A method for energy management and control of fuel cell vehicles considering multi-objective optimization

Abstract

This invention aims to address the single-minded focus on reducing hydrogen consumption in previous fuel cell vehicle energy management methods. While this improves fuel economy, it fails to guarantee optimal operation of the fuel cell and battery, thus hindering overall vehicle cost reduction. This invention proposes a multi-objective optimization-based energy management control method for fuel cell vehicles. It models the key components of the vehicle's powertrain and the mechanism of the fuel cell system, establishing models for fuel cell and battery lifespan degradation estimation and equivalent hydrogen consumption. The Pontryagin minimum principle is then used to optimize the multi-objective problem of fuel cell vehicle energy management. This invention satisfies the constraints of initial and final state of charge (SOC) balance and the driver's power demands, ensuring near-globally optimal hydrogen consumption while maintaining overall vehicle economy and reducing fuel cell and battery lifespan degradation.

Description

Technical Field

[0001] This invention belongs to the field of new energy vehicle control, specifically relating to a fuel cell vehicle energy management control method that considers multi-objective optimization. Background Technology

[0002] Compared to other types of new energy vehicles, fuel cell vehicles can achieve zero emissions at all stages and have advantages such as high efficiency, low noise, and short start-stop time, making them widely recognized as one of the important directions for the future development of the new energy vehicle industry. Currently, my country's hydrogen fuel cell vehicle technology still has shortcomings, with deficiencies in core hardware and software control strategies. Unlike fuel cells used in other fields, the load on vehicle fuel cells changes frequently depending on external driving conditions. If the vehicle's energy management strategy only aims to improve economy and reduce hydrogen consumption, while ignoring the rational distribution of energy between the power battery and fuel cell, it will easily lead to the fuel cell and power battery operating under harsh conditions for extended periods, severely damaging their lifespan. Given the high cost of both fuel cells and power batteries, developing a multi-objective optimized energy management strategy that simultaneously considers economy, power battery lifespan, and fuel cell lifespan can effectively reduce the total lifespan cost of the vehicle and extend the service life of fuel cells and power batteries. This is of great significance for accelerating the rapid development of the fuel cell vehicle industry and reducing energy consumption.

[0003] Among existing technologies, such as the invention patent authorized on September 16, 2022 (authorization announcement number: CN112776671B), which describes a method, system, and vehicle for energy management of fuel cell vehicles, this invention controls the working state of the power battery by obtaining the power demand of the entire vehicle and the rated power of the fuel cell. The aim is to ensure that the fuel cell operates at its rated power or stops supplying power. While this method can ensure that the fuel cell always operates within a relatively optimal range, it does not consider the usage conditions of the power battery. When the power demand of the entire vehicle is too high or too low, the power battery will undergo high-power "peak shaving and valley filling" charging and discharging, which will damage the lifespan of the power battery. Another example is the invention patent authorized on February 11, 2022 (authorization announcement number: CN113085665B), which describes a fuel cell vehicle energy management method based on the TD3 algorithm. Although this method considers a multi-objective optimization problem including energy consumption economy, fuel cell lifespan, and power battery lifespan, the control method used is more complex, computationally time-consuming, and the algorithm has lower efficiency. It also does not describe in detail how power is distributed and controlled. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a multi-objective optimization-based energy management and control method for fuel cell vehicles. It establishes mathematical models of the entire vehicle and core power components, a mechanistic model of the fuel cell system, a power battery life degradation model, and a fuel cell life degradation model. Utilizing the Pontryagin minimum principle to control the entire system, it can achieve near-global optimal hydrogen consumption for the entire vehicle while simultaneously considering the lifespan of the power battery and fuel cell, thereby improving the overall energy-saving quality of fuel cell vehicles.

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

[0006] 1. A fuel cell vehicle energy management and control method considering multi-objective optimization, characterized by comprising the following steps:

[0007] Step 1: Establish state equations and constraints

[0008] Pontryagin's minimum principle, based on the classical variational method, requires a mathematical description of the actual system. This method selects the battery charge (SOC) as the state variable, with an initial value of 0.7. The initial SOC of the actual battery charge is calibrated by adjustable parameters. Since both excessively high and low battery charge levels are detrimental to starting operation, this method restricts the battery charge (SOC(t)) at any given time to be at the highest allowable SOC for battery operation. max and minimum SOC min Within the interval, to ensure that the initial and final battery charge levels remain the same, this method sets the absolute difference in SOC between the initial and final battery charge levels to be less than 0.005 as a constraint condition for balance. SOC(t0) and SOC(t... f The initial and final states of charge of the power battery are represented by , respectively, as shown in Equation 1:

[0009]

[0010] Select fuel cell power P fc As a control variable of the system, in order to improve the lifespan of the fuel cell, this method limits the fuel cell power P at any given time. fc (t) are all located at the maximum power P fc_max and minimum power P fc_min Range; change in fuel cell power ΔP per unit time fc (t) is located at the maximum change ΔP fc_max and minimum change ΔP fc_min Within the range, as shown in Equation 2:

[0011]

[0012] The system state equation established by the method is in the form of Equation 3:

[0013]

[0014] Step two: Introduce costate variables, construct the objective function and Hamiltonian function.

[0015] Traditional classical variational methods are only suitable for solving unconstrained extremum problems. For the optimal extremum problem of practical systems, this method introduces a costate variable λ(t) to construct a Hamiltonian function H(x(t),u(t),λ(t),t), and further uses the Hamiltonian function to solve for the optimal control sequence u of the system. * (t), the optimal control sequence u of this method * (t) satisfies the following four conditions:

[0016] (1)u * (t) enables the Hamiltonian function to reach its minimum value during the entire algorithm;

[0017] H(x * (t),u * (t),λ * (t),t)≤H(x(t),u(t),λ(t),t) (4)

[0018] (2)u * (t) satisfies the costate equation, and the Hamiltonian function with respect to the state variable x(t) i Finding the partial differential equation yields the costate equation;

[0019]

[0020] (3)u * (t) satisfies the state equation, and the Hamiltonian function with respect to the costate variable λ(t) i Finding the partial differentials yields the state equations;

[0021]

[0022] (4)u * (t) satisfies the initial and final boundary conditions;

[0023] Equation 7 is the specific representation of the Hamiltonian function, where the first term L(x(t),u(t),t) represents the transition objective function of the system, characterizing the transition from the initial time t0 to the final time t. f The cost of system changes and the specific objective function are shown in Equation 8:

[0024] H(x(t),u(t),λ(t),t)=L(x(t),u(t),t)+λ(t)·f(x(t),u(t),t) (7)

[0025]

[0026] In the formula m fc m fc_equ m bat_equ These represent the total hydrogen consumption of the vehicle under actual operating conditions, the equivalent hydrogen consumption calculated based on fuel cell degradation, and the equivalent hydrogen consumption calculated based on power battery degradation, respectively. The total hydrogen consumption of the vehicle under actual operating conditions is calculated based on the vehicle dynamics equation and the calorific value of hydrogen. The equivalent hydrogen consumption calculated based on fuel cell life degradation and power battery life degradation are calculated based on Equations 9 and 10, respectively.

[0027]

[0028]

[0029] In the formula Δψ fc Indicates the percentage of fuel cell lifespan degradation; Δψ bat This indicates the percentage of battery life degradation. This method sets the lifespan of the fuel cell to terminate when the degradation reaches 10%; and the lifespan of the power battery to terminate when the degradation reaches 20%. M represents the percentage of battery lifespan degradation. fc For the price of fuel cell stacks, M bat For the price of power batteries, The objective function of this method is given by the price per unit volume of hydrogen, as shown in Equation 11. The first term of the objective function is the final-state cost function φ(SOC(t)). f ),t f The second term is the system transition objective function L(x(t),u(t),t) mentioned above, and the final-state cost function is calculated according to Equation 12:

[0030]

[0031]

[0032] In Equation 12 This represents the amount of electricity consumed, calculated based on the difference in SOC (State of Charge) of the initial and final battery charge, converted into the amount of hydrogen required to produce the same amount of electricity by burning hydrogen.

[0033] Step 3: Initialize costate variables and iteratively solve for the optimal control sequence u. * (t)

[0034] Initially assign values ​​to the costate variables, discretize the control variables within the allowable power range of the fuel cell, calculate the Hamiltonian function, and solve for the control variable u(t) in the current instantaneous state. i ), that is, the fuel cell power P at this time. fc (t iIf the SOC of the power battery satisfies the constraints of the state variables, then the desired control variable u(t) will be determined. i The optimal control value u in this instantaneous state is... * (t i If the constraints are not met, the costate variables are reassigned and the Hamiltonian function is calculated. This process is repeated to obtain the optimal control values ​​for all states. Finally, all optimal control values ​​are connected to form a curve to obtain the optimal control sequence u. * (t) trajectory.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. The energy management and control method for fuel cell vehicles proposed in this invention, which considers multi-objective optimization, can reduce the overall cost of the vehicle, extend the service life of the power battery and fuel cell, ensure the reasonable distribution of power among different power sources, and improve the energy-saving effect of the vehicle.

[0037] 2. This invention utilizes the Pontryagin minimum principle to solve for the optimal control sequence. The method is simple, can guarantee the constraint of basic balance between the initial and final SOC, and can follow the power demand of the whole vehicle. It has a small number of control variables and high algorithm execution efficiency. Attached Figure Description

[0038] The following description of the embodiments, taken in conjunction with the accompanying drawings, will make the embodiments readily understood, wherein:

[0039] Figure 1 This is a schematic diagram of the powertrain configuration of a fuel cell vehicle according to an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram of the energy management and control method for a fuel cell vehicle according to an embodiment of the present invention; Detailed Implementation

[0041] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0042] The following description, with reference to the accompanying drawings, illustrates a fuel cell vehicle energy management control method that considers multi-objective optimization; however, the invention is not limited to these embodiments.

[0043] Reference Appendix Figure 1The method targets key components of fuel cell vehicles, including fuel cells, power batteries, DC-DC converters, motor controllers, and drive motors, and establishes mathematical models of these components as the control objects.

[0044] Reference Appendix Figure 2 The implementation of this invention patent includes the following three steps: 1. A fuel cell vehicle energy management and control method considering multi-objective optimization, characterized by including the following steps:

[0045] Step 1: Establish state equations and constraints

[0046] Pontryagin's minimum principle, based on the classical variational method, requires a mathematical description of the actual system. This method selects the battery charge (SOC) as the state variable, with an initial value of 0.7. The initial SOC of the actual battery charge is calibrated by adjustable parameters. Since both excessively high and low battery charge levels are detrimental to starting operation, this method restricts the battery charge (SOC(t)) at any given time to be at the highest allowable SOC for battery operation. max and minimum SOC min Within the interval, to ensure that the initial and final battery charge levels remain the same, this method sets the absolute difference in SOC between the initial and final battery charge levels to be less than 0.005 as a constraint condition for balance. SOC(t0) and SOC(t... f The initial and final states of charge of the power battery are represented by , respectively, as shown in Equation 1:

[0047]

[0048] Select fuel cell power P fc As a control variable of the system, in order to improve the lifespan of the fuel cell, this method limits the fuel cell power P at any given time. fc (t) are all located at the maximum power P fc_max and minimum power P fc_min Range; change in fuel cell power ΔP per unit time fc (t) is located at the maximum change ΔP fc_max and minimum change ΔP fc_min Within the range, as shown in Equation 2:

[0049]

[0050] The system state equation established by the method is in the form of Equation 3:

[0051]

[0052] Step two: Introduce costate variables, construct the objective function and Hamiltonian function.

[0053] Traditional classical variational methods are only suitable for solving unconstrained extremum problems. For the optimal extremum problem of practical systems, this method introduces a costate variable λ(t) to construct a Hamiltonian function H(x(t),u(t),λ(t),t), and further uses the Hamiltonian function to solve for the optimal control sequence u of the system. * (t), the optimal control sequence u of this method * (t) satisfies the following four conditions:

[0054] (1)u * (t) enables the Hamiltonian function to reach its minimum value during the entire algorithm;

[0055] H(x * (t),u * (t),λ * (t),t)≤H(x(t),u(t),λ(t),t) (4)

[0056] (2)u * (t) satisfies the costate equation, and the Hamiltonian function with respect to the state variable x(t) i Finding the partial differential equation yields the costate equation;

[0057]

[0058] (3)u * (t) satisfies the state equation, and the Hamiltonian function with respect to the costate variable λ(t) i Finding the partial differentials yields the state equations;

[0059]

[0060] (4)u * (t) satisfies the initial and final boundary conditions;

[0061] Equation 7 is the specific representation of the Hamiltonian function, where the first term L(x(t),u(t),t) represents the transition objective function of the system, characterizing the transition from the initial time t0 to the final time t. f The cost of system changes and the specific objective function are shown in Equation 8:

[0062] H(x(t),u(t),λ(t),t)=L(x(t),u(t),t)+λ(t)·f(x(t),u(t),t) (7)

[0063]

[0064] In the formula m fc m fc_equ m bat_equThese represent the total hydrogen consumption of the vehicle under actual operating conditions, the equivalent hydrogen consumption calculated based on fuel cell degradation, and the equivalent hydrogen consumption calculated based on power battery degradation, respectively. The total hydrogen consumption of the vehicle under actual operating conditions is calculated based on the vehicle dynamics equation and the calorific value of hydrogen. The equivalent hydrogen consumption calculated based on fuel cell life degradation and power battery life degradation are calculated based on Equations 9 and 10, respectively.

[0065]

[0066]

[0067] In the formula Δψ fc Indicates the percentage of fuel cell lifespan degradation; Δψ bat This indicates the percentage of battery life degradation. This method sets the lifespan of the fuel cell to terminate when the degradation reaches 10%; and the lifespan of the power battery to terminate when the degradation reaches 20%. M represents the percentage of battery lifespan degradation. fc For the price of fuel cell stacks, M bat For the price of power batteries, The objective function of this method is given by the price per unit volume of hydrogen, as shown in Equation 11. The first term of the objective function is the final-state cost function φ(SOC(t)). f ),t f The second term is the system transition objective function L(x(t),u(t),t) mentioned above, and the final-state cost function is calculated according to Equation 12:

[0068] min J=min{φ(SOC(t f ),t f )+L(x(t),u(t),t)} (11)

[0069]

[0070] In Equation 12 This represents the amount of electricity consumed, calculated based on the difference in SOC (State of Charge) of the initial and final battery charge, converted into the amount of hydrogen required to produce the same amount of electricity by burning hydrogen.

[0071] Step 3: Initialize costate variables and iteratively solve for the optimal control sequence u. * (t)

[0072] Initially assign values ​​to the costate variables, discretize the control variables within the allowable power range of the fuel cell, calculate the Hamiltonian function, and solve for the control variable u(t) in the current instantaneous state. i ), that is, the fuel cell power P at this time. fc (t i If the SOC of the power battery satisfies the constraints of the state variables, then the desired control variable u(t) will be determined. iThe optimal control value u in this instantaneous state is... * (t i If the constraints are not met, the costate variables are reassigned and the Hamiltonian function is calculated. This process is repeated to obtain the optimal control values ​​for all states. Finally, all optimal control values ​​are connected to form a curve to obtain the optimal control sequence u. * (t) trajectory.

Claims

1. A fuel cell vehicle energy management control method considering multi-objective optimization, characterized by, The method comprises the following steps: Step one, establishing state equation and constraint condition Pontryagin's minimum principle, based on the classical variational method, requires a mathematical description of the actual system. This method selects the battery charge (SOC) as the state variable, with an initial value of 0.

7. The initial SOC of the actual battery charge is calibrated by adjustable parameters. Since both excessively high and low battery charge levels are detrimental to starting operation, this method restricts the battery charge (SOC(t)) at any given time to be at the highest allowable SOC for battery operation. max and minimum SOC min Within the interval, to ensure that the initial and final battery charge levels remain the same, this method sets the absolute difference in SOC between the initial and final battery charge levels to be less than 0.005 as a constraint condition for balance. SOC(t0) and SOC(t... f The initial and final states of charge of the power battery are represented by , respectively, as shown in Equation 1: Select fuel cell power P fc As a control variable of the system, in order to improve the lifespan of the fuel cell, this method limits the fuel cell power P at any given time. fc (t) are all located at the maximum power P fc_max and minimum power P fc_min Range; change in fuel cell power ΔP per unit time fc (t) is located at the maximum change ΔP fc_max and minimum change ΔP fc_min Within the range, as shown in Equation 2: The system state equation established by the method is in the form of formula 3: Step two, introducing co-state variable, constructing objective function and Hamilton function The traditional classical variation method is only suitable for solving the extreme value problem without constraint condition. For the optimal extreme value problem of actual system, the method introduces the covariant variable λ(t) to construct the Hamilton function H(x(t), u(t), λ(t), t), and further uses the Hamilton function to solve the optimal control sequence u * (t) of the system. The optimal control sequence u * (t) of the method satisfies the following four conditions: (1) u * (t) enables the Hamiltonian function to take a minimum value throughout the algorithm; H(x * (t),u * (t),λ * (t),t)≤H(x(t),u(t),λ(t),t) (4) (2) u * (t) satisfies the co-state equation, the Hamiltonian function is the partial derivative of the state variable x(t i ) with respect to the control variable u(t) (3) u * (t) satisfies the state equation, the partial derivative of the Hamiltonian function with respect to the co-state variable λ(t i ) is the state equation; (4) u * (t) satisfies initial and terminal boundary conditions; Formula 7 is a specific form of Hamilton function, wherein the first term L(x(t),u(t),t) represents an over-objective function of the system, representing a cost consumption of the system in a change process from a starting time t0 to a terminal time t f The cost consumption of the system in a change process is specifically shown in formula 8. H(x(t),u(t),λ(t),t)=L(x(t),u(t),t)+λ(t)·f(x(t),u(t),t) (7) wherein m fc , m fc_equ , m bat_equ respectively represent the hydrogen consumption of the whole vehicle under actual running conditions, the equivalent hydrogen consumption converted from the degradation of the fuel cell, and the equivalent hydrogen consumption converted from the degradation of the power battery, the hydrogen consumption of the whole vehicle under actual running conditions is calculated according to the automobile dynamics equation and the hydrogen heat value, and the equivalent hydrogen consumption converted from the degradation of the fuel cell and the equivalent hydrogen consumption converted from the degradation of the power battery are respectively calculated according to formula 9 and formula 10: where Δψ fc represents the percentage of fuel cell life degradation; Δψ bat represents the percentage of power battery life degradation, the method sets the fuel cell degradation degree to reach 10% when the life ends; the power battery degradation degree reaches 20% when the life ends, M fc is the price of fuel cell stack, M bat is the price of power battery, is the price of unit volume hydrogen, the objective function of the method is as shown in formula 11, the first item of the objective function is the terminal state cost function φ(SOC(t f ),t f ), and the second item is the system excess objective function L(x(t),u(t),t) described above, the terminal state cost function is calculated according to formula 12: minJ = min{φ(SOC(t f ),t f )+L(x(t),u(t),t)} (11) In formula 12 represents the power consumption calculated according to the initial and final power battery power SOC difference, which is converted into the amount of hydrogen required when burning hydrogen to emit the same amount of power; Step three, co-state variable initialization, loop iteration to solve optimal control sequence u * (t) The initial assignment of the co-state variable, the control variable is discretized in the power range allowed by the fuel cell and the Hamiltonian function is calculated, the control variable u(t) in the current instantaneous state is solved i , that is, the fuel cell power P fc (t i ) at this time, if the state of charge SOC of the power battery meets the constraint condition of the state variable, the control variable u(t) obtained i is taken as the optimal control value u * (t i ) in the instantaneous state, if the constraint condition is not met, the co-state variable is re-assigned and the Hamiltonian function is calculated, the above operation is repeated in a loop to obtain the optimal control value in all states, and the optimal control sequence u * (t) trajectory is obtained by connecting all the optimal control values.

Images