluid due to viscous shear friction can be highly simplified to a single-variable nonlinear function of the radial velocity gradient.

[0127] Within each millisecond-level extremely short control sampling period, the system utilizes the microprocessor's computing power to execute the following formula for calculating the local volumetric heat dissipation rate density:

[0128] ;

[0129] in,

[0130] This indicates the radial distance from the geometric axis of the pipe center. The local volumetric heat dissipation rate of a fluid particle at a given location, measured in watts per cubic meter (W / m³). );

[0131] This represents the local hydrodynamic viscosity obtained in step two using the modified Arrhenius equation, expressed in Pascal-seconds (Pa·s). );

[0132] Indicates radial coordinates The local fluid axial velocity gradient at a point, i.e., the relative shear deformation rate between adjacent fluid microlayers, is expressed in reciprocals of seconds. ).

[0133] Because the velocity gradient term in this physical formula has a square operation property, this means that in the extremely thin slip layer close to the outside of the cold core interface, as the flow velocity drops rapidly from the translational high pull to the zero velocity of the tube wall, the local velocity gradient increases steeply, and the viscous dissipation heat density will exhibit an extreme nonlinear amplification effect.

[0134] The microprocessor must extend this distributed microscopic local dissipation rate integral over the entire macroscopically set physical controlled pipe section volume to obtain the total macroscopic viscous dissipative heat power.

[0135] Considering that the fluid inside the central high-viscosity cold core region acts like a rigid solid plunger, undergoing translational motion without relative displacement under pressure difference, the relative velocity gradient between any adjacent particles within it is absolutely zero (i.e., ).

[0136] Therefore, the internal heat dissipation of the cold core region itself is physically defined as zero. The viscous frictional heat dissipation within the entire thermodynamic system is completely and extremely densely compressed within the radius of the cold core boundary. Extending to the physical inner radius of the pipe Within the annular slip layer space region.

[0137] The system along a length of ( For a pipe segment (where the characteristic physical pipe segment length is ), a cylindrical spatial volume integral is performed on the annular slip layer region. The specific spatial integral calculation expansion formula is as follows:

[0138] ;

[0139] In the above macroscopic power formula:

[0140] This represents the viscous dissipation and heat generation disturbance power generated by the intense shearing and stratification of the fluid within the entire controlled pipe section, measured in watts. );

[0141] The length of the characteristic physical pipe section is expressed in meters (m). );

[0142] For radial space variables in calculus calculations;

[0143] This refers to the axial spatial variable used in calculus calculations.

[0144] In the actual edge numerical calculation process of the in-vehicle system, continuous calculus operations must be discretized for execution by the digital chip. The microprocessor will allocate a dedicated array in memory to discretize the extremely thin radial thickness of the slip layer (i.e., The difference space is forcibly divided into tens or even hundreds of micrometer-level computational grid nodes. Because this is a strongly nonlinear multiphysics coupled system with strong internal positive feedback characteristics, the minute heat generated by intense fluid friction will cause a slight increase in the temperature of local micro-clusters. This temperature increase will immediately lead to a decrease in local fluid viscosity through the Arrhenius equation. The decrease in viscosity, in turn, changes the fluid's ability to resist shear, thereby reshaping the velocity gradient, which in turn changes the heat generation.

[0145] If a simple single-step forward Euler calculation method is used, this positive feedback loop is very likely to cause a severe numerical explosion in the microprocessor's calculation results within a few cycles.

[0146] To ensure the robustness of the system's computing power, the system must employ a "prediction-correction" numerical algorithm with absolute numerical convergence stability to approximate the actual physical power dissipation.

[0147] The specific underlying algorithm execution logic design is as follows:

[0148] To accommodate the limited computing power of automotive microprocessors and ensure numerical rigidity stability, this method forcibly limits the radial thickness of the slip layer (i.e., ...) before performing the prediction-correction iteration. The difference is discretized into N equidistant micro-grids (N is dynamically allocated according to the real-time main frequency of the microprocessor: N is 10 when the main frequency is below 200MHz; N is 30 when the main frequency is above 400MHz). Simultaneously, a fixed physical propagation time step is set. This step size is much smaller than the fluid thermal diffusion time constant, which is sufficient to guarantee the absolute stability of the implicit difference scheme.

[0149] The convergence criterion is set as follows: the convergence values ​​calculated in two consecutive global iterations. The relative rate of change is less than (Right now If the number of iterations exceeds the maximum safe number of iterations (50 times) set internally by the microprocessor, the iteration will be forcibly terminated and the historical power value of the previous control cycle will be used to prevent computing power overload.

[0150] In the initial stage of the current discrete time step, assuming that the local temperature field of the fluid inside the pipe remains frozen and unchanged within a very short period, an initial radial dynamic viscosity distribution is directly calculated using the temperature array from the previous computing cycle. Substituting this distribution into the finite difference equation, the initial array of velocity gradients for each grid is obtained. Then, based on the discretized area product formula, a preliminary estimate of the dissipated heat generation power is made.

[0151] During the calibration phase, the microprocessor treats this initially estimated heat generation power as an additional virtual volume heat source and forcibly introduces it into the transient microscopic heat conduction partial differential equation of the local fluid micro-particle. It then calculates the new local temperature state that the fluid micro-particle may reach at the end of this tiny time step due to the absorption of its own frictional heat.

[0152] Using this new temperature array with temperature rise feedback, the viscosity evolution equation is called again to correct the local fluid viscosity matrix for the second time. Based on the extremely realistic pressure drop boundary parameters on both sides of the pipeline obtained in step one, the spatial distribution of the velocity gradient is re-corrected.

[0153] Finally, based on the set convergence residual weights, the system performs a weighted average filtering of the estimated power value and the corrected power value, and uses this result as the final viscous dissipative heat disturbance power output during the current control interruption cycle. .

[0154] This nonlinear deep computing mechanism, based on numerical analysis and microrheology, ensures that the system can capture endogenous parasitic heating phenomena that are completely ignored by traditional systems under complex icy physical boundaries.

[0155] When the microprocessor calculates the viscous dissipation heat disturbance power When the value reaches or even exceeds the effective heat that can be removed by the meager convective heat transfer within a pipe section of the same geometry, this reveals a physical truth:

[0156] If the high speed of the liquid-cooled main water pump is maintained by relying on traditional feedback logic, it will not only fail to cool down the hot battery pack, but will also accelerate the irreversible collapse of the entire electrochemical thermodynamic system because the huge mechanical work applied by the water pump to the high-viscosity fluid is fully converted into viscous frictional heat.

[0157] The calculated and quantified value of the hidden endogenous perturbation thermal power will be sent to the thermoelectric coupling controller in the next step without delay as a crucial feedforward suppression signal, triggering active physical intervention across disciplines.

[0158] Step 4: Thermoelectric Coupling Feedforward Compensation Control

[0159] Step three quantifies the existence of severe viscous dissipative heat generation disturbances within the system and clarifies that the current thermal management system is nearly paralyzed in terms of physical flow. Therefore, traditional feedback control strategies based on temperature error elimination (such as simply adjusting the pump duty cycle or proportional solenoid valve opening using a PID algorithm) become completely meaningless. This is because, with the "high-viscosity cold core" still blocking the pipes, any control command attempting to increase the cooling medium flow rate from the fluid dynamics execution end will proportionally and drastically intensify the shear friction between the fluid and the pipe wall, leading to a dramatic increase in internal heat generation.

[0160] A cross-physics control architecture with significant "non-causal time advance" was established, which directly uses the reconstructed fluid microscopic impedance and endogenous frictional heat as feedforward suppression prior signals in the time domain. Before any physical temperature sensor detects the actual temperature spike (i.e. before the battery's massive thermal inertia hysteresis takes effect), the system has already predicted the inevitable trend of thermodynamic deterioration on the time axis and directly intervened in the underlying electrical safety boundary of the battery management system across domains.

[0161] The core of this engineering step lies in the real-time construction of a feedforward algebraic loop within the microprocessor, where transient thermal and electrical energy are mutually constrained and energy is conserved. Within any microsecond-level control sampling period, the system establishes an instantaneous energy inflow / outflow balance prediction evolution model for the controlled physical nodes. This model specifies:

[0162] The rate of change of total net heat accumulation at a local node of the battery system must be equal to the sum of the power of all internal heat sources within that space (including electrochemical heat generation and fluid parasitic frictional heat), minus the total heat dissipation power lost to the outside. The highest principle of system control is to ensure, by pre-adjusting electrical parameters, that the temperature rise slope mapped by this rate of net heat accumulation will never reach the safety threshold that would cause physical melting of the battery separator.

[0163] The system establishes the following core mathematical formula for the thermo-electric coupling feedforward time boundary in the computing kernel of the edge controller, in order to reverse-engineer the absolute safe operating boundary of the electrical system under the current operating conditions:

[0164] ;

[0165] In the above energy balance formula for thermal safety:

[0166] This represents the equivalent total thermal mass of the controlled local physical nodes, expressed in kilograms (kg). This thermal mass includes not only the mass of the battery cell itself, but also a precise calculation of the equivalent weighted mass of the local aluminum liquid cooling plate, thermal grease, and even the internal stagnant fluid;

[0167] This represents the equivalent overall specific heat capacity of the composite node physical material, expressed in joules per kilogram (Kelvin). );

[0168] Equivalent comprehensive thermal mass It is not about the quality of a single cell, but rather a lumped-parameter thermal capacity model constructed for controlled physical nodes; the specific calculation method is as follows: ,in For cell quality, The equivalent mass of the liquid-cooled plate participating in heat exchange. The mass of coolant retained within the pipe section; and These are the thermal participation coefficients (ranging from 0 to 1) pre-calibrated based on the ratio of thermal conductivity to convective thermal resistance.

[0169] Similarly, equivalent overall specific heat capacity The mass-weighted average method was used to obtain the following: The above-mentioned cell physical properties ( ) is obtained in real time through interpolation using a two-dimensional temperature-SOC lookup table pre-stored in the controller's read-only memory; metal and fluid properties ( , Then, based on the current pipe wall temperature and inlet temperature, linear interpolation is performed using the physical property database.

[0170] The maximum permissible transient temperature rise rate threshold for the battery pack, pre-defined for the dead zone in vehicle safety engineering specifications, expressed in Kelvin per second (kF / s). The fundamental physical purpose of setting this rate threshold is to prevent irreversible decomposition or internal short circuits caused by membrane thermal shrinkage due to a rapid increase in temperature at the electrochemical level.

[0171] This step attempts to solve the equation in reverse to determine the maximum safe physical limit current that the battery can output under the current extremely harsh heat dissipation conditions, expressed in amperes (A). This is the final cross-domain control target variable generated in this step;

[0172] The dynamic polarization and ohmic combined series internal resistance of the battery under the current temperature and state of charge, obtained by looking up the equivalent circuit and parameters in steps one and two, are given in ohms. );

[0173] This refers to the viscous dissipation heat disturbance power calculated nonlinearly through microscopic numerical integration in step three, expressed in watts (W). );

[0174] This is a significantly reduced estimate of the actual effective surface convection heat dissipation power provided by the extremely thin slip layer fluid on the outer side of the pipe wall, based solely on the system's assessment that the current pipeline is in a severely "cold core blockage" abnormal stratification state. The unit is watts. This value is not based on an estimate of the total circulation flow rate, but rather on a dynamic inversion based on the actual mass flow rate of the slip layer reconstructed in step two. Its specific calculation formula is: ;

[0175] in, The effective convective mass flow rate (not the total pumping flow rate of the water pump) is obtained by integrating the velocity profile of the slip layer region in step two. The instantaneous contact temperature of the inner surface of the tube wall is reconstructed in step two using the one-dimensional transient Fourier heat conduction law. The average temperature of fluid mixing is measured by a platinum resistance temperature sensor placed at the pipe outlet. When the cold core percentage... At its maximum, Approaching zero It then naturally decays to near zero, accurately reflecting the physical reality of heat dissipation failure.

[0176] By performing algebraic transformations on the core energy conservation formula in a microprocessor, the system can calculate the dynamic boundary evolution formula for cross-domain electrical current limiting at a microsecond-level response speed under extremely harsh thermodynamic boundary conditions, even where the fluid is nearly frozen.

[0177] ;

[0178] From the physical structure of the above equation, it can be clearly seen that:

[0179] Due to viscous dissipation, heat disturbance power is generated. It directly participates in the calculation as the "subtrahend" in the numerator term of the square root.

[0180] When the high-viscosity laminar flow state inside the pipe deteriorates rapidly, leading to The explosive increase may be due to limitations in effective convection cooling capacity caused by cold core blockage. When the value is extremely close to zero, the numerator of the formula is very easily dragged down below zero by the huge subtrahend, resulting in a negative result.

[0181] In response to this situation, a conditional decision branch is forcibly implanted within the microprocessor:

[0182] Before executing the time-consuming square root machine instruction, the numerator term is first processed. Perform polarity verification.

[0183] Once determined The microprocessor will directly bypass the square root operation and force the target limit current. The register value is assigned to an absolute 0 amperes.

[0184] The physical significance of this non-negative truncation logic is profound; it declares to the entire vehicle:

[0185] At this point, the frictional heat generated within the system has exceeded the limit allowed by the battery materials' safe heating rate. The thermal management system not only loses all heat dissipation capabilities but also becomes a fatal heat source. At this critical moment, it is imperative to unconditionally and immediately physically cut off all Joule heating currents within the cell to prevent thermal runaway. This combination of anti-collapse protection algorithms and physical foresight endows the control system with extremely high survivability.

[0186] To prevent undamped calculations by the microprocessor The instruction oscillates at high frequency between several adjacent control sampling cycles. The system must insert a nonlinear physical rate limiter and an inertial smoothing filter algorithm module with asymmetric characteristics into the final output channel of the target instruction generation.

[0187] The microcontroller's internal memory sets a hard physical constraint parameter for the maximum allowable rate of change of current. This limiter has asymmetric servo characteristics: when the latest calculated value is... When the value is significantly lower than the command in the previous calculation cycle (indicating that the pipeline micro-impedance is deteriorating and the thermodynamic environment is critical), the rate limiter will adaptively relax the downward rate of change constraint, allowing the derating current limiting command to take effect at the lowest point with an extremely fast response speed, so as to ensure emergency safety avoidance in dangerous situations.

[0188] Conversely, when viscous dissipation generates heat disturbance power The disturbance gradually decreases due to fluid thawing, making the calculated When the current begins to rise (indicating that the cold core inside the tube is melting and the effective heat dissipation capacity is in the recovery period), the rate limiter will switch strategies to extremely strictly limit the upward slope of the current command, forcing the entire electric drive system to slowly and smoothly restore high power output within a safe low temperature boundary, and resolutely preventing a secondary rebound of electrochemical heat.

[0189] The high-precision dynamic derating command data packet, after undergoing nonlinear asymmetric smoothing processing, is re-disassembled and re-encoded by the thermal management controller. The system, through a highly interference-resistant onboard high-speed differential communication network, assigns this command the highest preemption priority and broadcasts it across the entire network to the vehicle controller, chassis domain controller, and motor drive controller. This global collaborative strategy, based on underlying fluid dynamics hazard calculations, directly extracting parasitic heating signals, and feeding forward to higher-level electrical scheduling across physical boundaries, breaks the rigid situation of traditional vehicle subsystems operating independently and data silos.

[0190] Step 5: Reduced quota implementation and pipeline network follow-up

[0191] In step four, the microprocessor has successfully calculated and generated dynamic current-limiting instructions (i.e., the maximum permissible physical limit current) for suppressing endogenous heat sources based on time-advanced noncausal physical boundaries. This leads to the deeper stage of cross-subsystem collaborative execution, as well as the dynamic tracking and non-destructive recovery of the underlying thermodynamic and hydrodynamic states.

[0192] This step addresses the physical smoothness of control commands transitioning from the control domain to the execution domain, and how to safely and mechanically guide a pipeline filled with extremely high-viscosity fluid out of the abnormal non-isothermal stratified flow state of "cold core blockage" after the internal electrochemical heat source is forcibly suppressed, ultimately seamlessly transitioning to a normal, efficient convective heat transfer state. This entire process encompasses energy conservation torque servo mapping in the electrical dimension, and anti-surge and anti-water hammer / explosion-proof speed control in the fluid dynamics dimension.

[0193] At the physical execution level of the electrical domain derating command, when the vehicle controller and the distributed motor drive controller receive the highest priority broadcast from the thermal management microprocessor via the onboard high-speed differential communication bus... After receiving a safety boundary instruction, a forceful, hard disconnection strategy must never be employed. Considering the vehicle may be engaged in high-speed cruising, rapid acceleration for overtaking, or high-frequency energy recovery braking with strong mechanical inertia, abruptly cutting off the current directly on the DC bus would cause the drive wheels to instantly lose traction, resulting in extremely severe longitudinal mechanical jerking. This uncontrolled torque surge not only seriously affects driver safety but also generates a huge back electromotive force that could potentially damage the inverter's power module and trigger strong mechanical torsional vibrations in the chassis drive shafts. Therefore, the underlying software of the vehicle controller must embed a smooth torque servo mapping algorithm based on the law of energy conservation and first-order inertial delay.

[0194] Within the motor controller, the system establishes a set of macroscopic power constraint mapping physical formulas that follow the law of conservation of energy. This formula maps the limited DC power to the equivalent power of rotating machinery, thereby calculating the safe upper limit of torque on the AC drive side.

[0195] Its core derating torque is calculated as follows: ;

[0196] in,

[0197] Defined as the upper limit of safe mechanical torque that the motor rotor can currently output or absorb, derived from the battery thermal safety feedforward boundary, and its physical unit is Newton-meter (N·m). );

[0198] Defined as the real-time DC voltage observation value delivered from the high-voltage battery pack to the DC bus side of the motor inverter, in volts (V). );

[0199] This refers to the maximum permissible safe physical limit current calculated and issued in step four, in amperes ( ). );

[0200] Defined as a scalar value of the DC-to-AC combined power conversion efficiency of the motor inverter system at a specific temperature and switching frequency. This value is a dimensionless coefficient and is obtained by real-time interpolation by consulting the internal three-dimensional efficiency characteristic table.

[0201] Defined as the actual mechanical rotational angular velocity of the traction motor at the current moment, it is obtained by high-frequency decoding from the resolver sensor at the motor shaft end, and its physical unit is radians per second. ).

[0202] Through the above formula, the microprocessor ensures an impeccable balance in physical dimensions (watts) between the mechanical power (torque multiplied by angular velocity) on the left side of the equation and the limited electrical power (voltage multiplied by current multiplied by efficiency) on the right side. The motor controller receives... Then, it is used as a boundary limiter input into the internal magnetic field orientation control algorithm to smoothly reduce the quadrature axis current reference value.

[0203] Through this series of mappings, the system maintains the physical continuity of the vehicle's power output while ensuring the safety of heat conduction (limiting endogenous Joule heat). This allows the driver to experience a linear and gentle decrease in power at the pedal end, rather than a power outage due to a high-voltage system failure at the vehicle level.

[0204] While the electrical domain forcibly suppresses and takes over the electrochemical Joule heating and polarization heating, the underlying network physical state of the thermal management system is also undergoing subtle autonomous evolution. At this time, although the maximum charging and discharging current of the battery cells is strictly limited, the contactors are not physically disconnected, and the cells remain in a low-load continuous operating state.

[0205] The moderate heat generated by the battery, maintained within a safe heating threshold, is slowly and continuously transferred to the metal walls of the liquid cooling pipes via highly thermally conductive silicone grease. This controlled, gentle heat becomes the key energy source for dissolving the highly viscous fluid inside the pipes. By avoiding the additional frictional disturbances caused by the high-speed agitation of the water pump, the heat can stably and uniformly heat the fluid "slip layer" adhering to the pipe walls.

[0206] As the temperature of the fluid in the slip layer gradually increases, according to the Arrhenius modified physical equation pre-embedded in step two, the microscopic binding force between local fluid molecules is weakened, and the local dynamic viscosity exhibits a rapid, nonlinear decrease. Macroscopically, this manifests as the physical thickness of the slip layer steadily expanding towards the geometric axis of the pipe center, while the physical radius of the central "high-viscosity cold core" hindering flow decreases. It then began to gradually shrink and degenerate.

[0207] During this critical time window of fluid transition, the actuators (i.e., electronic water pumps) of the liquid cooling network must not be in a silent, stopped state, because stillness will cause a large amount of heat to accumulate locally on the pipe wall, creating physical hot spots.

[0208] However, it is absolutely crucial to avoid directly switching to full-power operation, as this would trigger the viscous dissipation heat generation surge mechanism described in step three. Therefore, the thermal management microprocessor controls the liquid-cooled water pump to enter a highly specialized "low-frequency micro-pulse flaw detection and servo-drive" mode. The pump drive module emits a low-duty-cycle electrical signal at a specific frequency, causing the liquid-cooled pump impeller to generate minute, periodic pressure fluctuations at extremely low speeds.

[0209] The amplitude of these minute pressure fluctuations is limited to a level insufficient to trigger any destructive viscous shear frictional heat generation, but it is sufficient to generate a small pressure drop at both ends of the pipeline network that can be captured by a high-precision differential pressure sensor. The system combines the real-time aligned ultrasonic flow meter data from step one to continuously detect and calculate the transient fluid impedance of the pipeline network (i.e., the macroscopic flow resistance characteristic parameters caused by the high-viscosity cold nuclei remaining in the pipeline).

[0210] As time progresses, the microprocessor continuously invokes the core algorithms of steps two and three within each background computing cycle, dynamically reconstructing the shrinking cold core radius. and the remaining real-time viscous dissipation heat disturbance power When the system's underlying preset "flow thawing release condition" is met, the system will prepare to fully restore the flow in the pipeline network. This flow thawing release condition is defined in the software logic as a composite Boolean logic discriminant containing two sub-conditions:

[0211] ;

[0212] in, Alignment is uniformly defined as the physical inner radius of the pipe; This is a pre-defined boundary value for the minimum proportion of the safe cold core in engineering applications using finite element simulation. This represents the viscous dissipative heat generation disturbance power reconstructed based on the microscopic velocity gradient at the current moment. This serves as the tolerable reference power dissipation noise baseline for system calibration.

[0213] When the above composite conditions are met simultaneously, and after a very short period of anti-shake filtering and noise reduction, the system state machine jumps, which indicates that the cooling medium inside the pipe has safely transitioned from an extremely pathological structure of high viscosity and strong stratification to a conventional Newtonian fluid state with uniform viscosity and good fluidity.

[0214] However, the instant after the fluid-phase transition is complete is a critical and extremely vulnerable period for the entire system. As the viscosity of all fluids within the pipe suddenly drops to an extremely low level, the system's macroscopic hydraulic resistance plummets instantly. If the controller were to simply and abruptly increase the pump's duty cycle to the target maximum speed at this moment, the high-speed rotating pump impeller would instantly inject enormous kinetic energy into the low-viscosity fluid.

[0215] The massive mass of coolant is rapidly accelerated in long and winding cooling channels, which can easily trigger a destructive hydrodynamic "water hammer effect" at pipe reduction points, proportional solenoid valve throttling orifices, or right-angle bends. This transient high-pressure shock wave caused by the rapid acceleration of the fluid can lead to irreversible plastic expansion or even rupture and leakage in flexible rubber connecting pipes, or directly cause the ceramic mechanical seals inside the electric water pump to shatter and fail due to impact.

[0216] The system measures the physical mass of the fluid within the pipeline as... Momentum and force analysis were performed on a rigid fluid column, and a partial differential equilibrium equation for inertial pressure drop was established:

[0217] ;

[0218] Due to macroscopic flow velocity Rotational angular velocity of the electric water pump There exists a definite volume mapping relationship. Substituting this into the above partial differential equation and extracting the angular acceleration term, the microprocessor finally calculates the hard physical constraint formula for the maximum allowable angular acceleration that the liquid-cooled water pump must not exceed during the recovery period: ;

[0219] in,

[0220] This indicates the maximum safe angular acceleration that the electric water pump rotor is allowed to perform. The controller uses this to hard limit the duty cycle ramp rate sent to the frequency converter at the pump end.

[0221] This refers to the physical threshold of the maximum transient shock wave pressure increment that pipeline metal joint materials and rubber seals can withstand without fatigue yielding, expressed in Pascals (Pa). );

[0222] The cross-sectional area of ​​the key flow channel that generates the water jet effect is measured in square meters. );

[0223] This represents the actual fluid density of the cooling medium after thawing at the current pipe wall temperature, expressed in kilograms per cubic meter. );

[0224] The length of the characteristic physical pipe section in the system that is highly susceptible to water column interruption and impact, expressed in meters (m). );

[0225] This is the theoretical flow-speed-volume mapping constant for the liquid-cooled water pump at the current head, expressed in cubic meters per radian (m³ / s). ).

[0226] The microprocessor calculates and integrates this physical constraint equation within each control step to generate a nonlinear speed ramp-up control curve with asymptotically smooth characteristics. Following the instructions of this curve, the controller gradually increases the actual speed of the pump, guiding the massive fluid mass to accelerate smoothly and ensuring the system safely crosses the hydrodynamic physical impact boundary caused by rheological abrupt changes.

[0227] When the actual flow rate of the liquid-cooled main pump safely increases under the guidance of the explosion-proof constraint curve and reaches the target heat dissipation requirement, and the data feedback from the distributed flexible temperature sensor array throughout the battery pack clearly indicates that the surface temperature field and local temperature difference of the battery pack have fallen back and stabilized within the normal safe operating range, the thermal management system microprocessor will send a "thermodynamic safety boundary release" digital handshake signal to the vehicle controller via the bus. After confirming the handshake is successful, the dynamic derating command, which is in an active state in step four, will then proceed. The torque limit, and its derivative, are smoothly removed by the system. At this point, the output limit of the vehicle's electric drive system returns to the nominal rated power boundary.

[0228] Example 2

[0229] The application scenario is illustrated by a pure electric heavy-duty commercial vehicle operating in extremely cold, high-latitude regions (such as winter temperatures as low as -30 degrees Celsius). After the vehicle has been left outdoors for an extended period to cool down, the driver starts the vehicle and immediately enters a highway entrance, aggressively accelerating to demand peak torque (at which point the battery pack instantly enters a high-rate discharge state, for example, an equivalent discharge rate of 3.0C). Under this extreme condition, a huge amount of electrochemical heat is instantly generated inside the battery pack, while the medium (alcohol-water coolant) inside its thermal management liquid cooling network is in a highly viscous, supercooled state similar to a gel.

[0230] Step 1: Realistic Response of Multi-Source Data Acquisition Steps under Extreme Operating Conditions

[0231] Upon vehicle power-on wake-up, the system first enters a forced silent calibration mechanism. At this time, the liquid-cooled main water pump has not yet received a start command, and the microprocessor forcibly reads the static residual value sequence of the differential pressure sensors distributed at the inlet and outlet of the liquid-cooled piping network.

[0232] The extreme cold of -30°C caused a change in the contraction stress of the metal diaphragm, resulting in a zero-point bias of 1.2 kPa for the sensor. The microprocessor calculates the average and writes it to non-volatile memory, ensuring that all subsequent pressure drop data are unconditionally subtracted from this compensation. As the vehicle accelerates rapidly, the high-power inverter generates strong electromagnetic interference.

[0233] Sensor nodes distributed inside and outside the battery pack capture raw data of total charging and discharging current, pressure drop, and total volumetric flow rate, which are mixed with a large amount of high-frequency electromagnetic noise. When processing this data with out-of-order delays, the microprocessor invokes a first-order inertial adaptive anti-aliasing filtering algorithm. At this point, the current observation is evaluated. Compared with historical forecasts The absolute value of the residual is extremely large, far exceeding the calibration reference noise variance. .

[0234] According to the formula The exponential term approaches zero, and the following smoothing coefficient... It rapidly approaches 1. The system establishes a rigid barrier in the digital domain, effectively blocking the contamination of the control system by inverter glitches, and finally outputs a smooth and high-confidence state estimation array to the next stage.

[0235] Step 2: True Response of the Pipe Layered Flow Pattern Reconstruction Step

[0236] After receiving the high-confidence state estimation array, the edge microprocessor combines the high-current data to calculate the current dynamic polarization and the ohmic combined series resistance. This led to the conclusion that extremely high local heat flux density was generated at the pipe wall. Simultaneously, the Arrhenius equation was applied. It was found that the dynamic viscosity of the fluid at the center of the pipe increased by hundreds of times compared to normal temperature.

[0237] To compute the flow regime under limited onboard computing power, the microprocessor discretizes the pipe space into a high-viscosity cold core region and a high-viscosity slip layer region. The mass conservation equations are then solved using a Newton-Raphson iterative optimization algorithm. After a few milliseconds of rapid directional gradient optimization, the system reconstructs a stunning geometric dimension: the unknown cold core physical radius. Reaching the physical inner radius of the pipe 90%.

[0238] This indicates that most of the fluid in the pipeline is in a rigid plunger state with an absolute zero velocity gradient, and only a very thin layer of fluid close to the pipe wall is undergoing rapidly decaying slip flow. The system has fallen into a serious dangerous situation of strong non-isothermal slip stratification.

[0239] Step 3: Actual Response of Viscous Dissipative Heat Generation Boundary Calculation Step

[0240] The system determined that the current ambient reference temperature (-30℃) was below the extreme cold fluid dynamic viscosity abrupt change threshold (-20℃), and the battery's real-time discharge rate (3.0C) far exceeded the high-rate heat generation threshold (2.0C). The nonlinear magnitude reversal trigger discrimination matrix was successfully activated.

[0241] The microprocessor immediately reconstructs the radial velocity gradient from step 2. Substitute into the local dissipation rate density function And apply the integral formula for the volume of a cylindrical space: .

[0242] Calculation results show that, due to the intense shear tearing within the thin layer, the viscous dissipation heat generated by the pump's work causes a disturbance in power. The wattage can reach several kilowatts. This hidden parasitic heat source is close to the battery cell, and instead of dissipating heat, it is accelerating the local temperature rise on the surface of the battery cell.

[0243] Step 4: Actual response of the thermoelectric coupling feedforward compensation control step

[0244] Before the physical temperature sensor shows a significant temperature rise (masked by the thermal inertia of its large mass), the microprocessor calculates the above... As a feedforward signal, it can be directly substituted into the energy boundary to derive the core mathematical formula: .

[0245] Due to the estimated effective surface convective heat dissipation power Extremely low, and Huge, numerator term inside the square root It was instantly dragged to a range less than zero.

[0246] The microprocessor's internal polarity check logic instantly detected this fatal danger, bypassing the square root operation and triggering non-negative protection cutoff. The system forcibly applied the maximum permissible physical limit current instruction. The value was assigned to an extremely conservative, tiny value (even close to zero amperes), which unconditionally cut off the source of Joule heat generated inside the cell before thermal runaway occurred.

[0247] Step 5: Actual Response of Rate Reduction Execution and Pipeline Follow-up Steps

[0248] The motor controller receives heavily compressed data via a high-speed bus. Following the instruction, the constraint mapping physics formula is invoked: .

[0249] Ensure that the upper limit of the mechanical torque of commercial vehicles is set while maintaining dimensional balance. Smoothly lowered pedal pressure. The driver will feel a linear limitation of power at the pedal end, rather than the mechanical jerking caused by a high-voltage power outage, ensuring the safety of the chassis drive shaft.

[0250] At the same time, the water pump drive module sends a low-frequency micro-pulse flaw detection electrical signal to the electronic water pump.

[0251] The excitation signal emitted by the water pump drive module is a square wave or sine wave voltage signal with a frequency of 1Hz~5Hz (preferably 2Hz) and a duty cycle of 5%~15% (preferably 10%), which drives the impeller to maintain a very low speed (<50RPM).

[0252] To ensure absolute safety, the high-speed comparator inside the microprocessor monitors the bus current in real time, and rigidly clamps the peak value of the tiny periodic pressure fluctuations generated by the impeller within a physical amplitude of ≤5kPa (which is only 1% of the burst pressure of the pipeline metal joint). This ensures that it can generate a tiny pressure drop at both the inlet and outlet of the pipeline network that is effectively captured by the high-precision differential pressure sensor, but is absolutely insufficient to cause any destructive viscous shear friction heat generation or mechanical water hammer impact.

[0253] The water pump impeller vibrates slightly, without inducing new viscous friction, while continuously monitoring the fluid resistance within the pipe. As the residual heat of the battery cell slowly melts the cold core, the microprocessor monitors the composite Boolean logic discriminant in real time. When the physical radius of the unknown cold nucleus is unknown. Shrinking to the boundary of the minimum proportion of safe cold nuclei Below, and Reduced to the tolerance reference power dissipation noise baseline The following conditions are met for the release of fluid thawing.

[0254] To prevent a sudden drop in viscosity from causing water hammer and rupturing the pipeline, the microprocessor strictly adheres to the physical constraint formula. Calculate the maximum permissible angular acceleration.

[0255] The electronic water pump follows the safety slope to nonlinearly increase the speed, guiding the massive fluid to accelerate smoothly and successfully cross the physical impact boundary brought about by the rheological abrupt change.

[0256] Ultimately, high-flow-rate convection heat transfer was restored, the battery pack temperature field returned to a safe range, the vehicle power limitation was lifted, and the commercial vehicle resumed full-power operation.

[0257] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An adaptive temperature control method for lithium battery packs based on the Internet of Things, characterized in that, include: The raw data of total charging and discharging current, pressure drop at the inlet and outlet of the liquid cooling pipeline network and total volumetric flow rate are obtained by the sensing nodes. After time stamp alignment and noise removal, a high-confidence state estimation array is output. Based on the high-confidence state estimation array, the local hydrodynamic viscosity, dynamic polarization and ohmic combined series internal resistance are calculated. A fluid momentum dimensionality reduction solution framework including the slip layer and the high-viscosity cold core region is established, and the unknown cold core physical radius and radial velocity gradient are iteratively reconstructed. A local dissipation rate density function is established based on the local fluid dynamic viscosity, the unknown cold core physical radius, and the radial velocity gradient. The viscous dissipation heat disturbance power generated by fluid shear friction is calculated by spatial volume integration. An energy balance algebraic loop is constructed by combining the viscous dissipation heat disturbance power, dynamic polarization, ohmic series resistance, and the maximum permissible transient limit temperature rise rate threshold. The maximum permissible physical limit current command is derived in reverse and non-negative protection cutoff is executed. The maximum permissible physical limit current command is mapped to the mechanical torque safety upper limit command for current limiting execution, and the maximum permissible angular acceleration is calculated based on the maximum transient shock wave pressure increment physical threshold to smoothly control the electronic water pump.

2. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, Before obtaining the raw data of total charging and discharging current, inlet and outlet pressure drop of the liquid cooling network, and total volumetric flow rate, a forced silent calibration mechanism is performed. The specific operation is as follows: During the silent time window before the liquid-cooled main water pump receives the start command, the microprocessor forcibly reads the static residual value sequence of the differential pressure sensor to calculate the mean value, and writes the mean value into the non-volatile memory as the zero-point offset compensation amount. In the subsequent dynamic operation, all the original pressure drop data collected are unconditionally subtracted from the zero-point offset compensation amount to restore the true flow pressure drop.

3. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 2, characterized in that, The timestamp alignment and noise stripping process employs a first-order inertial adaptive anti-aliasing filtering algorithm with a dynamic smoothing coefficient. The dynamic smoothing coefficient is calculated in real time based on the following exponential decay physical model: ； in, This indicates that the current discrete time step generates the dynamic smoothing coefficient; This indicates the raw observation value containing environmental noise captured at the current moment; This indicates the historical value of the state estimate generated at the previous discrete time step; This represents the system configuration forgetting factor scalar; This indicates that the preset physical channel has a calibrated reference noise variance; Within each calculation cycle, by evaluating the absolute value of the residual between the original observation value containing environmental noise captured at the current moment and the historical value of the state estimate generated at the previous discrete moment, a rigid barrier is established in the digital domain to block the pollution of the control system by glitch signals. Conversely, when the data variance returns to the range of stable physical variation, the follow-up smoothing coefficient is adaptively reduced and the weight of the current fresh observation data is increased again.

4. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, The framework for solving the fluid momentum dimension reduction problem, which includes the slip layer and the highly viscous cold core region, relies on the following discrete space partitioning rule: In order to adapt to the limited computing power of the vehicle edge controller and avoid the watchdog timer of the control system being frequently reset and restarted due to computing power overload, the microprocessor forcibly discretizes the fluid space inside the pipe into a high-viscosity cold core region that extends from the center to the unknown cold core physical radius and a high-viscosity slip layer region that extends close to the pipe wall to the unknown cold core physical radius. The fluid velocity gradient at any radial position within the high-viscosity cold core region is forcibly set to be absolutely zero. At the same time, the flow velocity in the high-viscosity slip layer region is set to decrease sharply from the translational velocity at the cold core interface to the absolute zero velocity at the pipe wall, satisfying the fluid dynamics no-slip physical boundary condition.

5. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 4, characterized in that, Also includes: The microprocessor calls the internal Newton-Raphson iterative optimization algorithm to perform mass conservation residual convergence calculation. The microprocessor assigns an initial safe guess value close to the physical inner radius of the pipe to the unknown cold core physical radius and iterates. It uses the gradient of the partial derivative of the residual with respect to the unknown cold core physical radius to quickly adjust the next iteration value and sets a maximum number of iterations limit. If the residuals on both sides of the equation fail to converge to the allowable small error band within the specified computing cycle, the microprocessor triggers the exception handling logic and temporarily freezes the current state update to continue using the safe physical size of the previous cycle.

6. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, The calculation of the viscous dissipation heat disturbance power generated by fluid shear friction requires the activation condition to satisfy the nonlinear order of magnitude reversal triggering discrimination matrix. The specific activation condition is as follows: Only when the system detects that the ambient reference temperature is lower than the threshold for abrupt change in the dynamic viscosity of the extremely cold fluid, and simultaneously detects that the real-time discharge rate of the battery is higher than the threshold for high-rate heat generation of the cell, forming a nonlinear intersection zone, can the enormous electrochemical heat only heat the extremely thin slip layer fluid while the central high-viscosity cold core remains as solid as a rock. This causes the liquid-cooled water pump to continuously output mechanical work, which is directly converted into internal heat energy of the system. Only then does the microprocessor activate the local dissipation rate density function to perform deep integration calculations to extract the values ​​of the parasitic heat source.

7. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, The reverse derivation of the maximum permissible physical limit current command relies on the following energy boundary derivation formula: ； in, This indicates the maximum permissible physical limit current command; This indicates that the composite node contains an equivalent combined thermal mass; This indicates that the composite node material exhibits an equivalent overall specific heat capacity; This represents the maximum permissible transient limiting temperature rise rate threshold; This represents the power of viscous dissipative heat generation disturbance; This represents an estimated value of the actual effective surface convection heat dissipation power. This represents the series resistance of dynamic polarization and Ohm's combined internal resistance; By using the above energy boundary derivation formula, before any physical temperature sensor detects a surge in the actual temperature and before the battery's massive thermal inertia hysteresis takes effect, the inevitable trend of thermodynamic deterioration can be predicted in advance, and direct cross-domain intervention can be made at the underlying electrical safety boundary of the battery management system.

8. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, The specific steps for performing non-negative protection cutoff are as follows: Before executing the square root machine instruction, the derivation formula for the energy boundary is performed. When polarity verification is performed, if the high-viscosity laminar flow state inside the tube deteriorates rapidly, causing the viscous dissipation heat generation disturbance power to increase explosively, making the molecule part easily dragged down below zero by a huge subtraction, once it is determined that the numerator term inside the square root is less than or equal to zero, the square root operation is bypassed directly, and the maximum permissible physical limit current command is forcibly assigned to an absolute zero ampere value, and all Joule heating current inside the cell is immediately and unconditionally cut off.

9. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 1, characterized in that, The calculation of the maximum permissible angular acceleration based on the physical threshold of the maximum transient shock wave pressure increment depends on the following physical constraint formula derived from momentum-force analysis: ； in, This indicates the maximum permissible angular acceleration that the electric water pump rotor is allowed to perform; This indicates the physical threshold value at which the piping component can withstand the maximum transient shock wave pressure increment. This indicates that the key flow channel cross-sectional area is required to generate the water jet effect; Represents the actual fluid density; This indicates the length of the pipe section with distinctive physical characteristics that are highly susceptible to water column impact. This represents the theoretical flow rate, rotational speed, and volumetric mapping constant of the liquid-cooled water pump. Within each control step, the physical constraint formula is calculated and integrated to generate a nonlinear speed ramp control curve, which guides the massive fluid mass to accelerate smoothly and ensures the system safely crosses the hydrodynamic physical impact boundary caused by rheological abrupt changes, preventing the internal ceramic mechanical seal of the electronic water pump from being crushed and failing due to impact.

10. The adaptive temperature control method for lithium battery packs based on the Internet of Things according to claim 9, characterized in that, The smooth control process of the electronic water pump includes low-frequency micro-pulse flaw detection and follow-up mode. The water pump drive module sends a preset frequency low duty cycle electrical signal, causing the liquid-cooled pump impeller to generate tiny periodic pressure fluctuations at extremely low speeds, which are insufficient to cause any destructive viscous shear friction heat generation. The transient fluid impedance of the pipeline network is continuously detected and calculated online. At the same time, the microprocessor determines whether the flow thawing release condition is met based on the following composite Boolean logic discriminant: ； in, Indicates the trigger flag for the release of the flow thawing condition; Indicates the unknown physical radius of the cold nucleus; Indicates the physical inner radius of the pipe; This indicates that a minimum safe cold core ratio boundary has been pre-defined; This represents the power of viscous dissipative heat generation disturbance; This represents the tolerance reference power dissipation noise baseline.

Images