Cold start thermal management method and system for extreme cold hydrogen fuel cell unmanned aerial vehicle

Abstract

The present application relates to hydrogen fuel cell thermal management and control technical field, especially to a kind of extremely cold hydrogen fuel cell unmanned aerial vehicle cold start thermal management method and system.The method includes: collecting environment, stack, air inlet, air duct data and fusion determination icing risk and target air inlet temperature;According to the partition lighting and purging rhythm driving cathode flow channel wall conformal positive temperature coefficient heating element in situ heating, obtain standard air temperature and humidity;Combined with duty information and initial heat flow upper limit configuration coupling path, implement high temperature difference coupling heat transfer and cooling liquid preheating;Set activation discharge and execute microchannel temperature equalization and throttling;Select phase change heat storage and controllable bypass and form timing;Based on load prediction reference trajectory, carry out joint closed-loop optimal control, generate pump speed and three-way valve opening degree and heat flow distribution instruction.The present application can effectively shorten the cold start time, improve temperature field consistency and extremely cold running stability.

Description

Technical Field

[0001] This invention relates to the field of thermal management and control technology for hydrogen fuel cells, and in particular to a cold start thermal management method and system for extremely cold hydrogen fuel cell drones. Background Technology

[0002] When hydrogen fuel cell drones operate in extremely cold environments, a combination of problems, including insufficient self-heating of the fuel cell stack, icing of the cathode intake and flow channels, uneven temperature field, and lag in the response of the liquid cooling circuit, makes it difficult to establish a self-sustaining temperature field during the cold start phase. Existing solutions mostly employ methods such as electric heating of the intake pipe, anode or cathode purging, and constant-value circulation of the liquid cooling circuit for preheating. In some scenarios, independent heat storage units or external insulation structures are introduced. However, under the constraints of the drone's limited size and mass, there is a significant contradiction between energy consumption, heat reuse, and component layout.

[0003] In engineering, most heating and heat exchange controls are based on fixed duty cycles or simple threshold switching, separating heating from purging, heat exchanger switching, and pump / valve regulation. This lack of feedforward information related to task load trends makes it difficult to maintain temporal consistency between heat supply and flow path status under conditions of load fluctuations and environmental disturbances. Cathode channel wall heating is often non-conformal, with the heat injection location disconnected from the icing source. Waste heat recovery from coupled high-temperature differential heat exchangers does not form a stable energy flow path with liquid-cooled preheating, and the temperature homogenization potential of microchannel cooling plates is not matched in real-time with the branch flow ratio. Consequently, localized overcooling or overheating is prone to occur in the transition zone from startup to activation.

[0004] Although phase change thermal storage units and bypass paths are used in some systems, their start-up and shutdown timing often depends on offline strategies or a single threshold, and is not aligned with load prediction reference trajectory parameters. The availability of thermal storage is not mapped to the constraint boundaries of pump speed and valve position, and the bypass execution command is not involved in the time window limitation of the upper and lower limits of the three-way valve opening. This results in a probability of conflict between valve and pump action and bypass switching, and the proportion of waste heat recovery and the degree of coolant preheating are difficult to stably implement according to the parameters configured along the coupling path. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a cold-start thermal management method for extremely cold-weather hydrogen fuel cell drones, comprising:

[0006] The system acquires ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point, and duct water retention signals, performs synchronous cleaning and fusion, and outputs icing risk index, target inlet temperature, and initial heat flow distribution upper limit.

[0007] Obtain the icing risk index and target intake temperature, perform zone lighting calculation and purging rhythm configuration, perform in-situ heating and anti-icing of cathode conformal positive temperature coefficient heating elements, and output qualified intake air temperature and humidity and duty information.

[0008] Obtain the inlet air temperature, humidity and occupancy information to meet the standards, configure the coupling path and implement high temperature difference coupling heat exchange and coolant preheating, and generate the preheated coolant temperature and waste heat recovery ratio;

[0009] Obtain the preheating coolant temperature and waste heat recovery ratio, set activation discharge parameters and perform safety checks, low-current activation and microchannel temperature control, calculate branch flow ratio and execute throttling configuration, and generate temperature uniformity index and activation temperature achievement mark.

[0010] Based on the temperature uniformity index and activation temperature achievement, the charging window and release window are constructed according to the configuration parameters of the phase change thermal storage unit, and the open-circuit window and closed-circuit window are constructed according to the bypass path parameters. The phase change thermal storage unit is selected and the bypass path is determined to generate the configuration parameters of the phase change thermal storage unit and the bypass path parameters. The phase change thermal storage and controllable bypass dynamic management is executed to generate the thermal storage availability and bypass execution command, and the energy flow allocation and timing arrangement are performed to generate the thermal storage and bypass timing parameters.

[0011] Acquire heat storage and bypass timing parameters, construct a load prediction model and generate a reference trajectory, execute load prediction joint closed-loop optimization control and heat flow distribution solution and command issuance, and generate pump speed command, three-way valve opening command and heat flow distribution command.

[0012] Furthermore, ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point, and duct water retention signals include:

[0013] The ambient temperature refers to the sequence of instantaneous temperature data continuously collected by an ambient temperature sensor deployed on the exterior of the UAV and directly exposed to the flight airspace; the ambient humidity refers to the sequence of instantaneous relative humidity data continuously collected by an ambient humidity sensor deployed on the exterior of the UAV and directly exposed to the flight airspace; the fuel cell stack inlet and outlet temperatures include inlet temperature data collected by a temperature detection element installed on the inlet side of the fuel cell stack and outlet temperature data collected by a temperature detection element installed on the outlet side of the fuel cell stack; the cathode inlet dew point refers to the inlet temperature and humidity data obtained by a temperature and humidity detection element installed in the cathode inlet pipeline; the duct water stagnation signal refers to a discrete or continuous state signal generated by a water stagnation detection element deployed in the airflow channel of the fuel cell system, used to indicate whether there is residual liquid water in the channel.

[0014] Furthermore, the process of performing phase change thermal storage and controlled bypass dynamic management to generate thermal storage availability and bypass execution commands also includes:

[0015] The system reads the configuration parameters and bypass path parameters of the phase change thermal storage unit, establishes two types of timing tables for each phase change unit in the phase change control unit: charging trajectory and releasing trajectory, and maps the valve position target and throttling section of each branch node to an executable opening trajectory. When the control clock reaches the charging start time of the phase change unit, the phase change unit and the cooling plate branch form a controlled heat exchange loop, and the heat exchange interface flux is gradually increased according to the configured target heat exchange intensity. When the control clock reaches the releasing start time, the phase change unit is switched to the cooling path near the low temperature region or to the upstream of the cathode air branch according to the configuration, so as to release the stored heat and complete the energy introduction in conjunction with the throttling section of the path.

[0016] Furthermore, controllable bypass dynamic management includes:

[0017] The branch valve position is driven to open and close according to the opening and closing thresholds of the bypass path parameters. A buffer period corresponding to the sequence preservation identifier is inserted before and after the path switching to perform gradual entry and exit processing on the opening trajectory. During the bypass open period, the minimum maintenance flow is maintained to remove condensate or slowly release local subcooling. During the bypass closed period, the opening is returned to the minimum discharge position to ensure that the path can be reactivated at any time.

[0018] Furthermore, the process of generating thermal storage availability also includes:

[0019] Read the shell temperature, interface temperature and cumulative energy count of each phase change unit, calculate the remaining chargeable capacity and remaining releaseable capacity of the phase change unit, and define the available capacity ratio of the phase change unit in this cycle by the smaller of the two values; weight the available capacity ratio of each phase change unit according to its target division of labor in the configuration parameters to obtain the thermal storage availability of this cycle.

[0020] Furthermore, the process of performing load prediction joint closed-loop optimization control and heat flow distribution solution and command issuance also includes:

[0021] Load prediction reference trajectory parameters are read sequentially from the solution buffer, and thermal storage availability and bypass execution instructions are loaded from the energy flow constraint buffer. At the same time, coupled path configuration parameters are read from the path execution buffer to constrain the solution space. The window identifiers and switching buffer segments in the reference trajectory are aligned with the bypass execution instructions one by one. The lower limit of the three-way valve opening is frozen in the bypass open window, and the upper limit of the opening is frozen in the bypass closed window. The thermal storage availability is mapped to the upper and lower bounds of the pump speed variation.

[0022] Furthermore, the process of solving the joint closed-loop optimization control also includes:

[0023] Combining the upper limit of flow rate and the number of stages that can be deployed in the coupling path configuration parameters, candidate pump speed sequences and candidate three-way valve opening sequences are generated on a rolling basis over time slices. In each time slice, real-time temperature and flow feedback at the inlet and outlet are collected, the deviation from the reference trajectory is calculated, and the deviation, along with the window identifier, is sent to the correction unit. After completing the rolling correction for all time slices, pump speed commands and three-way valve opening commands are output and packaged with time identifiers.

[0024] Furthermore, the operation of the calibration unit also includes:

[0025] If the deviation falls into the switching buffer section, the transition amplitude of the candidate sequence is suppressed by the principle of gradual rise and fall; if the deviation occurs at the bypass opening and closing boundary, the candidate sequences of adjacent time slices are finely adjusted according to the order preservation mark.

[0026] Furthermore, the two types of generated instructions are written into the instruction issuance buffer before dequeueing, and the trajectory segment index of their reference source is recorded synchronously.

[0027] Furthermore, a cold-start thermal management system for an extreme cold-weather hydrogen fuel cell drone, applied to any of the methods described above, includes:

[0028] The data acquisition and data synchronization unit is used to collect environmental, fuel cell stack, air intake, and air duct data, and to perform synchronous cleaning and fusion, outputting raw environmental, fuel cell stack, air intake, and air duct data.

[0029] The perception fusion and threshold calculation and cold start decision and parameter generation unit is used to perform perception fusion and threshold calculation on the raw data, generate the icing risk index, and generate the target intake temperature and the upper limit of the initial heat flow distribution.

[0030] The partition lighting and purging configuration, in-situ heating anti-icing and duty statistics and timing summary unit is used to perform partition lighting calculation and purging rhythm configuration based on the icing risk index and target intake air temperature, drive the cathode conformal positive temperature coefficient heating element to perform in-situ heating and anti-icing, generate compliant intake air temperature and humidity and output duty information.

[0031] The coupling path configuration and high temperature difference heat exchange and coolant preheating unit is used to perform coupling path configuration and flow limit constraints based on the qualified inlet temperature and humidity, duty information and initial heat flow distribution limit, implement high temperature difference coupled heat exchange and coolant preheating, and output the preheated coolant temperature and waste heat recovery ratio.

[0032] The activation discharge and microchannel temperature equalization control and throttling configuration unit is used to set activation discharge parameters and perform safety checks based on the preheating coolant temperature and waste heat recovery ratio, perform low-current activation and microchannel temperature equalization control, calculate the branch flow ratio and implement throttling configuration, and output temperature uniformity index and activation temperature achievement mark.

[0033] The phase change thermal storage and controllable bypass and timing arrangement unit is used to select phase change thermal storage units and determine bypass paths based on temperature uniformity index and activation temperature achievement mark, perform dynamic management and timing arrangement of phase change thermal storage and controllable bypass, and output thermal storage and bypass timing parameters.

[0034] The load prediction and reference trajectory and joint closed-loop optimization unit is used to construct a load prediction model and generate a reference trajectory based on the heat storage and bypass timing parameters and temperature uniformity index, execute load prediction joint closed-loop optimization control, and output pump speed command and three-way valve opening command.

[0035] The heat flow distribution and command issuance unit is used to perform heat flow distribution solution and command issuance based on pump speed command, three-way valve opening command and heat storage and bypass timing parameters, and output heat flow distribution command.

[0036] The key innovations of this invention include:

[0037] (1) The conformal positive temperature coefficient heating element of the cathode flow channel wall is coordinated with the zone lighting and purging rhythm to suppress icing in situ and is directly connected to the duty statistics and timing summary.

[0038] (2) The combined constraints of coupled path configuration and high temperature difference heat transfer integrate the inlet temperature and humidity, duty information and the upper limit of initial heat flow distribution into the decision basis for preheating coolant temperature and waste heat recovery ratio.

[0039] (3) The linkage mechanism between small current activation and microchannel temperature control forms an executable temperature control sequence in the activation range through branch flow ratio calculation and throttling configuration.

[0040] (4) The timing arrangement of phase change heat storage and controllable bypass is to establish a mapping relationship between the charging and releasing windows and the bypass opening and closing windows, generate heat storage and bypass timing parameters and participate in subsequent constraints.

[0041] (5) A joint closed-loop optimization control framework based on load prediction reference trajectory parameters is used to transform the heat storage availability, bypass execution command, and coupling path configuration parameters into time sequence commands of pump speed and three-way valve opening. Command consistency and solution-release closed loop are completed through window identification, switching buffer segment and sequence-preserving correction.

[0042] The following are its main beneficial effects:

[0043] (1) The present invention constructs a single time-based closed loop of acquisition-fusion-judgment-execution-recording, aligns the load prediction reference trajectory parameters, heat storage availability, bypass execution instructions and coupling path configuration parameters in the same solution space, drives the cathode flow channel wall conformal positive temperature coefficient heating element to implement in-situ heating and anti-icing through partition lighting and purging rhythm, and achieves coolant preheating in conjunction with coupling path configuration and high temperature difference heat exchange, and forms a stable temperature field by combining microchannel temperature control, branch flow ratio calculation and throttling configuration in the small current activation stage.

[0044] (2) Introduce the timing arrangement of phase change heat storage and controllable bypass, align the charging and releasing windows with the valve and pump actions, and avoid conflicts between bypass switching and valve and pump adjustment.

[0045] (3) In the joint closed-loop optimization control, the mathematical operation process of rolling solution, time window alignment and sequence-preserving correction is adopted to map the heat storage availability as pump speed change constraint, and the bypass execution command as three-way valve opening boundary. It is coordinated with the flow upper limit and the number of available stages in the coupling path configuration parameters to form pump speed command and three-way valve opening command packaged by time identifier. The command issuance is consistent with the heat flow distribution solution. The waste heat recovery ratio, the preheating degree of coolant and the zoning duty information are updated in the closed loop in the recording link, so as to realize the orderly timing and heat flow distribution between in-situ heating, coupled heat exchange, activation and bypass under extremely cold conditions. Attached Figure Description

[0046] Figure 1 A schematic flowchart illustrating a cold start thermal management method for an extremely cold hydrogen fuel cell drone provided in this application embodiment;

[0047] Figure 2 A structural block diagram of a cold start thermal management system for an extreme cold hydrogen fuel cell drone provided in this application embodiment;

[0048] Figure 3 This is a schematic diagram of a thermal management system for a hydrogen fuel cell drone in extremely cold environments, provided as an embodiment of this application. Detailed Implementation

[0049] The fuel cell thermal management system includes a fuel cell stack, a coupled thermal management and air supply unit, a control unit, and a heat storage material module. The fuel cell stack serves as the core power generation unit. The coupled thermal management and air supply unit includes: an integrated heated cathode cavity with positive temperature coefficient ceramic heating elements integrated into its flow channel walls for rapid heating of the air entering the stack and preventing water vapor freezing; a manifold coolant plate with a multi-manifold distribution structure embedded inside the stack to ensure uniform coolant distribution and avoid localized overheating or undercooling; a high-temperature differential heat exchanger connecting the coolant circuit and the cathode air circuit for recovering waste heat from the stack to preheat the intake air; and a controllable bypass branch for guiding the coolant to bypass the radiator and flow through the heat storage material module during the cold start phase. The control unit receives signals from temperature, pressure, and humidity sensors and executes the start-up logic and thermal management strategy described in steps S100 to S600. The heat storage material module, filled with phase change material, is arranged around the fuel cell stack or integrated into the coolant branch. It stores waste heat generated during the cold start phase and releases heat to buffer the temperature drop of the fuel cell stack during sudden load reductions. The coordinated operation of all components of this system provides the physical basis for achieving synchronous cleaning and fusion, zoned lighting calculation, high-temperature difference coupled heat transfer, phase change heat storage and controllable bypass dynamic management, and load prediction combined closed-loop optimization control.

[0050] Example 1: Refer to Figure 1 This is a flowchart illustrating a cold-start thermal management method for an extremely cold-weather hydrogen fuel cell drone provided in an embodiment of the present invention. The process may include at least steps S100-S600:

[0051] S100: Acquire ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point and duct water stagnation signals, perform synchronous cleaning and fusion, and output icing risk index, target inlet temperature and initial heat flow distribution upper limit.

[0052] S200: Obtain the icing risk index and target intake temperature, perform zone lighting calculation and purging rhythm configuration, perform in-situ heating and anti-icing of cathode conformal positive temperature coefficient heating elements, and output qualified intake air temperature and humidity and duty information.

[0053] S300: Obtain the compliant inlet air temperature, humidity and duty information, configure the coupling path and implement high temperature difference coupling heat exchange and coolant preheating, and generate the preheated coolant temperature and waste heat recovery ratio.

[0054] S400: Obtain the preheating coolant temperature and waste heat recovery ratio, set activation discharge parameters and perform safety checks, small current activation and microchannel temperature control, calculate branch flow ratio and execute throttling configuration, and generate temperature uniformity index and activation temperature achievement mark.

[0055] S500: Based on the temperature uniformity index and activation temperature, an inlet window and a release window are constructed according to the configuration parameters of the phase change heat storage unit, and an open-circuit window and a closed-circuit window are constructed according to the bypass path parameters. The phase change heat storage unit is selected and the bypass path is determined to generate the configuration parameters of the phase change heat storage unit and the bypass path parameters. Dynamic management of phase change heat storage and controllable bypass is performed to generate the heat storage availability and bypass execution command, and energy flow allocation and timing arrangement are performed to generate the heat storage and bypass timing parameters.

[0056] S600: Obtain heat storage and bypass timing parameters, construct load prediction model and generate reference trajectory, execute load prediction joint closed-loop optimization control and heat flow distribution solution and command issuance, and generate pump speed command, three-way valve opening command and heat flow distribution command.

[0057] Step S100 includes at least steps S110-S130:

[0058] S110: Acquire ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point and duct water retention signals, perform data synchronization and anomaly removal, and obtain raw data of environment, fuel cell, inlet and duct.

[0059] The ambient temperature refers to a continuous sequence of instantaneous temperature data collected by ambient temperature sensors deployed on the exterior of the UAV and directly exposed to the flight airspace. The sampling time is recorded using the master clock of the fuel cell stack controller as a unified time reference. The ambient humidity refers to a continuous sequence of instantaneous relative humidity data collected by ambient humidity sensors deployed on the exterior of the UAV and directly exposed to the flight airspace. The sampling time is kept at the same time reference as the ambient temperature. The fuel cell stack inlet and outlet temperatures include inlet temperature data collected by a temperature detection element installed on the inlet side of the fuel cell stack and outlet temperature data collected by a temperature detection element installed on the outlet side of the fuel cell stack. Both inlet and outlet temperature data are recorded using the master clock of the fuel cell stack controller as a unified time reference. The ambient temperature and humidity are aligned with a consistent time base to form a temperature pair sequence with a relative time relationship. The cathode inlet dew point refers to the estimated dew point temperature calculated by consulting a humid air property parameter table, based on the inlet air temperature and humidity data obtained by temperature and humidity detection elements installed in the cathode inlet pipeline, combined with real-time pressure information provided by pressure detection elements in the inlet pipeline. The duct water retention signal refers to a discrete or continuous state signal generated by water retention detection elements deployed in the airflow channel of the fuel cell system, used to indicate the presence of liquid water residue in the channel. This signal is aligned with the sampling time of the fan speed signal and the pipeline pressure difference signal and encapsulated in the same data frame structure. All the above signals are raw sampling sequences with source channel identifiers before the synchronization and anomaly removal processing.

[0060] Specifically, when acquiring ambient temperature and humidity, ambient temperature and humidity data are continuously collected by ambient temperature and humidity sensors installed on the outside of the unit, and the sampling time is recorded with a unified time base; when acquiring the inlet and outlet temperatures of the fuel cell stack, inlet temperature data and outlet temperature data are collected by temperature detection elements on the inlet side and outlet side of the fuel cell stack, respectively, and the time scale is aligned with the time base consistent with that of ambient temperature and humidity; when acquiring the cathode inlet dew point, inlet temperature and humidity data are generated by temperature and humidity detection elements in the inlet pipeline, and the dew point estimation result is read from the lookup table in combination with the pressure detection information of the inlet pipeline to form the cathode inlet dew point data; when acquiring the duct water stagnation signal, the duct water stagnation signal is generated by water stagnation detection elements deployed in the duct, and the signal is aligned with the sampling time of fan speed and pipeline pressure difference, maintaining the same data frame structure. During data synchronization, the time base of the fuel cell stack controller is used as the master clock. The sampling sequences of ambient temperature and humidity, fuel cell stack inlet and outlet temperatures, cathode inlet dew point, and duct water retention signals are interpolated and truncated to ensure that each sequence has a one-to-one corresponding sample pair at the same timestamp. During anomaly removal, range checks and abrupt change detection are performed on ambient temperature and humidity, sensor drift comparison and channel consistency checks are performed on fuel cell stack inlet and outlet temperatures, physical property constraint checks and compatibility checks between dew point and temperature are performed on cathode inlet dew point, and hysteresis debouncing and persistent threshold filtering are performed on duct water retention signals. In this way, missing samples, out-of-bounds samples, and transient jitter samples are removed, and the data frames at the corresponding times are covered with the latest compliant samples. After the above processing is completed, the ambient temperature, ambient humidity, fuel cell inlet / outlet temperature, cathode inlet dew point, and duct water retention signals, all under a unified time base and unit system, are sorted chronologically to obtain raw environmental, fuel cell, inlet, and duct data. This raw data serves as input for subsequent sensing fusion and threshold calculation, and is provided to subsequent steps on the same data bus. This allows subsequent steps to directly read the data content of the ambient temperature, ambient humidity, fuel cell inlet / outlet temperature, cathode inlet dew point, and duct water retention signals at the corresponding time points, achieving consistent input across modules. Understandably, after the synchronization and anomaly removal of the ambient temperature, ambient humidity, fuel cell inlet / outlet temperature, cathode inlet dew point, and duct water retention signals are completed in this step, raw environmental, fuel cell, inlet, and duct data with stable time scales and stable statistical properties are formed. This data serves as the sole data source for the sensing fusion and threshold calculation, ensuring that the data referenced in the sensing fusion and threshold calculation is unambiguous. Furthermore, when the raw environmental, fuel cell, intake, and duct data are written to the control buffer, they are accompanied by a source channel identifier and a valid duration identifier, which facilitates repeated referencing and consistency verification in subsequent steps within the same cycle.

[0061] S120. Perform perception fusion and threshold calculation from the raw data of the environment, fuel cell stack, air intake and air duct to obtain the icing risk index;

[0062] Specifically, the ambient temperature and humidity in the original data of the environment, fuel cell stack, air intake, and air duct are read to construct the environmental boundary condition input; the fuel cell inlet and outlet temperatures in the original data of the environment, fuel cell stack, air intake, and air duct are read to construct the fuel cell temperature difference input and the fuel cell temperature rise and fall trend input; the cathode air intake dew point in the original data of the environment, fuel cell stack, air intake, and air duct is read to construct the air intake dew point input; and the air duct water stagnation signal in the original data of the environment, fuel cell stack, air intake, and air duct is read to construct the air duct water stagnation input. During perception fusion, the data frames of the same period for environmental boundary condition input, fuel cell temperature difference input, fuel cell heating and cooling trend input, intake dew point input, and duct water retention input are first located using a unified timestamp. Then, missing fields are filled in according to the preset data priority. Specifically, environmental boundary condition input is preferentially read from external sensing channels. If external sensing channels are unavailable, short-term backfilling is performed from valid samples of the previous period. Fuel cell temperature difference input and fuel cell heating and cooling trend input are preferentially read from fuel cell measuring points. If fuel cell measuring points are unavailable, backfilling is performed from redundant measuring points and a backfilling flag is recorded. Intake dew point input is preferentially output according to the lookup table. If the lookup table is unavailable, the dew point of the previous period is maintained and a duration limit is applied. When the water retention detection element is unavailable, the duct water retention input is replaced by a combination of pressure difference and fan speed indication, and the replacement source is marked. After completing the above completion process, a consistency check is performed on the data frames of the same period. This is done through mutual verification using the compatibility relationship between ambient temperature and cathode inlet dew point, cross-verification using the relative relationship between the stack inlet and outlet temperatures, and logical verification using the covariance relationship between duct stagnation water input and pipeline pressure difference. Data frames that pass the verification proceed to the threshold calculation stage. During threshold calculation, the reference range for the external heat transfer coefficient is defined by the environmental boundary conditions input; the reference range for the degree of heat accumulation within the stack is defined by the stack temperature difference input and the stack temperature rise / fall trend input; the reference range for the inlet dew point input is defined by the inlet condensation boundary; and the reference range for the degree of liquid water residue within the duct is defined by the duct stagnation water input. Each reference range is mapped to a comparable risk segment according to the step sequence. Then, the risk segments are weighted and combined, and the ranges are merged to form a composite index on a single scale. This composite index is named the icing risk index in this step. After generating the icing risk index, the icing risk index, along with the source channel identifier and effective duration identifier of the corresponding period, is written into the control buffer to provide a traceable reference object for subsequent steps. Understandably, the icing risk index is directly referenced as a control switch and a higher-level constraint in subsequent steps. Specifically, the icing risk index is provided to the next step for cold start decision-making and parameter generation. At the same time, the icing risk index is also used as an input condition for zonal lighting calculation and purging rhythm configuration in the subsequent in-situ heating and anti-icing of the cathode conformal positive temperature coefficient heating element. Furthermore, the icing risk index is written into the constraint set of the optimization solver as a constraint boundary in the load prediction joint closed-loop optimization control, realizing unified solution with the subsequent pump speed command and three-way valve opening command.Through the above-mentioned perception fusion and threshold calculation, the icing risk index is generated in a way that corresponds one-to-one with the original data of the environment, fuel cell stack, air intake and air duct, and is explicitly associated with the data frame of the same period in the data structure to ensure the timing consistency when referenced in subsequent steps.

[0063] S130. Make a cold start decision on the icing risk index to generate the target intake temperature and the upper limit of the initial heat flow distribution.

[0064] Specifically, the icing risk index is read from the control buffer, and within the same cycle, the cathode inlet dew point and the fuel cell inlet / outlet temperature data from the original environmental, fuel cell stack, and intake / duct data are read to form a cold start criterion input set. When making a cold start decision, the icing risk index is first used to locate risk segments, and the cathode inlet dew point is used to construct the intake safety boundary, while the fuel cell inlet / outlet temperature is used to construct the in-pile temperature boundary. Then, based on the combination of risk segments and the two types of boundaries, candidate values ​​for the target intake temperature are generated. Subsequently, using the available energy flow of the reactor body as a higher-level constraint, the candidate values ​​for the target intake temperature are screened to eliminate those exceeding the energy flow carrying capacity. Candidate values ​​are used to obtain the target inlet air temperature. When deriving the initial heat flow allocation upper limit, the target inlet air temperature is used as the core objective, and the in-pile temperature difference boundary formed by the inlet and outlet temperatures of the fuel cell stack is used as the constraint condition. The heat flow allocation object is divided into three components: cathode air branch heat flow, coolant circuit heat flow, and bypass branch heat flow. The upper limit range of each component is determined based on the icing risk index, ensuring that the cathode air branch heat flow has a higher allocation upper limit in high-risk segments, a moderate allocation upper limit for the coolant circuit heat flow in low-risk segments, and an adjustable upper limit for the bypass branch heat flow in segments where load disturbances may occur. This generates the initial heat flow allocation upper limit. After generating the target inlet air temperature and the initial heat flow allocation upper limit, the target inlet air temperature is written to the cathode branch control cache and marked as directly readable by the partition lighting calculation and purging rhythm configuration. The initial heat flow allocation upper limit is written to the energy flow constraint cache and marked as directly readable by the coupling path configuration and flow rate upper limit constraint. To ensure seamless integration between steps, the target intake temperature is explicitly provided as a field to the in-situ heating and anti-icing zone lighting calculations of the cathode conformal positive temperature coefficient heating element while being written to the cache. This allows the zone lighting calculations to determine the on / off sequence for each heating zone within the same cycle. The initial heat flow allocation upper limit is also explicitly provided as a field to the coupling path configuration for high-temperature difference coupled heat exchange and coolant preheating. This allows the coupling path configuration to select series or parallel heat exchange topologies within the same cycle and to define the flow upper limit for each heat exchange branch. Simultaneously, the target intake temperature and the initial heat flow allocation upper limit are registered in the constraint management unit of the load prediction joint closed-loop optimization control, ensuring that subsequent pump speed commands, three-way valve opening commands, and heat flow allocation commands adhere to the same boundary conditions during solution processing. Furthermore, to support subsequent phase change heat storage and controllable bypass dynamic management, the bypass branch heat flow upper limit from the initial heat flow allocation upper limit is synchronously written to the bypass management cache for direct reading in subsequent steps during bypass ratio calculations, thus avoiding redundant estimations.Understandably, after the target intake temperature and the initial heat flow distribution upper limit are generated in this step, they are used as direct inputs for in-situ heating and anti-icing of the cathode conformal positive temperature coefficient heating element, as well as high temperature difference coupled heat exchange and coolant preheating, respectively. They are read and used in the calculation within the same control cycle. In the subsequent small current activation and microchannel temperature equalization control stages, the target intake temperature is used as an inlet-side setting to limit the intake conditions for small current activation, while the initial heat flow distribution upper limit is used as an energy flow boundary to limit the adjustable range of microchannel flow ratio and throttling configuration. In the subsequent phase change heat storage and controllable bypass dynamic management stages, the bypass branch heat flow upper limit in the initial heat flow distribution upper limit is used to limit the amplitude variation of the bypass execution command. In the final load prediction joint closed-loop optimization control stage, the target intake temperature and the initial heat flow distribution upper limit are written into the target set and constraint set to participate in the joint solution of pump speed command, three-way valve opening command and heat flow distribution command. In summary, S110 generates raw environmental, fuel cell, intake, and duct data; S120 generates an icing risk index from the raw environmental, fuel cell, intake, and duct data; and S130 generates a target intake temperature and an initial heat flow distribution limit based on the icing risk index. These two values ​​are then passed to subsequent closed-loop optimization control of in-situ heating of the cathode conformal positive temperature coefficient heating element, anti-icing and high temperature difference coupled heat exchange and coolant preheating, as well as load prediction. This constitutes a coherent closed loop from data synchronization and anomaly removal to perception fusion and threshold calculation, and then to cold start decision-making and parameter generation, ensuring that subsequent steps have clear boundaries and consistent timing for referencing input objects.

[0065] Step S200 includes at least steps S210-S230:

[0066] S210. Obtain the icing risk index and the target intake temperature, perform zone lighting calculation and purging rhythm configuration, and obtain zone lighting parameters and purging rhythm parameters.

[0067] Specifically, data frames from the same control cycle are read from the icing risk index and target inlet temperature generated and cached in the previous step. A partition index consistent with the cathode gas path geometry is established. The cathode conformal positive temperature coefficient heating element is divided into several continuous partitions according to the flow channel length direction and the sidewall bonding structure. A sequence identifier consistent with the inlet direction is established for each partition. In the partition lighting calculation, the risk segment is first located according to the icing risk index. Then, the target inlet temperature is mapped with the geometric heat capacity and bonding thermal conductivity of each partition to obtain the heating requirement of each partition. Subsequently, the lighting sequence calculation from upstream to downstream is performed with the heat transfer sequence in the inlet direction as a constraint to generate partition lighting parameters consisting of the opening time, maintenance duration, and closing time of each partition. To ensure the matching of partition lighting parameters with the residual water state in the flow channel, the time domain identifier of the duct water stagnation signal from the previous step is incorporated into the calculation process. For partitions in the water stagnation indication state, delayed lighting and extended airflow flushing period before closing are set. To generate the purging rhythm configuration, the on-time of the partition lighting parameters is first used as the reference time point. An upstream inlet period and a downstream venting period are introduced before and after this time point to form the rhythm's temporal framework. Then, the required air volume level is derived based on the target inlet temperature. Combined with the estimated flow channel pressure drop, the duration and interval of a single purging cycle are determined, ultimately yielding purging rhythm parameters that correspond one-to-one with the partition lighting parameters. After generation, the partition lighting parameters and purging rhythm parameters are written into the cathode branch control buffer and marked with a period identifier and a source field identifier for direct reading in subsequent steps. The partition lighting parameters are used in the next step to drive the on / off and proportion of in-situ heating and anti-icing of the cathode conformal positive temperature coefficient heating element. The purging rhythm parameters are used to set the timing of the fan speed and bypass valve position. Together, they determine the sequence and intensity of in-situ heating and airflow purging. Meanwhile, the partition lighting parameters, as the prior boundary of energy flow occupation, are copied to the energy flow constraint buffer of the coupling path configuration, so that the subsequent high temperature difference coupling heat exchange and coolant preheating are consistent with the partition lighting parameters when solving the path and flow limit, thus avoiding energy flow overlap in the same control cycle.

[0068] S220. The partition lighting parameters and the purging rhythm parameters are heated and iced in situ by cathode conformal positive temperature coefficient heating elements to generate compliant inlet air temperature and humidity and duct dryness indicators.

[0069] Specifically, within the same control cycle, the partition lighting parameters and purging rhythm parameters matching the partition index are read from the cathode branch control buffer. An on / off timing table and a power ratio table are established for each partition within the heating control unit. These two tables are interpolated and expanded according to the on-time, duration, and off-time in the partition lighting parameters to form a control trajectory with consistent resolution. In-situ heating and anti-icing are executed using a combination of close-fitting on / off control and partition power ratio control: when the on-time of any partition is reached, the positive temperature coefficient heating element for that partition enters an on / off control state, continuously supplying energy according to the ratio given in the power ratio table during the duration; if the duct water stagnation signal corresponding to that partition is still in the indicating state, an additional airflow scouring period is added before the off-time, the rhythm of which is directly provided by the purging rhythm parameters corresponding to that partition. When multiple zones are executed in parallel, to avoid local overheating caused by thermal coupling between adjacent zones, the target inlet air temperature is read and compared with the inlet side temperature detection data. When the inlet side temperature is significantly lower than the target inlet air temperature, priority is given to ensuring the power ratio of the upstream zone. When the inlet side temperature is close to or higher than the target inlet air temperature, the power ratio of the downstream zone is increased to ensure a continuous temperature rise zone along the process. Regarding the timing control of the fan and valves, the fan speed and bypass valve position are segmented based on the purging rhythm parameters. During each purging period, the fan is started, stopped, and its opening is switched according to the air volume level and interval length set by the rhythm parameters. A minimum maintenance flow rate is maintained during the purging intervals to remove localized condensation. To ensure seamless connection between in-situ heating and anti-icing and subsequent coupled heat exchange, a transition period is reserved between the closing time of each zone and the opening time of the downstream zone. This transition period is read in the next step by the coupling path configuration and is used to select the series or parallel heat introduction method. After all illuminated partitions reach their respective shutdown times, the temperature and humidity detection data from the inlet and outlet sides are aggregated. The corresponding intake temperature and humidity are calculated based on the same time marker. Inlet temperatures and humidity levels that meet the target intake temperature range and are not higher than the safety limit are marked as compliant intake temperature and humidity. Simultaneously, the continuous state of the duct water stagnation signal within the cycle and its state transition after flushing are statistically analyzed. If no water stagnation indication appears within several consecutive control segments, a duct drying flag is written at the end of the cycle. The compliant intake temperature and humidity and the duct drying flag are written to the coupling heat exchange trigger buffer and the activation inlet condition buffer after generation. The former is directly read by high-temperature difference coupling heat exchange and coolant preheating to enable coupling path configuration and flow limit constraints. The latter is directly read by small-current activation and microchannel temperature equalization control to set the activation discharge inlet conditions and the initial flow ratio of the microchannel.

[0070] S230. Perform duty cycle statistics and time sequence summary on the qualified inlet air temperature and humidity to generate duty cycle information of positive temperature coefficient heating element;

[0071] Specifically, within the same control cycle, the time stamps for achieving the compliant intake air temperature and humidity are read, and the on / off timing tables and power percentage tables for each zone are aligned. The on / off duration, number of on / off cycles, and average power percentage for each zone within the time segment determined to meet the compliant intake air temperature and humidity are statistically analyzed to form zone-level duty cycle statistics. The zone-level duty cycle statistics for all zones are then weighted and merged, with weights determined by the geometric air-receiving area, thermal conductivity, and purging rhythm participation of each zone within the cycle, resulting in the overall duty cycle statistics for the cathode conformal positive temperature coefficient heating element. To support unified constraints for subsequent energy flow allocation and topology selection, the overall duty cycle statistics are superimposed with the temporal sequence of each zone to generate a time-series summary. This time-series summary clearly marks the opening sequence of each zone, overlapping sections, and their correspondence with purging rhythm parameters, and adds transition markers to sections related to the coupling heat transfer transition. The overall duty cycle statistics and time-series summary are combined to form the duty cycle information of the positive temperature coefficient heating element, and written into the energy flow constraint buffer and the time-series coordination buffer. The former is used as a boundary condition for the flow limit and coupling timing in high-temperature difference coupled heat exchange and coolant preheating, restricting the introduction method of the coupling path configuration and the coolant flow limit in the peak duty cycle segment. The latter is used as the prior timing for valve and pump control in load prediction joint closed-loop optimization control, so that the pump speed command and three-way valve opening command avoid the duty-intensive segment when solving, reducing energy flow competition within the same cycle. Furthermore, in order to maintain the connection with phase change heat storage and controllable bypass dynamic management, the transition flag in the time-series summary is synchronously written into the bypass management buffer as a time-domain reference for subsequent bypass ratio calculation, ensuring that there is a switchable window between bypass triggering and partition shutdown. At this point, the partition lighting parameters and the purging rhythm parameters are mapped to time-corresponding inlet air temperature and humidity and duct dryness indicators through in-situ heating and anti-icing execution. These are then statistically summarized into positive temperature coefficient heating element duty cycle information. All the aforementioned fields are generated and distributed within the same control cycle, forming a closed-loop data chain from parameter calculation to execution and then to statistics. Understandably, through the above three-step process, without changing the configuration of external components, the synchronous driving of partition lighting and purging rhythm on in-situ heating and airflow scouring is achieved. This results in a unified data entry point that can be used to couple heat exchange and activation inlet and valve / pump optimization, demonstrating overall consistency in partition control, airflow management, and energy flow coordination.

[0072] Step S300 includes at least steps S310-S330:

[0073] S310. Obtain the qualified inlet air temperature and humidity, the duty information of the positive temperature coefficient heating element, and the initial heat flow distribution upper limit; perform coupling path configuration and flow upper limit constraint to obtain coupling path configuration parameters.

[0074] Specifically, within the same control cycle, the compliant inlet air temperature and humidity, consistent with the partition index, are read from the coupling heat exchange trigger buffer, and the positive temperature coefficient heating element duty cycle information is synchronously read from the energy flow constraint buffer. Simultaneously, the initial heat flow allocation upper limit is loaded from the energy flow constraint management unit. After verifying the consistency of the cycle identifiers and timestamps of the three, a coupling object set is established. This set includes the high temperature difference heat exchanger, coolant circuit, cathode air branch, and bypass branch. During the coupling path configuration process, firstly, based on the temperature and humidity ranges of the compliant inlet air temperature and humidity, the upper boundary of the allowable energy input to the air side is determined. Then, the overall duty cycle statistics and time-series summary in the positive temperature coefficient heating element duty cycle information are mapped to energy flow occupancy windows and prohibition windows, serving as constraints for coupling timing and flow rate limits. Finally, the initial heat flow allocation upper limit is unpacked according to the air side heat flow upper limit, coolant circuit heat flow upper limit, and bypass branch heat flow upper limit, forming a three-component heat flow boundary. Based on the aforementioned boundaries and windows, topology selection and path splicing are performed: when the prohibited window covers the upstream air side, a parallel coupling path leading the coolant circuit is selected; when both the energy flow occupancy window and the allowable energy input on the air side are satisfied, a series coupling path leading the air side is selected; if the two partially overlap, a transition period is set in the overlapping section and a sequence switching identifier is inserted. Furthermore, combining the available pump speed range of the coolant circuit and the number of stages that the high-temperature differential heat exchanger can be put into operation, flow upper limits are allocated to each branch according to the three-component heat flow boundaries, and the corresponding branch flow upper limit is set to zero within the prohibited window. After completing the above solution, coupling path configuration parameters are generated. These parameters include heat exchange topology selection field, inlet and outlet valve position target, coolant pump speed reference, air-side and coolant-side flow limit, number of stages that can be engaged, transition period identifier, and timing coordination identifier. The coupling path configuration parameters are written into the path execution buffer as the sole input for the next step of executing high temperature difference coupled heat exchange and coolant preheating. At the same time, the transition period identifier is synchronized to the timing coordination buffer for subsequent small current activation and microchannel temperature equalization control to maintain consistency between inlet conditions and local flow ratio.

[0075] S320. Perform high-temperature differential coupling heat exchange and coolant preheating on the coupling path configuration parameters to generate a preheated coolant temperature.

[0076] Specifically, the coupling path configuration parameters are read from the path execution buffer, and the heat exchange topology selection field is sequentially sent from the heat exchange control unit to the valve group controller. This executes the opening and closing and switching of the inlet and outlet valve positions, ensuring that the high-temperature differential heat exchanger, coolant circuit, and cathode air branch form a stable heat exchange channel according to the set series or parallel connection. Simultaneously, the coolant pump speed reference is written to the pump speed controller, and the heat exchange cores are activated step-by-step according to the available stages. When the transition period indicator appears, the inlet valve position is adjusted in a segmented manner to avoid transient backflow. Data acquisition after coupling is completed collaboratively by measurement points on both sides: on the air side, the inlet temperature, outlet temperature, and flow rate feedback are read at the same sampling frequency as the compliant inlet air temperature and humidity; on the coolant side, the coupled inlet temperature, coupled outlet temperature, and flow rate feedback are read and aligned with the pump speed feedback and valve position feedback for archiving. To utilize the high temperature difference under constrained conditions, the actual flow rates on both sides are adjusted and gradually released based on the upper limits of the air and coolant flow rates. A minimum maintenance flow strategy is implemented within the prohibited window, and a gradual increase / decrease strategy is implemented at the energy flow occupancy window boundaries to ensure compatibility with the time structure of the positive temperature coefficient heating element duty cycle information. When entering the transition period, the upper limit of the pilot-side flow rate is first reduced according to the timing coordination identifier, then the upper limit of the follow-side flow rate is increased, and finally restored to the configured steady-state upper limit after the valve position switch is completed. During the continuous execution process described above, the coupling outlet side of the coolant circuit is used as the preheating monitoring point. Temperature samples within the same cycle are denoised and averaged over time, and the preheating coolant temperature is generated based on the effective sample set of the path execution period. This preheating coolant temperature is written into the activation inlet condition buffer as the inlet temperature input for small-current activation and microchannel temperature equalization control. Simultaneously, this field is associated with and stored with the transition period identifier of the coupling path configuration parameters for reference in the next step of heat exchange efficiency calculation and energy flow accounting, specifying its source and effective segments.

[0077] S330. Calculate the heat exchange efficiency and energy flow of the preheated coolant temperature to generate the waste heat recovery ratio.

[0078] Specifically, the preheating coolant temperature, corresponding cycle coolant-side flow feedback, and air-side inlet and outlet temperatures are read within the energy metering unit. Simultaneously, the flow limit, available stages, and transition period identifiers in the coupling path configuration parameters are read. An energy flow recorder is established using time slices as the smallest metering unit. The air-side and coolant-side imported energy within each time slice is converted and accumulated using lookup tables. In time slices with transition period identifiers, the energy flow recorder is split and archived before and after the transition to avoid measurement confusion caused by different topologies. In time slices with prohibited windows, the energy flow is marked as zero, and the reason for the duty cycle shielding is recorded for subsequent valve and pump optimization to avoid duplicate deployment. After completing the full cycle accumulation, the coolant-side imported energy is merged into loop recovery energy, and the air-side imported energy is merged into air-side imported energy. This is then aligned and checked against the total distributable heat flow corresponding to the initial heat flow allocation limit for the corresponding cycle. If a timing gap exists, the gap segment is located using the timing coordination identifier, and the gap marker is filled according to the minimum sustaining flow principle to ensure the calculation boundary is closed. Subsequently, the waste heat recovery ratio is generated based on the proportion of energy recovered in the loop within the same cycle. The waste heat recovery ratio and the preheating coolant temperature are written together into the activation slope setting buffer for direct reading by low-current activation and microchannel temperature equalization control. This is used to set the heating rhythm and local flow ratio of activation discharge. At the same time, the waste heat recovery ratio, the upper limit of flow rate of coupling path configuration parameters, and the number of stages that can be put into operation are written together into the optimization constraint buffer as the energy flow constraint and switching penalty basis for load prediction joint closed-loop optimization control. The energy flow record segments related to bypass are synchronized to the bypass management buffer for phase change heat storage and controllable bypass dynamic management to determine the bypass ratio in subsequent cycles. Understandably, through the continuous execution of the coupled path configuration and flow limit constraint, the high temperature difference coupled heat exchange and coolant preheating, and the heat exchange efficiency calculation and energy flow accounting, a closed-loop data chain is formed, driven by the coupled path configuration parameters and with the preheated coolant temperature and waste heat recovery ratio as the core output. This ensures consistent input and output interfaces and timing coordination between the small current activation and microchannel temperature equalization control, phase change heat storage and controllable bypass dynamic management, and load prediction joint closed-loop optimization control. Overall, this achieves the connection and traceability of coupled configuration, heat exchange execution, and energy flow accounting within the same control cycle.

[0079] Step S400 includes at least steps S410-S430:

[0080] S410. Obtain the temperature of the preheated coolant and the proportion of waste heat recovery, set the activation discharge parameters and perform a safety check to obtain the activation discharge parameters.

[0081] Specifically, the preheating coolant temperature, which is identified at the same time as the coupling path configuration parameters, is read from the activation inlet condition buffer according to the control cycle. Simultaneously, the waste heat recovery ratio is read from the activation slope setting buffer. At the same time, the inlet condition field for the current cycle is extracted from the target inlet temperature generated in S130 and the compliant inlet temperature and humidity and duct dryness indicators generated in S220. Together with the ambient temperature and humidity provided in stage S100 and the fuel cell inlet and outlet temperatures, an input set for setting activation discharge parameters is established. During the activation discharge parameter setting process, the consistency of the inlet side is first verified using the preheating coolant temperature and the compliant inlet temperature and humidity to ensure that the inlet temperature and humidity are within the effective range of the same control cycle as the preheating coolant temperature. Then, the heating capacity boundary is constructed using the waste heat recovery ratio and the fuel cell inlet and outlet temperatures to limit the usable heat release level and the tolerable heating slope for a single cycle. Subsequently, the target inlet temperature is used as the inlet side setting to provide candidate ranges for three types of parameters: activation discharge current level, activation maintenance duration, and activation step interval. After the candidate interval is defined, a safety check is performed: The ambient temperature, humidity, and duct dryness indicators are reviewed for compliance. If the duct dryness indicator is not ready, the inlet flow rate is maintained and the start time of activation discharge is delayed within this cycle. The difference and trend of the inlet and outlet temperatures of the fuel cell stack are checked for out-of-bounds conditions. If an abnormal rise or fall occurs, the activation maintenance time is reduced and an anomaly is recorded. The waste heat recovery ratio is checked for saturation. If it exceeds the preset upper limit, the activation discharge current level is reduced to avoid energy flow competition with coupled heat exchange. After the above settings and checks, the selected activation discharge current level, activation maintenance time, and activation step interval are integrated into activation discharge parameters and written into the activation execution buffer according to the time identifier, serving as the sole input for subsequent low-current activation and microchannel temperature equalization control. Simultaneously, the inlet conditions and safety check results bound to the activation discharge parameters are archived for constraint reference and timing coordination in the load prediction joint closed-loop optimization control of the S600 stage, ensuring that subsequent pump speed commands, three-way valve opening commands, and heat flow distribution commands are consistent with the settings in this step during the solution process.

[0082] S420. Perform low-current activation and microchannel temperature control on the activation discharge parameters to generate temperature uniformity indicators.

[0083] Specifically, the activation discharge current level, activation duration, and activation step interval are read sequentially from the activation execution buffer and sent to the controlled branch on the DC side of the fuel cell stack in the activation control unit, so that the fuel cell stack enters a low-current activation state. During the entire activation duration, the discharge is performed in segments with the activation step interval as the time unit, so that the discharge amplitude and the interval between segments are aligned with the transition period of the coupling path formed in stage S320, thus being consistent with the time structure of the preheating coolant temperature. The microchannel temperature uniformity control is implemented as follows: First, the coolant temperature and flow feedback at the inlet and outlet sides of the multi-manifold microchannel cooling plate are read. A fixed initial branch allocation ratio is used to perform basic coolant supply to each branch, and a temperature sampling sequence along the path is constructed using a temperature detection array deployed inside the fuel cell stack. Then, the sampling sequence is segmented and summarized according to the activation step interval to form a temperature distribution record corresponding to each discharge segment. At the end of each segment, three types of statistics are calculated: temperature range within the segment, gradient along the path, and difference between different areas, as segment-level intermediate quantities. Next, a consistency check is performed on the segment set at the end of the cycle. Isolated abnormal segments caused by short-term fluctuations in inlet conditions are removed, and the statistics of the retained segments are aggregated into a single-scale temperature uniformity index according to preset weights. To ensure that the temperature uniformity index can be directly utilized later, its time correspondence with the activation discharge parameters is retained when the index is generated, and the index and the preheating coolant temperature of the corresponding cycle are written into the uniform temperature evaluation buffer. Simultaneously, short-term fluctuations and fragment rejections occurring during activation are recorded to help the phase change heat storage and controllable bypass dynamic management of the S500 stage identify possible transition windows. Before the end of the activation maintenance period, if continuous segments of unidirectional changes in the stack inlet and outlet temperatures are detected, and the aforementioned temperature distribution records show that the gradient along the path is in a low range, a stability marker is added to the temperature uniformity index at the end of this cycle to indicate that the next step can perform branch flow ratio calculation and throttling configuration within a relatively stable range.

[0084] S430. Based on the temperature uniformity index and the preheating coolant temperature, calculate the branch flow ratio and configure the throttling to obtain the branch flow ratio parameter and activation temperature achievement indicator.

[0085] Specifically, the temperature uniformity index and preheating coolant temperature corresponding to the current cycle are read from the temperature uniformity assessment buffer, and the latest inlet and outlet temperatures are extracted from the stack inlet and outlet temperatures in stage S100. The transition period identifier and coolant-side flow limit are read from the coupling path configuration parameters in stage S320, and a calculation input set is established within the branch allocation unit. First, the branch flow ratio calculation strategy set is selected based on the temperature uniformity index interval: when the temperature uniformity index is in a non-uniform state, resistance-increasing configuration is preferentially applied to branches near the high temperature value area, and resistance configuration is released to branches near the low temperature value area, with the start and end times of resistance-increasing and release constrained by the transition period identifier; when the temperature uniformity index is in a quasi-uniform state, a micro-adjustment strategy is adopted to make small-scale proportional corrections between branches to maintain balance along the path. After the strategy set is determined, the flow change amplitude of each branch is trimmed in combination with the preheating coolant temperature and coolant-side flow limit, prohibiting instantaneous jumps exceeding the limit, and setting an interlock sequence for simultaneous changes in adjacent branches to avoid reverse backflow before and after coupling path switching. After completing the above restrictions, the target proportion value of each branch and the target opening degree of the valve group are mapped to the branch flow ratio parameter, and sent to the throttling component at the branch node at the inlet of the cooling plate. The throttling configuration is executed according to the time identifier. During the execution of the throttling configuration, the continuous samples of the inlet and outlet temperatures of the fuel cell stack collected by the activation control unit are used as the judgment basis. If the inlet temperature and outlet temperature meet the activation temperature criterion within a certain number of consecutive sampling segments, and the aforementioned branch flow ratio parameter remains stable within the time period, an activation temperature achievement identifier is generated and written into the flow distribution buffer along with the branch flow ratio parameter. This allows the phase change heat storage and controllable bypass dynamic management in the S500 stage to directly read the upper-level constraint of the bypass ratio, and at the same time, it provides the load prediction joint closed-loop optimization control in the S600 stage as the boundary and initial value for solving the valve and pump. Furthermore, to ensure coordination with the in-situ heating and anti-icing of the cathode conformal positive temperature coefficient heating element in stage S200, the activation temperature achievement flag is synchronously written into the cathode branch control buffer. This allows the cathode conformal positive temperature coefficient heating element to implement power yielding and timing convergence based on this flag in subsequent cycles, thereby avoiding timing conflicts with the throttling configuration on the coolant side. Understandably, through the branch flow ratio calculation and throttling configuration in this step, the branch flow ratio parameter and the activation temperature achievement flag form a clear input-output link within the same control cycle: the temperature uniformity index and the preheated coolant temperature serve as inputs. After calculation and execution constrained by the transition period flag and the upper limit of the coolant side flow, structured parameters and status flags that can be directly consumed by subsequent modules are produced. Overall, this achieves continuous transmission and closed-loop updating from activation discharge parameters to temperature uniformity index, and then to branch flow ratio parameters and activation temperature achievement flags.In summary, S410 to S430, while maintaining consistent data interfaces with S320 and S330, sequentially execute the activation discharge parameter setting, low-current activation, microchannel temperature equalization control, branch flow ratio calculation, and throttling configuration. This forms a closed-loop link that starts with the preheating coolant temperature and waste heat recovery ratio, and ends with the branch flow ratio parameter and activation temperature as the identifier. This provides a directly applicable and time-consistent input basis for subsequent phase change heat storage, controllable bypass dynamic management, and load prediction combined closed-loop optimization control.

[0086] Step S500 includes at least steps S510-S530:

[0087] S510. Obtain the temperature uniformity index and the activation temperature achievement mark, select the phase change heat storage unit and determine the bypass path, and obtain the phase change heat storage unit configuration parameters and bypass path parameters.

[0088] Specifically, the temperature uniformity index, consistent with the current control cycle, is read from the temperature uniformity assessment buffer, and the activation temperature achievement flag is read synchronously from the flow distribution buffer to establish the entry criterion set for this cycle. After the entry criterion set is ready, the geometric position, thermal interface state, and historical charge / discharge records of each phase change unit in the phase change thermal storage module are summarized to form a list of units that can be put into operation. Then, using the temperature uniformity index as a spatial allocation criterion, phase change units near high temperature regions are marked with a priority charging order, and phase change units near low temperature regions are marked with a delayed charging order to ensure that the commissioning order of phase change units is compatible with the temperature distribution within the reactor. Subsequently, using the activation temperature achievement flag as a time admission criterion, the collaborative calculation of phase change thermal storage and bypass path is initiated only when the flag is ready. If the flag is not ready, the temperature uniformity index for this cycle is temporarily stored and candidate results for monitoring purposes only are generated. During the bypass path determination process, the system retrieves preset bypass branch nodes and loop connection relationships. Candidate paths are divided into several basic paths based on their coupling methods with radiator branches, cooling plate branches, and cathode air branches. Based on the interval values ​​of the temperature uniformity index, each basic path is labeled with an opening and closing threshold. Furthermore, the aforementioned spatial allocation criteria and time admission criteria are superimposed onto the basic path set. Paths that conflict with the phase change unit priority in the current cycle are eliminated, while paths consistent with the phase change unit deployment order and compatible with the activation temperature are retained. Based on this, the target heat transfer direction, target heat transfer intensity, and allowable number of switching operations are assigned to each phase change unit to form the phase change thermal storage unit configuration parameters. Simultaneously, a branch valve position target, throttling section, minimum allowable maintenance flow rate, and switching sequence identifier are given for each reserved path to form bypass path parameters. Finally, the phase change thermal storage unit configuration parameters and bypass path parameters are written into the bypass management buffer according to time identifiers, and the source of the referenced temperature uniformity index and activation temperature achievement identifier is explicitly recorded in the parameters for subsequent steps to verify the consistency of timing and source. The inputs to the above process are the temperature uniformity index and the activation temperature achievement identifier; the processing involves phase change thermal storage unit selection and bypass path determination; the outputs are the phase change thermal storage unit configuration parameters and bypass path parameters.

[0089] S520. Perform dynamic management of phase change heat storage and controllable bypass on the configuration parameters of the phase change heat storage unit and the bypass path parameters, and generate heat storage availability and bypass execution instructions.

[0090] Specifically, the configuration parameters of the phase change thermal storage unit and the bypass path parameters are read sequentially from the bypass management buffer. Two timing tables, one for charging trajectory and one for releasing trajectory, are established for each phase change unit within the phase change control unit. The valve position target and throttling section of each branch node are mapped to an executable opening trajectory. When the control clock reaches the charging start time of the phase change unit, the phase change unit and the cooling plate branch form a controlled heat exchange loop, and the heat exchange interface flux is gradually increased with the configured target heat exchange intensity. If the historical record of the phase change unit shows that it is close to the material phase change plateau, a slower increase is used in the initial charging stage, crossing the plateau and then returning to the target intensity to reduce short-term energy flow pulsations. When the control clock reaches the releasing start time, the phase change unit is switched to the cooling path near the low-temperature region or to the upstream of the cathode air branch, as configured, to release the stored heat and complete energy introduction in conjunction with the path throttling section. The controllable bypass dynamic management drives the opening and closing of the branch valve position according to the opening and closing thresholds of the bypass path parameters, and inserts a buffer period corresponding to the sequence preservation identifier before and after the path switching to perform gradual entry and exit processing on the opening trajectory, ensuring that the path switching does not compete with the throttling configuration of the branch node at the cooling plate inlet; during the bypass open period, the minimum maintenance flow rate is maintained to remove condensate or slowly release local subcooling; during the bypass closed period, the opening is returned to the minimum discharge position to ensure that the path can be re-engaged at any time. During the phase change and bypass linkage execution process, the shell temperature, interface temperature, and cumulative energy count of each phase change unit are periodically read. The remaining chargeable capacity and remaining releaseable capacity of the phase change unit are calculated, and the smaller of the two values ​​is used to define the available capacity ratio of the phase change unit in this cycle. The available capacity ratio of each phase change unit is weighted according to its target division of labor in the configuration parameters to obtain the thermal storage availability for this cycle. At the same time, the opening trajectory and switching time of each branch node in this cycle are compressed into an instruction sequence to obtain the bypass execution instruction. After generation, both are written into the energy flow constraint buffer and the timing coordination buffer for reference in the next step, and the source field is marked with the configuration parameters of the phase change thermal storage unit and the bypass path parameters on which it depends. The input of the above process is the configuration parameters of the phase change thermal storage unit and the bypass path parameters, the processing is phase change thermal storage and controllable bypass dynamic management, and the output is thermal storage availability and bypass execution instruction.

[0091] S530. Perform energy flow allocation and timing arrangement on the thermal storage availability and the bypass execution command to generate thermal storage and bypass timing parameters;

[0092] Specifically, the available heat storage capacity is read from the energy flow constraint buffer, the bypass execution command is read from the timing coordination buffer, and the upper limit of coolant flow and transition period identifier corresponding to the current cycle are synchronously read from the buffer related to coupled heat exchange and activation to ensure that the energy flow allocation is consistent with the existing timing. During the allocation phase, the upper limit of energy that the phase change unit can undertake in this cycle is first limited by the available heat storage capacity, and then the energy boundary that the bypass path can import or release in this cycle is limited by the opening trajectory and switching time in the bypass execution command. Subsequently, within the section covered by the transition period identifier, time windows are reserved for phase change charging and phase change releasing respectively, and the instruction segments of bypass opening and closing are aligned. After alignment, the share of each energy flow channel is allocated by time slice as the unit of measurement so that there is no contradiction between phase change charging, phase change releasing and bypass import in the same time slice. If a conflict occurs, the adjacent time slices are fine-tuned according to the order of the sequence identifier until the conflict is eliminated. In the timing arrangement phase, the allocated energy flow share is mapped to phase change charge identifiers, phase change release identifiers, and bypass opening targets for different time periods. Boundary labels consistent with transition period identifiers are added to each time period, enabling subsequent valve and pump control to perform gradual increases and decreases at the boundaries. Simultaneously, the time-segment master table containing phase change charge and release identifiers is compressed into two types of event sequences, and the time-segment master table corresponding to the bypass opening target is compressed into one type of event sequence. These three types of event sequences together constitute the heat storage and bypass timing parameters. After generation, these heat storage and bypass timing parameters are written into the bypass management buffer and the optimization constraint buffer. The bypass management buffer allows subsequent cycles to directly call the end event of the current cycle as an initial value before execution. The optimization constraint buffer is used by the load prediction joint closed-loop optimization control in step 600 to incorporate the prior occupancy and switching order of phase change and bypass in the joint solution of pump speed command, three-way valve opening command, and heat flow distribution command, forming boundary conditions consistent with the global solution. The inputs to the above process are the thermal storage availability and the bypass execution command, the processing is energy flow allocation and timing arrangement, and the output is thermal storage and bypass timing parameters. Understandably, through the continuous operation of steps S510 to S530, the temperature uniformity index and the activation temperature achievement identifier are mapped and transmitted as phase change thermal storage unit configuration parameters and bypass path parameters, further executed and quantified as thermal storage availability and bypass execution commands, and finally allocated and solidified as thermal storage and bypass timing parameters. This constitutes a closed-loop link based on spatial and temporal criteria, with energy flow and commands as the core, achieving overall timing consistency and data connectivity between phase change thermal storage, controllable bypass, and subsequent valve and pump optimization.

[0093] Step S600 includes at least steps S610-S630:

[0094] S610. Obtain the heat storage and bypass timing parameters and the temperature uniformity index, construct the load prediction model and generate the reference trajectory to obtain the load prediction reference trajectory parameters.

[0095] Specifically, the heat storage and bypass timing parameters consistent with the current control cycle are read from the optimization constraint buffer, and the temperature uniformity index is read synchronously from the temperature uniformity evaluation buffer. After time stamp verification, the two are used to establish an input set. To ensure continuity with the preceding steps, during the input set assembly process, the waste heat recovery ratio of the previous cycle is retrieved from the energy metering unit associated with S330, the preheating coolant temperature of the previous cycle is retrieved from the heat exchange control unit associated with S320, and the branch flow ratio parameter is retrieved from the flow distribution buffer generated by S430 to complete the prior characterization of the heat side and flow side. In the load prediction model construction stage, the temperature uniformity index is used to characterize the adjustability of the internal temperature distribution of the stack, and the heat storage and bypass timing parameters are used to characterize the relationship between phase change charging, phase change releasing, and bypass opening and closing on the available energy and channel available time of the heat side. Combined with the branch flow ratio parameter, waste heat recovery ratio, and preheating coolant temperature, a parameterized structure including temperature field state, valve and pump input, and heat storage bypass disturbance is established, and each input and disturbance is aligned with time slice as the basic granularity. Subsequently, reference trajectory generation is performed: First, the desired range and variation rhythm of the inlet and outlet temperature difference are set based on the temperature uniformity index. Then, the event sequence in the heat storage and bypass timing parameters is projected into heat-side importable and non-importable windows, and compatibility trimming is performed with the proportional boundary of the branch flow ratio parameter to form a target temperature difference trajectory that satisfies the time window and proportional boundary. At the same time, the target temperature trajectory and basic flow trajectory on the coolant side are given based on the waste heat recovery ratio and the preheating coolant temperature, and a switching buffer segment consistent with the bypass opening and closing is set for the three-way valve opening trajectory. The above-mentioned target temperature difference trajectory, target temperature trajectory, basic flow trajectory, and three-way valve opening switching buffer segment are uniformly encapsulated, along with their corresponding time identifiers and window identifiers, to generate load prediction reference trajectory parameters, which are written into the solution buffer for direct reading in the next step. Among them, the window identifier and switching buffer segment in the load prediction reference trajectory parameters are inherited from the heat storage and bypass timing parameters to ensure that the subsequent joint closed-loop optimization control follows the same time boundary during the solution.

[0096] S620. Perform load prediction joint closed-loop optimization control on the load prediction reference trajectory parameters to generate pump speed command and three-way valve opening command.

[0097] Specifically, load prediction reference trajectory parameters are read sequentially from the solution buffer, and the thermal storage availability and bypass execution instructions generated by S520 are loaded from the energy flow constraint buffer. Simultaneously, the coupled path configuration parameters generated by S310 are read from the path execution buffer to constrain the solution space. To ensure consistency with prior occupancy, the window identifiers and switching buffer segments in the reference trajectory are aligned with the bypass execution instructions one by one. The lower limit of the three-way valve opening is frozen within the bypass open window, and the upper limit of the opening is frozen within the bypass closed window. Then, the thermal storage availability is mapped to the upper and lower bounds of the pump speed variation, allowing the pump speed to increase flux within the phase change release window and limiting flux within the phase change charging window. The process then proceeds to the joint closed-loop optimization control solution: using the target temperature difference trajectory and the coolant side target temperature trajectory as references, and combining the upper limit of flow rate and the number of stages that can be engaged in the coupling path configuration parameters, candidate pump speed sequences and candidate three-way valve opening sequences are generated on a rolling basis over time slices; within each time slice, real-time temperature and flow feedback at the inlet and outlet are collected, the deviation from the reference trajectory is calculated, and the deviation, along with the window identifier, is sent to the correction unit. If the deviation falls into the switching buffer section, the transition amplitude of the candidate sequence is suppressed by the principle of gradual increase and decrease; if the deviation occurs at the bypass opening and closing boundary, the candidate sequences of adjacent time slices are finely adjusted according to the sequence preservation identifier to ensure that the valve and pump actions do not conflict with the bypass switching. After completing the rolling correction for all time slices, the pump speed command and the three-way valve opening command are output and packaged with time identifiers. Among them, the pump speed command prioritizes the common range of the target temperature difference trajectory and the target temperature trajectory on the coolant side, and the three-way valve opening command prioritizes the opening and closing boundaries and switching buffer segments of the bypass execution command. The two types of commands are written into the command release buffer before being dequeued, and the trajectory segment index of their reference source is recorded simultaneously for use in the next step of heat flow distribution solution and command issuance, ensuring the consistency of reference-solution-release.

[0098] S630. Based on the pump speed command, the three-way valve opening command, and the heat storage and bypass timing parameters, perform heat flow distribution solution and command issuance to generate heat flow distribution command;

[0099] Specifically, the pump speed command and three-way valve opening command within the same control cycle are read from the command issuance buffer, and the heat storage and bypass timing parameters are read synchronously from the optimization constraint buffer. A solution input set is established in the heat flow distribution unit. First, the event sequence of the three-way valve opening command and the heat storage and bypass timing parameters are aligned to determine the available channel set of the radiator branch, heat storage branch, and cooling plate branch in each time slice. The allowed import section is marked for the cathode air side high temperature difference heat exchanger in the bypass open window. Then, the time slice distribution of the total coolant flux is calculated based on the pump speed command, and the flux is allocated as branch candidate flux according to the channel set. If the candidate flux overlaps with the energy window of phase change charging or phase change release in a certain time slice, the allocation of the occupied window is reserved first according to the switching sequence preservation mark formed by S530, and the candidate flux of the remaining branches is reduced. After allocation, the candidate flux and flow limit, proportional boundary and switching buffer of each branch are uniformly trimmed to obtain the target flux of each branch in each time slice, and mapped to heat flow allocation instructions. In the instruction issuance stage, the heat flow allocation instructions are written to the valve pump actuator and the throttling component of the branch node in batches according to the time identifier, ensuring that they are executed synchronously with the pump speed instructions and the three-way valve opening instructions in the same control cycle. At the same time, the target flux before execution and the flow feedback and temperature feedback after execution are bound one by one and stored in the operation record buffer, which is used as one of the sources for data synchronization and anomaly removal in the S100 stage of the next control cycle. Furthermore, in order to maintain the closed-loop consistency with the preceding module, the time boundary of the issued heat flow allocation instructions is synchronized back to the timing coordination buffer, and the allocation segment related to bypass opening and closing is synchronized back to the bypass management buffer, so that the end state of this cycle can be directly reused as the initial value in subsequent cycles from S510 to S530. Understandably, through the continuous operation of steps S610 to S630, the load prediction reference trajectory parameters are solved into pump speed commands and three-way valve opening commands, and under the constraints of heat storage and bypass timing parameters, they are transformed into executable heat flow distribution commands, realizing the alignment and closed-loop release of reference trajectory, prior occupancy and real-time feedback within the same control cycle.

[0100] The cold start thermal management method for the aforementioned extreme cold hydrogen fuel cell drone can be divided into three main stages according to its inherent logic and temperature evolution process: preheating stage, self-sustaining heating stage, and steady-state operation and dynamic thermal management stage. In the preheating stage, when the control unit detects that the ambient temperature is lower than a preset threshold (e.g., -20°C), it activates the positive temperature coefficient heating element in the integrated heating cathode cavity to heat the air entering the fuel cell stack. It can also selectively activate the auxiliary heater in the coolant circuit to jointly heat the fuel cell stack body. Simultaneously, it controls the hydrogen supply system to circulate at low pressure and low flow rate to purge moisture from the anode. During the self-sustaining heating phase, when the average temperature of the fuel cell stack, calculated based on the inlet and outlet temperatures, rises to the first range (e.g., -5°C to 0°C), the control unit instructs the fuel cell stack to begin small-current discharge. In this phase, the fuel cell stack's own reaction heat and positive temperature coefficient (PTC) heating work synergistically, and the high-temperature differential heat exchanger transfers waste heat from the coolant circuit to the intake air, achieving internal energy recycling and causing the fuel cell stack temperature to rapidly rise to the activation temperature (e.g., above 10°C). During steady-state operation and dynamic thermal management, after the fuel cell stack enters normal power output, the PTC heating is shut off. The control unit, based on the load current and the internal temperature distribution of the fuel cell stack obtained by the microchannel temperature equalization control, dynamically adjusts the speed of the coolant pump and the opening of the three-way valve through load prediction combined closed-loop optimization control to control the coolant flow rate through the radiator, thereby maintaining the fuel cell stack temperature within the optimal operating range. Simultaneously, the heat storage material module absorbs excess heat during operation and releases heat when the load suddenly decreases, preventing water vapor condensation and freezing due to a sudden temperature drop.

[0101] Example 2: Figure 2 A structural block diagram of a cold-start thermal management system for an extreme cold-weather hydrogen fuel cell drone according to an embodiment of the present invention is shown. Figure 2 As shown, the structure may include:

[0102] The data acquisition and synchronization unit 01 is used to acquire environmental, fuel cell stack, intake, and duct data, and perform synchronous cleaning and fusion to output raw environmental, fuel cell stack, intake, and duct data. This raw data is provided to the perception fusion, threshold calculation, cold start decision-making, and parameter generation unit. Specifically, it receives time-stamped data streams from the environmental acquisition channel, fuel cell stack acquisition channel, intake acquisition channel, and duct acquisition channel. Alignment and missing data removal are performed according to a unified time base. Noise filtering and range verification are completed according to source labels, forming raw environmental, fuel cell stack, intake, and duct data grouped by channel. During data synchronization, the time stamp and source label of each batch are recorded, generating a corresponding raw data index. The raw environmental, fuel cell stack, intake, and duct data and the raw data index are written to the raw data cache and transmitted to the perception fusion, threshold calculation, cold start decision-making, and parameter generation unit as input. Simultaneously, the batch number and status are registered in the raw data log for subsequent retrieval and traceability.

[0103] The perception fusion and threshold calculation and cold start decision and parameter generation unit 02 is used to perform perception fusion and threshold calculation on the raw data, generate an icing risk index, and generate a target intake temperature and an initial heat flow allocation limit. The icing risk index and the target intake temperature are provided to the partition lighting and purging configuration and in-situ heating anti-icing and duty statistics and timing summary unit. The initial heat flow allocation limit is provided to the coupling path configuration and high temperature difference heat exchange and coolant preheating unit. Specifically, the system loads raw environmental, fuel cell, intake, and duct data from the raw data cache, performs frame-level alignment according to time signatures, removes abnormal segments and prunes intervals, and constructs a multi-channel fusion dataset. Based on the threshold table and trigger conditions, the system performs segment determination and critical boundary comparison on the fusion dataset, outputs the icing risk index, and generates the target intake temperature and initial heat flow allocation upper limit under the same time base. The icing risk index and target intake temperature are registered in the timing queue and transmitted to the partition lighting and purging configuration, in-situ heating anti-icing, duty cycle statistics, and timing summary unit. The initial heat flow allocation upper limit is registered in the parameter queue and transmitted to the coupling path configuration, high temperature difference heat exchange, and coolant preheating unit. At the same time, the current round index and judgment basis are stored in the decision record.

[0104] The partition lighting and purging configuration, in-situ heating anti-icing, duty cycle statistics and timing summary unit 03 is used to perform partition lighting calculation and purging rhythm configuration based on the icing risk index and the target intake air temperature, drive the cathode conformal positive temperature coefficient heating element for in-situ heating and anti-icing, generate compliant intake air temperature and humidity and output duty cycle information. The compliant intake air temperature and humidity and the duty cycle information, together with the initial heat flow distribution upper limit, are provided to the coupling path configuration and high temperature difference heat exchange and coolant preheating unit. Specifically, the system receives the icing risk index and target inlet air temperature, performs partition lighting calculations under the registered flow channel partition mapping, and generates partition opening and closing sequences and power ratios. Based on the purging strategy table, it configures purging rhythm parameters and switching intervals, drives the cathode conformal positive temperature coefficient heating element and purging actuators to execute synchronously according to the partition opening and closing sequence and purging rhythm, collects inlet air temperature and humidity in real time, and performs threshold determination to form compliant inlet air temperature and humidity. It performs duty cycle statistics and timing summaries on the partition opening and closing sequence, outputs duty cycle information, and writes the compliant inlet air temperature and humidity, duty cycle information, and initial heat flow allocation upper limit into the heat exchange input buffer for use by the coupling path configuration and high temperature difference heat exchange and coolant preheating units. Simultaneously, it registers the sequence number and execution segment in the partition and purging logs.

[0105] The coupling path configuration and high temperature difference heat exchange and coolant preheating unit 04 is used to perform coupling path configuration and flow limit constraints based on the qualified inlet air temperature and humidity and duty information and the initial heat flow distribution limit, implement high temperature difference coupled heat exchange and coolant preheating, and output the preheated coolant temperature and waste heat recovery ratio. The preheated coolant temperature and the waste heat recovery ratio are provided to the activation discharge and microchannel temperature equalization control and throttling configuration unit. Specifically, the system reads the compliant inlet temperature and humidity, duty cycle information, and initial heat flow allocation limit from the heat exchange input buffer. It then generates a coupling path configuration based on the registered heat exchanger coupling relationships in the path library, and sets the flow limit and switching boundary for each path according to the flow limit table. After the coupling path is activated, it performs coupled heat exchange on the high-temperature differential section, preheats the coolant circuit, writes the coolant inlet and outlet temperatures into the temperature record, calculates the preheated coolant temperature, and registers the waste heat recovery ratio according to the heat exchange section energy count. The preheated coolant temperature and waste heat recovery ratio are then transmitted to the activation discharge and microchannel temperature equalization control and throttling configuration unit. Simultaneously, the coupling path configuration and flow limit are registered in the path execution record for subsequent use.

[0106] The activation discharge and microchannel temperature equalization control and throttling configuration unit 05 is used to set activation discharge parameters and perform safety checks based on the preheating coolant temperature and waste heat recovery ratio, perform low-current activation and microchannel temperature equalization control, calculate the branch flow ratio and implement throttling configuration, and output temperature uniformity index and activation temperature achievement indicator. The temperature uniformity index and activation temperature achievement indicator are provided to the phase change heat storage and controllable bypass and timing arrangement unit. Specifically, the system receives the preheating coolant temperature and waste heat recovery ratio, sets activation discharge parameters according to the activation parameter library and performs safety checks, and triggers a small current discharge process according to the parameters; it collects temperature distribution data at the inlet and outlet of the microchannel cooling plate and compares it with the activation stage flux records, and performs uniform temperature control and local throttling adjustment; it generates branch flow ratios based on path execution records and flow limits, writes them into the throttling configuration table and sends them to relevant valves; it performs regional statistics and alignment of stage temperature distribution, outputs temperature uniformity indicators, and outputs activation temperature achievement indicators in the activation stage end determination; it transmits temperature uniformity indicators and activation temperature achievement indicators to the phase change heat storage and controllable bypass and timing arrangement unit, and registers parameters and execution status in the activation and throttling logs.

[0107] The phase change thermal storage and controllable bypass timing arrangement unit 06 is used to select phase change thermal storage units and determine bypass paths based on temperature uniformity indicators and activation temperature achievement indicators, perform dynamic management and timing arrangement of phase change thermal storage and controllable bypass, and output thermal storage and bypass timing parameters. These timing parameters and the temperature uniformity indicators are provided to the load prediction, reference trajectory, and joint closed-loop optimization unit. Specifically, it receives the temperature uniformity indicators and activation temperature achievement indicators, completes the selection of phase change thermal storage units and determination of bypass paths based on the phase change unit list and path list, and establishes a timing table for charging events, releasing events, and opening / closing events. During the phase change and bypass execution process, it records the start and end times of events, valve position targets, and switching buffer segments to form thermal storage and bypass timing parameters, and writes the corresponding spatial association information and time identifiers into the timing record. The thermal storage and bypass timing parameters, along with the temperature uniformity indicators, are transmitted to the load prediction, reference trajectory, and joint closed-loop optimization unit, while retaining the phase change and bypass execution log for subsequent alignment.

[0108] The load prediction and reference trajectory and joint closed-loop optimization unit 07 is used to construct a load prediction model and generate a reference trajectory based on the heat storage and bypass timing parameters and temperature uniformity index, execute load prediction joint closed-loop optimization control, and output pump speed commands and three-way valve opening commands. The pump speed commands and the three-way valve opening commands, together with the heat storage and bypass timing parameters, are provided to the heat flow distribution and command issuance unit. Specifically, it receives the heat storage and bypass timing parameters and temperature uniformity index, assembles the load prediction input set according to the time identifier, constructs the load prediction model and generates a reference trajectory, reads the window identifier, switching buffer segment and path constraints, executes joint closed-loop optimization control, and forms time-segmented pump speed commands and time-segmented three-way valve opening commands; writes the two types of commands into the command issuance buffer and records the trajectory segment index and time identifier, and at the same time, rewrites the heat storage and bypass timing parameters to the constraint segment of the same control cycle for the heat flow distribution and command issuance unit to call.

[0109] The heat flow distribution and command issuance unit 08 is used to perform heat flow distribution solution and command issuance based on pump speed command, three-way valve opening command, and heat storage and bypass timing parameters, and output heat flow distribution command. Specifically, it receives pump speed command, three-way valve opening command, and heat storage and bypass timing parameters, aligns the availability of channels in each time period according to the event queue, completes branch candidate flux allocation and boundary pruning, and forms branch target flux sequence; encapsulates the branch target flux sequence into heat flow distribution command and writes it into the execution interface, collects execution status, flow feedback and temperature feedback, registers them in the operation record, and sends the execution status back to the data acquisition and data synchronization unit and the perception fusion, threshold calculation and cold start decision and parameter generation unit for recording and updating.

[0110] Figure 3 A schematic diagram of a thermal management system for a hydrogen fuel cell drone in extremely cold environments is provided as an embodiment of this application. Figure 3As shown, the system uses a fuel cell stack as its core power generation unit. Its air passage is connected to the integrated heated cathode cavity, and its cooling passage is integrated with a manifold-type coolant plate. The coupled thermal management and air supply unit forms the critical path connecting the cathode intake air and the coolant circuit. The integrated heated cathode cavity (PTC) is used for rapid heating of the intake air, the manifold-type coolant plate is used to achieve uniform temperature within the stack, a high-temperature differential heat exchanger is connected between the coolant circuit and the cathode air circuit to achieve waste heat recovery, and a controllable bypass branch provides a flow path option that bypasses the radiator. A heat storage material module is arranged in this bypass branch or around the stack for heat exchange with the system. The control unit is connected to sensors for temperature, pressure, and humidity via signal lines to monitor the system status in real time; simultaneously, it is connected to actuators such as the PTC heating element, coolant pump, and three-way valve via control lines, thus forming a complete monitoring-control-execution closed loop. The schematic diagram clearly reveals the physical implementation path and energy and signal flow directions of each step, including synchronous cleaning and fusion, partitioned lighting calculation, high temperature difference coupling heat exchange, phase change heat storage and controllable bypass dynamic management, and load prediction combined closed-loop optimization control.

Claims

1. A cold start thermal management method for an extremely cold-weather hydrogen fuel cell drone, characterized in that, include: The system acquires ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point, and duct water retention signals, performs synchronous cleaning and fusion, and outputs icing risk index, target inlet temperature, and initial heat flow distribution upper limit. Obtain the icing risk index and target intake temperature, perform zone lighting calculation and purging rhythm configuration, perform in-situ heating and anti-icing of cathode conformal positive temperature coefficient heating elements, and output qualified intake air temperature and humidity and duty information. The execution of the partition lighting calculation specifically includes: Perform lighting sorting calculations from upstream to downstream to generate lighting parameters for each zone, consisting of the start time, duration, and stop time. During the calculation process, the time domain identifier of the duct water stagnation signal is input, and the zone in the water stagnation indication state is set with delayed lighting and extended airflow flushing period before stop. Obtain the inlet air temperature, humidity and occupancy information to meet the standards, configure the coupling path and implement high temperature difference coupling heat exchange and coolant preheating, and generate the preheated coolant temperature and waste heat recovery ratio; Obtain the preheating coolant temperature and waste heat recovery ratio, set activation discharge parameters and perform safety checks, low-current activation and microchannel temperature control, calculate branch flow ratio and execute throttling configuration, and generate temperature uniformity index and activation temperature achievement mark. Based on the temperature uniformity index and activation temperature achievement, the charging window and release window are constructed according to the configuration parameters of the phase change thermal storage unit, and the open-circuit window and closed-circuit window are constructed according to the bypass path parameters. The phase change thermal storage unit is selected and the bypass path is determined to generate the configuration parameters of the phase change thermal storage unit and the bypass path parameters. The phase change thermal storage and controllable bypass dynamic management is executed to generate the thermal storage availability and bypass execution command, and the energy flow allocation and timing arrangement are performed to generate the thermal storage and bypass timing parameters. Acquire heat storage and bypass timing parameters, construct a load prediction model and generate a reference trajectory, execute load prediction joint closed-loop optimization control and heat flow distribution solution and command issuance, and generate pump speed command, three-way valve opening command and heat flow distribution command.

2. The method according to claim 1, characterized in that, Ambient temperature, ambient humidity, fuel cell inlet and outlet temperatures, cathode inlet dew point, and duct water retention signals include: The ambient temperature refers to the sequence of instantaneous temperature data continuously collected by an ambient temperature sensor deployed on the exterior of the UAV and directly exposed to the flight airspace; the ambient humidity refers to the sequence of instantaneous relative humidity data continuously collected by an ambient humidity sensor deployed on the exterior of the UAV and directly exposed to the flight airspace; the fuel cell stack inlet and outlet temperatures include inlet temperature data collected by a temperature detection element installed on the inlet side of the fuel cell stack and outlet temperature data collected by a temperature detection element installed on the outlet side of the fuel cell stack; the cathode inlet dew point refers to the inlet temperature and humidity data obtained by a temperature and humidity detection element installed in the cathode inlet pipeline; the duct water stagnation signal refers to a discrete or continuous state signal generated by a water stagnation detection element deployed in the airflow channel of the fuel cell system, used to indicate whether there is residual liquid water in the channel.

3. The method according to claim 1, characterized in that, The process of performing phase change thermal storage and controlled bypass dynamic management to generate thermal storage availability and bypass execution commands also includes: The system reads the configuration parameters and bypass path parameters of the phase change thermal storage unit, establishes two types of timing tables for each phase change unit in the phase change control unit: charging trajectory and releasing trajectory, and maps the valve position target and throttling section of each branch node to an executable opening trajectory. When the control clock reaches the charging start time of the phase change unit, the phase change unit and the cooling plate branch form a controlled heat exchange loop, and the heat exchange interface flux is gradually increased according to the configured target heat exchange intensity. When the control clock reaches the releasing start time, the phase change unit is switched to the cooling path near the low temperature region or to the upstream of the cathode air branch according to the configuration, so as to release the stored heat and complete the energy introduction in conjunction with the throttling section of the path.

4. The method according to claim 1, characterized in that, Controllable bypass dynamic management includes: The branch valve position is driven to open and close according to the opening and closing thresholds of the bypass path parameters. A buffer period corresponding to the sequence preservation identifier is inserted before and after the path switching to perform gradual entry and exit processing on the opening trajectory. During the bypass open period, the minimum maintenance flow is maintained to remove condensate or slowly release local subcooling. During the bypass closed period, the opening is returned to the minimum discharge position to ensure that the path can be reactivated at any time.

5. The method according to claim 1, characterized in that, The process of generating thermal storage availability also includes: Read the shell temperature, interface temperature and cumulative energy count of each phase change unit, calculate the remaining chargeable capacity and remaining releaseable capacity of the phase change unit, and define the available capacity ratio of the phase change unit in this cycle by the smaller of the two values; weight the available capacity ratio of each phase change unit according to its target division of labor in the configuration parameters to obtain the thermal storage availability of this cycle.

6. The method according to claim 1, characterized in that, The process of performing load prediction joint closed-loop optimization control and heat flow distribution solution and command issuance also includes: Load prediction reference trajectory parameters are read sequentially from the solution buffer, and thermal storage availability and bypass execution instructions are loaded from the energy flow constraint buffer. At the same time, coupled path configuration parameters are read from the path execution buffer to constrain the solution space. The window identifiers and switching buffer segments in the reference trajectory are aligned with the bypass execution instructions one by one. The lower limit of the three-way valve opening is frozen in the bypass open window, and the upper limit of the opening is frozen in the bypass closed window. The thermal storage availability is mapped to the upper and lower bounds of the pump speed variation.

7. The method according to claim 1, characterized in that, The process of solving the joint closed-loop optimization control also includes: Combining the upper limit of flow rate and the number of stages that can be deployed in the coupling path configuration parameters, candidate pump speed sequences and candidate three-way valve opening sequences are generated on a rolling basis over time slices. In each time slice, real-time temperature and flow feedback at the inlet and outlet are collected, the deviation from the reference trajectory is calculated, and the deviation, along with the window identifier, is sent to the correction unit. After completing the rolling correction for all time slices, pump speed commands and three-way valve opening commands are output and packaged with time identifiers.

8. The method according to claim 7, characterized in that, The operation of the calibration unit also includes: If the deviation falls into the switching buffer section, the transition amplitude of the candidate sequence is suppressed by the principle of gradual rise and fall; if the deviation occurs at the bypass opening and closing boundary, the candidate sequences of adjacent time slices are finely adjusted according to the order preservation mark.

9. The method according to claim 7, characterized in that, The pump speed command and the three-way valve opening command are written into the command release buffer, and the trajectory segment index and time stamp are recorded synchronously.

10. A cold start thermal management system for an extreme cold hydrogen fuel cell drone, applied to the method described in any one of claims 1-9, characterized in that, include: The data acquisition and data synchronization unit is used to collect environmental, fuel cell stack, air intake, and air duct data, and to perform synchronous cleaning and fusion, outputting raw environmental, fuel cell stack, air intake, and air duct data. The perception fusion and threshold calculation and cold start decision and parameter generation unit is used to perform perception fusion and threshold calculation on the raw data, generate the icing risk index, and generate the target intake temperature and the upper limit of the initial heat flow distribution. The partition lighting and purging configuration, in-situ heating anti-icing and duty statistics and timing summary unit is used to perform partition lighting calculation and purging rhythm configuration based on the icing risk index and target intake air temperature, drive the cathode conformal positive temperature coefficient heating element to perform in-situ heating and anti-icing, generate compliant intake air temperature and humidity and output duty information. The coupling path configuration and high temperature difference heat exchange and coolant preheating unit is used to perform coupling path configuration and flow limit constraints based on the qualified inlet temperature and humidity, duty information and initial heat flow distribution limit, implement high temperature difference coupled heat exchange and coolant preheating, and output the preheated coolant temperature and waste heat recovery ratio. The activation discharge and microchannel temperature equalization control and throttling configuration unit is used to set activation discharge parameters and perform safety checks based on the preheating coolant temperature and waste heat recovery ratio, perform low-current activation and microchannel temperature equalization control, calculate the branch flow ratio and implement throttling configuration, and output temperature uniformity index and activation temperature achievement mark. The phase change thermal storage and controllable bypass and timing arrangement unit is used to select phase change thermal storage units and determine bypass paths based on temperature uniformity index and activation temperature achievement mark, perform dynamic management and timing arrangement of phase change thermal storage and controllable bypass, and output thermal storage and bypass timing parameters. The load prediction and reference trajectory and joint closed-loop optimization unit is used to construct a load prediction model and generate a reference trajectory based on the heat storage and bypass timing parameters and temperature uniformity index, execute load prediction joint closed-loop optimization control, and output pump speed command and three-way valve opening command. The heat flow distribution and command issuance unit is used to perform heat flow distribution solution and command issuance based on pump speed command, three-way valve opening command and heat storage and bypass timing parameters, and output heat flow distribution command.

Images