Energy management method for hydrogen fuel cell electric vehicle

Abstract

The present application belongs to the hydrogen fuel cell application technical field, especially to a kind of energy management method based on hydrogen fuel cell power generation car.The present application builds the energy management method of improved deterministic policy gradient algorithm, respectively in the target network and the current network of policy network and value network each increase one network, respectively select the smallest Q value in the training parameter process, to inhibit the overestimation bias problem existing in Q function itself, join the noise subject to normal distribution in the target network of value network, increase the estimation of action of policy function, make the estimated Q value more smooth, improve the stability of algorithm.Hydrogen fuel cell temperature control based on improved PID, introduce the error proportion, integral and differential operation after nonlinear change, since the gain parameter of controller changes with control error, thereby improve the adaptability of algorithm.

Description

Technical Field

[0001] This invention belongs to the field of hydrogen fuel cell application technology, and in particular relates to an energy management method based on a hydrogen fuel cell power generation vehicle. Background Technology

[0002] With the increasing prominence of energy and environmental issues, hydrogen fuel cell vehicles, with their advantages of high efficiency and near-zero emissions, are considered to have broad development prospects. The enthusiasm for researching hydrogen fuel cell vehicles is also growing worldwide, and significant progress has been made in some key technologies, ushering in a golden age for fuel cell vehicle research. Energy management technology is one of the key technologies for fuel cell vehicles and is currently one of the research focuses of researchers.

[0003] Furthermore, current research on energy management strategies for hydrogen fuel cell hybrid vehicles relies on relatively simple hydrogen fuel cell models based on empirical formulas, assuming a constant operating temperature. However, in reality, the temperature of a hydrogen fuel cell not only fluctuates continuously during operation but also constantly affects the voltage and output power of the fuel cell stack. Moreover, the temperature of the fuel cell stack requires regulation by components such as heat sinks to maintain it within a suitable range. Therefore, it is necessary to establish a thermal model for fuel cell systems that considers the effects of temperature.

[0004] Currently, most energy management strategies for fuel cell hybrid electric vehicles (FCEVs) aim to achieve optimal energy efficiency. However, considering the high manufacturing cost and relatively short lifespan of hydrogen fuel cells, designing an optimal energy management strategy that comprehensively considers both hydrogen fuel economy and lifespan economy is essential. In recent years, combining operating condition identification with traditional strategies to establish online energy management strategies with good operating condition adaptability has gradually become a research hotspot. However, research on hydrogen fuel cell hybrid electric vehicles (FCEVs) is still relatively limited. As a key type of new energy vehicle for future development, FCEVs deserve sufficient attention; therefore, it is necessary to establish energy management strategies for FCEVs with good operating condition adaptability. Summary of the Invention

[0005] This invention addresses the technical problems existing in the energy management of hydrogen fuel cell power generation vehicles by proposing a reasonable, simple, theoretically sound, and adaptable energy management method based on hydrogen fuel cell power generation vehicles that takes into account the effects of temperature.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is: an energy management method based on a hydrogen fuel cell power generation vehicle, comprising the following steps:

[0007] S1. Collect factors of the hydrogen fuel cell power generation vehicle, including vehicle mass, vehicle climbing slope, the frontal area of the vehicle facing the wind, and driving speed. The vehicle driving force formula is:

[0008] where m is the vehicle mass, g is the acceleration due to gravity, f is the rolling resistance coefficient, α is the vehicle climbing slope, A is the frontal area of the vehicle facing the wind, μ is the air resistance coefficient of the vehicle body shape, ρ is the air density, v is the driving speed, and δ is the rotational inertia coefficient;

[0009] S2. Build an energy management method based on the improved deterministic policy gradient algorithm. Add a network to each of the target network and the current network of the policy network and the value network respectively. Select the minimum Q value during the training parameter process. Add noise that follows a normal distribution to the target network of the value network. The action in the target network of the value network becomes:

[0010] a' t =β(S i+1 )|θ μ' +ε' t , where a' t is the action of adding noise that follows a normal distribution to the target network of the value network. β(S i+1 )|θ μ' is the optimal action, and ε' t is the noise that follows a normal distribution. Take the vehicle touch force as the state variable, and the power demand of the hydrogen fuel cell system and the maximum discharge power requirement of the power battery pack as the action variables. The reward at the current moment of the energy management method based on the improved deterministic policy gradient algorithm is:

[0011] where ω represents the weight factor, P D (t) is the maximum discharge power requirement of the power battery pack at the current moment, P T is the target value of the maximum discharge power requirement of the power battery pack, P H is the power demand of the hydrogen fuel cell system, and R t-1 is the reward at the previous moment;

[0012] S3. Hydrogen fuel cell temperature control based on the improved PID. The improved PID considers the non-linear change of temperature. The control formula of the improved PID is:

[0013] where I represents the coefficient of PID, 0 < a < 1 < b, e is the error between the current temperature and the set temperature, Δe is the change rate of the error between the current temperature and the set temperature, is the determination coefficient, and f() is the non-linear change function, and its formula is:

[0014] Where σ is the length interval;

[0015] S4. Complete the energy management of the hydrogen fuel cell power generation vehicle.

[0016] Preferably, in step S3, I represents the coefficients of the PID, including proportional, integral, and derivative coefficients.

[0017] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0018] This invention proposes an energy management method for hydrogen fuel cell generator vehicles. It establishes an improved deterministic policy gradient algorithm for energy management by adding a network to both the policy network and the target network of the value network, as well as a network to the current network. During parameter training, the minimum Q-value is selected to suppress overestimation bias in the Q-function. Normally distributed noise is added to the target network of the value network to increase the estimability of the policy function for actions, resulting in a smoother estimated Q-value and improved algorithm stability. Based on an improved PID control method for hydrogen fuel cell temperature control, nonlinear error proportional, integral, and derivative operations are introduced. Since the controller gain parameter changes with the control error, the algorithm's adaptability is improved. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram of a hydrogen fuel cell power generation vehicle provided in an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of the energy management method for the improved deterministic policy gradient algorithm provided in an embodiment of the present invention. Detailed Implementation

[0022] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0023] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0024] Examples, such as Figure 1 , Figure 2 As shown, current research on energy management strategies for fuel cell hybrid electric vehicles relies on relatively simple hydrogen fuel cell models based on empirical formulas, assuming a constant operating temperature. However, in reality, the temperature of a hydrogen fuel cell not only fluctuates continuously during operation but also constantly affects the voltage and output power of the fuel cell stack. Furthermore, the temperature of the fuel cell stack requires regulation by components such as heat sinks to maintain a suitable range. Therefore, a thermal model of the fuel cell system that considers the effects of temperature is needed.

[0025] Currently, most energy management strategies for fuel cell hybrid electric vehicles (FCEVs) aim to achieve optimal energy efficiency. However, considering the high manufacturing cost and relatively short lifespan of hydrogen fuel cells, designing an optimal energy management strategy that comprehensively considers both hydrogen fuel economy and lifespan economy is essential. In recent years, combining operating condition identification with traditional strategies to establish online energy management strategies with good operating condition adaptability has gradually become a research hotspot. However, research on hydrogen fuel cell hybrid electric vehicles (FCEVs) is still relatively limited. As a key type of new energy vehicle for future development, FCEVs deserve sufficient attention; therefore, it is necessary to establish energy management strategies with good operating condition adaptability. To this end, this invention proposes an energy management method based on a hydrogen fuel cell generator vehicle. First, factors related to the hydrogen fuel cell generator vehicle are collected, including vehicle mass, vehicle gradeability, frontal area, and speed. When describing vehicle motion, the vehicle is considered as a point mass in the longitudinal direction. Considering the resistance that the drive motor needs to overcome, including wheel rolling resistance, air resistance, gradient resistance, and acceleration resistance, the vehicle driving force formula is determined as follows:

[0026] Where m is the vehicle mass, g is the gravitational acceleration, f is the rolling resistance coefficient, α is the vehicle gradeability, A is the area of ​​the vehicle's frontal face, μ is the air resistance coefficient of the vehicle body shape, ρ is the air density, v is the driving speed, and δ is the rotational inertia coefficient.

[0027] Considering the development of neural networks, researchers have proposed deep reinforcement learning algorithms that combine deep learning and reinforcement learning. This algorithm combines the perceptual capabilities of deep learning with the decision-making capabilities of reinforcement learning, fully utilizing their advantages to successfully solve the control problem of learning capability management strategies from high-dimensional raw data. In recent years, deep deterministic policy gradient algorithms have received widespread attention. They apply behavior-evaluation algorithm network structures to solve deep reinforcement learning problems in continuous action spaces, making the output power smoother. However, due to the overestimation bias problem inherent in the Q-function, deep deterministic policy gradient algorithms still have some limitations. Therefore, this invention constructs an improved energy management method for deterministic policy gradient algorithms. It adds one network each to the policy network, the target network, and the current network of the value network, respectively, selecting the minimum Q-value during the training parameter process to suppress the overestimation bias problem inherent in the Q-function. Noise following a normal distribution is added to the target network of the value network to increase the estimability of the policy function for actions, making the estimated Q-value smoother and improving the stability of the algorithm. The actions in the target network of the value network become:

[0028] a' t =β(S i+1 )|θ μ' +ε' t , where a' t Adding normally distributed noise actions to the target network of the value network, β(S i+1 )|θ μ' For the optimal action, ε' t To accommodate normally distributed noise, the vehicle's kinetic energy is treated as a state variable, while the power demand of the hydrogen fuel cell system and the maximum discharge power requirement of the battery pack are treated as action variables. The reward for the current time step in the improved deterministic policy gradient algorithm's energy management method is:

[0029] Where ω represents the weighting factor, P D (t) represents the maximum discharge power requirement of the power battery pack at the current moment, P T P is the target value for the maximum discharge power requirement of the power battery pack. H For the power requirements of hydrogen fuel cell systems, R t-1 This is a reward from the previous moment.

[0030] In practice, during the operation of a hydrogen fuel cell, the temperature not only changes continuously but also affects the voltage and output power of the hydrogen fuel cell stack at all times. Moreover, the temperature of the hydrogen fuel cell stack needs to be regulated by components such as radiators to keep it within a suitable range. Therefore, a thermal model of the fuel cell system considering temperature effects needs to be established. Since the parameters of the PID control algorithm are fixed and unchanged, it is prone to cause overshoot in the system and cannot fully meet the stability requirements of the temperature control system. Therefore, non-linear factors are introduced based on the original PID, that is, the error proportional, integral, and differential operations after non-linear changes are introduced. Since the gain parameters of the controller change with the control error, the adaptability of the algorithm is improved. Therefore, for the temperature control of a hydrogen fuel cell based on the improved PID, the improved PID considers the non-linear change of temperature, and the control formula of the improved PID is:

[0031] where I represents the coefficient of PID, 0 < a < 1 < b, e is the error between the current temperature and the set temperature, and Δe is the rate of change of the error between the current temperature and the set temperature. is the coefficient of determination, and f() is a non-linear change function, and its formula is:

[0032] where σ is the length interval.

[0033] Finally, the energy management of the hydrogen fuel cell power generation vehicle is completed.

[0034] The above is only a preferred embodiment of the present invention, and it is not a limitation of the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An energy management method for a hydrogen fuel cell electric power generation vehicle, characterized by, Comprise the following steps: S1, collect hydrogen fuel cell power generation car factors, including vehicle mass, vehicle climbing degree, vehicle windward front area and driving speed, vehicle driving force formula is: Wherein, m is vehicle mass, g is gravity acceleration, f is rolling resistance coefficient, alpha is vehicle climbing degree, A is vehicle windward front area, mu is air resistance coefficient of body shape characteristics, rho is air density, v is driving speed, delta is rotational inertia coefficient; S2, build an energy management method based on improved deterministic policy gradient algorithm, add one network to the target network and current network of policy network and value network respectively, select the minimum Q value in the training parameter process, add noise subject to normal distribution to the target network of value network, and the action of the target network of value network becomes: a′ t = β(S i+1 )|θ μ′ +ε′ t where a′ t In the target network of the value network, the action is subjected to the noise of normal distribution, β(S i+1 )|θ μ′ is the optimal action, ε′ t is the noise subjected to the normal distribution, the vehicle power is taken as the state variable, the power demand of the hydrogen fuel cell system and the maximum discharge power requirement of the power battery pack are taken as the action variable, and the energy management method of the improved deterministic policy gradient algorithm is improved. The reward of the current time is: wherein ω represents a weight factor, P D (t) is the maximum discharge power requirement of the power battery pack at the current time, P T is a target value of the maximum discharge power requirement of the power battery pack, P H is the power demand of the hydrogen fuel cell system, R t-1 is the reward at the previous time S3, hydrogen fuel cell temperature control based on improved PID, the improved PID considers the nonlinear change of temperature, and the control formula of the improved PID is: where I represents the coefficient of PID, 0 < a < 1 < b, e is the error of the current temperature and the set temperature, Δe is the error change rate of the current temperature and the set temperature, is a coefficient, and f() is a nonlinear change function, and the formula is: Wherein sigma is length interval; S4, complete the energy management of hydrogen fuel cell power generation car.

2. The energy management method for a hydrogen fuel cell electric vehicle of claim 1, wherein, In the step S3, I represents the coefficient of PID, including proportional, integral and differential coefficient.

Images