A modular power supply startup and protection method and system for extreme low temperatures

Abstract

This invention provides a modular power supply startup and protection method and system under extreme low temperatures, relating to the fields of battery management and energy storage control technology. The method includes: acquiring multi-source temperature data of the battery module to construct an initial temperature field model, evaluating AC impedance and initial health status, and dividing the module into a ready core area and an extreme cold isolation area; disconnecting the dynamic switch matrix network to cut off the default physical connection between modules and locking the modules in the extreme cold isolation area; using the modules in the ready core area as an energy source, injecting differentiated high-frequency alternating current into the modules in the extreme cold isolation area based on historical experience, and optimizing according to infrared timing characteristics until all modules reach the ready standard; performing a consistency evaluation on the modules that have reached the ready standard to generate the best consistency sequence, and driving the dynamic switch matrix network to reconstruct the main output circuit by connecting the modules in the sequence in series.

Description

Technical Field

[0001] This application relates to the field of battery management and energy storage control technology, and more specifically, to a modular power supply startup and protection method and system under extreme low temperatures. Background Technology

[0002] With the increasing frequency of human activities in polar scientific expeditions, resource exploration in high-altitude and frigid regions, and deep space exploration, energy supply in extremely cold environments has become a core foundation for ensuring the smooth execution of various missions. In polar research stations or high-altitude microgrid energy storage power stations, energy storage systems are typically located in extreme cryogenic environments below -40°C or even -80°C. In this cryogenic dormancy state, the internal electrochemical kinetics of electrochemical energy storage media such as lithium-ion batteries are severely suppressed, specifically manifested as a sharp increase in electrolyte viscosity, a significant decrease in ionic conductivity, and an exponential increase in charge transfer impedance and ohmic resistance. When faced with high-rate power demands such as the startup of large-scale scientific equipment, radar array scanning, or large-scale heating for life support systems, the energy storage system must complete a cold start-up and provide high power output within a short time. However, existing battery management and startup technologies have many limitations in dealing with such extreme conditions.

[0003] Existing cold-start technologies for batteries in extremely cold environments generally rely on external heating or basic internal AC heating mechanisms. In this process, the system's perception of the battery's internal thermodynamic state is heavily dependent on distributed contact temperature sensors. In polar environments, due to intense cold winds and the temperature difference between the edges and center of the energy storage cabinet, a huge and extremely complex temperature gradient forms inside the battery system. Traditional point-contact temperature measurement methods have severe blind spots, failing to obtain a global, high-resolution thermal field distribution. This lack of perception accuracy leads to heating strategies often blindly outputting local temperature averages or extreme values, easily causing extremely uneven heat distribution; some modules face the risk of local thermal runaway due to overheating, while modules in the cold-end blind zone remain frozen. If a start-up is forced without a thorough understanding of the global thermal state, the low-temperature modules are highly susceptible to irreversible negative electrode lithium plating under charge and discharge excitation, severely reducing the battery's physical lifespan and even causing internal short circuits.

[0004] Furthermore, existing modular power systems generally employ a fixed, hard-connected topology (i.e., fixed series and parallel physical circuits). After experiencing deep freezing in polar regions, the health status and AC impedance of each battery module exhibit significant dispersion. Under this fixed topology, the output performance of the entire battery pack is strictly limited by the worst-performing module. When the system performs high-load drawdown in response to large-scale power demand, modules in extreme cold isolation or high-impedance states are forced to carry large currents, leading to deep over-discharge. The fixed physical circuitry cannot physically decouple and isolate high-risk modules during the cold start preparation phase, resulting in a persistently high failure rate in the initial startup phase.

[0005] Meanwhile, electrical loads in polar environments are often characterized by sudden onset and high-rate characteristics, and the external environment is highly variable. During continuous large-scale power consumption, existing systems lack dynamic topology reconfiguration capabilities and historical experience-based optimization mechanisms. If a module in the main power supply circuit experiences a sudden environmental change leading to a sharp drop in temperature and a surge in internal resistance, its voltage will rapidly fall below the consistency threshold. Since seamless micro-topology bypass replacement is not possible during operation, the performance failure of a single module will quickly spread along the fixed series circuit, ultimately causing the entire main power supply circuit to crash and trip, failing to guarantee power supply continuity and system safety in extremely cold environments. The deficiencies in existing technologies regarding multi-source thermal field sensing, dynamic physical topology decoupling, and high-load adaptive protection have become technical bottlenecks restricting the reliable application of polar energy storage systems. Summary of the Invention

[0006] This invention provides a method for starting and protecting a modular power supply under extreme low temperatures, the method comprising: Multi-source temperature data of the battery module are acquired to construct an initial temperature field model, assess AC impedance and initial health status, and divide the module into a ready core area and an extreme cold isolation area. Disconnect the dynamic switch matrix network to sever the default physical connections between modules and lock the modules within the extreme cold isolation zone; Using the modules in the ready core area as the energy source, and combining historical experience, differentiated high-frequency alternating currents are injected into the modules in the extreme cold isolation zone. Optimization is carried out based on infrared timing characteristics until all modules reach the ready standard. For modules that meet the readiness criteria, a consistency evaluation is performed to generate the best consistency sequence, which drives the dynamic switch matrix network to reconstruct the main output loop by connecting the modules in series within the sequence.

[0007] To acquire multi-source temperature data from the battery module, an initial temperature field model is constructed, and the AC impedance and initial health state are evaluated. This includes: acquiring any... The infrared equivalent temperature and contact absolute temperature of each battery module are used to calculate the first... The internal equivalent core temperature of each battery module ; Calculate AC impedance amplitude The initial health status of cryogenic cryogenics was assessed based on the AC impedance amplitude. : in, and These are the confidence weights for contact absolute temperature and infrared equivalent temperature, respectively. and These are contact absolute temperature and infrared equivalent temperature, respectively. and These are the three-dimensional spatial coordinate vectors of the geometric centers of the battery module and the battery cabinet, respectively. The Euclidean distance of the module from the center of the cabinet; The absolute temperature of the external environment; The preset penalty coefficient for thermal radiation leakage in extremely cold cabinets; This refers to the nominal AC impedance amplitude. This is the penalty factor for impedance attenuation in extreme cold.

[0008] Based on historical experience, differentiated high-frequency alternating currents are injected into modules within the extremely cold isolation zone, including historically experienced initial injection steps: Calling the global heating optimization model that was saved in the previous work cycle, extracting the first [model] belonging to the extreme cold isolation zone. Historical heat loss coefficient of the spatial location of each battery module; Based on the current internal equivalent core temperature of the battery module and its spatial distance from the geometric center of the battery cabinet, differentiated initial high-frequency alternating current weights are assigned to different modules within the extreme cold isolation zone. Determine the initial injected AC current amplitude: in, This is the historical heat loss coefficient; The preset wake-up-ready target temperature threshold; This is the internal equivalent core temperature; The absolute temperature of the external environment; This represents the farthest boundary Euclidean distance between each internal module and the geometric center; and These are the driving gain coefficients based on absolute temperature difference and spatial location thermal penalty, respectively.

[0009] Optimization is performed based on infrared temporal characteristics, including the infrared temporal characteristic extraction step: multiple frames of infrared imaging are taken continuously during the AC heating process, and the comprehensive apparent temperature change rate at the current moment is extracted based on the continuous imaging data. and the rate of cooling loss due to environmental heat conduction. The actual electrochemical heating rate was obtained. : in, For the first The infrared equivalent temperature of the step; The time sampling step size; and These are the specified sampling frame sequence number and duration, respectively.

[0010] Optimization based on infrared time-series characteristics also includes a dynamic adjustment step: calculating the actual heating rate error between the actual electrochemical heating rate of the battery module and the preset target safe heating rate. The extracted pure cooling loss rate is used as the feedforward compensation term for heat dissipation in extreme cold environments. Combined with the actual heating rate error, the AC current amplitude adjustment increment for this module is calculated. Based on the AC current amplitude adjustment increment, the target AC current amplitude for the next discrete time step is updated and lithium plating prevention safety limiting is implemented. The dynamic switching matrix network is adjusted to reconstruct the current energy space allocation.

[0011] In the dynamic adjustment process, the AC current amplitude adjustment increment and the updated target AC current amplitude The calculation formula is: in, For the first The actual heating rate error of the step; and These are the proportional and integral gain coefficients, respectively. This is the feedforward compensation coefficient for heat dissipation in extremely cold environments; This represents the absolute value of the cooling loss rate, which is always negative. The current amplitude at the current time step; This is the preset maximum permissible safe AC current threshold for a single module.

[0012] Until all modules reach the readiness standard, including the iterative steps of continuous feature extraction and adjustment: during the iterative process, the internal equivalent core temperature and AC impedance of each battery module in the extreme cold isolation zone are monitored in real time; when it is determined that the parameters of the battery module have crossed the preset safety threshold, the AC heating input to it is immediately cut off, and it is removed from the extreme cold isolation zone and returned to the ready core area in a bypass standby state; the above iterative process continues until the number of modules in the extreme cold isolation zone is zero, all modules reach the readiness standard and the heating process ends.

[0013] After the heating process ends, there is also a step to update and solidify the global heating optimization model. The data aggregation process is as follows: after the cycle ends, the global infrared feature extraction data of this startup process is aggregated; the counterbalancing ratio between the heat drawn away by the polar winds during this iteration of heating and the actual heat generated by internal electrochemical processes is calculated, and this is used as the basis for the next iteration. The actual comprehensive heat dissipation assessment value of each battery module during this cold start cycle : in, This represents the total number of sampling steps required to achieve full readiness during the heating process. The rate of heat loss during cooling; The actual internal electrochemical heating rate is the pure internal temperature rise rate.

[0014] The update and solidification steps for the global heating optimization model involve the following strategy: fusing and correcting the historical heat loss characteristics stored in the global heating optimization model from the previous cycle with the actual comprehensive heat loss assessment value obtained during this cold start, and calculating and generating the updated historical heat loss coefficient. And overwrite it to memory for use during the next startup after a deep freeze: in, This is the historical heat loss coefficient stored in the model from the previous cycle; It is a factor that leads to the forgetting of historical experiences in polar environments.

[0015] The present invention also provides a modular power supply startup and protection system for extreme low temperatures, the system comprising: Region division module: Acquire multi-source temperature data of battery module to construct initial temperature field model, evaluate AC impedance and initial health status, and divide the module into ready core area and extreme cold isolation area; Cold start optimization module: Disconnect the dynamic switch matrix network to cut off the default physical connection between modules and lock the modules in the extreme cold isolation zone; use the modules in the ready core area as the energy source, and inject differentiated high-frequency alternating current into the modules in the extreme cold isolation zone based on historical experience, and optimize according to infrared timing characteristics until all modules reach the ready standard. Link Reconstruction Module: Performs consistency evaluation on modules that have reached the readiness criteria to generate the best consistency sequence, and drives the dynamic switch matrix network to reconstruct the main output loop by connecting the modules in series within the sequence.

[0016] This application provides a modular power supply intelligent startup and protection method and system adapted to extreme low-temperature environments. In the cold-start preparation stage, the traditional indiscriminate equal heating strategy is abandoned, and a differentiated weight injection mechanism based on historical experience is introduced. The system can call upon the global heating optimization model stored in the previous working cycle at the initial heating state, directly extracting the historical heat loss characteristics of each module at a specific spatial location. Combining the core temperature of the current module with the spatial distance from the insulation center of the cabinet, the system assigns unequal initial high-frequency alternating current weights to modules at different physical locations. This differentiated allocation mechanism effectively counteracts the severe wind chill effect in polar edge regions, enabling edge cold-end modules to obtain greater initial activation energy, thereby greatly avoiding the time-consuming process of blindly exploring from scratch each cold start, and significantly compressing the system's global thermodynamic equilibrium and wake-up convergence time under extreme cold conditions. During the optimization execution of AC heating, this application designs a dynamic control algorithm that integrates feedforward compensation for cooling loss rate. The system decouples the internal pure electrochemical actual heat generation rate from the physical cooling loss rate caused by the external polar environment through high-frequency infrared time-series imaging. Faced with extreme weather conditions such as high-speed winds and blizzards in polar regions, traditional temperature feedback control often suffers from significant lag. This application, however, directly incorporates the extracted environmental cooling loss rate as a feedforward compensation term into the control logic. When a local windward module experiences a sudden surge in heat loss due to a sudden cold snap, the system can forcibly output a higher increment of positive heating current within the same cycle, achieving preemptive energy compensation. This mechanism ensures that, under extremely harsh and variable weather conditions, the current distribution can always accurately match the actual thermodynamic needs of the module, preventing localized overheating leading to thermal runaway or secondary freezing caused by heat loss.

[0017] After the cold start cycle, this application constructs a closed loop for updating the global heating optimization model and solidifying the strategy. The system re-evaluates and overwrites the environmental heat loss characteristics by summarizing the counter-current data of heat loss and internal heat generation during the cold start period. This self-learning mechanism retains long-term environmental icing thermal resistance memory while rapidly incorporating deviations caused by the latest climate change during the dormancy period, enabling the power system's cold start response capability to continuously approach the physical optimum. Once all components meet the readiness criteria, the system, targeting large-scale power demand, dynamically selects a set of modules with highly consistent physical and electrochemical states to reconstruct a highly consistent main power supply series circuit. This eliminates the consistency bottleneck of the solidified series circuit, which is constrained by the worst-performing individual unit, ensuring absolutely stable output voltage for large equipment during extreme cold and heavy load startup, and maximizing the safe release of the entire power output. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the modular power supply startup and protection process under extreme low temperatures according to the present invention. Figure 2 Infrared thermal imaging of the modular battery of this invention; Figure 3 This is a distribution diagram of the measured values ​​of the contact temperature sensor of the present invention; Figure 4 This is a temperature distribution diagram of the modular battery space in this invention. Detailed Implementation

[0019] This embodiment provides a modular power supply startup and protection system for extreme low temperatures, serving as the physical hardware foundation for implementing the modular power supply intelligent startup and protection method described in this application. This system and its execution environment are specifically designed for polar operating conditions, with specific application scenarios limited to polar research stations and high-altitude microgrid energy storage power stations. In these scenarios, the ambient temperature is consistently at or below -40°C, and the power system is in a deep-freeze dormant state. The physical architecture of this system is specifically designed to respond to the upcoming large-scale, high-rate power demand after the system has undergone deep freezing, such as the startup of large polar phased array radars or full-load heating of research station life support systems. It provides hardware support for system-level cold-state wake-up, thermodynamic equilibrium, and physical topology reconstruction. The system includes a high-protection-level insulated battery cabinet, a battery module array deployed inside the battery cabinet, a multimodal sensing network, a dynamic switch matrix network, and a control unit. Preferably, the battery module array consists of multiple independent standard battery modules arranged in a matrix on multiple insulating shelves inside the battery cabinet.

[0020] The multimodal sensing network includes a distributed array of miniature infrared thermal imaging modules and a group of contact sensors. Preferably, the distributed array of miniature infrared thermal imaging modules consists of multiple miniature infrared camera probes, arranged in a grid topology on the inner wall of the battery cabinet, the inside of the cabinet door, and the top edge of each insulating shelf. The installation angle and field of view of each miniature infrared camera probe overlap to construct a global blind-spot-free field of view covering the outer surface of all battery modules. In terms of component selection, the miniature infrared camera probes preferably use uncooled vanadium oxide long-wave infrared detectors, which have the characteristics of extreme cold resistance and wide temperature range operation, and are used to perform global continuous scanning of battery modules to collect non-contact thermal radiation data.

[0021] Preferably, the contact sensor group includes a temperature sensor and voltage and current sampling harnesses. The temperature sensor is preferably a low-temperature resistant surface-mount PT1000 platinum resistance thermometer or an NTC thermistor, which is tightly attached to the bottom of the positive and negative terminals of each battery module and the geometric center of the cell sidewall, respectively. The multi-source temperature data collected by multiple sets of PT1000 platinum resistance thermometers, along with the thermal radiation data collected by the infrared thermal imaging module array, are transmitted to the control unit to construct a high-precision initial three-dimensional thermal field and extract the initial health state and AC impedance of each module.

[0022] The dynamic switch matrix network is independently deployed in an isolated chamber on the side of the battery module array, used to disconnect or reconnect the physical connections between battery modules. Preferably, the dynamic switch matrix is ​​constructed from multiple bidirectional high-power solid-state semiconductor switching devices and high-voltage DC mechanical relays in a cascaded parallel topology. In terms of component selection, the bidirectional high-power solid-state semiconductor switching devices are preferably silicon carbide MOSFET modules or IGBT modules with anti-parallel diodes to meet the requirements of high-frequency switching operation and low conduction loss in extremely cold environments. In terms of physical connection, the positive and negative output terminals of each standard battery module in the battery cabinet are connected to the corresponding input node of the dynamic switch matrix network through independent large-section copper busbars or flexible busbars. The output nodes of the dynamic switch matrix network are respectively connected to the main output circuit, the backup circuit, and the internal AC bootstrap heating circuit. By controlling the turn-on and turn-off of the corresponding SiC MOSFET modules or IGBT modules through the gate drive circuit, the dynamic switch matrix network can independently achieve series connection, parallel connection, seamless bypass, and complete bipolar disconnection of any of the standard battery modules in the physical circuit.

[0023] Meanwhile, the internal AC bootstrap heating circuit is equipped with a bidirectional high-frequency inverter bridge arm. The DC side of the inverter bridge arm is coupled to the ready core area module in the battery module array through a dynamic switching matrix, and the AC side is coupled to the extreme cold isolation area module through a dynamic switching matrix, which is used to provide initial high-frequency alternating current excitation.

[0024] The control unit is installed inside the temperature control main control box of the battery cabinet. Preferably, the control unit adopts a heterogeneous computing hardware architecture including an industrial-grade field-programmable gate array and a multi-core digital signal processor. The DSP is used to perform timing iterative calculations for constructing the three-dimensional temperature field model and the global heating optimization model, while the FPGA is used to directly output microsecond-precision PWM gate drive signals to the dynamic switching matrix network.

[0025] The control unit establishes high-speed communication connections with the distributed miniature infrared thermal imaging module array, the contact sensor group, and the driving circuit of the dynamic switch matrix network through a low-temperature resistant isolated CAN bus and an industrial Ethernet harness, forming a closed-loop control circuit.

[0026] Next, this embodiment details a modular power supply startup and protection method under extreme low temperatures, which is based on a modular power supply startup and protection system under extreme low temperatures.

[0027] Step S100: Multimodal sensing and initial thermal field construction.

[0028] The control unit triggers a distributed miniature infrared thermal imaging module array and a contact sensor group to acquire multi-source temperature data; constructs an initial temperature field model, extracts the initial health status and AC impedance of each battery module; and divides the entire battery module into a ready core area and an extreme cold isolation area.

[0029] Step S101: Acquisition of multimodal physical quantities.

[0030] During polar hibernation, the significant temperature difference between the inside and outside of the battery cabinet makes its edges highly susceptible to severe heat loss due to the strong polar winds. Both infrared scanning data and point-contact data have limitations; no single data source can accurately reflect the core temperature of the battery cells inside the module.

[0031] The control unit issues a synchronous sampling command, activating the distributed miniature infrared thermal imaging module array deployed inside the battery cabinet to sample a total of [number missing]. Perform a global scan of the standard battery module to obtain any... One standard battery module ( The non-contact infrared surface temperature radiation matrix is ​​obtained, where N is the number of standard battery modules, and the infrared equivalent temperature of their geometric center region is extracted. Simultaneously, the control unit reads data from the contact sensor group to obtain information about the sensor attached to the first contact sensor. Contact absolute temperature at the bottom of the terminals of a standard battery module .

[0032] Step S102: Initial three-dimensional thermal field construction.

[0033] The control unit utilizes the acquired multi-source temperature data to construct an initial temperature field model based on the temperature gradient distribution characteristics inside the polar battery cabinet, in order to calculate the... The internal equivalent core temperature of a standard battery module The specific calculation formula is as follows: in, The confidence weight for contact absolute temperature; is the confidence weight of the infrared equivalent temperature, and satisfies . ; For the first The three-dimensional spatial coordinate vector of each standard battery module inside the battery cabinet is defined as follows: ; The three-dimensional spatial coordinate vector of the geometric center point inside the battery cabinet; Indicates the first Each standard battery module is offset from the center of the cabinet by an Euclidean distance; The absolute temperature of the polar, extremely cold environment outside the battery cabinet; This is the preset penalty coefficient for thermal radiation leakage in extremely cold environments. The spatial location penalty term in the formula is... The closer the module is to the center of the cabinet, the better the thermal insulation effect and the smaller the correction amount; the closer the edge module is to the insulating wall of the cabinet, the more severe the external extreme freezing wind disturbance, and the greater the gradient of the actual core temperature inside compared with the surface temperature of the pole, which completely eliminates the blind spot that traditional point temperature measurement cannot assess the internal state.

[0034] Step S103: AC impedance and initial health status assessment.

[0035] After extracting the temperature field reference, the control unit instructs the inverter bridge arm of the internal AC bootstrap heating circuit to inject a high-frequency probe AC ​​signal into each standard battery module to extract electrochemical information after dormancy. The high-frequency probe AC ​​signal can penetrate the frozen, high-viscosity electrolyte to detect charge transfer characteristics without inducing large-scale bulk lithium-ion insertion / extraction. The control unit acquires the first... The response voltage amplitude across each standard battery module Response current amplitude and phase angle difference Calculate the first AC impedance amplitude of a standard battery module : In extremely cold conditions, the loss of usable battery capacity is primarily limited by a surge in ohmic impedance and charge transfer impedance. The control unit uses the measured AC impedance amplitude to calculate the first... Initial health status of a standard battery module The calculation formula is: in, The AC impedance amplitude of this model of standard battery module at the standard reference temperature is measured at the factory standard period. This is the penalty coefficient for impedance attenuation in extreme cold conditions.

[0036] Step S104: Forced division of regions oriented towards the lithium plating prevention safety threshold.

[0037] The control unit is based on the calculated , and Perform multi-dimensional joint threshold determination, and Each standard battery module is strictly divided into a ready core area (defined as a set). ) and extreme cold isolation zones (set defined as ).

[0038] The partitioning logic is defined as follows: if and only if the first... A standard battery module is classified as such when it simultaneously meets the following three conditions. :Condition 1: ;Condition 2: ;Condition three: .

[0039] in, The preset critical temperature threshold for safe lithium intercalation in the graphite anode; This is the preset maximum allowable charge transfer impedance threshold; This is the preset minimum discharge health state threshold for extreme cold. If any one of the above three conditions is not met, then the... Each standard battery module is classified into .

[0040] Next, this embodiment will further describe in detail the battery module disconnection based on a dynamic switch matrix in the modular power supply startup and protection method under extreme low temperatures.

[0041] Step S200: Disconnect the battery module based on the dynamic switch matrix.

[0042] After completing the zoning, the control unit issues commands to disconnect the dynamic switch matrix network, severing the default physical series-parallel connections between all standard battery modules; and locks the modules within the extreme cold isolation zone, prohibiting them from participating in any DC load discharge. Specifically, this includes the following sub-steps: Step S201: Generate the target state vector for extremely cold working conditions.

[0043] In polar energy storage systems, the default physical connections are typically fixed series-parallel topologies. If a large-scale power demand is directly responded to after the system has returned to dormancy, the high-impedance modules in the extremely cold state will be forced to carry a large current from the main power supply circuit, potentially leading to severe irreversible lithium plating and deep over-discharge. To avoid this risk, the control unit must not use fixed hardware connections in the circuitry.

[0044] The control unit is for each of the first Each standard battery module defines a set of discretized switching state vectors. Its mathematical expression is: in, Enable the serial connection flag; Enable flag for parallel connection; This is the bypass enable flag; This is the physical isolation flag for bipolar disconnect. The value range of the above flags is... And they satisfy the mutual exclusion constraint condition: .

[0045] Based on the region forced partitioning result of step S100, the control unit generates the target state vector: For modules belonging to the extreme cold isolation zone (i.e.) The system forcibly locks its target state vector to... This means completely severing any physical connection between it and the main output circuit and the backup circuit, preventing it from being accidentally deloaded from the bottom layer.

[0046] For modules belonging to the ready core area (i.e.) To disconnect the default physical connection and prepare for subsequent AC heating reconfiguration, the system initially sets its target state vector to 0. This means it is in a bypass decoupling standby state.

[0047] Step S202: Temperature-compensated gate drive and switching execution.

[0048] The FPGA in the control unit is based on the generated It outputs PWM commands to the gate drive circuit of the dynamic switching matrix network.

[0049] In extreme low-temperature environments ranging from -40°C to -80°C, SiC MOSFET modules in dynamic switching matrix networks exhibit significant low-temperature semiconductor characteristic drift, with their threshold voltage increasing significantly as the temperature decreases. If a driving voltage at normal room temperature is used, the switching devices are prone to failing to fully turn on or off, resulting in high conduction losses or even device burnout.

[0050] Therefore, the control unit combines the first data obtained from the contact sensor group Contact absolute temperature of a standard battery module Dynamic temperature compensation is performed on the gate-source drive voltage of the SiC MOSFET module in the branch containing the extreme cold isolation zone module. The target gate-source voltage after compensation is... for: in, This is the rated gate-source voltage of this SiC MOSFET module under standard room temperature conditions; The preset standard room temperature reference absolute temperature is set to 298.15K. This is the extreme cold threshold voltage drift compensation coefficient for the calibrated semiconductor device, with units of V / K; This is the function for finding the maximum value.

[0051] Through temperature compensation, the FPGA output undergoes level conversion. It drives the corresponding solid-state semiconductor switching devices and high-voltage DC mechanical relays in the dynamic switching matrix network to operate, ensuring the absolute reliability of switching operation under extremely cold conditions.

[0052] Step S203: Isolation absoluteness verification based on current feedback.

[0053] In extremely cold environments, the mechanical contacts of high-voltage DC mechanical relays are highly susceptible to sticking due to low-temperature frost, leading to failure of physical isolation. To ensure the safety of the extremely cold isolation zone (… The standard battery module inside does not participate in DC discharge. After the control unit issues a disconnect command and delays the preset mechanical action time, it performs an isolation absolute verification.

[0054] The control unit collects data in real time via voltage and current sampling harnesses. One was identified as an extremely cold isolation zone ( The actual leakage current of the branch where the standard battery module is located. The control unit performs the following logical decision: in, The maximum safe leakage current threshold allowed by the system; It is a symbolic function.

[0055] like ,express If this is confirmed, the dynamic switch matrix topology decoupling and physical isolation of the module have been successfully executed, and the module is completely and securely locked.

[0056] like This indicates that there is a leakage current exceeding the threshold, meaning that the physical switch has mechanically stuck or solidified under extreme cold conditions. In this case, the control unit immediately triggers the backup fuse or redundant isolating contactor of the corresponding branch, forcibly disconnecting the physical link.

[0057] This step eliminates the risk of lithium plating or deep over-discharge in high-impedance modules under extremely cold conditions, and establishes an absolutely safe flexible disconnect topology before the system receives high-rate power extraction.

[0058] This embodiment further details the stage of global thermodynamic equilibrium and wake-up using the system's internal energy in the modular power supply startup and protection method under extreme low temperatures.

[0059] Step S300: Heating based on timing feedback and historical experience.

[0060] The control unit uses the modules in the ready core area as an energy source and, based on historical experience from the previous cycle, performs non-uniform differentiated high-frequency alternating current injection on the modules in the extreme cold isolation zone, while simultaneously using an infrared sensing network to extract timing features.

[0061] Step S301: Differentiated weight allocation based on historical experience initialization.

[0062] In extremely cold environments, the modules located in different areas of the battery cabinet are affected by polar cold fronts and heat dissipation to vastly different degrees. Injecting the same heating current into all modules in extreme cold conditions would cause the central modules to overheat and run away, while the peripheral modules would remain at low temperatures. Therefore, to avoid blindly applying uniform heating, historical experience is used to shorten the cold start-up convergence time of the system under extreme conditions.

[0063] The control unit's DSP retrieves the global heating optimization model (defined as a historical experience matrix) that was stored at the end of the previous complete working cycle from non-volatile memory. ). For any number belonging to an extremely cold isolation zone A standard battery module, namely Extract the historical heat loss coefficient of the spatial location of the module. .

[0064] The control unit extracts the internal equivalent core temperature based on step S100. And spatial distance characteristics, for the first Each standard battery module calculates the initial alternating current weight. : in, The preset wake-up-ready target temperature threshold; This represents the farthest boundary Euclidean distance between each module inside the battery cabinet and its geometric center. The driving gain coefficient is based on the absolute temperature difference; Let be the driving gain coefficient based on spatial location thermal penalty, and satisfy . .

[0065] Specifically, the first item The closer the module's current core temperature is to the external extremely cold environment temperature, the greater the initial activation energy required; the second item The module that is farther away from the insulation center of the cabinet and closer to the insulation wall has its initial weight forcibly amplified to counteract the severe wind chill effect in the edge area.

[0066] After calculating the weights, the control unit determines the first... Initial injected AC current amplitude of a standard battery module : in, The basic safety AC heating current amplitude constant set for the system.

[0067] At the hardware execution level, the control unit instruction dynamic switching matrix network will belong to the ready core area ( The modules are dynamically connected in parallel as the DC bus power supply terminal, and connected to the DC side of the bidirectional high-frequency inverter bridge arm in the internal AC bootstrap heating circuit; the inverter bridge arm, driven by the high-frequency PWM issued by the FPGA, inverts the DC power into high-frequency AC power, and precisely routes it to the extreme cold isolation zone through the dynamic switching matrix network. In each target module of the process, initial state stimulus injection based on historical experience is completed.

[0068] Step S302: Extraction of infrared temporal features during high-frequency excitation.

[0069] During AC heating, in order to accurately assess the dynamic antagonistic relationship between heating heat generation and heat dissipation in extremely cold environments, the control unit needs to extract time-series derivative features. The control unit drives a distributed miniature infrared thermal imaging module array with a preset time sampling step. Perform continuous multi-frame shooting. Let the current sampling step of the discretized time series be... The control unit extracts the first... The standard battery module is in the first infrared equivalent temperature of the step .

[0070] To separate the internal electrochemical heat generation from the physical cooling caused by the external environment, the control unit periodically inserts a measurement gap at the zero-crossing point of the AC waveform while maintaining the high-frequency alternating current injection.

[0071] Based on this, the control unit extracts the macroscopic heating rate and cooling rate respectively: Calculate the overall apparent rate of change of temperature at the current moment. : Secondly, the rate of cooling loss caused by heat conduction in the extremely cold environment within the measurement gap is extracted. : in, The specified sampling frame sequence number within the measurement gap; This is to measure the duration of the gap. In polar conditions, because the external ambient temperature is much lower than the temperature inside the cabinet, It remains a negative value. Finally, the control unit calculates the... The actual internal electrochemical heating rate of a standard battery module after removing environmental interference. : In this step, considering the extreme variability of the polar environment, if only the conventional total apparent temperature rise rate is relied upon, the system cannot determine whether the slow temperature rise of the module is due to the allocated AC current amplitude. The deficiency is still due to the sudden drop in temperature, which leads to a faster rate of heat loss. Sudden increase. By extracting infrared time-series features, the absolute thermodynamic state evolution trajectory of each module in the current time slice can be determined.

[0072] Step S303: Dynamic optimization based on heating feedback.

[0073] After extracting the infrared timing features, the control unit performs independent optimization control on each standard battery module within the extreme cold isolation zone. In the frigid polar environment, the external cold front wind speed exhibits strong nonlinear abrupt changes. If only temperature feedback control is used, the system response will suffer from severe lag, causing the heat inside the battery cells to be instantly depleted by the wind, leading to secondary freezing. Therefore, this application employs a dynamic control algorithm that integrates feedforward compensation for cooling loss rate.

[0074] Control unit calculates the first The standard battery module is in the first The actual heating rate error of the step : in, This threshold, defined by a preset target safe heating rate, indicates the maximum safe electrochemical heating limit that will not trigger internal localized thermal runaway. Subsequently, the control unit calculates the incremental adjustment of the AC current amplitude for this module. To combat the extremely severe dynamic wind chill effect, the algorithm extracts the environmental cooling loss rate from step S302. When directly introduced into the control law as a feedforward compensation term, its calculation model is as follows: in, This is the proportional gain coefficient; This is the integral gain coefficient; This is the feedforward compensation coefficient for heat dissipation in extremely cold environments; This represents the absolute value of the cooling loss rate, which is always negative. The first two terms of the formula are used to correct for deviations in the heating rate caused by changes in electrochemical impedance; while the third term... This constitutes a feedforward compensation unique to extremely cold environments. When a sudden blizzard in the polar region causes localized thermal resistance collapse of the enclosure and exacerbates the wind chill effect... The sudden surge in current forces the compensation algorithm to output a larger positive current increment, enabling modules with more severe wind chill effects to automatically preemptively acquire higher AC charging amplitudes within the same cycle of physical cooling.

[0075] The control unit updates the target AC current amplitude for the next discrete time step based on the adjustment increment. And implement lithium plating prevention safety limits: in, This is the preset maximum permissible safe AC current threshold for a single module. The control unit will calculate... The instruction is mapped to the duty cycle of the corresponding branch. The FPGA outputs a high-frequency drive signal to adjust the dynamic switching matrix network, thereby realizing the reconstruction of the current energy spatial distribution.

[0076] Step S304: Execute the update and solidification of the global heating optimization model in a loop.

[0077] The control unit continuously and cyclically executes the aforementioned initialization-infrared temporal feature extraction-dynamic optimization steps. During the loop, the control unit monitors each area in the extreme cold isolation zone in real time. The internal equivalent core temperature of the standard battery module within the module is... With AC impedance. When the first Once the absolute parameters of a standard battery module cross the safety threshold described in step S104, the control unit immediately cuts off the AC bootstrap heating circuit input to it and removes it from the extreme cold isolation zone ( Removed and reassigned to the ready core area. It is in a standby state.

[0078] The above cycle continues until the extreme cold isolation zone ( If the number of modules within a given range is zero, then all modules have reached the readiness standard, and the cold start thermodynamic equilibrium process of the entire system ends. Let the total number of cyclic sampling steps required for this closed-loop heating process to reach full readiness be... .

[0079] To eliminate the blind optimization time during startup after the next deep freeze, the control unit summarizes global data after the loop ends, updates and solidifies the global heating optimization model (i.e., the historical experience matrix). ).

[0080] In polar environments, the thickness of the ice and snow cover surrounding the cabinet and the prevailing wind direction exhibit short-term continuity within a specific scientific expedition cycle. Therefore, the heat dissipation intensity experienced by a standard battery module in space is autocorrelated over time. The control unit utilizes infrared data extracted during this startup process to calculate the... The actual comprehensive heat dissipation assessment value of each standard battery module during this cold start cycle : This ratio represents the balance between the heat lost from the physical location of the module by polar winds during this wake-up process and the actual heat generated by internal electrochemical processes. Subsequently, the control unit calculates and generates an updated historical heat loss coefficient. : in, This is the historical heat loss coefficient stored in the model from the previous cycle; It is a factor of historical experience forgetting in polar environments, and satisfies... .

[0081] The control unit updated all standard battery modules. The data is overwritten into the DSP's non-volatile memory, thus solidifying the global heating optimization model strategy. By incorporating a forgetting factor mechanism, the system not only remembers the long-term icing cover characteristics of the polar microenvironment (from...) (Retained), while quickly incorporating minor changes brought about by the latest sudden blizzard during the dormant period (by... (Correction). This machine learning and solidification mechanism ensures that when faced with the next unpredictable large-scale radar activation or heating load withdrawal, the polar large-capacity energy storage system can start up directly in the optimal initial state that is closest to the physical reality, compressing the global temperature difference smoothing time to the limit.

[0082] Once the cyclic heating is complete, the system officially transitions from the cold-state wake-up phase to the high-load operation phase. During this phase, the system must rely on a flexible physical topology to ensure the absolute safety and continuity of power supply in the polar regions.

[0083] Step S400: Dynamic homogeneous stringing for large-scale power consumption.

[0084] After the cycle is completed, all standard battery modules have reached the preset readiness criteria, moved from the extreme cold isolation zone to the ready core zone, and are in bypass standby mode in the dynamic switch matrix network. At this time, the control unit executes dynamic isomorphic stringing in response to the upcoming high-rate power demand in the polar region, specifically including the following execution process: The control unit reads in real time the current absolute temperature, current AC impedance, and updated health status values ​​of all standard battery modules within the ready core area. The control unit performs a consistency deviation assessment on this three-dimensional data, searching for a set of modules across the entire set whose deviations in these three key physical quantities are all within a preset minimum tolerance range. This generates the optimal consistency sequence with the highest electrochemical and thermodynamic consistency. The remaining ready modules not selected for this sequence are designated as a backup circuit sequence.

[0085] The control unit sends a topology reconfiguration command to the dynamic switch matrix network. The FPGA drives the bidirectional high-power solid-state semiconductor switching devices in the corresponding branches to seamlessly switch each standard battery module in the optimal consistency sequence from bypass state to series connection state, thereby constructing a main output circuit capable of withstanding high-rate DC discharge.

[0086] In traditional hard-connected systems, even after global heating is achieved, the differences in internal resistance and temperature between modules are amplified instantaneously under the impact of a large current of hundreds of amperes. This causes modules with slightly lower temperatures or slightly higher internal resistance to be preferentially over-consumed, triggering a bottleneck effect in the entire series circuit. This application addresses this by using dynamic isomorphic stringing to reconstruct a highly uniform output array specifically designed for heavy-load output, avoiding the bottleneck effect and ensuring absolutely stable system output voltage during large-scale load withdrawal, thus maximizing the safe rated power of the polar battery system.

[0087] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0088] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0089] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for starting and protecting a modular power supply under extreme low temperatures, characterized in that, The method includes: Multi-source temperature data of the battery module are acquired to construct an initial temperature field model, assess AC impedance and initial health status, and divide the module into a ready core area and an extreme cold isolation area. Disconnect the dynamic switch matrix network to sever the default physical connections between modules and lock the modules within the extreme cold isolation zone; Using the modules in the ready core area as the energy source, and combining historical experience, differentiated high-frequency alternating currents are injected into the modules in the extreme cold isolation zone. Optimization is carried out based on infrared timing characteristics until all modules reach the ready standard. For modules that meet the readiness criteria, a consistency evaluation is performed to generate the best consistency sequence, which drives the dynamic switch matrix network to reconstruct the main output loop by connecting the modules in series within the sequence.

2. The method according to claim 1, characterized in that, To acquire multi-source temperature data from the battery module, an initial temperature field model is constructed, and the AC impedance and initial health state are evaluated. This includes: acquiring any... The infrared equivalent temperature and contact absolute temperature of each battery module are used to calculate the first... The internal equivalent core temperature of each battery module ; Calculate AC impedance amplitude The initial health status of cryogenic cryogenics was assessed based on the AC impedance amplitude. : ； ； in, and These are the confidence weights for contact absolute temperature and infrared equivalent temperature, respectively. and These are contact absolute temperature and infrared equivalent temperature, respectively. and These are the three-dimensional spatial coordinate vectors of the geometric centers of the battery module and the battery cabinet, respectively. The Euclidean distance of the module from the center of the cabinet; The absolute temperature of the external environment; The preset penalty coefficient for thermal radiation leakage in extremely cold cabinets; This is the nominal AC impedance amplitude; This is the penalty factor for impedance attenuation in extreme cold conditions.

3. The method according to claim 1, characterized in that, Based on historical experience, differentiated high-frequency alternating currents are injected into modules within the extremely cold isolation zone, including historically experienced initial injection steps: Calling the global heating optimization model that was saved in the previous work cycle, extracting the first [model] belonging to the extreme cold isolation zone. Historical heat loss coefficient of the spatial location of each battery module; Based on the current internal equivalent core temperature of the battery module and its spatial distance from the geometric center of the battery cabinet, differentiated initial high-frequency alternating current weights are assigned to different modules within the extreme cold isolation zone. Determine the initial injected AC current amplitude: ； in, This is the historical heat loss coefficient; The preset wake-up-ready target temperature threshold; This is the internal equivalent core temperature; The absolute temperature of the external environment; This represents the farthest boundary Euclidean distance between each internal module and the geometric center; and These are the driving gain coefficients based on absolute temperature difference and spatial location thermal penalty, respectively.

4. The method according to claim 1, characterized in that, Optimization is performed based on infrared temporal characteristics, including the infrared temporal characteristic extraction step: multiple frames of infrared imaging are performed continuously during the AC heating process, and the comprehensive apparent temperature change rate at the current moment is extracted based on the continuous imaging data. and the rate of cooling loss due to environmental heat conduction. The actual electrochemical heating rate was obtained. : ； ； ； in, For the first The infrared equivalent temperature of the step; The time sampling step size; and These are the specified sampling frame sequence number and duration, respectively.

5. The method according to claim 4, characterized in that, Optimization based on infrared time-series characteristics also includes a dynamic adjustment step: calculating the actual heating rate error between the actual electrochemical heating rate of the battery module and the preset target safe heating rate. The extracted pure cooling loss rate is used as the feedforward compensation term for heat dissipation in extreme cold environments. Combined with the actual heating rate error, the AC current amplitude adjustment increment for this module is calculated. Based on the AC current amplitude adjustment increment, the target AC current amplitude for the next discrete time step is updated and lithium plating prevention safety limiting is implemented. The dynamic switching matrix network is adjusted to reconstruct the current energy space allocation.

6. The method according to claim 5, characterized in that, In the dynamic adjustment process, the AC current amplitude adjustment increment and the updated target AC current amplitude The calculation formula is: ； ； in, For the first The actual heating rate error of the step; and These are the proportional and integral gain coefficients, respectively. This is the feedforward compensation coefficient for heat dissipation in extremely cold environments; This represents the absolute value of the cooling loss rate, which is always negative. The current amplitude at the current time step; This is the preset maximum permissible safe AC current threshold for a single module.

7. The method according to claim 1, characterized in that, Until all modules reach the readiness standard, including the iterative steps of continuous feature extraction and adjustment: during the iterative process, the internal equivalent core temperature and AC impedance of each battery module in the extreme cold isolation zone are monitored in real time; when it is determined that the parameters of the battery module have crossed the preset safety threshold, the AC heating input to it is immediately cut off, and it is removed from the extreme cold isolation zone and returned to the ready core area in a bypass standby state; the above iterative process continues until the number of modules in the extreme cold isolation zone is zero, all modules reach the readiness standard and the heating process ends.

8. The method according to claim 7, characterized in that, After the heating process ends, there is also a step to update and solidify the global heating optimization model. The data aggregation process is as follows: after the cycle ends, the global infrared feature extraction data of this startup process is aggregated. The ratio of heat drawn away by polar winds during this iteration of heating of the battery module to the actual heat generated by internal electrochemical processes is calculated, and this ratio is used as the basis for the calculation of the first iteration. The actual comprehensive heat dissipation assessment value of each battery module during this cold start cycle : ； in, This represents the total number of sampling steps required to achieve full readiness during the heating process. The rate of heat loss during cooling; The actual internal electrochemical heating rate is the pure internal temperature rise rate.

9. The method according to claim 8, characterized in that, The update and solidification steps for the global heating optimization model involve the following strategy: fusing and correcting the historical heat loss characteristics stored in the global heating optimization model from the previous cycle with the actual comprehensive heat loss assessment value obtained during this cold start, and calculating and generating the updated historical heat loss coefficient. And overwrite it to memory for use during the next startup after a deep freeze: ； in, This is the historical heat loss coefficient stored in the model from the previous cycle; It is a factor that leads to the forgetting of historical experiences in polar environments.

10. A modular power supply startup and protection system for extreme low temperatures, characterized in that, The system includes: Region division module: Acquire multi-source temperature data of battery module to construct initial temperature field model, evaluate AC impedance and initial health status, and divide the module into ready core area and extreme cold isolation area; Cold start optimization module: Disconnect the dynamic switch matrix network to cut off the default physical connection between modules and lock the modules in the extreme cold isolation zone; use the modules in the ready core area as the energy source, and inject differentiated high-frequency alternating current into the modules in the extreme cold isolation zone based on historical experience, and optimize according to infrared timing characteristics until all modules reach the ready standard. Link Reconstruction Module: Performs consistency evaluation on modules that have reached the readiness criteria to generate the best consistency sequence, and drives the dynamic switch matrix network to reconstruct the main output loop by connecting the modules in series within the sequence.

Images