A power system and control method for a light-hydrogen-electric vehicle

Abstract

The application discloses a kind of light-hydrogen-electricity power system and control method for car, it is related to light-hydrogen-electricity hybrid power system for car and energy management field;The application integrates fuel cell, power battery and photovoltaic cell, forms the multi-energy power system for car, gives three kinds of energy optimal output ratio calculation method, and uses depth deterministic policy gradient algorithm to explore optimal control strategy.The algorithm fully considers the influence of external variables such as road condition, illumination intensity, traffic flow on energy management, to reduce control error and improve strategy reliability.Network information layer interacts with vehicle system in real time, comprehensively analyzes road condition, illumination, traffic flow, hydrogen consumption, bus stability and power battery durability and other multidimensional factors, so as to realize the best balance of hydrogen consumption and energy stability.

Description

Technical Field

[0001] This invention relates to the field of hybrid power systems and energy management for photovoltaic-hydrogen-electric vehicles, and particularly to a power system and control method for photovoltaic-hydrogen-electric vehicles. Background Technology

[0002] With the increasing global demand for renewable and clean energy, hybrid electric vehicles (HEVs) have received widespread attention as a highly energy-efficient mode of transportation with low emissions. Photovoltaic cells and fuel cells, as two important renewable energy technologies, are increasingly used in HEVs. Photovoltaic cells directly convert solar energy into electricity, providing a clean power source for vehicles. The application of photovoltaic cells can reduce dependence on traditional fuels, lower carbon emissions, and thus mitigate the impact of climate change. However, the output of photovoltaic cells is affected by factors such as weather conditions, solar radiation intensity, and panel orientation, resulting in unstable energy output. Therefore, effective energy management strategies are needed to maximize energy utilization. Meanwhile, fuel cells, as a highly efficient and clean energy technology, also play a crucial role in HEVs. Fuel cells generate electricity through the chemical reaction of hydrogen and oxygen, producing no emissions or pollution, and possessing extremely high energy density and environmental friendliness. However, challenges remain regarding fuel cell stability, hydrogen storage, and distribution, requiring comprehensive consideration of vehicle system design and energy management strategies. The power battery, as the core component of a HEV, is responsible for storing and releasing electrical energy and is crucial for achieving efficient energy management of the vehicle. Energy management strategies for power batteries involve charging, discharging, and switching between energy-saving modes. They require comprehensive consideration of factors such as the vehicle's power requirements, the battery's health status, and the source of electrical energy to achieve optimal energy utilization efficiency and vehicle performance.

[0003] Therefore, this invention proposes a power system and control method for a photovoltaic-hydrogen-electric vehicle. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a power system and control method for a photovoltaic-hydrogen-electric vehicle.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A photovoltaic-hydrogen-electric vehicle power system includes a fuel cell, a power battery, a photovoltaic cell, a drive motor, a DC / DC converter, and a DC / AC converter. The DC / DC converter is a non-isolated, multi-phase interleaved variable boost DC / DC converter. The photovoltaic panels of the photovoltaic cell are arranged in the roof area, and the fuel cell and the power battery are located at the rear of the roof.

[0007] Preferably, the rated output power of the fuel cell should account for 62% to 68% of the rated power of the vehicle motor, while the rated output power of the power battery should account for 32% to 38% of the rated power of the vehicle motor. When working at night, the total rated output power of the fuel cell and the power battery is 100%.

[0008] Preferably: the longitudinal dynamics model of the photovoltaic-hydrogen-electric vehicle power system is as follows: f is the rolling resistance coefficient, M is the curb weight, g is the acceleration due to gravity, α is the slope, which is generally set to 0 in energy management, and ρ a It represents air density, A represents the vehicle's frontal area, and C... D δ is the air resistance coefficient, v is the vehicle speed, δ is the mass coefficient (set to 1), and a is the vehicle acceleration.

[0009] Preferably: the model of the fuel cell is E represents the open-circuit voltage of the fuel cell, V fc and i fc R represents the output voltage and output current. ohm Let i represent the internal resistance, i0 represent the exchange current, and a, c2, c3, and i max These are all empirical constants.

[0010] Preferably: the equivalent circuit model of the power battery is τ is the time constant, τ = C pol .R pol The state of charge (SOC) of a battery is an important parameter characterizing the battery's state; its instantaneous changes are expressed as current and battery capacity Q. cap The ratio of .

[0011] Preferably, the model of the drive motor is η. m =f(w m ,T m ), When the motor speed w m and torque T m Once determined, the efficiency η of the drive motor can be obtained. m n m This represents the motor speed.

[0012] Preferably: the model of the photovoltaic cell is I sc For short-circuit current, V oc I is the open-circuit voltage. m V is the maximum power point current. m This is the voltage at the maximum power point.

[0013] Preferably, the DC / DC converter is logic-defined as the rate of change of vehicle demand current α and the rate of change of vehicle reference current α. ref The operating mode of the fuel cell multiphase interleaved variable boost DC / DC converter is switched according to the value of α, and the conversion formula is as follows: P max P min These are the maximum and minimum values ​​of the load power; V arg The average voltage is the operating voltage under normal operating conditions; f(state) is the operating condition influence coefficient, which reflects the change rate α of the reference current under different operating conditions. ref Make corrections;

[0014] The conversion conditions of the DC / DC converter are as follows:

[0015] ①: When α<0, disconnecting switches S7 and S8 is equivalent to an interleaved parallel DC / DC converter;

[0016] ②: When 0 < α < α ref When switch S8 is disconnected, it is equivalent to a three-phase interleaved boost DC / DC converter;

[0017] ③: When α>α ref When switch S8 is disconnected, it is equivalent to a four-phase interleaved boost DC / DC converter; and considering the influence of the actual power of the system, a power factor k is introduced. pf (0.9~1.0) is used to determine the final parameters of the fuel cell. The formula for calculating the final parameters of the fuel cell after introducing the power coefficient is as follows: P pf,fuel =k pf ·P fuel P total =P pf,fuel +P bat +P pv .

[0018] A control method for a power system for a photovoltaic-hydrogen-electric vehicle includes the following steps:

[0019] S1: The vehicle's sensor system measures the vehicle's physical signals, including vehicle status information, power battery information, fuel cell information, and photovoltaic system information. Vehicle status information includes: vehicle speed v, acceleration a, and vehicle position. Power battery information includes: remaining power battery charge (SOC) and power battery charge / discharge efficiency η. ba The power battery voltage and current, internal resistance, and fuel cell information include: fuel cell output power P. FC Efficiency η FC And hydrogen consumption rate ΔH2, photovoltaic system information includes: irradiance PV, weather temperature T, photovoltaic cell output power Ppv ;

[0020] S2: The photovoltaic system uses the maximum power point tracking algorithm (MPPT) to control the photovoltaic cells to ensure that the output power of the photovoltaic cells is maximized;

[0021] S3: Extract vehicle speed v, acceleration a, hydrogen consumption rate △H2, and motor power demand P from the CAN signal in the vehicle and feed them back as state variables to the deep deterministic policy gradient algorithm network as observation values; obtain the vehicle's geographical location in real time through the GPS module and upload it to the cloud server. After processing by the cloud server, the vehicle's surrounding road signals, including road length, road slope, road curvature, road traffic information, real-time light intensity, temperature and future changes in the area where the vehicle is located, are fed back to the vehicle.

[0022] S4: Upload vehicle-related data to the cloud server. After the delay time of the feedback data exceeds the threshold T, use a built-in RBF neural network to predict the input data required by DDPG.

[0023] S5: Collect data on the operation of the array of vehicles, including road length, road slope, road curvature, road traffic information, and future changes in light intensity and temperature in the area where the vehicle is located. Select 70% of the data for training, 15% for validation, and 15% for testing. Then set the relevant parameters of the RBF neural network and train the network. Finally, embed the trained network into the DDPG network.

[0024] S6: Preprocess the information collected by the vehicle and the information fed back from the cloud, since the Actor network, Critic network and the built-in RBF neural network used when there is a data feedback delay have the same input;

[0025] S7: Select road condition K F Acceleration a, hydrogen consumption rate ΔH2, motor power demand P M Bus power variation ΔP, Illumination intensity PV t Next moment PV t+Δt The temperature parameter T is defined, the rate of change of light intensity β is defined, and some parameters are selected to form the state space S. S = [K F ,a,△H2,P M ,△P,β,T];

[0026] The pre-allocated output power of the fuel cell and the power battery are selected as action variables, and the action space a is formed:

[0027] a = [P] FC P ba ]

[0028] Four optimization objectives were identified: minimum hydrogen consumption, maximum battery life, maintained battery SOC, and stable bus output power. A reward function R was constructed based on these four objectives.

[0029] R=w1.△H2+w2.l+w3.△SOC+w4.smooth

[0030] Where w1, w2, w3, and w4 are the weighting coefficients for hydrogen consumption rate, power battery life, power battery SOC, and bus smoothness, respectively. l represents power battery life, and smoothness represents bus smoothness.

[0031] S8: The power of the photovoltaic cells is redistributed to the bus in the form of power redistribution, as shown in the following formula:

[0032]

[0033] P b ′ a =P pv -P f ′ c

[0034] Where ε1, ε2, ε3, and ε4 are the correlation coefficients for photovoltaic cell power redistribution, and P f ′ c P represents the amount of power shared by the photovoltaic cells for the fuel cell. b ′ a P represents the amount of power that photovoltaic cells contribute to the power battery. ba P fc These are the pre-allocated output power for the power battery and the fuel cell, respectively.

[0035] P fc-d =P fc -P f ′ c

[0036] P pv-d =P ba -P b ′ a

[0037] Where P fc-d P represents the final output power of the fuel cell in the system. pv-d This refers to the final output power of the system's power battery.

[0038] The gradient descent method is used for computation and updates.

[0039] S9: Call the DDPG algorithm in the computer, create and initialize the training environment, construct the training set from the acquired test data, add random noise to the action variables to realize the function of exploring the environment, and constrain the probability distribution of the actions. Then start training the network so that the network can obtain the optimal action based on the input state variables.

[0040] Preferably, in step S9, the DDPG algorithm includes A2C and an experience pool, where A is the policy function, responsible for generating continuous actions and interacting with the environment, and C is the value function, providing gradient information, evaluating the performance of actions, and providing feedback for updates. Its policy function is represented as a. t =μ(S) t ,|θ Q The value function is represented as Q(S) t ,a t |θ Q );

[0041] Step S9 specifically includes the following steps:

[0042] S91: Collect multiple state-action pairs e t =(S t ,a t ,R t ,S t+1 ), obtain experience pool D t ={e1,e2,…e t};

[0043] S92: When the algorithm learns, it can directly sample from the experience pool;

[0044] S93: Value function label value calculation and error update to y i =r i +γQ′(s i+1 ,μ′(s i+1 |θ u′ )|θ Q′ )

[0045] L(θ)=∑ i (y i -Q(s i ,α i |θ Q )) 2 ;;

[0046] S94: For policy networks, according to... Gradients are used to update the network;

[0047] S95: For both A2C and the experience pool, corresponding evaluation and target networks are set up to achieve delayed updates. The evaluation network is used to select actions, update parameters, and pass the parameters to the target network using a small approximation method. The target network is used to calculate the learning objective, which is the label value y. i θ Q′ ←τθ Q +(1-τ)θ Q ;θ μ′ ←τθ μ +(1-τ)θ μ′ Where τ is the update coefficient, and the random noise added to the output action is a Laplace process. b t This indicates the rate attenuation of noise as learning progresses.

[0048] Beneficial effects:

[0049] This invention integrates fuel cells, power batteries, and photovoltaic cells to form a multi-energy power system for vehicles. It presents three methods for calculating the optimal energy output ratio and explores the optimal control strategy using a deep deterministic strategy gradient algorithm. This algorithm fully considers the impact of external variables such as road conditions, light intensity, and traffic flow on energy management to reduce control errors and improve strategy reliability. The network information layer interacts with the vehicle system in real time, comprehensively analyzing multiple dimensions of factors such as road conditions, light intensity, traffic flow, hydrogen consumption, bus stability, and power battery durability, thereby achieving the optimal balance between hydrogen consumption and energy stability. Attached Figure Description

[0050] Figure 1 This is a flowchart of a power system and control method for a photohydrogen electric vehicle provided by an embodiment of the present invention;

[0051] Figure 2 This is an energy layout diagram of a power system for a photohydrogen electric vehicle provided by an embodiment of the present invention;

[0052] Figure 3 This is the main circuit of the hydrogen fuel cell four-phase interleaved DC / DC converter provided in the embodiments of the present invention;

[0053] Figure 4 This is a topology diagram of the electric power system for a photovoltaic-hydrogen-electric hybrid vehicle provided in an embodiment of the present invention;

[0054] Figure 5 This is an energy management strategy framework based on DDPG provided by an embodiment of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.

[0056] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "setting" should be interpreted broadly. For example, they can refer to a fixed connection or setting, a detachable connection or setting, or an integral connection or setting. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0057] like Figure 1 As shown, this invention relates to a power system and control method for a photovoltaic-hydrogen-electric vehicle. The specific method steps are as follows:

[0058] First, weather information is acquired, specifically including solar irradiance (Pv) and temperature (T). A maximum power point tracking (MPPT) algorithm is used on the photovoltaic cells to enable them to output maximum power in real time under varying solar irradiance conditions. Because the lifespan of hydrogen fuel cells is significantly affected by ripple current, and because they inherently possess characteristics such as slow dynamic response and low-voltage high-current operation, conventional Boost converters are insufficient to meet the power output requirements of hydrogen fuel cells. Therefore, in one embodiment of this patent, a method is proposed as follows: Figure 3 The diagram shows a non-isolated, four-phase interleaved variable boost DC / DC converter, which uses a multi-phase current inner loop control method to control the current of each branch separately.

[0059] In one embodiment, the vehicle's operation involves different energy demands on the power system due to varying conditions such as normal operating conditions, maximum drive power operating conditions, and braking. Therefore, the rate of change of vehicle demand current α is defined as follows, and the operating mode of the fuel cell four-phase interleaved boost DC / DC converter is switched according to the value of α.

[0060]

[0061]

[0062] In the formula α ref P is the rate of change of the vehicle's reference current. max P min These are the maximum and minimum values ​​of the load power; V arg The average voltage is the operating voltage under normal operating conditions; f(state) is the operating condition influence coefficient, which reflects the change rate α of the reference current under different operating conditions. ref Corrections are made. When α < 0, switches S7 and S8 are disconnected, equivalent to an interleaved parallel DC / DC converter; when 0 < α < α ref When switch S8 is disconnected, it is equivalent to a three-phase interleaved boost DC / DC converter;

[0063] When α>α refWhen switch S8 is disconnected, it is equivalent to a four-phase interleaved boost DC / DC converter.

[0064] The above-mentioned four-phase interleaved boost DC / DC converter for fuel cells can bring advantages such as reducing the output current ripple of the fuel cell, improving the service life of the fuel cell, and reducing the voltage stress of each switching transistor.

[0065] In one embodiment, according to the hybrid power system topology described in this patent, the vehicle's energy configuration is as follows: photovoltaic panels are arranged in the roof area, while the fuel cell and power battery are located at the rear of the roof, such as... Figure 2 As shown. This system is suitable for vehicles with a length greater than 6.5 meters and a width greater than 2.1 meters. The system is designed with a fuel cell as the core energy source, where the rated output power of the fuel cell should account for 62% to 68% of the rated power of the vehicle's electric motor, and the rated output power of the power battery should account for 32% to 38% of the rated power of the vehicle's electric motor. To ensure that the total rated output power of the fuel cell and power battery remains at 1% during nighttime operation when the photovoltaic cells cannot function properly. Considering the impact of the actual system power, a power factor k is introduced. pf (0.9~1.0) is used to determine the final parameters of the fuel cell. The specific calculation formula is as follows.

[0066] P pf,fuel =k pf ·P fuel

[0067] P total =P pf,fuel +P bat +P pv

[0068] This configuration not only ensures the vehicle's power performance but also maximizes the potential of photovoltaic power generation, thereby optimizing energy efficiency.

[0069] In one embodiment, the vehicle's sensor system measures the vehicle's physical signals, primarily including vehicle status information, power battery information, fuel cell information, and photovoltaic system information. The vehicle status information includes: vehicle speed v, acceleration a, and vehicle position; the power battery information includes: remaining power battery charge state of charge (SOC), and power battery charge / discharge efficiency (η). ba、 The power battery voltage and current, internal resistance, etc.; the fuel cell information includes: fuel cell output power P. FC Efficiency η FC The photovoltaic system information includes: irradiance PV, ambient temperature T, and photovoltaic cell output power P. The information also includes hydrogen consumption rate ΔH2. pv ;

[0070] The vehicle dynamics model is established, mainly involving the longitudinal dynamics model of the vehicle; the hydrogen consumption model of the fuel cell, the equivalent circuit model of the power battery, and the photovoltaic cell model are established; the drive motor, DC / DC, and DC / AC models are established; and the corresponding connections are made according to the topology of the hybrid system.

[0071] In one embodiment, vehicle speed v, acceleration a, hydrogen consumption rate ΔH2, and motor power demand P are extracted from the vehicle's CAN signal and fed back as state variables to the deep deterministic policy gradient algorithm network as observation values. The vehicle's geographical location is acquired in real time through the GPS module and uploaded to the cloud server. After processing by the cloud server, the surrounding road signals, including road length, road slope, road curvature, road traffic information, real-time light intensity, temperature, and future changes in the area where the vehicle is located, are fed back to the vehicle.

[0072] When uploading vehicle-related data to the cloud server, there is a certain feedback delay due to network fluctuations, which prevents the vehicle from updating parameters in real time, thus affecting the power tracking of the bus. Therefore, when the delay time of the feedback data exceeds the threshold T, a built-in RBF neural network is used to predict the input data required by DDPG.

[0073] In one embodiment, during RBF training, data on the operation of an array of vehicles is first collected, including road length, road slope, road curvature, road traffic information, and future changes in light intensity and temperature in the vehicle's location. 70% of this data is selected for training, 15% for validation, and 15% for testing. Then, the relevant parameters of the RBF neural network are set and the network is trained. Finally, the trained network is integrated into the DDPG network.

[0074] The information collected from vehicles and fed back from the cloud is preprocessed. Since the Actor network, Critic network, and the built-in RBF neural network used to handle data feedback delays all have the same input, they share a common preprocessing network for feature extraction. This improves the network's data preprocessing speed, thereby accelerating the network's convergence speed.

[0075] Select road condition K F Acceleration a, hydrogen consumption rate ΔH2, motor power demand P M Bus power variation ΔP, Illumination intensity PV t Next moment PV t+Δt Parameters such as temperature T are used to define the rate of change of light intensity β, and some parameters are selected to form the state space S.

[0076]

[0077] S = [K F ,a,△H2,PM ,△P,β,T]

[0078] Vehicle-mounted photovoltaic (PV) cells exhibit significant power output fluctuations compared to fixed-installation PV cells. This is because vehicle-mounted PV panels are often shaded by objects on either side of the road, leading to substantial fluctuations in sunlight intensity. Therefore, this patent introduces the variable of sunlight intensity variation rate to more accurately analyze the performance of vehicle-mounted PV systems.

[0079] When the bus power demand is high and the photovoltaic cell output power is low, the power battery alone may not be able to meet the demand. Therefore, it is necessary to significantly increase the output power of the fuel cell to maintain balance. To this end, the energy management strategy proposed in this patent comprehensively considers information such as changes in road conditions and the rate of change in light intensity to predict changes in bus power demand and photovoltaic cell output power. Furthermore, by increasing the opening of the hydrogen valve in advance, it avoids insufficient hydrogen supply when the fuel cell needs to output higher power, thereby increasing the output power of the fuel cell to cope with high power demand conditions.

[0080] The pre-allocated output power of the fuel cell and the power battery are selected as action variables, and the action space a is formed:

[0081] a = [P] FC P ba ]

[0082] This invention has four optimization objectives: minimum hydrogen consumption, maximum battery life, maintained battery SOC, and stable bus output power. A reward function R is constructed based on these four optimization objectives:

[0083] R=w1.△H2+w2.l+w3.△SOC+w4.smooth

[0084] Where w1, w2, w3, and w4 are the weighting coefficients for hydrogen consumption rate, power battery life, power battery SOC, and bus smoothness, respectively. l represents power battery life, and smoothness represents bus smoothness.

[0085] Due to the numerous uncontrollable factors and significant power fluctuations of photovoltaic cells, introducing them into a DDPG would substantially increase the network training time. Therefore, this invention employs power redistribution to direct the power flow from the photovoltaic cells to the bus. The output power of the fuel cell is limited by factors such as hydrogen consumption rate and battery SOC; therefore, when considering the photovoltaic cell output, these factors must be taken into account for allocation. The formula is as follows:

[0086]

[0087] P′ ba =P pv -P′fc

[0088] Where ε1, ε2, ε3, and ε4 are the correlation coefficients for photovoltaic cell power redistribution, and P f ′ c P represents the amount of power shared by the photovoltaic cells for the fuel cell. b ′ a P represents the amount of power that photovoltaic cells contribute to the power battery. ba P fc These are the pre-allocated output power for the power battery and the fuel cell, respectively.

[0089] P fc-d =P fc -P f ′ c

[0090] P pv-d =P ba -P b ′ a

[0091] Where P fc-d For the final fuel cell output power, P pv-d This refers to the final output power of the power battery.

[0092] The gradient descent method is used here for calculation and update. This algorithm can achieve real-time, accurate, and optimal updates of the coefficients in the photovoltaic cell power redistribution formula.

[0093] The DDPG algorithm is invoked within the computer, and a training environment is created and initialized. The acquired test data is used to construct the training set. Since the DDPG network has only a unique output for a given input, it lacks the ability to explore the environment. Therefore, random noise is added to the action variables to achieve the function of exploring the environment. Simultaneously, the probability distribution of its actions is constrained to ensure that the entropy of the distribution is not too small, in order to try different actions and avoid the phenomenon of local optima. Then, the network is trained to enable it to obtain the optimal action based on the input state variables.

[0094] The vehicle longitudinal dynamics model established as described in claim 1 is as follows:

[0095]

[0096] Where f is the rolling resistance coefficient, M is the curb weight, g is the gravitational acceleration, α is the slope (generally set to 0 in energy management), and ρ is the rolling resistance coefficient. a It represents air density, A represents the vehicle's frontal area, and C... D δ is the air drag coefficient, v is the vehicle speed, δ is the mass coefficient (set to 1), and a is the vehicle acceleration. The specific form of the fuel cell model is as follows:

[0097]

[0098] Where τ is the time constant, τ = C pol .R pol The state of charge (SOC) of a battery is an important parameter characterizing the battery's state; its instantaneous changes are expressed as current and battery capacity Q. cap The ratio of .

[0099] The specific form of the drive motor is as follows:

[0100] η m =f(w m ,T m )

[0101]

[0102] When the motor speed w m and torque T m Once determined, the efficiency η of the drive motor can be obtained. m n m This represents the motor speed.

[0103] The specific form of the photovoltaic cell model is as follows:

[0104]

[0105] In the formula I sc For short-circuit current, V oc I is the open-circuit voltage. m V is the maximum power point current. m This is the voltage at the maximum power point.

[0106] The method as described in claim 2, characterized in that: the DDPG algorithm includes an actor and a critic (A2C), and an experience pool; wherein A is a policy function responsible for generating continuous actions and interacting with the environment, and C is a value function providing gradient information, evaluating the performance of actions, and providing feedback for updates. The policy function is represented as a... t =μ(S) t ,|θ Q The value function is represented as Q(S) t ,a t |θ Q The performer network, after training, can enable the critic network to output the maximum Q value.

[0107] Since samples are not independent in deep learning problems, and there may be obvious probabilistic relationships between states, the empirical replay technique is used to collect multiple state-action pairs e. t=(S t ,a t ,R t ,S t+1 ), obtain experience pool D t ={e1,e2,…e t During algorithm learning, samples can be directly taken from the experience pool, which smooths the data distribution, solves the problem of excessive correlation between data, and improves the algorithm's generalization ability. This invention employs priority replay in experience replay, as samples with larger TD errors have a greater impact on the network's backpropagation, making them more likely to be sampled, thus accelerating the convergence speed during algorithm training.

[0108] The label value calculation and error update for the value function are as follows:

[0109] y i =r i +γQ′(s i+1 ,μ′(s i+1 |θ u′ )|θ Q′ )

[0110] L(θ)=∑ i (y i -Q(s i ,α i |θ Q )) 2 ;

[0111] For the policy network, the network is updated according to the following gradient:

[0112]

[0113] For both the Actor and Critic, corresponding evaluation and target networks are set up to implement delayed updates. The evaluation network is used to select actions, update parameters, and pass the parameters to the target network using a small approximation method. The target network is used to calculate the learning objective, i.e., the label value y. i .

[0114] θ Q′ ←τθ Q +(1-τ)θ Q

[0115] θ μ′ ←τθ μ +(1-τ)θ μ′

[0116] Where τ is the update coefficient.

[0117] The random noise added to the output action is a Laplace process, specifically in the form of the following equation:

[0118]

[0119] Among them, b t This indicates the rate attenuation of noise as learning progresses.

[0120] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A control method for a photovoltaic-hydrogen-electric vehicle power system, comprising a photovoltaic-hydrogen-electric vehicle power system, characterized in that, The power system includes a fuel cell, a power battery, a photovoltaic cell, a drive motor, a DC / DC converter, and a DC / AC converter. The DC / DC converter is a non-isolated, multi-phase interleaved variable boost DC / DC converter. The photovoltaic panels of the photovoltaic cell are arranged in the roof area, and the fuel cell and the power battery are located at the rear of the roof. The DC / DC converter is based on the logic of defining the rate of change of vehicle demand current. Vehicle reference current change rate and according to The numerical value is used to convert the operating mode of the fuel cell multiphase interleaved variable boost DC / DC converter. The conversion formula is as follows: , , It is a moment The current required by the vehicle at that time yes The current required by the vehicle at that time It is the interval between two moments. These are the maximum and minimum values ​​of the load power; This is the average voltage under normal operating conditions; It is the operating condition influence coefficient, which realizes the rate of change of the reference current under different operating conditions. Make corrections; The conversion conditions of the DC / DC converter are as follows: ①: When When switches S7 and S8 are disconnected, it is equivalent to an interleaved parallel DC / DC converter. ②: When When switch S8 is disconnected, it is equivalent to a three-phase interleaved boost DC / DC converter; ③: When When switch S8 is closed, it is equivalent to a four-phase interleaved boost DC / DC converter; and considering the influence of the actual power of the system, a power factor is introduced. (0.9~1.0) is used to determine the final parameters of the fuel cell. The calculation formula for the final parameters of the fuel cell after introducing the power coefficient is as follows: , This indicates the final output power of the fuel cell after DC / DC boosting. Indicates the total power of the system bus. The output power of the power battery, This refers to the output power of the photovoltaic cells. It also includes the following steps: S1: The vehicle's sensor system measures the vehicle's physical signals, including vehicle status information, power battery information, fuel cell information, and photovoltaic system information. Vehicle status information includes: vehicle speed v, acceleration a, and vehicle position. Power battery information includes: remaining battery charge (SOC), power battery charge / discharge efficiency (ba), power battery voltage and current, and internal resistance. Fuel cell information includes: fuel cell output power P. FC ,efficiency fc And hydrogen consumption rate ΔH2, photovoltaic system information includes: irradiance PV, weather temperature T, photovoltaic cell output power. ; S2: The photovoltaic system uses the maximum power point tracking algorithm (MPPT) to control the photovoltaic cells to ensure that the output power of the photovoltaic cells is maximized; S3: Extract vehicle speed v, acceleration a, hydrogen consumption rate △H2, and motor power demand P from the CAN signal in the vehicle and feed them back as state variables to the deep deterministic policy gradient algorithm network as observation values; obtain the vehicle's geographical location in real time through the GPS module and upload it to the cloud server. After processing by the cloud server, the vehicle's surrounding road signals, including road length, road slope, road curvature, road traffic information, real-time light intensity, temperature and future changes in the area where the vehicle is located, are fed back to the vehicle. S4: Upload vehicle-related data to the cloud server. After the delay time of the feedback data exceeds the threshold T, use a built-in RBF neural network to predict the input data required by DDPG. S5: Collect data on the operation of the array of vehicles, including road length, road slope, road curvature, road traffic information, and future changes in light intensity and temperature in the area where the vehicle is located. Select 70% of the data for training, 15% for validation, and 15% for testing. Then set the relevant parameters of the RBF neural network and train the network. Finally, embed the trained network into the DDPG network. S6: Preprocess the information collected by the vehicle and the information fed back from the cloud, since the Actor network, Critic network and the built-in RBF neural network used when there is a data feedback delay have the same input; S7: Select Road Conditions acceleration Hydrogen consumption rate ΔH2, motor power demand Bus power variation Light intensity The next moment ,temperature Parameters, defining the rate of change of light intensity And select some parameters to form the state space S. ; The pre-allocated output power of the fuel cell and the power battery are selected as action variables, and the action space a is formed: ； Four optimization objectives were identified: minimum hydrogen consumption, maximum battery life, maintained battery SOC, and stable bus output power. A reward function R was constructed based on these four objectives. R= ； in, , , , These are the weighting coefficients for hydrogen consumption rate, power battery life, power battery SOC, and bus smoothness, respectively. For the lifespan of the power battery, The smoothness of the busbar; S8: The power of the photovoltaic cells is redistributed to the bus in the form of power redistribution, as shown in the following formula: ； ； in, Indicates the output power of the photovoltaic cell. This represents the change in the output power of a photovoltaic cell over a fixed time interval. , , , The correlation coefficient for photovoltaic cell power redistribution. The amount of power shared by photovoltaic cells in fuel cells. The amount of power that photovoltaic cells contribute to the power battery. , These are the pre-allocated output power for the power battery and the fuel cell, respectively. ； ； in For the final fuel cell output power of the system, This refers to the final output power of the system's power battery. The gradient descent method is used for computation and updating. S9: Call the DDPG algorithm in the computer, create and initialize the training environment, construct the training set from the acquired test data, add random noise to the action variables to realize the function of exploring the environment, and constrain the probability distribution of the actions. Then start training the network so that the network can obtain the optimal action based on the input state variables.

2. The control method for a photovoltaic-hydrogen-electric vehicle power system according to claim 1, characterized in that, The rated output power of the fuel cell should account for 62% to 68% of the rated power of the vehicle motor, while the rated output power of the power battery should account for 32% to 38% of the rated power of the vehicle motor. When working at night, the total rated output power of the fuel cell and the power battery is 100%.

3. The control method for a power system for a photovoltaic-hydrogen-electric vehicle according to claim 1, characterized in that, The longitudinal dynamics model of the photovoltaic-hydrogen-electric vehicle power system is as follows: , It is the rolling resistance coefficient. It is the curb weight. It is gravitational acceleration. It refers to the slope, which is typically set to 0 in energy management. It represents air density, and A represents the vehicle's frontal area. It is the air drag coefficient. It's the vehicle speed. is the mass coefficient, set to 1, and a is the vehicle acceleration.

4. The control method for a power system for a photovoltaic-hydrogen-electric vehicle according to claim 1, characterized in that, The model of the fuel cell is as follows: E represents the open-circuit voltage of the fuel cell. and Indicates output voltage and output current. Indicates internal resistance. Indicates alternating current. , , , These are all empirical constants.

5. The control method for a photovoltaic-hydrogen-electric vehicle power system according to claim 1, characterized in that, The equivalent circuit model of the power battery is as follows: , Indicates the battery open-circuit voltage. It is the battery current. It is the battery's internal resistance. It is the polarization current. Indicates polarization resistance. It is a capacitor that characterizes the change of open-circuit voltage as a function of the integral of current. Indicates the output voltage. This represents the output power, and τ is the time constant. The state of charge (SOC) of a battery is an important parameter characterizing the battery's state; its instantaneous changes are expressed as current and battery capacity. The ratio of .

6. The control method for a power system for a photovoltaic-hydrogen-electric vehicle according to claim 1, characterized in that, The equivalent model of the drive motor is , When the motor speed and torque Once determined, the efficiency of the drive motor can be obtained. , This represents the motor speed.

7. The control method for a photovoltaic-hydrogen-electric vehicle power system according to claim 1, characterized in that, The model of the photovoltaic cell is as follows: , To provide current output for photovoltaic cells, The output voltage of the photovoltaic cell. This is the short-circuit current. Open circuit voltage, The maximum power point current, This is the voltage at the maximum power point.

Images