Fuel cell-lithium battery hybrid power dynamic distribution system and medium

Abstract

The invention relates to the technical field of fuel cell hybrid power supply, in particular to a fuel cell-lithium battery hybrid power dynamic distribution system and a medium, and the system comprises a fuel cell stack, a lithium battery energy storage module, a three-way valve cooling subsystem, a dynamic power compensation module and a central processing unit. Dynamically controlling a cooling path, selectively preheating a lithium battery energy storage module through a heat exchanger using fuel cell waste heat data, and monitoring bus voltage, load current, and battery state, performing power distribution based on thermal constraints, and when an overheating risk is predicted, performing power distribution on the lithium battery energy storage module. And dynamically redistributing the power request from the fuel cell to the lithium battery energy storage module, and during power distribution calculation, optimizing the operation mode of the cooling system and the corresponding parasitic energy consumption as dynamic variables to generate an optimized power distribution scheme. The service life of the battery is prolonged through intelligent thermal management and dynamic power distribution.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cell hybrid power supply, and in particular to a fuel cell-lithium battery hybrid power dynamic distribution system and medium. Background Art

[0002] Fuel cell hybrid systems, as efficient and clean energy solutions, show broad application prospects in areas such as backup power for communication base stations, data centers, and new energy vehicles. These systems typically use a fuel cell stack as the primary power source, supplemented by energy storage units such as lithium batteries to cope with load fluctuations and enhance the system's dynamic response capabilities.

[0003] However, when the load power exceeds the set output of the fuel cell, the lithium battery unconditionally supplements the power difference. This method ignores the state of the lithium battery itself, especially when its state of charge is low. Frequent or deep discharge will cause irreversible damage to its internal chemical structure, significantly exacerbating the degradation of the electrode material and causing a rapid decrease in capacity. In addition, the aging rate of lithium batteries varies greatly in different state of charge ranges, temperatures and discharge depths. Traditional control systems often sacrifice the long-term service life of the battery in order to meet instantaneous power requirements.

[0004] Therefore, this mode of operation directly leads to premature aging and performance degradation of the battery, which not only increases the system's life cycle cost, but also reduces the long-term reliability of the entire hybrid system.

[0005] How to optimize the power distribution strategy in the hybrid system and, under the premise of ensuring stable system operation, minimize the working range that may damage the lithium battery, thereby effectively slowing down the battery aging rate and extending its service life, is a key technical issue that needs to be urgently addressed in the current field of fuel cell hybrid technology.

[0006] Therefore, a fuel cell-lithium battery hybrid power dynamic distribution system and medium are proposed. Summary of the Invention

[0007] The purpose of the present invention is to provide a fuel cell-lithium battery hybrid power dynamic distribution system and medium, which extends the battery life through intelligent thermal management and dynamic power distribution. It includes a fuel cell stack, a lithium battery energy storage module, a three-way valve cooling subsystem, a dynamic power compensation module and a central processing unit. It uses the real-time temperature field data of the fuel cell stack to dynamically control the cooling path, uses the fuel cell waste heat data to selectively preheat the lithium battery energy storage module through a heat exchanger, and monitors the bus voltage, load current and battery status to perform power distribution based on thermal constraints. When the risk of overheating is predicted, the power request is dynamically reallocated from the fuel cell to the lithium battery energy storage module. When performing power distribution calculations, the operating mode of the cooling system and the corresponding parasitic energy consumption are optimized as dynamic variables to generate an optimized power distribution plan. The present invention also discloses a computer medium using a fuel cell-lithium battery hybrid power dynamic distribution system.

[0008] To achieve the above object, the present invention provides the following technical solutions: A fuel cell-lithium battery hybrid power dynamic distribution system, comprising: The fuel cell stack serves as the main power supply, and its output power is connected to the DC bus through a DC-DC converter; The lithium battery energy storage module is connected to the DC bus through a bidirectional DC-DC converter to provide transient power and store energy; The three-way valve cooling subsystem uses the temperature field data of the fuel cell stack to dynamically control the cooling path and selectively preheat the lithium battery energy storage module through the heat exchanger. The dynamic power compensation module monitors the bus voltage, load current, and battery status, and reallocates power between the fuel cell stack and the lithium battery energy storage module when it predicts that the fuel cell stack temperature exceeds a preset safety range; The central controller optimizes the parasitic energy consumption corresponding to the operating mode of the three-way valve cooling subsystem as a dynamic variable and generates an optimized allocation plan. The dynamic power compensation module executes the optimized allocation plan. The compensation mode is determined according to the real-time state of charge of the lithium battery energy storage module. The external load prediction system is connected to perform power prediction. When a high load event is predicted, the three-way valve cooling subsystem is instructed to start pre-cooling.

[0009] Preferably, the process of dynamically controlling the cooling path includes: Based on the real-time temperature field data obtained from the three-way valve cooling subsystem, the temperature gradient between each area is calculated; and according to the comparison result of the temperature gradient with the preset high threshold and the preset low threshold, the corresponding control instruction is generated. When the temperature gradient is greater than the high threshold, the enhanced cooling path is opened, and when the temperature gradient is less than the low threshold, it is switched to the energy-saving bypass circulation path.

[0010] Preferably, the control instructions are generated based on dynamic calculations of fuzzy weighted fusion of multiple basic control matrices, including: taking the state of charge of the lithium battery energy storage module and the current load power as inputs of the fuzzy controller, and calculating the dynamic fusion weights corresponding to each basic control matrix through a structured fuzzy rule base composed of multiple parallel sub-rule bases; the sub-rule base includes: an energy-saving strategy sub-rule base, a performance strategy sub-rule base and a balancing strategy sub-rule base; the temperature gradient calculated in real time is queried in all basic control matrices to obtain multiple groups of candidate instruction sets, and these candidate instruction sets are weighted averaged using the dynamic fusion weights independently calculated by each sub-rule base to calculate the control instructions.

[0011] Preferably, the selective preheating of the lithium battery energy storage module by passing the outlet coolant through the heat exchanger includes: monitoring the real-time temperature of the lithium battery energy storage module through the dynamic power compensation module; comparing the real-time temperature with a preset lower threshold value of the optimal operating temperature range of the lithium battery; when it is determined that the real-time temperature is lower than the lower threshold value, utilizing a waste heat recovery and utilization circuit to selectively guide the high-temperature coolant flowing out of the fuel cell stack through the heat exchanger to heat the lithium battery energy storage module until the temperature returns to within the optimal operating temperature range.

[0012] Preferably, the reallocation of power between the fuel cell stack and the lithium battery energy storage module includes: the central controller predicts the future temperature trajectory under the current power output through a predictive thermal model based on the temperature change rate of the fuel cell stack; when the predictive thermal model determines that the future temperature trajectory will reach the safety upper limit threshold within a predetermined time, the power allocation strategy based on thermal constraints is triggered, and the central controller generates and issues a collaborative control instruction, which includes a derating instruction for actively reducing the power output of the fuel cell stack and a compensation instruction for synchronously instructing the bidirectional DC-DC converter to increase the power drawn from the lithium battery energy storage module; the dynamic power compensation module generates a corrected control amount through a dynamic gain regulator according to the collaborative control instruction to execute thermal limitation on the fuel cell stack.

[0013] Preferably, the dynamic gain adjuster includes: An online system identification unit injects a pseudo-random binary sequence excitation signal into the control instructions of the bidirectional DC-DC converter; collects the excitation signal and the system's current response, and estimates the system's impulse response function in real time using a cross-correlation operation; performs a Fourier transform on the impulse response function to calculate a broadband dynamic impedance spectrum of the lithium battery energy storage module's current state; a state and disturbance decoupling unit, which compares the broadband dynamic impedance spectrum with a reference impedance spectrum of the lithium battery energy storage module in a baseline healthy state, quantifies the internal impedance deviation caused by battery state aging, and separates the external disturbance component caused by external load changes from the overall system response; A state feedback control law unit receives the internal impedance deviation and the external disturbance component, and calculates an original control instruction through a preset feedback and feedforward gain matrix; The control execution layer receives the original control instructions, updates them, and generates the final corrected control quantity.

[0014] Preferably, the process of optimizing the operating mode of the three-way valve cooling subsystem and the corresponding parasitic energy consumption as dynamic variables includes: determining the current operating mode of the three-way valve cooling subsystem; obtaining the corresponding parasitic energy consumption value from a multidimensional lookup table according to the operating mode; subtracting the parasitic energy consumption value from the total power generated by the system in real time to obtain the real-time system net output power as the optimization target; continuously monitoring the response changes of the real-time system net output power caused by the disturbance by injecting a periodic disturbance signal into the power allocation instruction of the system; adaptively adjusting the basic operating point of the power allocation by performing correlation demodulation and integration processing on the disturbance signal and the net power response change, and continuously tracking and converging to the current maximum value point of the real-time system net output power.

[0015] Preferably, the central controller adopts a hierarchical target optimization strategy when generating an optimized power distribution plan, including: when all system indicators are within the normal range, maximizing the system's net output power is the primary optimization goal, and power distribution is adjusted first to improve the overall operating efficiency; when a drastic change in load power greater than a preset threshold is detected, temporarily switching to maintaining DC bus voltage stability as the primary optimization goal, and prioritizing the lithium battery to perform rapid power compensation to ensure power quality; when the temperature or temperature gradient of the fuel cell stack is detected to exceed the safety threshold, temporarily switching to ensuring the thermal management safety of the stack as the primary optimization goal, and prioritizing derating and enhanced cooling.

[0016] Preferably, determining the compensation mode according to the real-time state of charge of the lithium battery energy storage module includes: classifying the real-time state of charge of the lithium battery energy storage module obtained from the dynamic power compensation module into one of a plurality of preset intervals including a forced charging interval, an adaptive compensation interval, and a limited power compensation interval, and determining the current compensation mode; generating a control instruction through a feedforward-feedback composite control algorithm to adjust the bidirectional DC-DC converter in the dynamic power compensation module.

[0017] Preferably, an active electromagnetic disturbance suppression module is further included, which collects the characteristic spectrum generated by the switching frequency and harmonic components of the bidirectional DC-DC converter in real time through a near-field probe arranged at the bidirectional DC-DC converter; when the disturbance amplitude at a specific frequency point is monitored to exceed a preset threshold, the active electromagnetic disturbance suppression module executes a dynamic spectrum optimization strategy based on a spread spectrum clock, periodically modulates the core switching frequency of the bidirectional DC-DC converter, broadens the concentrated spectrum energy and disperses it into a wider frequency band, thereby reducing the peak electromagnetic emission intensity.

[0018] A medium stores computer program instructions, which, when executed by a processor, implements the steps of a fuel cell-lithium battery hybrid power dynamic distribution system.

[0019] Compared with the prior art, the present invention has the following beneficial effects: 1. By dynamically controlling the cooling path, local overheating of the stack is effectively avoided, ensuring its long-term reliability. By employing a control command generation method based on fuzzy weighted fusion, the system can smoothly and continuously adjust the cooling strategy based on global conditions such as the lithium battery SOC and load power, rather than simply switching between thresholds, achieving a higher level of intelligence. Furthermore, the system utilizes waste heat from the fuel cell to selectively preheat the lithium battery, improving the performance of the energy storage unit in cold environments, achieving cascaded energy utilization, and enhancing the system's adaptability to all climates.

[0020] 2. This invention uses coordinated control instructions to reduce fuel cell power while simultaneously providing equal power compensation from the lithium battery, thereby achieving preventative protection for the fuel cell stack without interrupting power supply to the load. During this process, an advanced dynamic gain regulator employs adaptive control based on Lyapunov theory, ensuring extremely stable bus voltage during power transfer and guaranteeing high-quality power supply.

[0021] 3. This invention can online balance the three core objectives of system net output power (economy), bus voltage stability (power quality), and stack temperature uniformity (long-term life). To maximize net power, the system also uses the parasitic energy consumption of the cooling system as a dynamic variable, reducing system operating costs.

[0022] 4. This invention monitors the electromagnetic spectrum generated by the bidirectional DC-DC converter in real time and proactively implements a dynamic spectrum optimization strategy based on spread-spectrum clocking when disturbances exceed specified limits. By spreading the concentrated spectral energy, the system significantly reduces peak electromagnetic emission intensity and improves its electromagnetic compatibility, enabling safe and reliable deployment in locations with stringent electromagnetic environment requirements, such as communication base stations and data centers. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1A schematic structural diagram of a fuel cell-lithium battery hybrid power dynamic distribution system provided by an embodiment of the present invention; Figure 2 A schematic diagram of the allocation strategy flow of the dynamic power compensation module provided in an embodiment of the present invention; Figure 3 A schematic flow chart of a dynamic power compensation method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] See also Figures 1 to 3 The present invention provides a fuel cell-lithium battery hybrid power dynamic distribution system and medium, and the technical solution is as follows: A fuel cell-lithium battery hybrid power dynamic distribution system, the specific structure is as follows Figure 1 Shown, including: The fuel cell stack serves as the main power supply, and its output power is connected to the DC bus through a DC-DC converter; The lithium battery energy storage module is connected to the DC bus through a bidirectional DC-DC converter to provide transient power and store energy; The three-way valve cooling subsystem uses the temperature field data of the fuel cell stack to dynamically control the cooling path and selectively preheat the lithium battery energy storage module through the heat exchanger. The dynamic power compensation module monitors the bus voltage, load current, and battery status, and reallocates power between the fuel cell stack and the lithium battery energy storage module when it predicts that the fuel cell stack temperature exceeds a preset safety range; The central controller optimizes the parasitic energy consumption corresponding to the operating mode of the three-way valve cooling subsystem as a dynamic variable and generates an optimized allocation plan. The dynamic power compensation module executes the optimized allocation plan. The compensation mode is determined according to the real-time state of charge of the lithium battery energy storage module. The external load prediction system is connected to perform power prediction. When a high load event is predicted, the three-way valve cooling subsystem is instructed to start pre-cooling.

[0026] It also includes a three-way valve, a circulation pump, a heat dissipation assembly connected to the fuel cell stack, and temperature sensors arranged in different areas of the stack (for example, the water inlet, the water outlet, and at least two core reaction areas).

[0027] In this embodiment, the following control parameters are preset for the central controller: The high threshold of temperature gradient \(T_H = 5^{\circ}C\), the low threshold of temperature gradient \(T_L = 2^{\circ}C\), and the control period \(t_c = 200ms\).

[0028] Furthermore, the process of dynamically controlling the cooling path includes: Based on the real-time temperature field data obtained from the three-way valve cooling subsystem, calculate the temperature gradient between regions; and generate corresponding control instructions according to the comparison results of the temperature gradient with the preset high threshold and the preset low threshold. When the temperature gradient is not less than the high threshold, activate the enhanced cooling path, and when the temperature gradient is not greater than the low threshold, switch to the energy-saving bypass circulation path.

[0029] At the beginning of each control period (\(t_c\)), the central controller reads the values of the temperature sensors arranged at all monitoring points of the fuel cell stack in real time through the communication bus, obtains a complete and instantaneous snapshot of the temperature field data. The central controller processes the collected temperature field data, calculates the maximum temperature gradient in the entire stack at the current moment by comparing the temperature values between all monitoring points, and records it as \(\Delta T_{max}\).

[0030] The central controller compares the calculated \(\Delta T_{max}\) with the preset high and low thresholds and executes the following logical judgment: Case 1: If \(\Delta T_{max}\geq T_H(5^{\circ}C)\), it indicates that there is significant temperature non-uniformity inside the stack and there is a risk of local overheating. At this time, the controller determines that the target path is the enhanced cooling path. The controller generates a set of control instructions, instructing the three-way valve to switch to the path that directs the coolant completely to the heat dissipation assembly, and at the same time instructing the circulation pump to operate at a high speed to achieve maximum heat dissipation.

[0031] Case 2: If \(\Delta T_{max}\leq T_L(2^{\circ}C)\), it indicates that the temperature inside the stack is very uniform and the system is in a stable or low heat generation state. At this time, the controller determines that the target path is the energy-saving bypass circulation path. The controller generates a set of control instructions, instructing the three-way valve to switch to the bypass path that completely bypasses the heat dissipation assembly, and at the same time instructing the circulation pump to operate at a lower energy-saving speed to save energy consumption.

[0032] Case 3: If \(T_L < \Delta T_{max}< T_H\) (that is, the temperature difference is between \(2^{\circ}C\) and \(5^{\circ}C\)), it indicates that the system state is within the acceptable range. To avoid oscillations caused by frequent switching of the control system at the threshold boundary, the controller will maintain the control instructions of the previous cycle unchanged and keep the current cooling path and the pump speed.

[0033] The central controller sends the generated corresponding control instructions to the actuator of the three-way valve and the driver of the circulation pump through control signals to drive the hardware to complete path switching and speed adjustment.

[0034] After completing an instruction execution, the system waits for the arrival of the next control cycle (after 200ms) and repeats the entire process. Through this high-speed, continuous closed-loop control, the system ensures that the temperature gradient of the fuel cell stack is always controlled within the target range.

[0035] The dynamic cooling path control process also includes an adaptive stabilization mechanism to suppress control oscillations. This mechanism performs a lock operation after each cooling path switch to avoid frequent switching at threshold boundaries. The lock duration is correlated with the changing trend of the system's thermal state. When the central controller determines that the temperature gradient is changing rapidly, indicating an unstable thermal state, the lock duration is shortened to maintain faster re-responsiveness. Conversely, when the temperature gradient is changing slowly and the system is in a steady state, the lock duration is extended to prioritize operational smoothness and reduce actuator wear. This adaptive stabilization mechanism intelligently balances response speed and stability, extending hardware life while ensuring safe system response under critical operating conditions.

[0036] Through the method described in this embodiment, the present invention can perform real-time and intelligent cooling path selection according to the actual thermal state of the fuel cell stack, effectively preventing local overheating, improving temperature uniformity, and significantly reducing energy consumption when the system load is low.

[0037] Furthermore, the generation of the control instructions is based on dynamic calculation of fuzzy weighted fusion of multiple basic control matrices, including: taking the charge state of the lithium battery energy storage module and the current load power as the input of the fuzzy controller, and calculating the dynamic fusion weights corresponding to each basic control matrix through a structured fuzzy rule base composed of multiple parallel sub-rule bases; the sub-rule base includes: an energy-saving strategy sub-rule base, a performance strategy sub-rule base and a balancing strategy sub-rule base; the temperature gradient calculated in real time is queried in all basic control matrices to obtain multiple groups of candidate instruction sets, and these candidate instruction sets are weighted averaged using the dynamic fusion weights independently calculated by each sub-rule base to calculate the control instructions.

[0038] The central controller's fuzzy controller takes as input two continuously changing system state variables: the lithium battery energy storage module's state of charge and the current load power. Based on these inputs, the fuzzy controller's internal structured fuzzy rule base calculates the dynamic fusion weights corresponding to each fundamental matrix in real time, with the sum of all weights always being 1.

[0039] The logic of the weight calculation is determined by its internal sub-rule base, for example: The energy-saving strategy sub-rule base ensures that when the system SOC decreases or the load power decreases, the weight will increase accordingly; The performance strategy sub-rule base ensures that when the system SOC increases or the load power increases, the weight will increase accordingly; The balance strategy sub-rule base ensures that the weights dominate in the intermediate state.

[0040] The central controller calculates the maximum temperature gradient of the fuel cell stack, ΔT_max, in real time. Using this value, the controller queries all basic control matrices (energy saving, balance, and performance) in parallel, generating multiple independent candidate instruction sets representing different strategies.

[0041] The controller uses the calculated dynamic fusion weights to perform a weighted average calculation on the resulting sets of candidate commands. The output of this calculation serves as the coordinated control command. For example, the final circulating pump speed command is calculated by multiplying the speed values ​​in each candidate command by their corresponding fusion weights and then summing them.

[0042] The central controller sends the calculated final coordinated control instructions to actuators such as three-way valves and circulation pumps for execution. The entire process is carried out at high speed, thus achieving smooth, continuous and optimal dynamic adjustment of the cooling system.

[0043] Through this implementation method, the system no longer switches abruptly between several fixed modes. Instead, it can dynamically and continuously integrate the advantages of multiple control strategies according to the continuous changes in the system power state, thereby finding the optimal balance between performance and energy consumption under any operating conditions.

[0044] Furthermore, the selective preheating of the lithium battery energy storage module through the heat exchanger using the fuel cell waste heat data includes: monitoring the real-time temperature of the lithium battery energy storage module through the dynamic power compensation module; comparing the real-time temperature with a preset lower threshold value of the optimal operating temperature range of the lithium battery; and when it is determined that the real-time temperature is lower than the lower threshold value, utilizing a waste heat recovery and utilization loop to selectively guide the high-temperature coolant flowing out of the fuel cell stack through the heat exchanger to heat the lithium battery energy storage module until the temperature returns to within the optimal operating temperature range.

[0045] Specifically, during system operation, the central controller, through data acquisition units such as the dynamic power compensation module, simultaneously monitors two key real-time data points: the real-time temperature of the lithium-ion battery energy storage module and the real-time temperature of the high-temperature coolant flowing out of the fuel cell stack outlet, representing the waste heat of the fuel cell. To achieve precise control, the controller has preset a lower threshold for the optimal operating temperature range of the lithium-ion battery, as well as a temperature threshold for effectively utilizing the waste heat.

[0046] The decision to initiate the preheating process is a dual-conditional process. First, the controller compares the real-time temperature of the monitored lithium-ion battery with a preset lower threshold. Second, the controller compares the monitored waste heat data (i.e., coolant temperature) with a preset available heat source threshold. Only when the real-time temperature of the lithium-ion battery is below the lower threshold of its optimal operating range and the fuel cell is indeed generating sufficiently high-quality waste heat will the controller determine that the preheating conditions are met. This dual-conditional mechanism avoids inefficient preheating attempts when the fuel cell itself is not yet fully warmed up and the waste heat temperature is insufficient.

[0047] Once the preheating conditions are met, the controller will instruct an electronically controlled valve to open through the waste heat recovery and utilization circuit, selectively directing the high-temperature coolant flowing out of the fuel cell stack through a heat exchanger connected to the lithium battery. The high-temperature coolant continuously heats the lithium battery energy storage module here, and the central controller continues to monitor the battery temperature in real time. The heating process will continue until the temperature of the lithium battery energy storage module returns to its preset optimal operating temperature range. The controller will then instruct the valve to close, terminating the preheating process.

[0048] By utilizing otherwise wasted energy and replacing the additional parasitic losses associated with traditional electric heaters, the system's overall energy efficiency is directly improved. More importantly, it ensures that lithium batteries remain within their optimal operating temperature range even in cold environments, safeguarding their transient power compensation performance and effectively extending their lifespan by avoiding low-temperature damage. This preheating strategy, intelligently informed by waste heat data, significantly enhances the system's adaptability and operational reliability in all climates.

[0049] Furthermore, the central controller predicts the future temperature trajectory under the current power output based on the temperature change rate of the fuel cell stack through a predictive thermal model; when the predictive thermal model determines that the future temperature trajectory will reach the safety upper limit threshold within a predetermined time, the power allocation strategy based on thermal constraints is triggered, and the central controller generates and issues a collaborative control instruction, which includes a derating instruction for actively reducing the power output of the fuel cell stack and a compensation instruction for synchronously instructing the bidirectional DC-DC converter to increase the power drawn from the lithium battery energy storage module; the dynamic power compensation module generates a corrected control amount through a dynamic gain regulator based on the collaborative control instruction to execute thermal restriction on the fuel cell stack. The specific process is as follows Figure 2 shown.

[0050] Specifically, a predictive thermal model is solidified in the central controller, which describes the thermal capacitance and thermal resistance characteristics of the fuel cell stack. To implement this strategy, in this embodiment, the control parameters set in the controller include: the upper safety temperature threshold of the fuel cell stack is set to 85°C, and the prediction time window is set to 60 seconds.

[0051] During system operation, the central controller continuously provides real-time fuel cell stack temperature change rate, current power output, and other operating data as input to the first-order lumped parameter thermal model. Based on these inputs, the model calculates and generates a 60-second temperature trajectory prediction curve, depicting the likely temperature change trend of the stack. The central controller continuously analyzes this predicted trajectory. If it determines that the trajectory will reach the safety upper limit of 85°C within a predetermined time, it will determine that the system is at imminent risk of overheating and immediately trigger a thermal constraint-based power allocation strategy.

[0052] Once the strategy is triggered, the central controller immediately generates and issues a set of coordinated control instructions. This instruction set consists of two parts: a derating instruction sent to the fuel cell power controller, which proactively lowers the fuel cell's power output limit to reduce heat generation at the source; and a compensation instruction sent simultaneously to the dynamic power compensation module, which drives the bidirectional DC-DC converter to draw power from the lithium battery energy storage module equal to the fuel cell's derating value, thereby ensuring that the total power supply to the terminal load is not affected.

[0053] When the dynamic power compensation module performs power transfer according to this set of coordinated control instructions, its internal dynamic gain regulator plays a key role, generating precise correction control variables to drive the DC-DC converter to smoothly and stably complete power compensation, effectively completing preventive thermal limitation of the fuel cell without affecting the power supply quality.

[0054] The collaborative control instructions also include imposing a dynamic, asymmetric time phase relationship between the control instructions issued to the two converters. Specifically, when power transfer is required, the central controller will issue an instantaneous compensation instruction to the faster-responding lithium battery converter based on the inherent dynamic response speed difference between the two converters, and issue a derating instruction with a preset delay or smooth slope to the relatively slow-responding fuel cell converter. Through this active time misalignment and waveform shaping, it is ensured that the power changes of the two converters can complement each other in real time during the entire dynamic transfer process, thereby minimizing the instantaneous fluctuations of the DC bus voltage. By actively coordinating the instruction timing, seamless power transfer is achieved, improving power quality and system stability.

[0055] After predicting overheating risks using thermal models, the system reduces fuel cell power while simultaneously instructing the lithium battery to provide equal and seamless power compensation. This process is ensured to be smooth and stable by a dynamic gain regulator, mitigating the risk of thermal runaway in advance without compromising power supply continuity and power quality to the terminal loads. This also extends equipment life by avoiding high-temperature shocks.

[0056] Furthermore, the dynamic gain adjuster includes: An online system identification unit injects a pseudo-random binary sequence excitation signal into the control instructions of the bidirectional DC-DC converter; collects the excitation signal and the system's current response, and estimates the system's impulse response function in real time using a cross-correlation operation; performs a Fourier transform on the impulse response function to calculate a broadband dynamic impedance spectrum of the lithium battery energy storage module's current state; a state and disturbance decoupling unit, which compares the broadband dynamic impedance spectrum with a reference impedance spectrum of the lithium battery energy storage module in a baseline healthy state, quantifies the internal impedance deviation caused by battery state aging, and separates the external disturbance component caused by external load changes from the overall system response; A state feedback control law unit receives the internal impedance deviation and the external disturbance component, and calculates an original control instruction through a preset feedback and feedforward gain matrix; The control execution layer receives the original control instructions, updates them, and generates the final corrected control quantity.

[0057] Specifically, the excitation signal uses a 1023-bit pseudo-random binary sequence generated by a 10-bit linear feedback shift register. The sequence's update clock is synchronized with the system's main control clock and set to 10 kHz. The logic level of the pseudo-random binary sequence output is mapped into two different values. After multiplying by a fixed amplitude coefficient, the values ​​are digitally superimposed on the original control command output by the state feedback control law unit. The amplitude coefficient is selected so that the voltage ripple it ultimately causes at the output of the bidirectional DC-DC converter does not exceed the allowable tolerance band of the bus voltage under steady-state operation.

[0058] To achieve real-time estimation of the impulse response function, the system identification unit includes two 1023-byte circular data buffers, one for storing the excitation signal sequence from the most recent acquisition cycle and the other for storing the system's current response sequence. The impulse response function is estimated by multiplying the real-time current response sequence with the excitation signal sequences from the same time period but with varying time delays, then summing the results. This process is repeated for each possible delay point to obtain the complete impulse response function.

[0059] Next, the impedance spectrum solver receives the discrete impulse response function sequence calculated by the online system identification unit and uses a 1024-point fast Fourier transform algorithm to transform the impulse response function sequence, thereby obtaining the system's discrete frequency response. This discrete frequency response is the desired wideband dynamic impedance spectrum representing the current state of the lithium battery energy storage module. The result includes impedance amplitude and phase angle information corresponding to a series of frequency points within the frequency range of 0 to 5 kHz.

[0060] Subsequently, the state and disturbance decoupling unit analyzes the calculated impedance spectrum. The core of this unit is a reference impedance spectrum lookup table stored in the non-volatile memory of the controller. The lookup table pre-stores the impedance spectrum data of the lithium battery in the baseline health state. A deviation spectrum is obtained by vector subtraction of the complex value of the dynamic impedance spectrum calculated in real time and the complex value of the reference impedance spectrum corresponding to the current working condition obtained by interpolation from the lookup table. The shape and amplitude changes of the deviation spectrum in the medium and high frequency range are quantified as internal state deviations that reflect the deterioration of the current health state of the battery or temperature changes. At the same time, the unit identifies the deviation of the deviation spectrum in the DC and extremely low frequency range close to DC as an equivalent DC bias caused by a step change in the external load, and separates and quantifies it into an external disturbance component.

[0061] Finally, the state feedback control law unit receives the internal impedance deviation and external disturbance component provided by the previous stage. The final original control instruction is formed by the linear superposition of the feedforward compensation part and the feedback correction part. The feedforward compensation part is obtained by multiplying the estimated external disturbance component by a preset feedforward gain coefficient, which is used to actively and quickly offset the external load disturbance. The feedback correction part is obtained by calculating the quantized internal impedance deviation through a preset mapping function and feedback gain matrix, which is used to correct the dynamic response deviation caused by the change in the internal state of the battery. The preset feedback and feedforward gain matrix is ​​calculated in the offline design stage using the linear quadratic regulator optimal control algorithm in modern control theory.

[0062] First, based on the system's state-space model, linearization processing is performed for multiple typical operating points. Second, a quadratic objective function is defined to measure the system's overall performance. This function weightedly considers the tracking error of key state variables (such as bus voltage deviation and battery internal state deviation) and the energy consumption of the control output command. Finally, by solving the corresponding algebraic Riccati equation, a set of optimal feedback gain matrices that minimize the value of this quadratic objective function can be uniquely calculated.

[0063] The feedforward gain is calculated after the feedback gain is determined, based on the principle of eliminating steady-state errors. This design approach, based on optimal control theory, yields a gain matrix that ensures fast and stable dynamic response while mathematically optimizing control costs (i.e., the power converter's switching losses and energy scheduling), achieving the inherent unity of high performance and high efficiency.

[0064] The control execution layer receives the digitized original control instructions and converts them into high-frequency gate drive signals with a frequency of 20 kHz to drive the power switches of the bidirectional DC-DC converter through the digital pulse width modulation module, thus completing the entire closed-loop control.

[0065] The entire dynamic gain regulator forms a complete closed-loop adaptive system, realizing a dynamic control system that can online sense the internal state changes of the lithium battery and external load disturbances, and perform active, rapid and precise compensation, significantly improving the stability and robustness of the hybrid system.

[0066] Furthermore, the process of optimizing the operating mode of the cooling system and the corresponding parasitic energy consumption as dynamic variables includes: determining the current operating mode of the three-way valve cooling subsystem; obtaining the corresponding parasitic energy consumption value from a multidimensional lookup table according to the operating mode; subtracting the parasitic energy consumption value from the total power generated by the system in real time to obtain the real-time system net output power as the optimization target; continuously monitoring the response changes of the real-time system net output power caused by the disturbance by injecting a periodic disturbance signal into the power distribution instruction of the system; adaptively adjusting the basic operating point of the power distribution by performing correlation demodulation and integration processing on the disturbance signal and the net power response change, and continuously tracking and converging to the current maximum value point of the real-time system net output power.

[0067] Specifically, the central controller executes a real-time optimization algorithm to continuously track and operate at the maximum point of the real-time system net output power. The net output power is defined as the total power generated by the system minus the parasitic energy consumption of auxiliary systems such as cooling.

[0068] First, the central controller determines the current operating mode of the three-way valve cooling subsystem, as well as the operating status of the circulating pump and fan, in real time. Based on these statuses, the controller retrieves the precise parasitic energy consumption value for the current operating conditions from an internally stored multidimensional lookup table. Simultaneously, the controller obtains the total power generated by the system through a power sensor. The controller then subtracts the total power from the total power to obtain the real-time net system output power, which serves as the optimization target.

[0069] Next, the optimization algorithm kicks in. The controller injects a preset, small, periodic disturbance signal into the base operating point of the current system's power allocation instructions. This disturbance signal is a sine wave with a specific amplitude and frequency. The amplitude is small enough not to affect normal system operation, and the frequency is well below the system's primary dynamic response frequency. While injecting this disturbance signal, the controller continuously monitors the real-time response changes in the system's net output power caused by the disturbance.

[0070] Finally, the algorithm module within the controller performs correlation demodulation and integration processing on the two signals: the input disturbance signal and the output net power response change. The core of this processing is to estimate the gradient direction of the net output power with respect to the power allocation operating point through demodulation: if the net output power response is in phase with the disturbance signal, it means that the current operating point is to the left of the maximum point and the operating point should be increased; if it is out of phase, it means that it is to the right and the operating point should be decreased. The demodulated gradient estimate signal is processed by an integrator, and its output serves as the adaptive adjustment amount for the basic power allocation operating point. Through this closed-loop process, the system operating point is driven to continuously climb, and ultimately it can continuously track and converge to the current maximum point of the real-time system net output power.

[0071] By dynamically accounting for the parasitic energy consumption of the cooling system and optimizing net output power, the system automatically tracks and operates at the true maximum value of available power. This adaptive optimization approach ensures the system maintains peak efficiency under all operating conditions and aging conditions, maximizing fuel utilization and ultimately significantly reducing long-term operating costs.

[0072] Furthermore, when generating an optimized power distribution plan, the central controller adopts a hierarchical target optimization strategy, including: when all system indicators are within the normal range, maximizing the system's net output power is the primary optimization goal, and power distribution is adjusted first to improve the overall operating efficiency; when a drastic change in load power greater than a preset threshold is detected, temporarily switching to maintaining DC bus voltage stability as the primary optimization goal, and prioritizing the lithium battery to perform rapid power compensation to ensure power quality; when the temperature or temperature gradient of the fuel cell stack is detected to exceed the safety threshold, temporarily switching to ensuring the thermal management safety of the stack as the primary optimization goal, and prioritizing derating and enhanced cooling.

[0073] Specifically, in this embodiment of the present invention, the central controller implements a hierarchical target optimization strategy that enables it to dynamically adjust its control priority in different operating scenarios. This strategy includes three main operating modes: Normal operation mode: When the system is operating stably, that is, when the load power fluctuations are smooth, the DC bus voltage is stable, and the fuel cell temperature and temperature gradient are within the ideal range, the system operates in this mode by default. At this time, the primary optimization goal of the central controller is to maximize the net output power of the system. To achieve this goal, the controller will execute a real-time net power optimization algorithm: it will use the current parasitic energy consumption of the cooling system as a cost item, subtract it from the total power, and obtain the net output power; and by injecting small periodic disturbances into the power distribution instructions, it will continuously and adaptively adjust the power distribution ratio between the fuel cell and the lithium battery to track and converge to the true maximum value of the net output power, thereby prioritizing the improvement of the overall operating efficiency.

[0074] In transient response mode, when the central controller detects a dramatic change in load power greater than a preset threshold (for example, a load step exceeding 30% of rated power within 100 milliseconds), the system immediately switches from normal operation to a temporary state where maintaining DC bus voltage stability is the primary optimization objective. In this mode, the controller suspends net power optimization calculations and prioritizes directing the lithium battery energy storage module to perform rapid power compensation via a bidirectional DC-DC converter. The lithium battery responds instantaneously, absorbing or releasing large currents to offset the impact of the sudden load change, ensuring that bus voltage fluctuations are kept to a minimum, thereby ensuring power quality. Once the load surge ends and the bus voltage stabilizes, the system automatically switches back to normal operation.

[0075] In safety protection mode, when the central controller detects that the temperature or temperature gradient of the fuel cell stack exceeds a preset safety threshold (for example, an average temperature above 85°C or a gradient greater than 10°C), the system prioritizes safety over all other objectives and immediately switches to ensuring the safety of the stack thermal management as the primary optimization goal. In this mode, the controller prioritizes a set of coordinated actions: first, it issues a derating command to the fuel cell, actively reducing its power output to reduce heat generation; second, it issues a command to the three-way valve cooling subsystem to switch to an enhanced cooling path; and simultaneously, it instructs the lithium battery to perform power compensation to maintain power supply to the load. The system exits this mode only when both the stack temperature and temperature gradient return to a safe range.

[0076] Through this hierarchical target optimization strategy, the central controller can act like an intelligent decision-making brain, dynamically and seamlessly switching between the three core tasks of "pursuing efficiency," "ensuring stability," and "ensuring safety" based on the priority of real-time working conditions. This enables the entire system to combine efficient economy, excellent dynamic performance, and strong operational reliability.

[0077] Furthermore, the central controller executes a dynamic power compensation method, including: classifying the real-time state of charge of the lithium battery energy storage module obtained from the dynamic power compensation module into one of a plurality of preset intervals including a forced charging interval, an adaptive compensation interval, and a limited power compensation interval, and determining the current compensation mode accordingly; and generating a control instruction through a feedforward-feedback composite control algorithm to adjust the bidirectional DC-DC converter in the dynamic power compensation module. The specific process is as follows: Figure 3 shown.

[0078] Specifically, in this embodiment of the present invention, to achieve a balanced management of lithium battery health and system dynamic performance, the central controller implements an adaptive dynamic power compensation method. The core of this method is to dynamically adjust the operating mode and control strategy of the lithium battery energy storage module based on the real-time state of charge (SOC). The controller pre-defines three specific SOC operating ranges, with different compensation modes set for each range.

[0079] During system operation, the central controller continuously obtains real-time SOC values ​​from the dynamic power compensation module and classifies them into one of the following three preset ranges to determine the current compensation mode to be executed: Adaptive compensation range (SOC is 30% to 80%): In this mode, the lithium-ion battery energy storage module acts as a high-performance power buffer, fully utilizing its potential. The controller utilizes a feedforward-feedback composite control algorithm to regulate the bidirectional DC-DC converter. The feedforward control loop monitors load power changes in real time and provides advance compensation to address sudden changes, while the feedback control loop precisely stabilizes the bus voltage. This ensures optimal system dynamic response while maintaining a healthy SOC range.

[0080] Power limit compensation range (SOC less than 30%): When the battery charge is low, the system's control strategy shifts from performance to battery protection. In this mode, while the feedforward-feedback control algorithm continues to operate to stabilize voltage, the controller imposes an additional power limit. The maximum discharge power of the lithium-ion battery energy storage module is strictly limited to a safe, low level, effectively preventing irreversible damage to the battery due to deep discharge and extending its service life.

[0081] Forced charging range (SOC higher than 80%): When the battery charge is sufficiently high, the system's primary task is to prevent overcharging and prepare for subsequent power compensation. In this mode, the lithium battery's discharge function is suppressed. If the fuel cell's output power exceeds the current load requirement, the controller adjusts the bidirectional DC-DC converter to instruct the lithium battery energy storage module to absorb the excess energy until it is fully charged. This ensures efficient system energy utilization and maintains optimal backup status for the energy storage unit.

[0082] By dynamically switching modes based on SOC intervals, the present invention ensures that the system can achieve an intelligent and optimal balance between high-performance response and long-life protection under different operating conditions.

[0083] Furthermore, it also includes an active electromagnetic disturbance suppression module, which collects the characteristic spectrum generated by its switching frequency and harmonic components in real time through a near-field probe arranged at the bidirectional DC-DC converter; when the disturbance amplitude at a specific frequency point is monitored to exceed a preset threshold, the active electromagnetic disturbance suppression module executes a dynamic spectrum optimization strategy based on a spread spectrum clock, periodically modulates the core switching frequency of the bidirectional DC-DC converter, broadens the concentrated spectrum energy and disperses it into a wider frequency band, thereby reducing the peak electromagnetic emission intensity.

[0084] Specifically, in this embodiment of the present invention, to ensure that the system can operate reliably in electromagnetically sensitive locations such as communication base stations, the system integrates an active electromagnetic disturbance suppression module. The core components of this module include a near-field probe (such as a magnetic field loop antenna) installed above the power inductor of the bidirectional DC-DC converter, a spectrum analysis unit, and a dynamic clock modulation unit. The module presets the frequency points that need to be monitored, namely the core switching frequency of the DC-DC converter (100kHz in this embodiment) and its key second and third harmonic frequencies, and sets the upper limit of the disturbance amplitude for these frequencies that complies with relevant electromagnetic compatibility standards (such as CISPR25) as a preset threshold.

[0085] During system operation, the near-field probe continuously samples the electromagnetic field signal radiated by the DC-DC converter. This signal is fed into the spectrum analysis unit, which generates the system's characteristic spectrum for that frequency band in real time by performing a fast Fourier transform (FFT). The central controller continuously analyzes this spectrum and compares the actual disturbance amplitude at the monitored frequency with a preset threshold.

[0086] When the controller determines that the disturbance amplitude at any monitored frequency exceeds its corresponding preset threshold, it immediately triggers a dynamic spectrum optimization strategy based on spread-spectrum clocking. At this point, the dynamic clock modulation unit changes the clock signal it outputs to the DC-DC converter's gate driver. Instead of outputting a fixed 100kHz square wave, it periodically modulates the core switching frequency with a triangular wave. Specifically, the switching frequency is linearly and periodically swept within a preset frequency offset range (e.g., ±2.5kHz) from a center frequency of 100kHz at a lower modulation frequency (e.g., 10kHz), varying between 97.5kHz and 102.5kHz.

[0087] The periodic modulation of the switching frequency can broaden and disperse the single, sharp spectrum energy that was originally highly concentrated at 100kHz and its harmonic frequencies into a wider frequency band. Although the total radiation energy remains basically unchanged, its peak electromagnetic emission intensity is significantly weakened, enabling the system to meet electromagnetic compatibility standards.

[0088] Furthermore, the dynamic spectrum optimization strategy implemented by the active electromagnetic disturbance suppression module includes a feedforward triggering mechanism based on pre-determined operating conditions. If the central controller determines that the system is about to enter a high-power output or drastic load change impact condition, the active electromagnetic disturbance suppression module will preemptively activate the spread-spectrum clock, proactively suppressing and avoiding the foreseeable increase in electromagnetic noise under such conditions. This ensures that the system consistently meets electromagnetic compatibility standards at all times, especially under dynamic impacts, achieving a higher level of operational reliability.

[0089] Through real-time spectrum monitoring and a dynamic spread-spectrum clock strategy, intelligent, on-demand suppression is achieved. It actively modulates the switching frequency only when disturbances exceed the specified threshold, broadening the spectrum energy. This significantly reduces peak electromagnetic emissions without sacrificing normal operating efficiency, ensuring the system meets electromagnetic compatibility standards. This not only ensures reliable application in sensitive environments such as communication base stations, but also helps reduce reliance on bulky external shielding and filters, resulting in a more integrated and cost-effective product.

[0090] A dynamic power distribution system that integrates intelligent thermal management and hierarchical optimization control effectively resolves the contradiction between instantaneous performance and long-term life in traditional hybrid systems. Instead of passively replenishing power with lithium batteries, active and coordinated power allocation is performed based on the fuel cell's predicted thermal risk and the lithium battery's real-time state of charge (SOC). This not only prevents fuel cell overheating through power transfer, but also limits the output of the lithium battery when the battery is low to avoid deep discharge damage. The system can intelligently switch priorities between the three goals of ensuring power quality, improving net output efficiency, and ensuring component safety based on operating conditions. At the same time, it uses the waste heat from the fuel cell to preheat the lithium battery, achieving cascaded energy utilization. This comprehensive solution extends the service life of core components and improves overall operating economy while ensuring reliable power supply for the system.

[0091] Example 2: The application scenario of this embodiment is a fuel cell-lithium battery hybrid power system for powering a communication base station, which is equipped with a fuel cell-lithium battery hybrid power dynamic distribution system provided by the present invention.

[0092] During a power grid failure, the base station's backup power system was immediately activated. When the system started, the central controller detected through the dynamic power compensation module that the performance of the lithium battery energy storage module was limited due to the low ambient temperature. To ensure that the energy storage unit could respond to subsequent power fluctuations at any time, the controller immediately performed a selective preheating function: after monitoring that the coolant temperature at the outlet of the fuel cell stack (i.e., waste heat data) reached the available threshold, it directed this part of the high-temperature coolant to flow through the heat exchanger to heat the lithium battery until its temperature returned to the optimal operating range.

[0093] When the base station's network traffic is at a low point, the load is stable and the power is low. At this point, the system automatically operates in a mode where maximizing the system's net output power is the primary optimization goal. Based on the lithium battery's state of charge and low-load operating conditions, the controller uses a fuzzy weighted fusion algorithm to calculate the weights favoring an energy-saving strategy, instructing the three-way valve cooling subsystem to switch to an energy-saving bypass loop. Simultaneously, the system initiates a disturbance observation-based optimization algorithm. By injecting tiny periodic disturbance signals into the power allocation instructions and monitoring the net output power response changes caused by the disturbances, the system adaptively adjusts the power allocation point between the fuel cell and lithium battery to track and converge on the true maximum value of the net output power, prioritizing economical operation.

[0094] When a large number of users begin to access the network, the base station load changes dramatically in a short period of time. The system detects that the load impact is greater than the preset threshold and immediately switches the control strategy temporarily, turning to maintaining DC bus voltage stability as the primary optimization goal. During this period, net power optimization is suspended, and the controller prioritizes instructing the lithium battery energy storage module to perform fast, high-power compensation through a bidirectional DC-DC converter, instantly responding to sudden load changes and ensuring that bus voltage fluctuations are suppressed to a very small range, thereby ensuring high-quality power supply and uninterrupted communication services.

[0095] During the daytime peak service period, the base station must cope with prolonged, sustained high power output, placing the fuel cells under constant high load. At this point, the predictive thermal model built into the central controller, based on the temperature change rate and current power, predicts that the stack temperature trajectory will soon reach the safety limit. The system's optimization objectives shift again, temporarily prioritizing stack thermal management safety. The controller immediately issues coordinated control commands, proactively reducing the fuel cell power output (derating) to reduce temperature at the source, while simultaneously increasing the lithium battery power output to compensate for the shortfall and maintain total power supply to the base station. Throughout the seamless power transfer process, the dynamic gain regulator, through online identification and decoupling of the battery's dynamic impedance spectrum, enables precise feedforward compensation for load disturbances. Combined with state feedback, the controller synchronously adjusts to changes in the battery's internal state, maintaining high bus voltage stability even during the intense power transfer. Simultaneously, the cooling system switches to an enhanced cooling path based on real-time temperature gradient monitoring to ensure stack safety.

[0096] Throughout continuous operation, the system's active electromagnetic disturbance suppression module operates continuously in the background. Using near-field probes, it continuously collects the characteristic spectrum generated by the bidirectional DC-DC converter in real time. If the disturbance amplitude at a specific frequency exceeds a preset threshold (especially during periods of drastic load fluctuations), the module proactively implements a dynamic spectrum optimization strategy based on spread-spectrum clocking. This strategy spreads the concentrated spectrum energy and significantly reduces peak electromagnetic emission intensity, ensuring that the hybrid system itself does not interfere with sensitive RF communication equipment within the base station, achieving both high reliability and high electromagnetic compatibility.

[0097] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0098] Through intelligent thermal management and dynamic power allocation, the system achieves refined protection and efficient utilization of core components. The system not only uses predictive thermal models to proactively mitigate fuel cell overheating risks, but also utilizes waste heat to preheat lithium batteries, extending equipment life and enhancing all-weather adaptability. Regarding power allocation, the system intelligently switches and optimizes between multiple objectives, including ensuring bus voltage stability, improving net output efficiency, and ensuring operational safety, based on real-time operating conditions. Combined with active electromagnetic disturbance suppression technology, the system is particularly suitable for scenarios with extremely high reliability and power quality requirements, such as communication base stations and data centers.

[0099] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A fuel cell-lithium battery hybrid power dynamic distribution system, characterized in that: include: The fuel cell stack serves as the main power supply, and its output power is connected to the DC bus through a DC-DC converter; The lithium battery energy storage module is connected to the DC bus through a bidirectional DC-DC converter to provide transient power and store energy; The three-way valve cooling subsystem uses the temperature field data of the fuel cell stack to dynamically control the cooling path and selectively preheat the lithium battery energy storage module through the heat exchanger. The dynamic power compensation module monitors the bus voltage, load current, and battery status, and reallocates power between the fuel cell stack and the lithium battery energy storage module when it predicts that the fuel cell stack temperature exceeds a preset safety range; The central controller optimizes the parasitic energy consumption corresponding to the operating mode of the three-way valve cooling subsystem as a dynamic variable and generates an optimized allocation plan; the dynamic power compensation module executes the optimized allocation plan; The compensation mode is determined based on the real-time state of charge of the lithium battery energy storage module. An external load prediction system is connected to perform power prediction. When a high load event is predicted, the three-way valve cooling subsystem is instructed to start pre-cooling.

2. A fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: The process of dynamically controlling the cooling path includes: Based on the real-time temperature field data obtained from the three-way valve cooling subsystem, the temperature gradient between each area is calculated; and according to the comparison result of the temperature gradient with the preset high threshold and the preset low threshold, the corresponding control instruction is generated. When the temperature gradient is greater than the high threshold, the enhanced cooling path is opened, and when the temperature gradient is less than the low threshold, it is switched to the energy-saving bypass circulation path.

3. A fuel cell-lithium battery hybrid power dynamic distribution system according to claim 2, characterized in that: The generation of the control instructions is based on dynamic calculation of multiple basic control matrices through fuzzy weighted fusion, including: taking the charge state of the lithium battery energy storage module and the current load power as the input of the fuzzy controller, and calculating the dynamic fusion weights corresponding to each basic control matrix through a structured fuzzy rule base composed of multiple parallel sub-rule bases; the sub-rule base includes: an energy-saving strategy sub-rule base, a performance strategy sub-rule base and a balancing strategy sub-rule base; the temperature gradient calculated in real time is queried in all basic control matrices to obtain multiple groups of candidate instruction sets, and these candidate instruction sets are weighted averaged using the dynamic fusion weights independently calculated by each sub-rule base to calculate the control instructions.

4. A fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: The selective preheating of the lithium battery energy storage module by passing the outlet coolant through the heat exchanger includes: monitoring the real-time temperature of the lithium battery energy storage module through the dynamic power compensation module; comparing the real-time temperature with a preset lower threshold value of the optimal operating temperature range of the lithium battery; and when it is determined that the real-time temperature is lower than the lower threshold value, utilizing a waste heat recovery and utilization circuit to selectively guide the high-temperature coolant flowing out of the fuel cell stack through the heat exchanger to heat the lithium battery energy storage module until the temperature returns to within the optimal operating temperature range.

5. The fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: The redistribution of power between the fuel cell stack and the lithium battery energy storage module includes: the central controller predicts the future temperature trajectory under the current power output based on the temperature change rate of the fuel cell stack through a predictive thermal model; when the predictive thermal model determines that the future temperature trajectory will reach the safety upper limit threshold within a predetermined time, the power allocation strategy based on thermal constraints is triggered, and the central controller generates and issues a collaborative control instruction, which includes a derating instruction for actively reducing the power output of the fuel cell stack and a compensation instruction for synchronously instructing the bidirectional DC-DC converter to increase the power drawn from the lithium battery energy storage module; the dynamic power compensation module generates a corrected control amount through a dynamic gain regulator according to the collaborative control instruction to execute thermal limitation of the fuel cell stack.

6. A fuel cell-lithium battery hybrid power dynamic distribution system according to claim 5, characterized in that: The dynamic gain adjuster comprises: An online system identification unit injects a pseudo-random binary sequence excitation signal into the control instructions of the bidirectional DC-DC converter; collects the excitation signal and the system's current response, and estimates the system's impulse response function in real time using a cross-correlation operation; performs a Fourier transform on the impulse response function to calculate a broadband dynamic impedance spectrum of the lithium battery energy storage module's current state; a state and disturbance decoupling unit, which compares the broadband dynamic impedance spectrum with a reference impedance spectrum of the lithium battery energy storage module in a baseline healthy state, quantifies the internal impedance deviation caused by battery state aging, and separates the external disturbance component caused by external load changes from the overall system response; A state feedback control law unit receives the internal impedance deviation and the external disturbance component, and calculates an original control instruction through a preset feedback and feedforward gain matrix; The control execution layer receives the original control instructions, updates them, and generates the final corrected control quantity.

7. The fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: The process of optimizing the operating mode and corresponding parasitic energy consumption of the three-way valve cooling subsystem as dynamic variables includes: determining the current operating mode of the three-way valve cooling subsystem; obtaining the corresponding parasitic energy consumption value from a multidimensional lookup table according to the operating mode; subtracting the parasitic energy consumption value from the total power generated by the system in real time to obtain the real-time system net output power as the optimization target; continuously monitoring the response changes of the real-time system net output power caused by the disturbance by injecting a periodic disturbance signal into the power distribution instruction of the system; adaptively adjusting the basic operating point of the power distribution by performing correlation demodulation and integration processing on the disturbance signal and the net power response change, and continuously tracking and converging to the current maximum value point of the real-time system net output power.

8. The fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: When generating an optimized power distribution plan, the central controller adopts a hierarchical target optimization strategy, including: when all system indicators are within the normal range, maximizing the system's net output power is the primary optimization goal, and power distribution is adjusted first to improve the overall operating efficiency; when a drastic change in load power greater than a preset threshold is detected, the strategy is temporarily switched to maintaining DC bus voltage stability as the primary optimization goal, and the lithium battery is preferentially instructed to perform rapid power compensation to ensure power quality; when the temperature or temperature gradient of the fuel cell stack is detected to exceed the safety threshold, the strategy is temporarily switched to ensuring the thermal management safety of the stack as the primary optimization goal, and derating and enhanced cooling are prioritized.

9. The fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: The method of determining the compensation mode according to the real-time state of charge of the lithium battery energy storage module includes: classifying the real-time state of charge of the lithium battery energy storage module obtained from the dynamic power compensation module into one of multiple preset intervals including a forced charging interval, an adaptive compensation interval, and a limited power compensation interval to determine the current compensation mode; and generating control instructions through a feedforward-feedback composite control algorithm to adjust the bidirectional DC-DC converter in the dynamic power compensation module.

10. The fuel cell-lithium battery hybrid power dynamic distribution system according to claim 1, characterized in that: It also includes an active electromagnetic disturbance suppression module, which collects the characteristic spectrum generated by the switching frequency and harmonic components of the bidirectional DC-DC converter in real time through a near-field probe arranged at the bidirectional DC-DC converter; when the disturbance amplitude at a specific frequency point exceeds a preset threshold, the active electromagnetic disturbance suppression module executes a dynamic spectrum optimization strategy based on a spread spectrum clock, periodically modulates the core switching frequency of the bidirectional DC-DC converter, broadens the concentrated spectrum energy and disperses it into a wider frequency band, thereby reducing the peak electromagnetic emission intensity.

11. A medium having computer program instructions stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a fuel cell-lithium battery hybrid power dynamic distribution system as described in any one of claims 1 to 10 are implemented.

Images