WO2022241898A1

WO2022241898A1 - Hierarchical energy-saving driving method for fuel cell vehicle

Info

Publication number: WO2022241898A1
Application number: PCT/CN2021/101441
Authority: WO
Inventors: 魏小栋; 孙超; 孙逢春; 任强; 周飞鲲
Original assignee: 北京理工大学; 广州汽车集团股份有限公司
Priority date: 2021-05-18
Filing date: 2021-06-22
Publication date: 2022-11-24
Also published as: CN113085666A; CN113085666B

Abstract

A hierarchical energy-saving driving method for a fuel cell vehicle. In upper-layer calculation, road information of all road sections, such as speed limits, road gradients, signal light positions, and signal timing of the road sections, is acquired; a fuel cell vehicle longitudinal dynamics model and a cost function which comprises vehicle required power and driving time are constructed; and then a vehicle optimal driving trajectory under global vehicle speed planning is calculated by means of an optimization algorithm. In lower-layer calculation, a fuel cell system model, a motor system model, and a power battery system model are established according to the power structure of the fuel cell vehicle. The model is convexified to establish a vehicle dynamics model meeting the standard paradigm of a convex optimization algorithm. An optimal vehicle speed trajectory calculated on the upper layer is input as a vehicle driving condition to solve the energy management problem of the vehicle. According to the calculation result, vehicle speed trajectories, power battery SOC trajectories, and fuel cell output power trajectories of all road sections are obtained. For the problem of optimal ecological driving of a fuel cell vehicle, a vehicle speed trajectory optimization method using the minimum required power of the vehicle and the arrival time of the vehicle as optimization goals is systematically provided; secondly, a standard paradigm of a convex optimization algorithm for solving fuel cell energy management is established, and the problem of energy management is quickly solved by means of iteration, thereby laying a foundation for implementing model-based prediction control.

Description

A hierarchical fuel cell vehicle energy-saving driving method

technical field

The invention relates to the technical field of vehicle speed planning and energy management of a fuel cell vehicle, in particular to a method for energy-saving driving of a stratified fuel cell vehicle.

Background technique

In recent years, with the development of new technologies in the field of electronic information, new technologies such as the Internet of Things, cloud computing, and big data are penetrating into traditional industries. In the automotive industry, connected and automated vehicles have gradually become a research hotspot and are causing tremendous changes in the industry. The networking of automobiles can allow vehicles to better perceive traffic information, such as road speed limits, traffic signal information, etc., and improve driving safety and convenience. The automation of the vehicle can improve the computing power of the on-board computer, better combine traffic information and vehicle dynamics to realize vehicle speed prediction, planning and energy management of the vehicle, reduce vehicle energy consumption and improve driving economy.

Energy-saving driving is to reduce the total power demand of the vehicle by rationally planning the driver's driving behavior and vehicle speed trajectory, thereby reducing fuel efficiency. According to different traffic information, vehicle energy-saving driving methods can be divided into two categories: one is energy-saving driving methods that do not consider traffic signal information, this method only needs to consider basic information of the road (speed limit, slope, etc.) and the vehicle's own conditions ( engine and motor torque limit and speed limit, etc.), combined with vehicle dynamics, use some optimization algorithms to achieve energy-saving driving of the vehicle, which is suitable for road sections without traffic lights; Obtain the phase and timing information of the traffic light signal, consider the driving situation of the vehicle on the road section with the signal light, and use the optimization algorithm to obtain the optimal speed trajectory of the vehicle at different arrival times, so as to minimize the energy consumption of the vehicle.

In the prior art, ecological driving methods for traditional fuel vehicles, electric vehicles, and hybrid vehicles based on different traffic information have all been proposed. However, no one has been involved in the energy-saving driving of fuel cell vehicles based on signal light information.

Contents of the invention

Aiming at the defects in the prior art, the invention provides a layered fuel cell vehicle energy-saving driving method, which solves the defects in the prior art.

The technical names defined in the present invention are as follows:

SoC: The ratio of the remaining capacity of the battery to the capacity of the fully charged state, and its value range is 0-1.

Global vehicle speed planning: refers to the speed planning for the whole process of a certain vehicle from the starting point to the end point.

Convex conversion: refers to the convex processing of the established mathematical model through some technical means, so that the functions and feasible regions in the model meet the basic requirements of the convex optimization algorithm, so as to achieve the purpose of using the convex optimization algorithm to solve the problem.

Model Predictive Control: At each sampling instant, the current control solution is obtained by solving a finite time-domain open-loop optimal control problem, and the optimized control idea is completed through rolling time-domain optimization.

In order to realize above object of the invention, the technical scheme that the present invention takes is as follows:

The invention proposes a layered optimization algorithm to solve the optimal ecological driving problem and the vehicle energy management problem when the fuel cell vehicle passes through traffic lights.

In the upper layer, the optimization algorithm is used to solve the optimal vehicle speed planning problem when the vehicle passes through the traffic lights. Firstly, the road information of the whole road section is obtained based on technical means such as the Internet of Vehicles, such as the speed limit of each road section, road slope, signal light position, and signal timing. Then combined with vehicle kinematics, a fuel cell vehicle longitudinal dynamics model and a cost function including vehicle demand power and travel time are constructed. Finally, the optimization algorithm is used to obtain the optimal driving trajectory of the vehicle and the output power trajectory of the fuel cell under the global vehicle speed planning.

In the lower layer, a convex optimization algorithm is used to solve the energy management problem of the vehicle. First, according to the power structure of the fuel cell vehicle, the fuel cell system model, the motor system model and the power battery system model are established. Secondly, the fuel cell vehicle dynamics model is transformed into a convex process, and a vehicle dynamics model that satisfies the standard paradigm of convex optimization algorithm is established. Then, the optimal vehicle speed trajectory calculated by the upper layer is input as the vehicle driving condition to solve the vehicle energy management problem. Finally, according to the calculation results, the vehicle speed trajectory, power battery SOC trajectory and fuel cell power trajectory of the whole road section are obtained.

Specific steps are as follows:

S1. Based on the Internet of Vehicles and other technical means to obtain road information of the entire road section, such as road length, speed limit, number of signal lights, phase, timing, etc., to establish a road model, the specific steps are as follows:

When driving on a road with a total length of S _tol and there are N traffic lights, assuming that the i-th light is at a distance from the starting point of S _i , then S _i ∈ [0, S _tol ], i={1, 2, 3 , 4...n}; In the road, the cycle of each signal light is independent and always remains unchanged, and the cycle of each signal light can be set according to the needs; define the cycle of the i-th traffic light as T ⁱ ∈ R ⁺ , Each cycle includes three phases of red light, green light, and yellow light, and the duration is respectively

but

use

Indicates the running time of the i-th traffic light in its own cycle when the vehicle departs; the absolute time when the vehicle passes the i-th red street light is

Calculate the time of the traffic light in its own cycle when the vehicle passes the i-th traffic light

but

S2. Combined with vehicle longitudinal kinematics, construct a fuel cell vehicle longitudinal dynamics model and a cost function including vehicle demand power and travel time.

The demanded power in the cost function in this step can be the sum of the total demanded power of the vehicle, the total positive demanded power of the vehicle, the total negative demanded power of the vehicle, and the absolute value of the demanded power of the vehicle at each moment; take the total positive demanded power as an example :

st v _min (s)≤v(s)≤v _max (s)

a _min (s)≤a(s)≤a _max (s)

t(S _tol )≤t _ref

v(0)=v(S _tol )=v _min (s)

Among them, λ is the penalty coefficient, S is the driving distance, v _min is the minimum speed of the vehicle, v _max is the maximum speed of the vehicle, a _min is the maximum deceleration of the vehicle, a _max is the maximum acceleration of the vehicle, and t _ref is the latest arrival time.

S3. Using the optimization algorithm to obtain the optimal driving trajectory of the vehicle under the global vehicle speed planning;

The optimization algorithm in this step may be dynamic programming, distribution estimation, minimum value principle, genetic algorithm and the like. Taking dynamic programming as an example, record the traveling distance of the vehicle as S, select the vehicle speed and traveling time as the state variable x=[v(s), t(s)] ^T during the solution process, and select the acceleration as the control variable u=a(s); Assuming that the vehicle can only drive forward during the whole process, the vehicle speed v≥0; the state transition equation is v(s+1)=v(s)+a(s), t(s+1)=t(s)+1 /v(s+1).

S4. The speed planning results in the upper-level calculation can be presented in the following ways: one is to form the time-distance curve of the vehicle; the other is to form the speed-distance curve of the vehicle; the third is to combine the first two curves to generate the time-speed of the vehicle curve. The above three results can be input as the driving conditions of the underlying energy management.

S5. Combining the power system structure of the fuel cell vehicle, establish a fuel cell vehicle dynamics model, mainly including the fuel cell system model, the motor system model and the power battery system model, and the energy transfer relationship between the three systems.

S6. Perform convex conversion processing on the established fuel cell system model, motor system model and power battery system model, and establish a standard paradigm of convex optimization algorithm. This process must not only maintain the accuracy of the model, but also enable the model to satisfy the properties of convex functions. The convex conversion methods of the three system models are as follows:

The conversion method of the fuel cell system model is to combine the working efficiency of the fuel cell system, and fit the hydrogen combustion release power of the fuel cell into a quadratic function about the output power of the fuel cell system;

The convex conversion method of the motor system is to combine the working efficiency and characteristic curve of the motor, and fit the output power of the motor to a quadratic function about the input power of the motor and the motor speed;

The convex conversion method of the power battery system is to simplify the battery into a first-order equivalent circuit, and fit the output power of the battery to a quadratic function related to the chemical energy of the battery.

S7. Using a convex optimization algorithm to solve the energy management problem of the vehicle under the optimal driving trajectory. The optimization algorithm in this step can be alternate direction multiplier method, interior point method, etc.

S8. According to the result calculated in step S5, the vehicle speed trajectory, power battery SOC trajectory and fuel cell power trajectory of the whole road section are obtained.

Compared with the prior art, the present invention has the advantages of:

Aiming at the optimal ecological driving problem of fuel cell vehicles, the present invention systematically proposes a vehicle speed trajectory optimization method with the minimum required power of the whole vehicle and the vehicle arrival time as optimization targets. Secondly, a standard paradigm for solving fuel cell energy management with convex optimization algorithm is established, and the energy management problem is quickly solved through iteration, which lays the foundation for the realization of model predictive control.

Description of drawings

Fig. 1 is a schematic flow chart of the fuel cell vehicle energy-saving driving method of the present invention;

Fig. 2 is the phase and signal timing situation of embodiment traffic lights;

Fig. 3 is the power topological diagram of the fuel cell vehicle power system of the embodiment;

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and examples.

As shown in Figure 1, a hierarchical fuel cell vehicle energy-saving driving method includes the following steps:

1. Establish the road model where the vehicle is about to travel

Based on the Internet of Vehicles and other technical means to obtain road information of the entire road section, such as the speed limit of each road section, road slope, signal light position, and signal timing, etc., to establish a road model;

(1) Get road information

The road information of the entire road section is obtained based on technical means such as the Internet of Vehicles. Specifically, the road model used for global vehicle speed planning includes the following information:

(a) the total length of the travel section;

(b) the speed limit of vehicles in each driving section;

(c) slope information of each road section in the road to be driven;

(d) The position information of each signal light on the road to be driven from the starting point;

(e) The phase and signal timing of each signal light, etc.

(2) Establish road model

When driving on a road with a total length of S _tol and there are N traffic lights, assuming that the i-th light is at a distance from the starting point of S _i , then:

S _i ∈ [0, S _tol ], i={1, 2, 3, 4...n} (1)

On the road, the cycle of each signal light is independent and always remains unchanged, and the cycle of each signal light can be set according to the needs. Define the cycle of the i-th traffic light as T ⁱ ∈ R ⁺ , each cycle includes three stages of red light, green light, and yellow light, and the durations are respectively

As shown in Figure 2, then:

use

Indicates the running time of the i-th traffic light in its own cycle when the vehicle starts. The absolute time when the vehicle passes the i-th red street light is

Therefore, we can calculate the time of the traffic light in its own cycle when the vehicle passes the i-th traffic light

2. Global speed planning

A fuel cell vehicle longitudinal dynamics model and a cost function including vehicle demand power and travel time are constructed. Using the optimization algorithm to obtain the optimal vehicle trajectory under the global vehicle speed planning;

(1) Establish vehicle longitudinal dynamics model

According to the longitudinal dynamics of the vehicle, the required power of the vehicle can be expressed as:

Among them, m is the vehicle mass, v is the vehicle speed, ρ _a is the air density, A is the windward area, c _d is the drag coefficient, Cr is the rolling resistance coefficient, _g is the gravity acceleration, and θ is the road slope.

(2) Establish a cost function that includes vehicle demand power and travel time

When solving the vehicle speed planning problem, the power system structure of the vehicle is not considered, the vehicle is regarded as a mass point, and the total required power and the time it takes to reach the destination are taken as the optimization objectives. During the driving process of the vehicle, generally the smaller the required power is, the lower the fuel consumption will be. The following cost function is established:

(3) Use the optimization method to solve the speed planning problem

Alternative optimization methods include dynamic programming, distribution estimation, minimum principle, genetic algorithm, etc. Let's take dynamic programming as an example.

During the solution process, the vehicle speed and travel time are selected as the state variables, and the acceleration is selected as the control variable.

Assume that the vehicle can only move forward during the whole process, and the vehicle speed v≥0. The state transition equation is:

3. Establish vehicle power system model

(1) Establish the energy transfer relation of fuel cell vehicle power system

Figure 3 shows the power system structure of a fuel cell vehicle, and the drive energy of the vehicle is jointly provided by the fuel cell and the power battery. Among them, P _drv is the required power of the whole vehicle, and the fuel cell system converts the chemical energy P _fcs generated by the reaction of hydrogen and oxygen into electrical energy P _fc for external output. The chemical energy _Pb of the power battery is transmitted in the form of electrical energy _Pc . P _c and P _fc are coupled to form P _d to act on the motor, and the motor then transmits the mechanical power P _em to the output shaft. During this process, the energy of the fuel cell can only be output and transferred to the power battery and the motor, and the energy of the power battery and the motor can be converted into each other.

P _dem (t) = P _em (t) (8)

P _mot (t) = P _fc (t) + P _c (t) (9)

P _d =P _fc +P _c (10)

(2) Establish fuel cell system model

The fuel cell system converts chemical energy P _fcs into electrical energy P _fc through the electrochemical reaction of hydrogen and oxygen, and then transfers it to the motor and power battery in the form of electrical energy. Considering the efficiency of the fuel cell system itself, the following relationship exists in the energy conversion process:

P _fc =P _fcs η _fc (11)

Among them, _ηfc is the fuel cell system efficiency.

Then calculate the instantaneous hydrogen consumption by the low calorific value of hydrogen LHV _H2 = 120MJ/kg

(3) Establish the motor system model

The output shaft speed ω _out and the motor speed ω _mot are calculated according to the transmission ratio τ of the vehicle and the wheel radius R _wheel .

ω _out =v/R _wheel (13)

ω _mot = τω _out (14)

During the running of the vehicle, when the required power P _drv is positive, the coupled energy P _d is delivered to the motor by the fuel cell and the power battery. When the required power P _drv is negative, the motor transfers the energy P _em generated by braking to the power battery.

Among them, η _mot is the motor system efficiency.

(4) Establish a power battery system model

Fuel cell vehicles are usually matched with a power battery. First, it can make up for the shortcoming of the fuel cell’s slow dynamic response. Second, it can recover the energy generated when the vehicle brakes to ensure the safety and reliability of the system. For the convenience of research, the power battery system can be simplified into a first-order equivalent circuit, and the relationship between variables is as follows:

Among them, V _oc is the open circuit voltage of the power battery, V _batt is the load voltage, I _batt is the circuit current, and R is the internal resistance of the power battery. According to the above formula, the relationship between P _b and P _c can be deduced as:

4. Carry out convex processing on the fuel cell power system model

The object of the convex optimization algorithm is a mathematical model in which the feasible region is a convex set and the objective function is a convex function. Therefore, before using the convex algorithm, the fuel cell energy management model should be turned convex. In the original model, there are mainly three systems involved in energy conversion, namely the fuel cell system, the motor system and the power battery system. The following is the convexization process of the three system models.

(1) Convexation of the fuel cell system

The fuel cell system power P _{fcs, k} at time k is fitted to a quadratic function related to the output power P _{fc, k} through the fuel cell system efficiency diagram:

Among them, α ₂ , α ₁ , and α ₀ are constants, which do not change with the rotational speed like a fuel engine.

(2) Convexation of the motor system

The power sum P _d,k of the fuel cell and the power battery at time k is fitted to a quadratic function related to the motor power P _em,k and the motor speed ω _em,k through the motor system efficiency diagram:

Because the efficiency curves of the motor are different at different speeds, β ₂ , β ₁ , and β ₀ are constantly changing with the speed of the motor.

(3) Convexation of the power battery system

According to the open-circuit voltage V _oc of the power battery and the internal resistance R of the battery, the chemical energy P _b,k of the power battery at time k is fitted as a function related to the electric energy P _c,k . In order to maintain the convexity of the function, it is assumed that the open circuit voltage V _oc and the internal resistance R of the power battery are constant.

5. Solving vehicle energy management problems

(1) Establish the objective function

According to the power structure of the vehicle, the driving state can be divided into three situations.

Table 1 Three situations of vehicle motion state

In the above three cases, the fuel cell will be turned on only when P _drv,k ≥ 0, and the output power of the fuel cell can be expressed as follows:

Combining (18) and (21) the fuel cell system power can be expressed as:

Combining (12) and (21), the total hydrogen consumption of the vehicle, that is, the objective function, can be expressed as follows:

Among them, P _{b, max} and P _{b, min} are the maximum power and minimum power of the power battery respectively, P _{fcs, max} and P _{fcs, min} are the maximum power and minimum output power of the fuel cell system, P _{em, max} and P _{em , min} is the maximum power and minimum power of the motor.

(2) Use convex optimization algorithm to solve the problem

Alternative convex optimization methods include alternating direction multiplier method, interior point method, etc.

The Alternating Direction Multiplier Algorithm can be used to solve distributed convex optimization problems. Its standard form is as follows:

Where f and g are convex functions, and x and z are two sets of variables.

Write the objective function (23) as a convex function about the power of the power battery P _b :

The constraints corresponding to Ax+Bz=c in formula (24) are:

Among them, I is a unit vector, φ is an N-dimensional column vector whose elements are all 1, and ψ is a lower triangular matrix of N×N. The relationship between power battery energy E, SOC and power Q _batt is as follows:

The scaling form of function (25) is:

Among them, ρ ₁ and ρ ₂ are penalty coefficients, both of which are greater than 0.

In order to avoid the overrun of power battery power P _b and SOC in the iterative process, the following function is defined:

When the ADMM algorithm is used to solve the problem, the specific iterative process of each parameter is as follows:

E ^j+1 ＝∏ ^k [φE ₀ +ψζ ^j+1 +λ ^j ] (32)

λ ^j+1 ＝λ ^j +φE ₀ +ψζ ^j+1 -E ^j+1 (33)

It can be seen from the power structure that the higher the output power of the power battery, the smaller the output power of the fuel cell system, so the initial value of the iteration is set as:

The original and dual residuals are:

The threshold ε ^primal of the original residual and the threshold ε ^dual of the dual residual can be adjusted according to the requirement of simulation accuracy.

6. Obtain vehicle speed trajectory, power battery SOC trajectory and fuel cell power trajectory of the whole road section.

Through the first and second steps in the above method, the optimal vehicle speed trajectory of the vehicle on the upcoming road section can be calculated, including the time-distance curve, speed-distance curve and time-speed curve of the vehicle.

The SOC-time curve, SOC-speed curve and SOC-distance curve of the power battery when the vehicle is traveling at the optimal speed trajectory in the road section can be calculated through steps 3, 4 and 5 of the above method. In addition, fuel cell power-time curves, fuel cell power-speed curves and fuel cell power-distance curves can be calculated.

Those skilled in the art will appreciate that the embodiments described here are to help readers understand the implementation method of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical revelations disclosed in the present invention without departing from the essence of the present invention, and these modifications and combinations are still within the protection scope of the present invention.

Claims

A method for energy-saving driving of a stratified fuel cell vehicle, comprising the following steps:

S1. Obtain road information of the entire road section based on technical means such as Internet of Vehicles, and establish a road model.

S2. Combined with vehicle longitudinal kinematics, construct a fuel cell vehicle longitudinal dynamics model and a cost function including vehicle demand power and travel time;

S3. Using the optimization algorithm to obtain the optimal driving trajectory of the vehicle under the global vehicle speed planning;

S4. The speed planning results in the upper-level calculation can be presented in the following ways: one is to form the time-distance curve of the vehicle; the other is to form the speed-distance curve of the vehicle; the third is to combine the first two curves to generate the time-speed of the vehicle Curve; the above three results can be input as the driving conditions of the underlying energy management;

S5. Combining the power system structure of the fuel cell vehicle, establish a fuel cell vehicle dynamics model, mainly including the fuel cell system model, the motor system model and the power battery system model, and the energy transfer relationship between the three systems;

S6. Perform convex conversion processing on the established fuel cell system model, motor system model and power battery system model, and establish a standard paradigm of convex optimization algorithm;

S7, using a convex optimization algorithm to solve the energy management problem of the vehicle under the optimal driving trajectory;

S8. According to the result calculated in step S5, the vehicle speed trajectory, power battery SOC trajectory and fuel cell power trajectory of the whole road section are obtained.
A method for energy-saving driving of a stratified fuel cell vehicle according to claim 1, wherein the specific steps of step S1 are as follows:

When driving on a road with a total length of S tol and there are N traffic lights, assuming that the i-th light is at a distance from the starting point of S i , then S i ∈ [0, S tol ], i={1, 2, 3 , 4...n}; In the road, the cycle of each signal light is independent and always remains unchanged, and the cycle of each signal light can be set according to the needs; define the cycle of the i-th traffic light as T i ∈ R + , Each cycle includes three phases of red light, green light, and yellow light, and the duration is respectively
but

use
Indicates the running time of the i-th traffic light in its own cycle when the vehicle departs; the absolute time when the vehicle passes the i-th red street light is
Calculate the time of the traffic light in its own cycle when the vehicle passes the i-th traffic light
but
A hierarchical fuel cell vehicle energy-saving driving method according to claim 2, characterized in that the optimization algorithm in step S3 is one of dynamic programming, distribution estimation, minimum value principle, and genetic algorithm.
A kind of hierarchical fuel cell vehicle energy-saving driving method according to claim 2, characterized in that the optimization algorithm in step S3 is dynamic programming, record the vehicle travel distance as S, and select the vehicle speed and travel time as the state in the solution process Variable x=[v(s), t(s)] T , select acceleration as the control variable u=a(s); assume that the vehicle can only move forward during the whole process, and the vehicle speed v≥0; the state transition equation is

v(s+1)=v(s)+a(s), t(s+1)=t(s)+1/v(s+1).
A method for energy-saving driving of a hierarchical fuel cell vehicle according to any one of claims 1 to 4, characterized in that, in step S6, the turning-convex processing methods of the three system models are as follows:

The conversion method of the fuel cell system model is to combine the working efficiency of the fuel cell system, and fit the hydrogen combustion release power of the fuel cell into a quadratic function about the output power of the fuel cell system;

The convex conversion method of the motor system is to combine the working efficiency and characteristic curve of the motor, and fit the output power of the motor to a quadratic function about the input power of the motor and the motor speed;

The convex conversion method of the power battery system is to simplify the battery into a first-order equivalent circuit, and fit the output power of the battery to a quadratic function related to the chemical energy of the battery.