Power and heat coordinated control method for connected fuel cell vehicles

Abstract

The present invention is applicable to the field of fuel cell technology and provides a method for coordinated power and heat control of a networked fuel cell vehicle. The method comprises: constructing a speed prediction model and a vehicle dynamics system model; establishing a power battery system model; establishing a fuel cell thermal management system model; establishing a passenger compartment thermal model; analyzing the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system; and designing a power-heat layered optimization framework to achieve efficient energy utilization and real-time control through a layered coordination mechanism. The present invention can effectively address the comprehensive operating cost optimization problem of an integrated thermal management system considering energy and heat coupling, achieving efficient energy utilization and real-time system response. This method improves the real-time responsiveness and stability of the system, ensuring efficient operation under various conditions.

Description

Technical Field

[0001] The present invention relates to the field of fuel cell technology, and in particular to a method for coordinated power and heat control of a networked fuel cell vehicle. Background Art

[0002] The energy management and passenger compartment-fuel cell coupled thermal management of fuel cell hybrid vehicles are two interrelated and mutually influential systems. Because fuel cells serve as both a power source and a heat source, they are key components in both the passenger compartment-fuel cell coupled thermal management system and the energy management system. In the energy management system, the fuel cell's operating strategy directly affects the efficiency of the fuel cell and power battery, as well as the power battery's state of charge (SOC). In the passenger compartment-fuel cell coupled thermal management system, waste heat from the fuel cell is the primary heat source, and its operating strategy directly influences the characteristics of the vehicle's thermal management system. Therefore, considering the coupled relationship between the passenger compartment-fuel cell coupled thermal management system and energy management, and designing a fuel cell energy management method that incorporates integrated thermal management from the perspective of vehicle power and heat sources, is of great significance for improving energy utilization.

[0003] Disadvantages of existing technology:

[0004] 1. Current fuel cell hybrid vehicle energy management strategies are often based on the assumption of ideal thermal conditions, failing to fully consider the dual nature of fuel cells as both a power source and a heat source in low-temperature scenarios. Traditional approaches treat energy management and thermal management as independent subsystems, failing to establish a dynamic coupling between the two.

[0005] 2. The power-heat coupling system of fuel cell vehicles has significant multi-time scale characteristics. If only a single algorithm is used, it is difficult to meet the system's multi-time scale requirements and cannot effectively coordinate the multi-objective requirements of power distribution and waste heat utilization, resulting in a sharp increase in heat pump air conditioning energy consumption in low-temperature environments and a significant decrease in vehicle range.

[0006] 3. Currently, most research on the coordinated optimization of energy and thermal management focuses on fuel-powered or pure electric vehicles, with relatively little research on fuel cell hybrid vehicles. Because fuel cell hybrid vehicles differ significantly from traditional vehicles in terms of energy structure, thermal management requirements, and optimization approaches, existing layered architectures cannot be directly adapted. Summary of the Invention

[0007] The present invention provides a power and heat coordinated control method for a networked fuel cell vehicle to solve the technical problems mentioned in the background technology.

[0008] A power-heat coordinated control method for a networked fuel cell vehicle, the method comprising:

[0009] S1. Build a speed prediction model and a vehicle dynamics system model;

[0010] S2. Establish a power battery system model;

[0011] S3. Establish a fuel cell thermal management system model;

[0012] S4. Establish a passenger cabin thermal model coupled with fuel cell waste heat;

[0013] S5. Analyze the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system;

[0014] S6. Design a power-heat hierarchical optimization framework to achieve efficient energy utilization and real-time control through a hierarchical collaborative mechanism.

[0015] As a further technical solution of the present invention, in step S1, the steps of constructing the speed prediction model and the vehicle dynamics system model include:

[0016] Based on historical standard operating speed data and real-time navigation information, a long short-term memory network is used to predict future vehicle speed. ; Use the vehicle power balance formula, ignoring the slope resistance, to obtain the vehicle's required driving power :

[0017] ;

[0018] Where: - vehicle quality, - rolling resistance coefficient, - air resistance coefficient, - windward area, - Rotational mass conversion factor, is the gravitational acceleration of the object.

[0019] The fuel cell system and power battery must work together to provide the full power required to meet the vehicle's driving needs. At the same time, the power battery must store regenerative energy. The energy management system allocates the output power of the fuel cell and power battery to minimize the vehicle's operating costs. The balance between the power of each power source and the required power is shown below:

[0020] ;

[0021] Where, is the fuel cell output power, Provides required driving power to the power battery.

[0022] As a further technical solution of the present invention, in step S2, the step of establishing a power battery system model includes:

[0023] The power battery model is established based on the equivalent circuit model. The state of charge (SOC) of the power battery is a key parameter of the power battery, which reflects the remaining available capacity of the power battery. The ampere-hour integration method is used to calculate the SOC value of the power battery at the current moment. The calculation formula is as shown in the formula:

[0024] ;

[0025] ;

[0026] Where, Represents the internal resistance of the power battery, is the power battery current, is the open circuit voltage, is the total charge of the power battery, for Initial value, Provides required driving power for the power battery, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system, and All are battery state of charge ( ) and temperature;

[0027] therefore The dynamic equation can be expressed as .

[0028] As a further technical solution of the present invention, in step S3, the step of establishing a fuel cell thermal management system model includes:

[0029] The dynamic equation for fuel cell temperature is:

[0030] ;

[0031] Where, The heat generated inside the battery pack, The heat removed by the radiator, To transfer heat to the external environment, The heat removed by the heat exchanger, is the mass of the fuel cell, is the specific heat capacity of the fuel cell, is the fuel cell temperature;

[0032] When the battery is working, the output open circuit voltage There are three kinds of irreversible potentials between the electromotive force in thermodynamic theory:

[0033] ;

[0034] Where: is the theoretical electromotive force of thermodynamics; is the activation polarization electromotive force; is the ohmic polarization electromotive force; is the concentration polarization electromotive force, the unit is V; these irreversible electromotive forces are generated in the form of heat energy, so the battery generates heat inside for:

[0035] ;

[0036] Where: is the number of cells in the battery; is the current;

[0037] The heat exchange between the fuel cell and the outside world is mainly related to the fuel cell operating temperature and the ambient temperature. The heat exchanged with the outside environment can be expressed as:

[0038] ;

[0039] Where, is the fuel cell thermal resistance, is the ambient temperature;

[0040] The heat removed by the heat exchanger is expressed as:

[0041] ;

[0042] Where, and are the specific heat capacity and mass flow rate of the coolant, and are the heat transfer coefficient and heat transfer area of ​​the heat exchanger, and are the coolant temperatures at the inlet and outlet of the heat exchanger, respectively, and are approximately = , is the temperature of the heat exchanger;

[0043] The dynamic equation for the heat exchanger temperature is:

[0044] ;

[0045] ;

[0046] ;

[0047] ;

[0048] Where, is the heat capacity of the heat exchanger, The heat transferred from the refrigerant to the heat exchanger, and is the specific heat capacity and mass flow rate of the refrigerant; and are the inlet enthalpy and outlet enthalpy of the heat exchanger, respectively, and a and about The relationship function of

[0049] The water pump regulates the mass flow of the coolant, and the fan enhances the heat dissipation capacity of the radiator. By managing the coolant flow and heat dissipation capacity, the optimal operating temperature of the fuel cell is maintained;

[0050] The heat removed by the radiator is expressed as:

[0051] ;

[0052] Water pump power Expressed as:

[0053] ;

[0054] Radiator power Expressed as:

[0055] ;

[0056] Energy consumption of fuel cell cooling components Expressed as:

[0057] ;

[0058] Energy consumption of fuel cell cooling components Expressed as:

[0059] ;

[0060] Where, is the flow rate of coolant; is the maximum flow rate of the coolant; is the maximum power of the water pump; is the specific heat capacity of air; is the air flow rate of the radiator, which is related to the fan speed and is the control variable of the present invention; is the maximum air inlet volume of the radiator; is the air density; is the radiator outlet air temperature, which is approximately the radiator inlet temperature, that is, the heat exchanger outlet temperature ; is the ambient temperature;

[0061] As a further technical solution of the present invention, in step S4, the step of establishing a passenger compartment thermal model coupled with fuel cell waste heat includes:

[0062] The heat load of the passenger compartment mainly comes from inside and outside the cabin: the external heat load mainly causes heat conduction through the heat transfer characteristics of the car body material and solar radiation The internal heat load is transferred to the passenger compartment by the heat generated by the vehicle occupants , heat from ventilation systems , heat transferred by the air conditioning system composition, ignoring the heat from solar radiation;

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] ;

[0068] ;

[0069] Where, is the convective heat transfer coefficient between the vehicle body shell and the passenger compartment, is the effective area of ​​convective heat transfer for each component, is the ambient temperature, is the cabin temperature, and are the thickness and thermal conductivity of each body shell material; represents the convective heat transfer coefficient between the roof and the interior passenger compartment, represents the convective heat transfer coefficient between the roof and the external environment, Predict the speed of the vehicle in the future time domain, is the number of passengers, is the airflow mass passing through the evaporator, represents the air recirculation coefficient, is the specific heat capacity of the cabin air; is the air density, is the air volume in the passenger compartment;

[0070] The in-car heat exchanger acts as a condenser to release heat into the passenger compartment during heating. Its main function is to discharge the heat absorbed in the heat exchanger and return the refrigerant to liquid form. The compressor is the power source of the automotive heat pump air conditioning system, driving the refrigerant circulation throughout the heat pump system. The thermal model of the in-car heat exchanger is:

[0071] ;

[0072] Where, is the heat capacity of the heat exchanger in the vehicle, is the temperature of the heat exchanger inside the vehicle, The heat transferred from the air to the heat exchanger in the car, The heat transferred from the refrigerant to the heat exchanger in the vehicle;

[0073] The heat transferred from the air to the heat exchanger in the vehicle is expressed as:

[0074] ;

[0075] Where, is the specific heat capacity of air; is the mass flow rate of air, which is controlled by the blower; and are the air temperatures at the inlet and outlet of the heat exchanger in the vehicle, respectively; is the heat transfer coefficient of the in-vehicle heat exchanger; is the heat exchange area of ​​the heat exchanger in the vehicle;

[0076] The heat transferred from the refrigerant to the heat exchanger in the vehicle is expressed as:

[0077] ;

[0078] ;

[0079] Where, is the specific heat capacity of the refrigerant, is the mass flow rate of the refrigerant, which is controlled by the power of the compressor to regulate; and are the inlet and outlet enthalpies of the in-vehicle heat exchanger, respectively; is the maximum power of the compressor; is the maximum flow rate of refrigerant.

[0080] The heat transferred from the air conditioning system to the passenger compartment is expressed as:

[0081] .

[0082] As a further technical solution of the present invention, in step S5, the step of analyzing the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system includes:

[0083] The present invention aims to consider the coupling relationship between the fuel cell-passenger compartment coupled thermal management system and the energy management system, and optimize the vehicle operating cost while meeting the passenger compartment temperature requirements. The optimization objective function can be expressed as:

[0084] ;

[0085] Where, is the hydrogen consumption, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system compressor, and is the weight value, in the present invention and are the unit prices of hydrogen and electricity, is the vehicle operating cost; the optimization problem can be described by minimizing the hydrogen consumption and electricity consumption during operation:

[0086] ；

[0087] :

[0088] ;

[0089] ;

[0090] ;

[0091] ;

[0092] ;

[0093] ;

[0094] ;

[0095] ;

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] in, is the fuel cell temperature, is the fuel cell output power, is the predicted future time domain speed of the vehicle, is the temperature of the heat exchanger, is the ambient temperature, is the coolant flow rate; is the air intake volume of the radiator; Represents fuel cell temperature, passenger compartment temperature, heat exchanger temperature and battery The dynamic equations of Initialize the battery , set to 0.7; is the initial temperature of the heat exchanger; is the initial temperature of the fuel cell; Initial temperature of the passenger compartment; and are the minimum and maximum operating temperatures of the fuel cell respectively; and They are the minimum and maximum temperatures to ensure passenger cabin comfort; and Power batteries The minimum and maximum values ​​of and are the minimum and maximum values ​​of power provided to the fuel cell respectively; and are the minimum and maximum values ​​of the coolant flow rate of the fuel cell system, respectively; and are the minimum and maximum values ​​of the compressor power respectively; and are the minimum and maximum air inlet volumes to the radiator, respectively.

[0104] As a further technical solution of the present invention, in step S6, the steps of designing a power-heat hierarchical optimization framework and achieving efficient energy utilization and real-time control through a hierarchical coordination mechanism include:

[0105] The design power-heat hierarchical optimization framework includes an upper global optimization layer and a lower real-time control layer. The upper layer Deep Deterministic Policy Gradient (DDPG) algorithm generates the fuel cell output power according to the predicted operating conditions within the long prediction time domain. and fuel cell target temperature The lower layer model predictive control (MPC) uses a dynamic model to predict the thermal state (fuel cell temperature and passenger compartment temperature) in the short term in the future, and calculates the coolant flow in real time through online optimization. , Radiator air intake and compressor power etc. to ensure fuel cell temperature tracking The cabin temperature is maintained at around 25°C while strictly meeting physical constraints such as flow limits and compressor power constraints. Finally, the real-time operating status is fed back to the upper layer to form a closed-loop optimization.

[0106] In the hierarchical control strategy, the upper-level deep deterministic policy gradient algorithm focuses on global economy, and globally optimizes the fuel cell's output power and thermal management target temperature based on future driving conditions predicted in the long-term domain; the lower-level model predictive control focuses on ensuring the suppression of transient disturbances and the satisfaction of constraints, and uses its rolling optimization and constraint processing capabilities to quickly suppress temperature fluctuations; the two form a real-time interactive collaborative optimization framework through target transfer and dynamic feedback.

[0107] As a further technical solution of the present invention, step S6 includes:

[0108] S6.1, lower-level MPC rolling optimization control strategy:

[0109] Based on the optimization objectives provided by the upper layer, the lower layer adopts model predictive control to use dynamic models to predict the behavior of the system in the short term in the future and adjust the coolant flow in real time. , Radiator air intake and compressor power etc. to meet the current thermal management requirements while ensuring that system constraints are not violated;

[0110] The lower layer solves the optimal coolant flow rate based on the current fuel cell temperature and passenger compartment temperature , Radiator air intake and compressor power ;

[0111] ;

[0112] :

[0113] ;

[0114] ;

[0115] ;

[0116] ;

[0117] ;

[0118] ;

[0119] ;

[0120] ;

[0121] ;

[0122] ;

[0123] ;

[0124] Where, is the prediction step length; k represents the current time; Indicates the target temperature of the passenger compartment, which is 25°C;

[0125] S6.2. Approximate response model of MPC:

[0126] The role of the approximate response model is to help DDPG fully evaluate the impact of target temperature on actual energy consumption, safety constraints, and dynamic performance when generating target temperature by predicting the dynamic response behavior of the lower-level control system during the upper-level decision-making process.

[0127] First, based on a high-fidelity power-thermal coupling model, input and output data under various operating conditions were recorded to form a training dataset containing samples from various operating conditions. The data was then preprocessed using Kalman filtering and operating condition classification to improve the stability of model training. The model uses a neural network architecture. The input layer receives the target temperature of the fuel cell, the current temperature of the fuel cell, and the current temperature of the passenger compartment. The hidden layer consists of fully connected layers, and the output layer displays the coolant flow rate, radiator air intake, and compressor power.

[0128] The trained model is embedded into the upper DDPG decision process. When DDPG generates the target temperature Then, the MPC approximate response model calculates the coolant flow rate corresponding to the fuel cell target temperature. , Radiator air intake and compressor power , thereby dynamically adjusting the fuel cell target temperature according to the predicted lower layer output results;

[0129] S6.3. Upper-layer DDPG control strategy:

[0130] The upper layer optimizes the fuel cell's output power and the overall strategy of the thermal management system based on predicted future driving conditions to minimize vehicle operating costs. DDPG can learn from its environment and obtain the optimal control strategy.

[0131] Select the state variable as: ;

[0132] Select fuel cell output power and the target temperature of the fuel cell As an action variable:

[0133] ;

[0134] The reward function is related to the optimization objective and is set as:

[0135] ；

[0136] :

[0137] ;

[0138] ;

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] In the reward function, the weight factor and Set as the unit price of hydrogen and electricity respectively;

[0144] The DDPG algorithm uses the actor-critic algorithm as its basic architecture and a deep neural network as an approximation of the policy function and action-value function. The DDPG algorithm establishes an actor network and a critic network , respectively and Parameterized, where the actor network outputs an action based on the current state The output of the actor network is the input of the critic network. The critic network will evaluate the value of the output action based on the output action and the current state. At the same time, in order to make the learning process of the algorithm more stable, corresponding target actor networks are established for the actor network and the critic network respectively. Target critic network , respectively and parameterization;

[0145] S6.4. Offline training:

[0146] The agent interacts with the environment to generate experience data samples , will be stored in the experience pool, and batch data samples will be extracted for training, where and are the system status at the current moment and the system status at the next moment respectively. is the action output, Rewards for system control process;

[0147] The critic network outputs the action value function of the current state, which is used to evaluate the current strategy. The network parameters are updated by performing gradient descent on the loss function. The loss function is expressed as:

[0148] ;

[0149] ;

[0150] In the formula is the reward during the system control process, is the learning rate of the critic network, is the number of data samples drawn from the experience pool, Estimate the action value function of the target critic at the next moment, Estimate the next action output for the target actor network;

[0151] The Actor network outputs the action of the current state and updates the network parameters by performing gradient descent on the action-value function of the critic network:

[0152] ;

[0153] In the formula is the learning rate of the actor network;

[0154] Target network parameters and Participate in the soft update mechanism to improve the stability of learning:

[0155] ;

[0156] ;

[0157] Where, is the soft update factor;

[0158] S6.5. Online Application:

[0159] The optimal control strategy and optimal evaluation function obtained by offline training , for the current state , the optimal action To express , the cost corresponding to the optimal action Expressed as , the upper DDPG will be the current state Input to the trained actor network, output fuel cell output power and fuel cell target temperature ; The lower MPC is based on the current fuel cell temperature and cabin temperature and the fuel cell target temperature set by the upper layer , using the dynamic model to predict the system state in the future time domain and solve the coolant flow online , Radiator air intake and compressor power The optimal control quantity is obtained and the first step optimization result is sent to the actuator.

[0160] Beneficial effects achieved by the present invention:

[0161] To address the challenges of thermal-power synergistic control for fuel cell vehicles in low-temperature environments, this paper proposes a control strategy based on a layered architecture. This strategy achieves efficient coordination of energy and thermal management through the dynamic collaboration of upper-layer deep reinforcement learning and lower-layer model predictive control. In low-temperature scenarios, the upper-layer Deep Deterministic Policy Gradient (DDPG) globally optimizes the fuel cell output power and target temperature based on driving condition predictions and ambient temperature information. Its core objective is to dynamically balance hydrogen consumption costs with waste heat utilization potential. The lower-layer Model Predictive Control (MPC) receives instructions from the upper layer and uses a coupled fuel cell-passenger compartment thermodynamic model to predict temperature trends in the future time domain. It then solves for coolant flow, radiator air volume, and compressor power in real time, strictly adhering to physical constraints while satisfying fuel cell temperature tracking errors and maintaining a stable passenger compartment temperature around 25°C. This improves the system's real-time responsiveness and stability, ensuring efficient operation under various conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0162] Figure 1 This is a schematic structural diagram of the power-heat coupling system for a fuel cell vehicle provided by the present invention.

[0163] Figure 2 This is a flow chart of the working mode of the power-heat coupling system provided by the present invention.

[0164] Figure 3 MAP diagram of fuel cell power, temperature and current.

[0165] Figure 4 This is the control block diagram for power-heat stratification optimization. DETAILED DESCRIPTION

[0166] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0167] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0168] The present invention mainly solves the following technical problems:

[0169] 1. Current fuel cell hybrid vehicle energy management strategies are often based on the assumption of ideal thermal conditions, failing to fully consider the dual nature of fuel cells as both a power source and a heat source in low-temperature scenarios. Traditional approaches treat energy management and thermal management as independent subsystems, failing to establish a dynamic coupling between the two.

[0170] 2. The power-heat coupling system of fuel cell vehicles has significant multi-time scale characteristics. If only a single algorithm is used, it is difficult to meet the system's multi-time scale requirements and cannot effectively coordinate the multi-objective requirements of power distribution and waste heat utilization, resulting in a sharp increase in heat pump air conditioning energy consumption in low-temperature environments and a significant decrease in vehicle range.

[0171] 3. Currently, most research on the coordinated optimization of energy and thermal management focuses on fuel-powered or pure electric vehicles, with relatively little research on fuel cell hybrid vehicles. Because fuel cell hybrid vehicles differ significantly from traditional vehicles in terms of energy structure, thermal management requirements, and optimization approaches, existing layered architectures cannot be directly adapted.

[0172] In view of the above problems, the present invention conducts in-depth research on a new type of fuel cell vehicle power-heat coupling system. Figure 1 As shown in Figure 2, the system integrates multiple subsystems, including the fuel cell, passenger compartment, and power battery. These subsystems are subject to complex thermal and energy coupling. To overcome the limitations of traditional hierarchical control schemes in coordinating these subsystems, this paper proposes a power-heat hierarchical optimization control solution.

[0173] In order to improve the low temperature adaptability of fuel cells, the present invention selects the following working modes for research: the radiator dissipates heat for the fuel cell and the power battery together, and the passenger compartment is heated. , , , ,like Figure 2 As shown, is the temperature of the power battery, is the temperature of the fuel cell, is the temperature of the passenger compartment, The waste heat from the fuel cell system is transferred to the passenger compartment system through a heat exchanger. This design reduces the power consumption of the heat pump air conditioner in low-temperature environments.

[0174] An embodiment of the present invention provides a method for coordinated power and heat control of a networked fuel cell vehicle, the method comprising:

[0175] S1. Build a speed prediction model and a vehicle dynamics system model;

[0176] S2. Establish a power battery system model;

[0177] S3. Establish a fuel cell thermal management system model;

[0178] S4. Establish a passenger cabin thermal model coupled with fuel cell waste heat;

[0179] S5. Analyze the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system;

[0180] S6. Design a power-heat hierarchical optimization framework to achieve efficient energy utilization and real-time control through a hierarchical collaborative mechanism.

[0181] In step S1 of this embodiment, the steps of constructing the speed prediction model and the vehicle dynamics system model include:

[0182] Based on historical standard operating speed data and real-time navigation information, a long short-term memory network (LSTM) is used to predict future time domain vehicle speed. , providing high-precision speed information input to the system; through the vehicle power balance formula, ignoring the slope resistance, the vehicle's required driving power can be obtained :

[0183] ;

[0184] Where: - vehicle quality, - rolling resistance coefficient, - air resistance coefficient, - windward area, - Rotating mass conversion factor, is the gravitational acceleration of the object.

[0185] The fuel cell system and power battery must work together to provide the full power required to meet the vehicle's driving needs. At the same time, the power battery must store regenerative energy. The energy management system allocates the output power of the fuel cell and power battery to minimize the vehicle's operating costs. The balance between the power of each power source and the required power is shown below:

[0186] ;

[0187] Where, is the fuel cell output power, Provides required driving power to the power battery.

[0188] In step S2 of this embodiment, the step of establishing a power battery system model includes:

[0189] Power batteries can make up for the shortcomings of fuel cells, such as slow response speed and inability to recover braking energy. A power battery model is established using a method based on an equivalent circuit model. The state of charge (SOC) of a power battery is a key parameter of the power battery, reflecting the remaining available capacity of the power battery. This invention uses the ampere-hour integration method to calculate the current SOC value of the power battery. The calculation formula is shown as follows:

[0190] ;

[0191] ;

[0192] Where, Represents the internal resistance of the power battery, is the power battery current, is the open circuit voltage, is the total charge of the power battery, for Initial value, Provides required driving power for the power battery, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system, and All are battery state of charge ( ) and temperature, which are usually modeled by experimentally generated MAP tables (i.e., lookup tables);

[0193] therefore The dynamic equation can be expressed as .

[0194] In step S3 of this embodiment, the step of establishing a fuel cell thermal management system model includes:

[0195] The dynamic equation for fuel cell temperature is:

[0196] ;

[0197] Where, The heat generated inside the battery pack, The heat removed by the radiator, To transfer heat to the external environment, The heat removed by the heat exchanger, is the mass of the fuel cell, is the specific heat capacity of the fuel cell, is the fuel cell temperature;

[0198] Fuel cells are devices that convert chemical energy into electrical energy. The relationship between fuel cell power, temperature and current is as follows: Figure 3 As shown;

[0199] When the battery is working, the output open circuit voltage There are three kinds of irreversible potentials between the electromotive force in thermodynamic theory:

[0200] ;

[0201] Where: is the theoretical electromotive force of thermodynamics; is the activation polarization electromotive force; is the ohmic polarization electromotive force; is the concentration polarization electromotive force, the unit is V; these irreversible electromotive forces are generated in the form of heat energy, so the battery generates heat inside for:

[0202] ;

[0203] Where: is the number of cells in the battery; is the current;

[0204] The heat exchange between the fuel cell and the outside world is mainly related to the fuel cell operating temperature and the ambient temperature. The heat exchanged with the outside environment can be expressed as:

[0205] ;

[0206] Where, is the fuel cell thermal resistance, is the ambient temperature;

[0207] The heat removed by the heat exchanger is expressed as:

[0208] ;

[0209] Where, and are the specific heat capacity and mass flow rate of the coolant, and are the heat transfer coefficient and heat transfer area of ​​the heat exchanger, and are the coolant temperatures at the inlet and outlet of the heat exchanger, respectively, and are approximately = , is the temperature of the heat exchanger;

[0210] The dynamic equation for the heat exchanger temperature is:

[0211] ;

[0212] ;

[0213] ;

[0214] ;

[0215] Where, is the heat capacity of the heat exchanger, The heat transferred from the refrigerant to the heat exchanger, and is the specific heat capacity and mass flow rate of the refrigerant; and are the inlet enthalpy and outlet enthalpy of the heat exchanger, respectively, and can be fitted to obtain a and about The relationship function of

[0216] The water pump regulates the mass flow of the coolant, and the fan enhances the heat dissipation capacity of the radiator. By managing the coolant flow and heat dissipation capacity, the optimal operating temperature of the fuel cell is maintained;

[0217] The heat removed by the radiator can be expressed as:

[0218] ;

[0219] Water pump power Expressed as:

[0220] ;

[0221] Radiator power Expressed as:

[0222] ;

[0223] Energy consumption of fuel cell cooling components Expressed as:

[0224] ;

[0225] Where, is the flow rate of coolant; is the maximum flow rate of the coolant; is the maximum power of the water pump; is the specific heat capacity of air; is the air flow rate of the radiator, which is related to the fan speed and is the control variable of the present invention; is the maximum air inlet volume of the radiator; is the air density; is the radiator outlet air temperature, which is approximately the radiator inlet temperature, that is, the heat exchanger outlet temperature ; is the ambient temperature;

[0226] There is a direct relationship between the hydrogen consumption of a fuel cell and its power output. Generally, as the power output increases, the hydrogen consumption of the fuel cell will also increase accordingly. Temperature also has a certain impact on the hydrogen consumption of the fuel cell. Generally speaking, as the temperature increases, the chemical reaction rate inside the fuel cell will accelerate, improving the efficiency of the fuel cell. However, excessively high temperatures can damage the fuel cell and reduce efficiency. Therefore, the hydrogen consumption of the fuel cell is affected by both the power output and the operating temperature of the fuel cell. Most previous studies have only focused on exploring the relationship between hydrogen consumption and power under the optimal operating temperature conditions of the fuel cell. In contrast, this study more comprehensively considered the hydrogen consumption of the fuel cell at different operating temperatures. and power and temperature The relationship between: .

[0227] In step S4 of this embodiment, the step of establishing a passenger compartment thermal model coupled with fuel cell waste heat includes:

[0228] The heat load of the passenger compartment mainly comes from inside and outside the cabin: the external heat load mainly causes heat conduction through the heat transfer characteristics of the car body material and solar radiation The internal heat load is transferred to the passenger compartment by the heat generated by the vehicle occupants , heat from ventilation systems , heat transferred by the air conditioning system composition, ignoring the heat from solar radiation;

[0229] ;

[0230] ;

[0231] ;

[0232] ;

[0233] ;

[0234] ;

[0235] Where, is the convective heat transfer coefficient between the vehicle body shell and the passenger compartment, is the effective area of ​​convective heat transfer for each component, is the ambient temperature, is the cabin temperature, and are the thickness and thermal conductivity of each body shell material; represents the convective heat transfer coefficient between the roof and the interior passenger compartment, represents the convective heat transfer coefficient between the roof and the external environment, is the predicted future time domain speed of the vehicle, is the number of passengers, is the airflow mass passing through the evaporator, represents the air recirculation coefficient, is the specific heat capacity of the cabin air; is the air density, is the air volume in the passenger compartment;

[0236] The in-car heat exchanger acts as a condenser to release heat into the passenger compartment during heating. Its main function is to discharge the heat absorbed in the heat exchanger and return the refrigerant to liquid form. The compressor is the power source of the automotive heat pump air conditioning system, driving the refrigerant circulation throughout the heat pump system. The thermal model of the in-car heat exchanger is:

[0237] ;

[0238] Where, is the heat capacity of the heat exchanger in the vehicle, is the temperature of the heat exchanger inside the vehicle, The heat transferred from the air to the heat exchanger in the car, The heat transferred from the refrigerant to the heat exchanger in the vehicle;

[0239] The heat transferred from the air to the heat exchanger in the vehicle is expressed as:

[0240] ;

[0241] Where, is the specific heat capacity of air; is the mass flow rate of air, which is controlled by the blower; and are the air temperatures at the inlet and outlet of the heat exchanger in the vehicle, respectively; is the heat transfer coefficient of the in-vehicle heat exchanger; is the heat exchange area of ​​the heat exchanger in the vehicle;

[0242] The heat transferred from the refrigerant to the heat exchanger in the vehicle is expressed as:

[0243] ;

[0244] ;

[0245] Where, is the specific heat capacity of the refrigerant, is the mass flow rate of the refrigerant, which is controlled by the power of the compressor to regulate; and are the inlet and outlet enthalpies of the in-vehicle heat exchanger, respectively; is the maximum power of the compressor; is the maximum flow rate of refrigerant.

[0246] The heat transferred from the air conditioning system to the passenger compartment is expressed as:

[0247] ;

[0248] To simplify the system model, it is assumed in the present invention that the flow rate of the air heated by the air conditioning system and entering the passenger compartment remains unchanged, that is, the power of the blower remains unchanged.

[0249] In step S5 of this embodiment, the step of analyzing the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system includes:

[0250] This invention focuses on optimizing energy consumption in fuel cell vehicles operating in low-temperature environments. To ensure thermal comfort in the passenger compartment in cold conditions, the heat pump air conditioning system must supply a large amount of heat energy to maintain the cabin temperature, reducing the vehicle's range. Fuel cells not only serve as a power source to supply the energy required for vehicle operation, but also, during the electrochemical reaction, nearly half of the energy generated is released as heat rather than electricity. Therefore, the fuel cell can also serve as a heat source to assist in heating the passenger compartment, effectively reducing the energy consumption burden of the heat pump air conditioning system. By integrating the fuel cell thermal management system with the passenger compartment thermal management system, waste heat generated by the fuel cell can be efficiently utilized to heat the passenger compartment, significantly reducing energy consumption. When the fuel cell operating temperature is low, the heat dissipation system must operate at a relatively high power level. Simultaneously, due to the low fuel cell temperature, the heat transferred to the passenger compartment is relatively low, resulting in the air conditioning system consuming more energy to maintain a comfortable cabin temperature. When the fuel cell target temperature is high, the heat dissipation system can operate at a lower power level. Furthermore, the fuel cell transfers more heat to the passenger compartment, reducing the air conditioning energy required to maintain the cabin temperature.

[0251] The present invention aims to consider the coupling relationship between the fuel cell-passenger compartment coupled thermal management system and the energy management system, and optimize the vehicle operating cost while meeting the passenger compartment temperature requirements. The optimization objective function can be expressed as:

[0252] ;

[0253] Where, is the hydrogen consumption, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system compressor, and is the weight value, in the present invention and are the unit prices of hydrogen and electricity, is the vehicle operating cost; the optimization problem can be described by minimizing the hydrogen consumption and electricity consumption during operation:

[0254] ；

[0255] :

[0256] ;

[0257] ;

[0258] ;

[0259] ;

[0260] ;

[0261] ;

[0262] ;

[0263] ;

[0264] ;

[0265] ;

[0266] ;

[0267] ;

[0268] ;

[0269] ;

[0270] ;

[0271] in, is the fuel cell temperature, is the fuel cell output power, is the predicted future time domain speed of the vehicle, is the temperature of the heat exchanger, is the ambient temperature, is the coolant flow rate; is the air intake volume of the radiator; Represents fuel cell temperature, passenger compartment temperature, heat exchanger temperature and battery The dynamic equations of Initialize the battery , set to 0.7; is the initial temperature of the heat exchanger; is the initial temperature of the fuel cell; Initial temperature of the passenger compartment; and are the minimum and maximum operating temperatures of the fuel cell respectively; and They are the minimum and maximum temperatures to ensure passenger cabin comfort; and Power batteries The minimum and maximum values ​​of and are the minimum and maximum values ​​of power provided to the fuel cell respectively; and are the minimum and maximum values ​​of the coolant flow rate of the fuel cell system, respectively; and are the minimum and maximum values ​​of the compressor power respectively; and are the minimum and maximum air inlet volumes to the radiator, respectively.

[0272] In step S6 of this embodiment, the steps of designing a power-heat hierarchical optimization framework and achieving efficient energy utilization and real-time control through a hierarchical coordination mechanism include:

[0273] The design power-heat hierarchical optimization framework includes an upper global optimization layer and a lower real-time control layer, such as Figure 4 The upper layer Deep Deterministic Policy Gradient (DDPG) algorithm generates fuel cell output power according to the predicted operating conditions within the long prediction time domain. and fuel cell target temperature The lower layer model predictive control (MPC) uses a dynamic model to predict the thermal state (fuel cell temperature and passenger compartment temperature) in the short term in the future, and calculates the coolant flow in real time through online optimization. , Radiator air intake and compressor power etc. to ensure fuel cell temperature tracking The cabin temperature is maintained at around 25°C while strictly meeting physical constraints such as flow limits and compressor power constraints. Finally, the real-time operating status is fed back to the upper layer to form a closed-loop optimization.

[0274] In the hierarchical control strategy, the upper-level deep deterministic policy gradient algorithm focuses on global economy, and globally optimizes the fuel cell output power and thermal management target temperature based on future driving conditions predicted in the long-term domain; the lower-level model predictive control focuses on ensuring the suppression of transient disturbances and the satisfaction of constraints, and uses its rolling optimization and constraint processing capabilities to quickly suppress temperature fluctuations; the two form a real-time interactive collaborative optimization framework through target transfer and dynamic feedback; for example, when the operating conditions suddenly change, the upper layer adjusts the power and fuel cell target temperature, and the lower layer responds by dynamically optimizing control variables such as coolant flow, radiator air intake and compressor power.

[0275] Step S6 of this embodiment includes:

[0276] S6.1, lower-level MPC rolling optimization control strategy:

[0277] Based on the optimization objectives provided by the upper layer, the lower layer adopts model predictive control to use dynamic models to predict the behavior of the system in the short term in the future and adjust the coolant flow in real time. , Radiator air intake and compressor power etc. to meet the current thermal management requirements while ensuring that system constraints are not violated;

[0278] The lower layer solves the optimal coolant flow rate based on the current fuel cell temperature and passenger compartment temperature , Radiator air intake and compressor power ;

[0279] ;

[0280] :

[0281] ;

[0282] ;

[0283] ;

[0284] ;

[0285] ;

[0286] ;

[0287] ;

[0288] ;

[0289] ;

[0290] ;

[0291] ;

[0292] Where, is the prediction step length; k represents the current time; represents the target temperature of the passenger compartment, which is set to 25°C in the present invention;

[0293] S6.2. Approximate response model of MPC:

[0294] The role of the approximate response model is to help DDPG fully evaluate the impact of target temperature on actual energy consumption, safety constraints, and dynamic performance when generating target temperature by predicting the dynamic response behavior of the lower-level control system during the upper-level decision-making process.

[0295] First, based on a high-fidelity power-thermal coupling model, input and output data under various operating conditions were recorded to form a training dataset containing samples from various operating conditions. The data was then preprocessed using Kalman filtering and operating condition classification to improve the stability of model training. The model uses a neural network architecture. The input layer receives the target temperature of the fuel cell, the current temperature of the fuel cell, and the current temperature of the passenger compartment. The hidden layer consists of fully connected layers, and the output layer displays the coolant flow rate, radiator air intake, and compressor power.

[0296] The trained model is embedded into the upper DDPG decision process. When DDPG generates the target temperature Then, the MPC approximate response model calculates the coolant flow rate corresponding to the fuel cell target temperature. , Radiator air intake and compressor power , thereby dynamically adjusting the fuel cell target temperature according to the predicted lower layer output results;

[0297] S6.3. Upper-layer DDPG control strategy:

[0298] The upper layer optimizes the fuel cell's output power and the overall strategy of the thermal management system based on predicted future driving conditions to minimize vehicle operating costs. DDPG can learn from its environment and obtain the optimal control strategy.

[0299] Select the state variable as: ;

[0300] Select fuel cell output power and the target temperature of the fuel cell As an action variable:

[0301] ;

[0302] The reward function is related to the optimization objective and is set as:

[0303] ；

[0304] :

[0305] ;

[0306] ;

[0307] ;

[0308] ;

[0309] ;

[0310] ;

[0311] In the reward function, the weight factor and Set as the unit price of hydrogen and electricity respectively;

[0312] The DDPG algorithm uses the actor-critic algorithm as its basic architecture and a deep neural network as an approximation of the policy function and action-value function. The DDPG algorithm establishes an actor network and a critic network , respectively and Parameterized, where the actor network outputs an action based on the current state The output of the actor network is the input of the critic network. The critic network will evaluate the value of the output action based on the output action and the current state. At the same time, in order to make the learning process of the algorithm more stable, corresponding target actor networks are established for the actor network and the critic network respectively. Target critic network , respectively and parameterization;

[0313] S6.4. Offline training:

[0314] The agent interacts with the environment to generate experience data samples , will be stored in the experience pool, and batch data samples will be extracted for training, where and are the system status at the current moment and the system status at the next moment respectively. is the action output, Rewards for system control process;

[0315] The critic network outputs the action value function of the current state, which is used to evaluate the current strategy. The network parameters are updated by performing gradient descent on the loss function. The loss function is expressed as:

[0316] ;

[0317] ;

[0318] In the formula is the reward during the system control process, is the learning rate of the critic network, is the number of data samples drawn from the experience pool, Estimate the action value function of the target critic at the next moment, Estimate the next action output for the target actor network;

[0319] The Actor network outputs the action of the current state and updates the network parameters by performing gradient descent on the action-value function of the critic network:

[0320] ;

[0321] In the formula is the learning rate of the actor network;

[0322] Target network parameters and Participate in the soft update mechanism to improve the stability of learning:

[0323] ;

[0324] ;

[0325] Where, is the soft update factor;

[0326] S6.5. Online Application:

[0327] The optimal control strategy and optimal evaluation function obtained by offline training , for the current state , the optimal action To express , the cost corresponding to the optimal action Expressed as , the upper DDPG will be the current state Input to the trained actor network, output fuel cell output power and fuel cell target temperature ; The lower MPC is based on the current fuel cell temperature and cabin temperature and the fuel cell target temperature set by the upper layer , using the dynamic model to predict the system state in the future time domain and solve the coolant flow online , Radiator air intake and compressor power The optimal control quantity is obtained and the first step optimization result is sent to the actuator.

[0328] It should be noted that, in this document, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or apparatus that includes a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0329] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for coordinated power and heat control of a networked fuel cell vehicle, characterized in that: The method comprises: S1. Build a speed prediction model and a vehicle dynamics system model; S2. Establish a power battery system model; S3. Establish a fuel cell thermal management system model; S4. Establish a passenger cabin thermal model coupled with fuel cell waste heat; S5. Analyze the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system; S6. Design a power-heat hierarchical optimization framework to achieve efficient energy utilization and real-time control through a hierarchical coordination mechanism; In step S5, the step of analyzing the coupling relationship between the energy management system and the fuel cell-passenger compartment coupled thermal management system includes: Considering the coupling relationship between the fuel cell-passenger cabin coupled thermal management system and the energy management system, the vehicle operating cost is optimized while meeting the passenger cabin temperature requirements. The optimization objective function is expressed as: ； Where, is the hydrogen consumption, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system compressor, Provides required driving power for the power battery, and are the unit prices of hydrogen and electricity, The operating cost of the vehicle is minimized. The optimization problem is to minimize the hydrogen consumption and electricity consumption during operation: ； ： ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； in, is the fuel cell temperature, is the fuel cell output power, is the predicted future time domain speed of the vehicle, is the temperature of the heat exchanger, is the temperature of the passenger compartment, is the ambient temperature, is the coolant flow rate; is the air intake volume of the radiator; Represents fuel cell temperature, passenger compartment temperature, heat exchanger temperature and battery The dynamic equations of Initialize the battery , set to 0.7; is the initial temperature of the heat exchanger; is the initial temperature of the fuel cell; Initial temperature of the passenger compartment; and are the minimum and maximum operating temperatures of the fuel cell respectively; and They are the minimum and maximum temperatures to ensure passenger cabin comfort; and Power batteries The minimum and maximum values ​​of and are the minimum and maximum values ​​of power provided to the fuel cell respectively; and are the minimum and maximum values ​​of the coolant flow rate of the fuel cell system, respectively; and are the minimum and maximum values ​​of the compressor power respectively; and are the minimum and maximum air inlet volumes to the radiator, respectively.

2. The power-heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: In step S1, the steps of constructing the speed prediction model and the vehicle dynamics system model include: Based on historical standard operating speed data and real-time navigation information, a long short-term memory network is used to predict future vehicle speed. ; Use the vehicle power balance formula, ignoring the slope resistance, to obtain the vehicle's required driving power : ； Where: - vehicle quality, - rolling resistance coefficient, - air resistance coefficient, - windward area, - Rotating mass conversion factor, is the gravitational acceleration of the object; The fuel cell system and power battery must work together to provide the full power required to meet the vehicle's driving needs. At the same time, the power battery must store regenerative energy. The energy management system allocates the output power of the fuel cell and power battery to minimize the vehicle's operating costs. The balance between the power of each power source and the required power is shown below: ； Where, is the fuel cell output power, Provides required driving power to the power battery.

3. The power-heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: In step S2, the steps of establishing a power battery system model include: The power battery model is established based on the equivalent circuit model. The state of charge (SOC) of the power battery is a key parameter of the power battery, which reflects the remaining available capacity of the power battery. The ampere-hour integration method is used to calculate the SOC value of the power battery at the current moment. The calculation formula is as shown in the formula: ； ； Where, Represents the internal resistance of the power battery, is the power battery current, is the open circuit voltage, is the total charge of the power battery, for Initial value, Provides required driving power for the power battery, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system, and Battery state of charge and temperature functions; therefore The dynamic equation is expressed as .

4. The power and heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: In step S3, the steps of establishing a fuel cell thermal management system model include: Fuel cell temperature The dynamic equation is: ； Where, The heat generated inside the battery pack, The heat removed by the radiator, To transfer heat to the external environment, The heat removed by the heat exchanger, is the mass of the fuel cell, is the specific heat capacity of the fuel cell, is the fuel cell temperature; When the battery is working, the output open circuit voltage There are three kinds of irreversible potentials between the electromotive force in thermodynamic theory: ； Where: is the theoretical electromotive force of thermodynamics; is the activation polarization electromotive force; is the ohmic polarization electromotive force; is the concentration polarization electromotive force, the unit is V; these irreversible electromotive forces are generated in the form of heat energy, so the battery generates heat inside for: ； Where: is the number of cells in the battery; is the current; The heat exchange between the fuel cell and the outside world is related to the fuel cell operating temperature and the ambient temperature. The heat exchanged with the outside environment can be expressed as: ； Where, is the fuel cell thermal resistance, is the ambient temperature; The heat removed by the heat exchanger is expressed as: ； Where, and are the specific heat capacity and mass flow rate of the coolant, and are the heat transfer coefficient and heat transfer area of ​​the heat exchanger, and are the coolant temperatures at the inlet and outlet of the heat exchanger, respectively, and are approximately = , is the temperature of the heat exchanger; The dynamic equation for the heat exchanger temperature is: ； ； ； ； Where, is the heat capacity of the heat exchanger, The heat transferred from the refrigerant to the heat exchanger, and is the specific heat capacity and mass flow rate of the refrigerant; and are the inlet enthalpy and outlet enthalpy of the heat exchanger, respectively, and a and about The relationship function of The water pump regulates the mass flow of the coolant, and the fan enhances the heat dissipation capacity of the radiator. By managing the coolant flow and heat dissipation capacity, the optimal operating temperature of the fuel cell is maintained; Heat removed by the radiator Expressed as: ； Water pump power Expressed as: ； Radiator power Expressed as: ； Energy consumption of fuel cell cooling components Expressed as: ； Where, is the flow rate of coolant; is the maximum flow rate of the coolant; is the maximum power of the water pump; is the specific heat capacity of air; The air volume entering the radiator is related to the fan speed; is the maximum air inlet volume of the radiator; is the maximum power of the radiator; is the air density; is the radiator outlet air temperature, which is approximately the radiator inlet temperature, that is, the heat exchanger outlet temperature ; is the ambient temperature.

5. The power and heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: In step S4, the step of establishing a passenger compartment thermal model coupled with fuel cell waste heat includes: The heat load of the passenger compartment comes from inside and outside the cabin: the external heat load causes heat conduction through the heat transfer characteristics of the car body material and solar radiation The internal heat load is transferred to the passenger compartment by the heat generated by the vehicle occupants , heat from ventilation systems , heat transferred by the air conditioning system composition, ignoring the heat from solar radiation; ； ； ； ； ； ； Where, is the convective heat transfer coefficient between the vehicle body shell and the passenger compartment, is the effective area of ​​convective heat transfer for each component, is the ambient temperature, is the cabin temperature, and are the thickness and thermal conductivity of each body shell material; represents the convective heat transfer coefficient between the roof and the interior passenger compartment, represents the convective heat transfer coefficient between the roof and the external environment, is the predicted future time domain speed of the vehicle, is the number of passengers, is the airflow mass passing through the evaporator, represents the air recirculation coefficient, is the specific heat capacity of the cabin air; is the air density, is the air volume in the passenger compartment; The in-car heat exchanger acts as a condenser to release heat into the passenger compartment during heating. Its function is to discharge the heat absorbed in the heat exchanger and return the refrigerant to liquid form. The compressor is the power source of the automotive heat pump air conditioning system, driving the refrigerant circulation throughout the heat pump system. The thermal model of the in-car heat exchanger is: ； Where, is the heat capacity of the heat exchanger in the vehicle, is the temperature of the heat exchanger inside the vehicle, The heat transferred from the air to the heat exchanger in the car, The heat transferred from the refrigerant to the heat exchanger in the vehicle; The heat transferred from the air to the heat exchanger in the vehicle is expressed as: ； Where, is the specific heat capacity of air; is the mass flow rate of air, which is controlled by the blower; and are the air temperatures at the inlet and outlet of the heat exchanger in the vehicle, respectively; is the heat transfer coefficient of the in-vehicle heat exchanger; is the heat exchange area of ​​the heat exchanger in the vehicle; The heat transferred from the refrigerant to the heat exchanger in the vehicle is expressed as: ； ； Where, is the specific heat capacity of the refrigerant, is the mass flow rate of the refrigerant, which is controlled by the power of the compressor to regulate; and are the inlet and outlet enthalpies of the in-vehicle heat exchanger, respectively; is the maximum power of the compressor; is the maximum flow rate of refrigerant; The heat transferred from the air conditioning system to the passenger compartment is expressed as: 。 6. The power-heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: In step S6, the steps of designing a power-heat hierarchical optimization framework and achieving efficient energy utilization and real-time control through a hierarchical coordination mechanism include: The design power-heat hierarchical optimization framework includes an upper global optimization layer and a lower real-time control layer. The upper layer deep deterministic policy gradient algorithm generates the fuel cell output power according to the predicted operating conditions within the long prediction time domain. and fuel cell target temperature The lower layer model predictive control uses the dynamic model to predict the thermal state in the short term in the future, and calculates the control variables in real time through online optimization to ensure the fuel cell temperature tracking. The cabin temperature is maintained at 25°C while strictly meeting physical constraints. Finally, the real-time operating status is fed back to the upper layer, forming a closed-loop optimization loop. In the hierarchical control strategy, the upper-level deep deterministic policy gradient algorithm focuses on global economy, and globally optimizes the fuel cell's output power and thermal management target temperature based on future driving conditions predicted in the long-term domain; the lower-level model predictive control focuses on ensuring the suppression of transient disturbances and the satisfaction of constraints, and uses its rolling optimization and constraint processing capabilities to quickly suppress temperature fluctuations; the two form a real-time interactive collaborative optimization framework through target transfer and dynamic feedback.

7. The power and heat coordinated control method for a networked fuel cell vehicle according to claim 1, characterized in that: Step S6 includes: S6.1, lower-level MPC rolling optimization control strategy: Based on the optimization objectives provided by the upper layer, the lower layer uses model predictive control to use dynamic models to predict the system's behavior in the short-term future and adjust the control variables in real time to meet the current thermal management needs while ensuring that the system constraints are not violated; The lower layer solves the optimal coolant flow rate based on the current fuel cell temperature and passenger compartment temperature , Radiator air intake and compressor power ; ； ： ； ； ； ； ； ； ； ； ； ； ； Where, is the prediction step length; k represents the current time; Indicates the target temperature of the passenger compartment, which is 25°C; is the fuel cell temperature, is the fuel cell output power, is the predicted future time domain speed of the vehicle, is the temperature of the heat exchanger, is the temperature of the passenger compartment, is the ambient temperature, is the coolant flow rate; is the air intake volume of the radiator; Represents fuel cell temperature, passenger compartment temperature, heat exchanger temperature and battery The dynamic equations of Initialize the battery , set to 0.7; is the initial temperature of the heat exchanger; is the initial temperature of the fuel cell; Initial temperature of the passenger compartment; and are the minimum and maximum operating temperatures of the fuel cell respectively; and They are the minimum and maximum temperatures to ensure passenger cabin comfort; and Power batteries The minimum and maximum values ​​of and are the minimum and maximum values ​​of power provided to the fuel cell respectively; and are the minimum and maximum values ​​of the coolant flow rate of the fuel cell system, respectively; and are the minimum and maximum values ​​of the compressor power respectively; and are the minimum and maximum values ​​of the air inlet volume of the radiator respectively; S6.

2. Approximate response model of MPC: The role of the approximate response model is to help DDPG evaluate the impact of target temperature on actual energy consumption, safety constraints, and dynamic performance by predicting the dynamic response behavior of the lower-level control system during the upper-level decision-making process. First, based on a high-fidelity power-thermal coupling model, input and output data under various operating conditions were recorded to form a training dataset containing samples from various operating conditions. The data was then preprocessed using Kalman filtering and operating condition classification to improve the stability of model training. The model uses a neural network architecture. The input layer receives the target temperature of the fuel cell, the current temperature of the fuel cell, and the current temperature of the passenger compartment. The hidden layer consists of fully connected layers, and the output layer displays the coolant flow rate, radiator air intake, and compressor power. The trained model is embedded into the upper DDPG decision process. When DDPG generates the target temperature Then, the MPC approximate response model calculates the coolant flow rate corresponding to the fuel cell target temperature. , Radiator air intake and compressor power , thereby dynamically adjusting the fuel cell target temperature according to the predicted lower layer output results; S6.

3. Upper-layer DDPG control strategy: The upper layer optimizes the fuel cell's output power and the overall strategy of the thermal management system based on predicted future driving conditions to minimize vehicle operating costs. DDPG learns interactively with the environment to obtain the optimal control strategy. Select the state variable as: ; Select fuel cell output power and the target temperature of the fuel cell As an action variable: ； The reward function is related to the optimization objective and is set as: ； ： ； ； ； ； ； ； Where, is the predicted future time domain speed of the vehicle, is the temperature of the heat exchanger, is the temperature of the passenger compartment, is the ambient temperature, is the coolant flow rate; is the air intake volume of the radiator; Represents fuel cell temperature, passenger compartment temperature, heat exchanger temperature and battery The dynamic equations of Initialize the battery , set to 0.7; is the initial temperature of the fuel cell; Initial temperature of the passenger compartment; and are the minimum and maximum operating temperatures of the fuel cell respectively; and They are the minimum and maximum temperatures to ensure passenger cabin comfort; and Power batteries The minimum and maximum values ​​of and are the minimum and maximum values ​​of power provided to the fuel cell respectively; and are the minimum and maximum values ​​of the coolant flow rate of the fuel cell system, respectively; and are the minimum and maximum values ​​of the compressor power respectively; and are the minimum and maximum values ​​of the air inlet volume of the radiator respectively; is the hydrogen consumption, is the energy consumption of the fuel cell cooling components, is the energy consumption of the air conditioning system compressor, The required driving power provided to the power battery; In the reward function, the weight factor and Set as the unit price of hydrogen and electricity respectively; The DDPG algorithm uses the actor-critic algorithm as its basic architecture and a deep neural network as an approximation of the policy function and action-value function. The DDPG algorithm establishes an actor network and a critic network , respectively and Parameterized, where the actor network outputs an action based on the current state The output of the actor network is the input of the critic network. The critic network will evaluate the value of the output action based on the output action and the current state. At the same time, in order to make the learning process of the algorithm more stable, corresponding target actor networks are established for the actor network and the critic network respectively. Target critic network , respectively and parameterization; S6.

4. Offline training: The agent interacts with the environment to generate experience data samples , will be stored in the experience pool, and batch data samples will be extracted for training, where and are the system status at the current moment and the system status at the next moment respectively. is the action output, Rewards for system control process; The critic network outputs the action value function of the current state, which is used to evaluate the current strategy. The network parameters are updated by performing gradient descent on the loss function. The loss function is expressed as: ； ； In the formula is the reward during the system control process, is the learning rate of the critic network, is the number of data samples drawn from the experience pool, Estimate the action value function of the target critic at the next moment, Estimate the next action output for the target actor network; The Actor network outputs the action of the current state and updates the network parameters by performing gradient descent on the action-value function of the critic network: ； In the formula is the learning rate of the actor network; Target network parameters and Participate in the soft update mechanism to improve the stability of learning: ； ； Where, is the soft update factor; S6.

5. Online Application: The optimal control strategy and optimal evaluation function obtained by offline training , for the current state , the optimal action To express , the cost corresponding to the optimal action Expressed as , the upper DDPG will be the current state Input to the trained actor network, output fuel cell output power and fuel cell target temperature ; The lower MPC is based on the current fuel cell temperature and cabin temperature and the fuel cell target temperature set by the upper layer , using the dynamic model to predict the system state in the future time domain and solve the coolant flow online , Radiator air intake and compressor power The optimal control quantity sends the first step optimization result to the actuator.

Images