CN111332126A - Vehicle braking energy recovery control method and device, vehicle and storage medium - Google Patents

Abstract

The invention provides a vehicle braking energy recovery control method, a vehicle braking energy recovery control device, a vehicle and a storage medium, wherein the method comprises the steps of calculating the braking force required by the vehicle according to the current state of the vehicle when a vehicle controller receives a braking signal to obtain three control variables of front wheel friction braking torque, rear wheel friction braking torque and motor regenerative braking torque, and calculating the three control variables based on an improved genetic algorithm of a prediction model; executing a genetic algorithm under the framework of the model predictive control; adopting a multi-population combination iteration and average distribution method; after the optimal control sequence is output, recalculating three control variable values of the optimal control sequence at the next moment according to the vehicle state; and according to the calculated motor regenerative braking torque of the optimal control sequence at each moment, the vehicle control unit sends control signals to the motor and the controller thereof. The invention ensures the safety of the whole vehicle under the emergency braking working condition and improves the braking recovery energy under the conventional braking working condition.

Description

Vehicle braking energy recovery control method, device, vehicle and storage medium

technical field

The invention relates to a vehicle braking energy recovery control method, in particular to a braking energy recovery control method for a hybrid bus equipped with a coaxial parallel electromechanical coupling system and an electronically controlled pneumatic mechanical braking system.

Background technique

In recent years, with the continuous growth of car ownership, environmental pollution and energy shortage problems have become increasingly serious, and vehicle electrification has become an inevitable trend in the development of the automotive industry. As a representative technology of vehicle electrification, hybrid vehicles have gradually become a hot research topic in the automotive industry. Braking energy recovery technology, as one of the key technologies to achieve energy saving in hybrid electric vehicles, can convert kinetic energy during braking into electrical energy, thereby greatly improving the fuel economy of the vehicle. At present, the braking system of a hybrid vehicle is mainly divided into two configurations, one is a parallel configuration, which does not change the original mechanical system of the vehicle, but only adds the braking energy recovery torque to the original mechanical system. The braking function is completed together with the mechanical braking force provided by the system. The other is a series configuration, in which the brake pedal and the original mechanical braking system have been decoupled, and the total braking force is distributed by the mechanical braking system and the braking energy recovery system according to the control method, such as Liang Li , YuanboZhang,Chao Yang,et al.Model Predictive Control-based Efficient EnergyRecovery Control Strategy for Regenerative Braking System of Hybrid ElectricBus[J].Energy Conversion and Management,2016,111:299-314. When the motor is involved in the braking system, how can the braking torque required by the whole vehicle be properly distributed between the braking energy recovery system and the mechanical braking system in a variety of complex urban, suburban and even extreme working conditions to ensure the entire vehicle The optimal balance between vehicle stability and economy has become an urgent problem to be solved.

For electric vehicles, Li Liang et al. proposed a motor compensation braking control method based on sliding mode control theory. Using the characteristics of fast response speed of the motor, when the ABS system is triggered, the hydraulic machinery of the driving wheel can be quickly compensated by the motor torque. The braking torque change demand improves the stability of the system (LI Liang, LI Xujian, WANG Xiangyu et al. Transient switching control strategy from regenerative braking to anti-lock braking with a semi-brake-by-wire system [J]. Veh Syst Dyn. 2016, 54(2):231-257).

For HEVs, Yang Yang et al. designed a braking energy recovery and hydraulic braking coordination control system based on anti-lock braking control system hardware, and verified the effectiveness and feasibility of the scheme through simulation. For electric vehicles, TKBera proposed an anti-lock braking system and a braking energy recovery system coordinated controller based on synovial control theory, which ensures that the wheel slip rate remains at its optimal slip rate during emergency braking. ; For the braking control problem of front-wheel drive electric vehicles, Kanarachos designed an integrated control method based on the state Riccati equation.

Based on electric vehicles equipped with in-wheel motors, Wang Junmin proposed a nonlinear model predictive braking energy recovery control method. Based on the hybrid braking system of electric vehicles, based on the uncertainty model predictive control theory, Liu Wei et al. also proposed an uncertainty model predictive hybrid braking control method, which improved the economy and robustness of the vehicle.

In order to improve the braking energy as much as possible while satisfying the stability of the whole vehicle, Kim uses the optimal control method based on the conventional genetic algorithm to obtain the optimal problem of torque distribution between the braking energy recovery force and the hydromechanical braking force. , the braking energy recovery control method is designed and its effectiveness is verified.

Based on the above research, for the bus equipped with coaxial parallel electromechanical coupling system and electronically controlled pneumatic mechanical braking system, such as Yang C, Jiao X, Li L, et al. A robust H∞ control-based hierarchical modetransition control system for plug- In hybrid electric vehicle[J].MechanicalSystems and Signal Processing,2018,99:326-344, in a variety of complex urban, suburban and even extreme working conditions, the torque of the motor, battery and mechanical braking system is comprehensively considered And power constraints, on the premise of ensuring vehicle braking safety, the predictive control method based on improved genetic algorithm for efficient braking energy recovery is still blank.

SUMMARY OF THE INVENTION

The problem solved by the present invention is how to comprehensively consider the motor, battery and mechanical brake under various complex urban, suburban and even extreme working conditions for a passenger car equipped with a coaxial parallel electromechanical coupling system and an electronically controlled pneumatic mechanical braking system The torque and power constraints of the system are based on the improved genetic algorithm to efficiently predict the braking energy recovery to ensure the braking safety of the whole vehicle.

In order to solve the above problems, in the first aspect, the present invention proposes a vehicle braking energy recovery control method, which includes a wire-controlled pneumatic mechanical braking system composed of an air compressor, a cylinder and a brake valve, a vehicle controller, a motor A braking energy recovery control system consisting of its controller, gearbox, battery and its management unit, accelerator pedal position sensor, brake pedal position sensor, and vehicle speed sensor, in the wire-controlled pneumatic mechanical braking system for each wheel are equipped with air pressure regulating valve, the wire-controlled pneumatic mechanical brake system is used to individually adjust and control the wheel cylinder pressure of each wheel,

When the vehicle controller receives the braking signal, it calculates the required braking force of the vehicle according to the current state of the vehicle. The required braking force of the vehicle is distributed to three control variables of the front and rear axles, and the three control variables are the front and rear axles. wheel friction braking torque, rear wheel friction braking torque and motor regenerative braking torque, the three control variables are calculated by an improved genetic algorithm based on the prediction model;

Execute the genetic algorithm under the framework of the model predictive control, that is, obtain the values of the three control variables of the optimal control sequence by solving the optimal problem in the limited time domain at the current moment;

Multi-group combination iterative and average distribution methods are used to improve computational efficiency and prevent it from converging to local optimal solutions;

After the optimal control sequence is output, the values of the three control variables of the optimal control sequence are recalculated at the next moment according to the vehicle state, so as to realize the rolling optimization in the whole braking process;

According to the calculated motor regenerative braking torque of the optimal control sequence at each moment, the vehicle controller sends a control signal to the motor and its controller, so that the motor and its controller control The motor outputs the corresponding braking torque.

Further, the vehicle braking energy recovery control method further includes:

Using the control architecture of the model predictive control, at each moment based on the current state of the vehicle and previous historical information or expected vehicle speed, the genetic algorithm is used to calculate the optimal control sequence in the limited prediction time domain and the control time domain, Obtain the optimal control quantity at the current moment;

The three control variables are placed in different subpopulations, and when the prediction calculation is performed, the individuals of the different subpopulations are combined, and then the maximum fitness of each individual in all its combinations is used as its fitness value, and finally the Individuals of the different subpopulations are iteratively updated;

Then, using the method of uniform distribution of the initial population, for each of the different subpopulations, the available area that satisfies the constraints is divided into several average parts, and the boundary points of the several average parts are selected as the Individual values for the different subpopulations.

Further, the vehicle braking energy recovery control method further includes:

generating the different subpopulations of the three control variables within the constraints of the mean distribution;

permuting and combining individuals in the different subpopulations of the three control variables;

Use the prediction model to predict the result of the combined control variable sequence, and calculate the fitness of each individual based on the fitness function;

When the end condition is reached, the calculation ends and the value of the first control period of the three control variables in the optimal control combination is output;

When the end condition is not reached, the different subpopulations are selected according to the selection process of the genetic algorithm, and the selected individuals are crossed and mutated under the constraints to generate the next generation population individuals.

Further, the basic operator of the genetic algorithm includes a selection operator, and the selection operator is:

The individuals in the different subpopulations are evenly divided into the first level, the second level, the third level and the fourth level according to their fitness values. The probability of selection is 0.4, the probability of selection of the individual in the second level is 0.3, the probability of selection of the individual in the third level is 0.2, and the probability of selection of the individual in the fourth level is 0.1, Each of the selections is used to favor the selection of the individual that is superior in the selection of paternal and maternal species.

Further, the basic operator of the genetic algorithm further includes a crossover operator, and the crossover operator is:

After the parent body and the mother body are determined, the next generation individual needs to be generated according to its genes, and the crossover operator is determined according to the first formula and the second formula;

The first formula is:

u _i,j (t+1)=P ₁ u _ik (t)+P ₂ u _ih (t);

The second formula is:

u _i,j+1 (t+1)=P ₂ u _ik (t)+P ₁ u _ih (t);

Among them, P ₁ is a randomly generated value between 0 and 1, P ₂ is the difference between 1 and P ₁ , u _ik (t) and u _ih (t) are the parents selected in the t generation, u _{i, j} (t+1) and ui _,j+1 (t+1) are the offspring after cross-inheritance in the t+1 generation, i represents the number of subpopulations, and j represents the jth individual in the i population.

Further, the basic operator of the genetic algorithm also includes a mutation operator, and the mutation operator is:

In the process of generating the next generation of new individuals, a random number between 0 and 10 is randomly generated at the same time. If the random number is less than 8, the individual is not mutated; if the random number is greater than or equal to 8 , then the individual mutates, and the value it carries is randomly generated within the constraints.

Further, the vehicle braking energy recovery control method further includes:

In the iterative calculation process, the method of keeping the optimal is adopted, that is, the optimal individual in the previous generation population is retained in the next generation population, the individual has the greatest fitness, and its genes are retained for faster and more efficient calculation. convergence.

In a second aspect, the present invention also provides a vehicle braking energy recovery control device comprising:

The calculation unit is used for calculating the required braking force of the vehicle according to the current state of the vehicle when the vehicle controller receives the braking signal, the required braking force of the vehicle is distributed to three control variables of the front and rear axles, the three The three control variables are the front wheel friction braking torque, the rear wheel friction braking torque and the motor regenerative braking torque, and the three control variables are calculated by using an improved genetic algorithm based on the prediction model;

an execution unit, configured to execute the genetic algorithm under the framework of the model predictive control, that is, to obtain the values of the three control variables of the optimal control sequence by solving the optimal problem in the finite time domain at the current moment;

An optimization unit that uses a multi-group combination iterative and average distribution method to improve computational efficiency and prevent it from converging on a local optimum;

an output unit, configured to recalculate the values of the three control variables of the optimal control sequence at the next moment according to the vehicle state after outputting the optimal control sequence, so as to realize rolling optimization in the entire braking process;

A sending unit, configured to send a control signal to the motor and its controller according to the motor regenerative braking torque of the optimal control sequence obtained at each moment by the vehicle controller, so that the motor Its controller controls the motor to output the corresponding braking torque.

In a third aspect, the present invention also provides a vehicle, including a wire-controlled pneumatic mechanical brake system composed of an air compressor, a cylinder and a brake valve, and a vehicle controller, a motor and its controller, a gearbox, a battery and A braking energy recovery control system consisting of its management unit, accelerator pedal position sensor, brake pedal position sensor, and vehicle speed sensor, the wire-controlled pneumatic mechanical braking system is equipped with an air pressure regulating valve for each wheel. The air-controlled mechanical brake system is used to individually adjust and control the wheel cylinder pressure of each wheel, and also includes a computer-readable storage medium and a processor in which a computer program is stored, when the computer program is read and executed by the processor. , to realize the above-mentioned vehicle braking energy recovery control method.

In a fourth aspect, the present invention also provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is read and executed by a processor, the vehicle system as described above is implemented. Kinetic energy recovery control method.

The experimental results show that the proposed control method can ensure the safety of the whole vehicle under emergency braking conditions, and at the same time, under conventional braking conditions, it is the same as the rule-based braking energy commonly used in the current coaxial-parallel hybrid electric bus. Compared with the regenerative control method, the braking regenerative energy can be increased by 15%.

Description of drawings

FIG. 1 is a schematic diagram of the configuration of a coaxial parallel bus driving and braking system in an embodiment of the present invention.

FIG. 2 is a MAP diagram of motor efficiency in an embodiment of the present invention.

FIG. 3 is a characteristic diagram of the air brake system in the embodiment of the present invention.

FIG. 4 is a framework step diagram of a braking energy recovery control method in an embodiment of the present invention.

FIG. 5 is a simulation experiment result of a sandstone pavement in an embodiment of the present invention.

FIG. 6 is an experimental result of the braking energy recovery control method of the present invention under standard operating conditions in an embodiment of the present invention.

FIG. 7 is a comparison experiment result of a regular braking energy recovery control method under standard operating conditions in an embodiment of the present invention.

FIG. 8 is a hardware-in-the-loop experiment result of the braking energy recovery control method in the embodiment of the present invention.

Description of reference numbers:

1- air compressor, 2- dryer, 3- cylinder, 4- four-circuit switch valve, 5- cylinder, 6- brake pedal, 7- brake pedal stroke simulator, 8- brake valve, 9- Air pressure regulating valve, 10-engine, 11-clutch, 12-ISG motor, 13-inverter, 14-battery pack, 15-intermediate transmission mechanism, 16-brake wheel cylinder, 17-brake, 18-wheel.

Detailed ways

In order to make the above objects, features and advantages of the present invention more clearly understood, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

It should be noted that, in the following description, suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meaning per se. Thus, "module", "component" or "unit" may be used interchangeably.

A vehicle braking energy recovery control method according to an embodiment of the present invention firstly combines the mechanical structure and dynamic characteristics of the passenger car braking system to build a 7-DOF longitudinal dynamic model for the braking process of the hybrid electric bus; then combines the braking system The high nonlinearity of the tire in the critical stability field and the multi-objective characteristics of the performance requirements such as stability and economy in the braking process, the genetic algorithm is selected to optimize the mechanical braking torque of the front and rear axles and the braking torque of the motor in the limited time domain. The optimal allocation problem is predicted and solved, and the method of rolling optimization of the control cycle is adopted to realize the optimal control of the entire braking process. The method of table and nearest point implements the real-time processing of this control method.

vehicle model

For a bus equipped with a coaxial parallel electromechanical coupling system, a braking energy recovery control method based on an improved genetic algorithm is proposed. The configuration of the bus braking system is shown in Figure 1.

The drive system is composed of engine, clutch, motor and other components, and the braking system is divided into two parts, one is the braking energy recovery system composed of the motor, gearbox, battery and other components, and the other is the air compressor, cylinder and brake. The wire-controlled pneumatic mechanical brake system composed of valves, etc. Each wheel is equipped with an air pressure regulating valve in the control-by-wire pneumatic mechanical braking system, which can individually adjust and control the wheel cylinder pressure of each wheel.

The longitudinal dynamics model of the vehicle with seven degrees of freedom is established, and the main parts involved in the braking system are described as follows.

Vehicle Dynamics Model

Considering the characteristics of the suspension, a vehicle dynamics model is established. The front of the logo represents the forward direction of the vehicle, that is, the OX direction, the vehicle mass point is set to O, the vertical motion direction of the vehicle is represented by OZ, and the lateral motion direction is represented by OY. The model mainly considers Longitudinal and vertical motion characteristics of the car.

The specific formula in the model is as follows:

Longitudinal movement formula:

In the formula, m represents the mass of the passenger car (kg); m _s represents the mass of the sprung part of the passenger car (kg); F _x1 and F _x2 represent the driving force (N) provided by the ground facing the front and rear wheels respectively; F _resist represents the resistance (N ); D ₀ represents the distance (m) from the center of mass to the center axis of the vehicle's pitching motion, where the resistance is expressed as follows:

F _resist =F _f +F _w +mgi

In the formula, F _f represents rolling resistance (N); F _w represents air resistance (N); F _i represents slope resistance; f ₁ , f ₂ represent rolling resistance coefficient; C _D represents air resistance coefficient; The windward area (m ² ) of , i represents the slope.

The vertical motion formula of the car:

In the formula, F _s1 and F _s2 respectively represent the elastic forces (N) of the front and rear axle suspensions on the sprung part of the passenger car.

The pitch motion formula:

In the formula, J _y represents the moment of inertia of the pitching motion of the vehicle (kg·m ² ); a and b are the longitudinal distances from the center of mass of the vehicle to the front and rear axles (m); l represents the longitudinal distance between the front and rear axles (m).

Tire Movement Formula:

F _z1 =m _s gb/2l+m ₁ gK _b1 Z ₁

F _z2 =m _s ga/2l+m ₂ gK _b2 Z ₂

In the formula, Z ₁ and Z ₂ respectively represent the vertical displacement of the front and rear wheels (m); m ₁ and m ₂ respectively represent the wheel mass (kg) of the front and rear wheels; K _b1 and K _b2 respectively represent the vertical stiffness of the front and rear wheels (N /m); J ₁ , J ₂ respectively represent the moment of inertia of the front and rear wheels (kg·m ² ); R ₁ , R ₂ represent the wheel radius (m) of the front and rear wheels respectively; T _b1 , T _b2 represent the The mechanical braking torque (Nm); _Treb represents the braking energy recovery torque (Nm) provided by the motor. F _z1 and F _z2 represent the vertical loads (N) of the front and rear wheels to the ground, respectively. ω ₁ and ω ₂ are the angular velocities of the front and rear wheels, respectively.

The suspension motion formula is as follows

In the formula, C ₁ and C ₂ respectively represent the damping of the front and rear axle suspensions; K ₁ and K ₂ respectively represent the stiffness (N/m) of the front and rear axle suspensions.

Combining the above formulas, the model mainly considers the longitudinal velocity of the vehicle, the vertical displacement and pitch angle of the body above the suspension, the vertical displacement of the front and rear wheels, and the rotational angular velocity of 7 degrees of freedom.

tire model

There are many modeling schemes for tires. The present invention selects a commonly used magic formula to simulate tire characteristics. The specific formula is as follows:

μ _i =σD sin(Ctan ^-1 {BS _xi -E[BS _xi -tan ^-1 (BS _xi )]})

F _xi = μ _i F _Zi i∈{1,2}

In the formula, μ represents the friction coefficient, s represents the slip rate of the tire, σ represents the ground adhesion coefficient, B, C, D and E are the relevant parameters in the magic formula, and their specific meanings are B: stiffness factor; C: curve shape factor, which determines the shape characteristics of the curve; D: peak factor, which represents the maximum value of the curve; E: curve curvature factor, which determines the shape near the maximum value of the curve.

battery model

For the battery model, the present invention adopts a simple battery internal resistance model. The capacity parameter of the lithium battery is set to 80Ah, and the specific formula is as follows:

I ² ×R _int -V _oc ×I+P=0

In the formula, R _int is the internal resistance (Ω); I is the current (A); V _oc is the battery starting voltage (V); P is the load power (kw).

Where Q is the rated power (C); SOC ₀ is the initial SOC.

motor model

The maximum torque that the selected ISG motor can output is 750Nm, and the rated power and peak power of the motor are 92Kw and 121Kw respectively. As shown in Figure 2, the motor efficiency is mainly obtained through the calibrated motor efficiency MAP map.

Pneumatic Mechanical Brake System Model

The structure of the air brake system is shown in Figure 1. The system is powered by an air compressor, while the wheel cylinder pressure can be controlled by regulating valves on each wheel. In practice, the response curve of the air brake system can be appropriately simplified, as shown in Figure 3. The percentage (20%, 50%, 100%) represents the opening degree of the regulating servo valve. When a pressure command is given by the controller, the air brake system will firstly have a short braking time, and then the actual brake air pressure will increase approximately linearly, and finally reach the target pressure p _t . The rate of change of brake air pressure varies with the characteristics of the regulating valve. The response characteristics to reach the target pressure can be expressed as follows:

In the formula, p _t represents the target brake pressure, _ux represents the rate of change of the brake pressure, and τ ₀ represents the working time of the air brake system.

The relationship between brake air pressure and air brake torque is as follows:

T _b =k _pb p

Among them, T _b represents the pneumatic braking torque acting on the wheel, p represents the gas pressure of the brake wheel cylinder, and the coefficient k _pb can be calibrated by repeated tests of the pneumatic braking system.

The parameters used in the model are shown in Table 1.

Table 1 Some parameters in the simulation model

Brake energy recovery control method

Wheels and other components in the braking system have highly nonlinear characteristics in the critical instability region. Genetic algorithms can directly use nonlinear equations that are more relevant to the actual situation to solve the fitness, and select the best fitness among all historical individuals through continuous optimization and iteration. individual, so genetic algorithm is used to solve the optimal control sequence. In order to improve the effectiveness and reliability of the algorithm, the algorithm has been improved in a targeted manner. First, the algorithm is put under the framework of model predictive control, that is, the optimal problem is obtained by solving the optimal problem in the limited time domain at the current moment. After outputting the sequence, the optimal problem is recalculated according to its vehicle state at the next moment, so as to realize the rolling optimization in the whole braking process. converges to a local optimal solution. The genetic algorithm has a huge amount of calculation, and it is difficult to meet the real-time demand. In view of this defect, a multi-dimensional table is made according to the input of the genetic algorithm, and the real-time processing is carried out by the method of the nearest point.

In the braking process, multiple objectives such as economy and safety need to be considered. The purpose of the present invention is to follow the driving intention as much as possible and recover the braking energy to the greatest extent under the premise of ensuring the safety of automobile braking, so as to be within the allowable range of other indicators. Improve vehicle economy. In order to improve the effectiveness of the genetic algorithm, the present invention improves it. As shown in FIG. 7 , the process of the braking energy recovery control method of the improved genetic algorithm of the present invention is divided into five steps. The specific process is shown in Table 2:

Table 2 Control method flow

computational model

In order to simplify the calculation, the 7-DOF dynamic model is no longer used in the genetic algorithm, but the 3-DOF model is used to predict the driving state of the vehicle under different control sequences. Compared with the 7-DOF dynamic model, the 3-DOF model no longer considers the impact of the suspension system on the driving state of the vehicle. In the braking energy recovery control method, the vehicle dynamics model with 3 degrees of freedom is selected as the prediction model to predict the future state of the vehicle. The specific formula is as follows:

ma ₁ (k)=F _x1 (k)+F _x2 (k)-F _resist (k)

α ₁ (k)=(T _b1 (k)-R ₁ F _x1 (k))/J ₁

ω ₁ (k+1)=ω ₁ (k)+α ₁ (k)Ts

α ₂ (k)=(T _b2 (k)+T _reb −R ₂ F _x2 (k))/J ₂

ω ₂ (k+1)=ω ₂ (k)+α ₂ (k)Ts

The variables used in the above formula can be calculated by:

F _resist (k)=F _f (k)+F _w (k)

F _f (k)=mg(f ₁ +f ₂ v(k))

S ₁ (k)=(ω ₁ (k)R ₁ -v(k))/v(k)

S ₂ (k)=(ω ₂ (k)R ₂ -v(k))/v(k)

μ ₁ (k)=σDsin[Ctan ^-1 (BS ₁ (k)-E{BS ₁ (k)-tan ^-1 [BS ₁ (k)]})]

μ ₂ (k)=σDsin[Ctan ^-1 (BS ₂ (k)-E{BS ₂ (k)-tan ^-1 [BS ₂ (k)]})]

F _x1 (k)=F _z1 (k) μ ₁ (k)

F _x2 (k)=F _z2 (k)μ ₂ (k)

where k is the number of time steps; v is the vehicle speed; Ts is the sampling time; a ₁ is the acceleration; α ₁ and α ₂ are the angular accelerations of the front and rear wheels, respectively.

constraint

Combined with the system characteristics, the constraint conditions for setting the braking energy recovery torque are as follows: first, combined with the motor MAP map, the maximum torque that the motor can output under the current state is obtained by looking up the current motor speed; second, combined with Battery status, get the maximum power that the current battery can output, and then combine the converted power and speed to get the maximum braking energy recovery torque; thirdly, based on the set maximum braking energy recovery torque change rate limit and upper The current braking energy recovery torque limit is obtained from the braking energy recovery torque at a moment. Since the pneumatic braking torque that the system can provide is relatively large, the constraints on the pneumatic braking torque mainly consider the limit of the maximum torque change rate set by the system. The specific description formula is as follows:

T _reb ω _m η _{motor ≤P batt_lim}

|T _b2 (k+1)-T _b2 (k)|≤T _pchange,max ·Ts

|T _b1 (k+1)-T _b1 (k)|≤T _pchange,max ·Ts

where η _motor is the motor efficiency, ω _m is the motor speed, P _{batt_lim} is the maximum charging power limit of the battery in the braking energy recovery system in the current state; T _reb,max is the maximum torque limit of the motor in the current state; T _rechange,max is the limit of the maximum rate of change of the braking energy recovery torque; T _pchange,max is the limit of the maximum rate of change of the pneumatic braking torque at the current moment.

The constraints of different subgroups are different. The constraints of the subgroup representing the regenerative braking torque of the motor are mainly the constraints of the maximum braking torque of the motor and the maximum charging power of the battery at the current speed, which represent the constraints of the mechanical friction braking torque of the front and rear axles. The condition is mainly the rate of change of the braking torque.

fitness function

During the braking process, the braking stability should be ensured first, the driving intention of the upper layer should be realized secondly, and the economy should be improved as much as possible under the above premise. The present invention divides the braking state into a general control mode and an anti-lock braking mode based on the slip ratio by comprehensively weighing the above performance.

General Control Mode

The general condition for starting the control mode is that the slip ratio of the front and rear wheels does not exceed the set value L1 during the braking process. In this case, since the slip ratio is low, the main consideration is to realize the driving intention and improve the economy. In addition, it should be noted that as the slip ratio increases, the anti-lock braking control mode should not be triggered as much as possible by adjusting the braking force distribution ratio between the front and rear wheels. Its fitness function is as follows:

where:

e _i (k)=v _ref (k+i|k)-v(k+i|k), i=1,...,h _p

η _i =η _trans η _motor , i=1,...,h _p

where _vref is the desired vehicle speed based on driving intent. e _i is the deviation between the expected vehicle speed and the predicted vehicle speed by calculation. h _c represents the control domain under the framework of the model prediction algorithm. h _p represents the prediction domain under the framework of the model prediction algorithm. η _i is the efficiency of the braking energy recovery system during braking. η _trans is the efficiency of the drive train from the motor to the wheels. J is fitness. w _x , w _y and w _z are the weighting factors of driving intention, economy and braking stability, respectively, and w _z is zero when the maximum slip rate of the front and rear wheels is less than L2, that is, due to the low slip rate at this time, The influence of slip rate on stability is no longer considered. After the slip rate is greater than L2, in order not to trigger the anti-lock braking control mode as much as possible, the weighting factor increases rapidly with the increase of slip rate.

Anti-lock Braking (ABS) Control Mode

The condition for the activation of the anti-lock braking control mode is that the slip ratio of any one wheel exceeds the set value L1. In this mode, due to the large slip rate, there is a risk of wheel locking and thus causing the vehicle body movement to be in an unstable area, so the main consideration in this mode is the vehicle body stability. Its fitness function is as follows:

where S _xrefer is the expected wheel slip rate.

end condition

The end condition of the genetic algorithm is generally the iterative algebra or the performance reaches a certain level. The end condition of this method is set to a certain iterative algebra, and the algorithm ends when the iterative algebra is reached.

Improvement of Genetic Algorithm Based on Model Predictive Control Framework

Considering the stability and economy of the whole vehicle, a new improved genetic algorithm based on the prediction model is used to solve the problem of braking torque distribution in the braking process. The specific description is as follows:

First, the control architecture of model predictive control is adopted. Based on the current state variables of the vehicle and previous historical information at each moment, the genetic algorithm is used to calculate the optimal control sequence in the finite prediction time domain and the control time domain, and then output the most optimal control sequence at the current moment. To optimize the control amount, repeat the process at each subsequent control time node, and finally realize the rolling optimization control of the entire braking process.

Secondly, in the genetic algorithm with limited number of iterations, if all control variables are placed in the same population, it may cause mutual interference between different control variables and affect the optimization efficiency of the algorithm. In order to solve this problem, the present invention puts three control variables in different sub-populations. When predicting, individuals of different sub-populations are combined, and then the best fitness of each individual in all its combinations is used as its fitness. value, and finally the individuals of the subpopulation are iteratively updated.

Then, using the method of uniform distribution of the initial population, for each population, the available area that satisfies the constraints is divided into several average parts, and the boundary points of each part are selected as the individual values of the initial population.

Then, according to the specific conditions of the optimization problem, the basic operators in the genetic algorithm process, including the selection operator, the crossover operator and the mutation operator, are set as follows:

Selection operator: The individuals in the population are equally divided into four levels according to their fitness values. In each selection, the selection probability of individuals in the first level is 0.4, the second level is 0.3, the third level is 0.2, and the last level It is 0.1, so it is more inclined to choose the better individual when choosing the parent and the mother.

Crossover operator: When the parent and mother are determined, the next generation of individuals needs to be generated according to their genes. The equation of the crossover operator is as follows, and the value P ₁ between 0 and 1 is randomly generated, and P ₂ is the difference between 1 and P ₁ difference.

u _i,j (t+1)=P ₁ u _ik (t)+P ₂ u _ih (t)

u _i,j+1 (t+1)=P ₂ u _ik (t)+P ₁ u _ih (t)

Among them: u _ik (t), u _ih (t) are the parent bodies selected in the t generation, u _i,j (t+1), u _i,j+1 (t+1) are passed in the t+1 generation. The offspring after cross inheritance, i represents the number of subpopulations, and j represents the jth individual in the i population.

Mutation operator: In the process of generating the next generation of new individuals, a number between 0 and 10 is randomly generated at the same time. If the random number is less than 8, the individual does not mutate; if the random number is greater than or equal to 8, the individual mutates, and the value it carries is randomly generated within the constraint range.

Finally, in the iterative process of the genetic algorithm, the optimal method is adopted, that is, the optimal individual in the previous generation population is retained in the next generation population. The individual has good fitness, and retaining its genes will help the algorithm to be faster and more effective. convergence.

real-time method

The braking energy recovery control method based on genetic algorithm has a large amount of calculation, so the calculation efficiency is low. Aiming at this problem, an equivalent control method of the control method is designed and proposed. The real-time performance of the strategy is realized by selecting the control amount of the point closest to the current state in the table.

The above-mentioned control method based on the improved genetic algorithm is used to perform off-line operations to generate a multi-dimensional table based on the input set. Due to the limited storage space of the actual controller, the input set is simplified, and it is no longer necessary to input the air brake torque at the previous moment, that is, the change rate of air brake torque is no longer considered in the table, but the limit is placed After the multi-dimensional table, it is used as the output upper and lower limit hard constraints. The simplified input set W _v and the interval Δw between points in each dimension of the table are as follows, where the value of the front and rear wheel speeds is the equivalent value after multiplying the tire radius.

W=[v,ω ₁ ,ω ₂ ,v _ref ,z] ^T ∈R ⁵

W _v =[0,v _define -10,v _define -10,v _define -10Ts,0.1] ^T ≤w≤[90,v _define +1,v _define +1,v _define +Ts,0.9] ^T

Δw=[1,1,1,Ts,0.1] ^T

The required input of the method of the present invention includes: historical information is mainly used for vehicle speed prediction of the vehicle, if the vehicle speed prediction module is not included in the algorithm, that is, the vehicle speed in the future period is counted as a known quantity given by other controllers (such as Now the unmanned driving technology will carry out the vehicle speed planning for a period of time in the future), then the historical information is replaced by the expected vehicle speed; the vehicle status mainly includes the wheel speed of each wheel, the vehicle speed, the friction braking torque of the front and rear wheels, and the regenerative braking torque, and Pavement adhesion coefficient. The friction braking torque of the front and rear wheels and the regenerative braking torque are mainly used to calculate the constraints of the optimal control sequence, because the algorithm restricts the torque change rate considering the system vibration and the mechanical characteristics of the actuator.

The control sequence of the present invention refers to the optimal control sequence of the control variables calculated by the algorithm in the control domain. The braking force distribution method of the present invention mainly involves three control variables: the friction braking torque of the front wheels, the friction braking torque of the rear wheels , regenerative braking torque. If the control domain is 5 control cycles, the control sequence is a 3*5 matrix, that is, the output values of the three variables in each control cycle are combined into a control sequence, and only the first row of the control sequence is output during control, that is, only the first row of the control sequence is output. Output the value in the first control cycle, and after the next control cycle, the system will recalculate the corresponding optimal control sequence to perform rolling optimization.

Experimental verification

In order to verify the braking safety and energy recovery efficiency of the proposed control method and its real-time equivalent strategy, simulation and hardware-in-the-loop experiments are carried out respectively in the present invention. The verification consists of three parts: (1) vehicle braking stability verification; (2) vehicle braking energy recovery capability verification; (3) hardware-in-the-loop experiment.

Vehicle Brake Stability Verification

The simulation conditions of braking safety are set as follows: the initial vehicle speed is 80Km/h, the expected braking deceleration is -0.7g, the simulation is carried out on the gravel road, and the adhesion coefficient is set to 0.6.

First verify car safety on gravel roads. The simulation results are shown in Figure 5. In order to trigger the ABS mode, the expected deceleration is set to 0.7g, and the road adhesion coefficient is set to 0.6, so in Figure 5(a), the actual vehicle speed does not completely follow the expected vehicle speed, but it can be seen that in the full use of the gravel road Its braking deceleration is also approximately 0.6g after the adhesion. As shown in Figure 5(b), since the ABS mode is triggered, in order to prevent the wheels from locking, the braking process will no longer shift gears, the gears will be locked in high gear, and the transmission is relatively small, so relative to the mechanical braking torque , the braking energy recovery torque provided by the motor is small, but after calculation, it can be obtained that the motor has output its maximum braking torque, and the difference between the required braking torque and the braking energy recovery torque is given by Mechanical brake torque compensation. In Figure 5(c), it can be seen that the slip ratio is around 0.15, the wheels are not locked during the whole braking process, and the safety of the whole vehicle is guaranteed. As shown in Figure 5(d), the braking energy recovered during the braking process is kJ. Since the emergency braking process takes a short time and the main goal of the process is to ensure braking safety, the braking energy recovery efficiency is higher than Low.

Brake Energy Recovery Capability Verification

The simulation condition of the braking energy recovery efficiency is set to the Chinese national standard condition. The driving process adopts a common rule-based strategy. The benchmark strategy used as a comparison in the braking process is also a rule-based control method. This strategy is briefly described. As follows: When the brake pedal is lower than the threshold value C, the required braking torque is all provided by the motor. When it is higher than the threshold value C, the required braking torque is proportionally distributed between the front and rear axles (with the braking As the pedal increases, the proportion of braking torque distributed by the drive shaft gradually decreases, and finally the front and rear axles tend to their load proportions), in which the braking torque distributed on the rear axle is preferentially provided by the motor. The pneumatic mechanical braking torque is compensated, and the braking torque distributed on the front axle is all provided by the pneumatic mechanical braking torque.

The simulation results are shown in Figure 6 and Figure 7. In Figure (a), it can be seen that the vehicle speed following effect of the two strategies is very good, and the actual vehicle speed and the expected vehicle speed almost coincide. As shown in Figure (b), the proposed strategy has a larger motor torque during the braking process than the comparison strategy. Due to the small braking deceleration, when the braking energy recovery torque is not enough to meet the demand, the remaining The braking torque is almost entirely compensated by the rear wheel mechanical braking torque, which is not listed here. As shown in Figure (c), since there is no emergency braking in the whole working condition, the wheel slip ratio is within 0.1, which meets the predetermined requirements. Finally, as shown in Fig. (d), compared with the comparison strategy, the braking energy recovered by the braking energy recovery control method proposed in the present invention is greatly improved in the working condition.

hardware-in-the-loop experiment

The hardware-in-the-loop system mainly includes DSPACE and its host computer, controller and its host computer, etc. The DSPACE host computer has a vehicle model other than the control system, which communicates with DSPACE through the DSPACE dedicated line. In the host computer of the controller, the control method model generates the core control program through automatic code generation technology, and then the program is combined with some peripheral low-level communication programs to form a complete controller program. The host computer 2 and the controller pass the CAN line For communication, it is convenient to program programs to the controller. During the hardware-in-the-loop experiment, the whole vehicle and the environment are simulated in DSPACE, the controller is the actual controller, and the hardware-in-the-loop simulation is realized through CAN communication between DSPACE and the actual controller.

The working condition of the hardware-in-the-loop experiment is set as the emergency braking condition on the gravel road, which is the same as the simulation experiment of braking safety verification. The experimental results are shown in Figure 8. The curve is basically similar to Figure 5. However, because the control sequence in the equivalent strategy is obtained by looking up the table according to the state of the vehicle body, and the data changes greatly, there are certain fluctuations in the curve, and the overall control effect is slightly poor. Recovered braking energy is also less. However, it can be seen from Figure 8(c) that this strategy also ensures that the wheel slip ratio is within a safe range.

Table 3 gives the data comparison of the verification test in detail. As shown in the table, in the emergency braking condition, the model-in-the-loop result on the gravel road is better, and the maximum slip rate is 0.19. In the hardware-in-the-loop experiment , because the equivalent strategy is adopted, the control effect is not as good as the model-in-the-loop experiment, the maximum slip rate is 0.3055, and the wheels are not locked in all experiments, indicating that the proposed control method can ensure the safety of the vehicle during braking. Under the national standard operating conditions, the braking energy recovered by the rule-based control method is 4686.2kJ, while the braking energy recovered by the genetic algorithm-based braking energy recovery control method proposed in the present invention is 5397.8kJ, which is 5397.8kJ compared to the rule-based control method. The method increases by 15.19%, and the proportion of recovered energy in the total braking energy reaches 60.27%, thereby verifying the braking energy recovery efficiency of the present invention.

Table 3 Comparison of simulation results between the control method of the present invention and the benchmark method

In another embodiment of the present invention, a wheel braking energy recovery control device includes:

The calculation unit is used to calculate the braking force required by the vehicle according to the current state of the vehicle when the vehicle controller receives the braking signal, and the required braking force of the vehicle is distributed to the three control variables of the front and rear axles, and the three control The variables are the front wheel friction braking torque, the rear wheel friction braking torque and the motor regenerative braking torque, and the three control variables are calculated by an improved genetic algorithm based on the prediction model;

an execution unit, configured to execute the genetic algorithm under the framework of the model predictive control, that is, to obtain the values of the three control variables of the optimal control sequence by solving the optimal problem in the finite time domain at the current moment;

An optimization unit that uses a multi-group combination iterative and average distribution method to improve computational efficiency and prevent it from converging on a local optimum;

an output unit, configured to recalculate the values of the three control variables of the optimal control sequence at the next moment according to the vehicle state after outputting the optimal control sequence, so as to realize rolling optimization in the entire braking process;

A sending unit, configured to send a control signal to the motor and its controller according to the motor regenerative braking torque of the optimal control sequence obtained at each moment by the vehicle controller, so that the motor Its controller controls the motor to output the corresponding braking torque.

In another embodiment of the present invention, a vehicle includes a wire-controlled pneumatic mechanical brake system composed of an air compressor, a cylinder and a brake valve, and a vehicle controller, a motor and its controller, a gearbox, a battery and its A braking energy recovery control system consisting of a management unit, an accelerator pedal position sensor, a brake pedal position sensor, and a vehicle speed sensor, the wire-controlled pneumatic mechanical braking system is equipped with an air pressure regulating valve for each wheel, and the wire-controlled pneumatic mechanical braking system is provided with an air pressure regulating valve. The pneumatic mechanical brake system is used to individually adjust and control the wheel cylinder pressure of each wheel, and further comprises a computer-readable storage medium and a processor in which a computer program is stored, when the computer program is read and executed by the processor, The vehicle braking energy recovery control method as described above is realized.

In another embodiment of the present invention, a computer-readable storage medium stores a computer program, and when the computer program is read and executed by a processor, the above-mentioned braking energy is realized recycling control methods.

Although the present invention is disclosed above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be based on the scope defined by the claims.

Claims (10)

1. A vehicle braking energy recovery control method, comprising a wire-controlled pneumatic mechanical braking system composed of an air compressor, a cylinder and a brake valve, a vehicle controller, a motor and its controller, a gearbox, a battery and its A braking energy recovery control system consisting of a management unit, an accelerator pedal position sensor, a brake pedal position sensor, and a vehicle speed sensor, the wire-controlled pneumatic mechanical braking system is equipped with an air pressure regulating valve for each wheel, and the wire-controlled pneumatic mechanical braking system is provided with an air pressure regulating valve. The pneumatic mechanical braking system is used to individually adjust and control the wheel cylinder pressure of each wheel, and is characterized in that, When the vehicle controller receives the braking signal, it calculates the required braking force of the vehicle according to the current state of the vehicle. The required braking force of the vehicle is distributed to three control variables of the front and rear axles, and the three control variables are the front and rear axles. wheel friction braking torque, rear wheel friction braking torque and motor regenerative braking torque, the three control variables are calculated by an improved genetic algorithm based on the prediction model; Execute the genetic algorithm under the framework of the model predictive control, that is, obtain the values of the three control variables of the optimal control sequence by solving the optimal problem in the limited time domain at the current moment; Multi-group combination iterative and average distribution methods are used to improve computational efficiency and prevent it from converging to local optimal solutions; After the optimal control sequence is output, the values of the three control variables of the optimal control sequence are recalculated at the next moment according to the vehicle state, so as to realize the rolling optimization in the whole braking process; According to the calculated motor regenerative braking torque of the optimal control sequence at each moment, the vehicle controller sends a control signal to the motor and its controller, so that the motor and its controller control The motor outputs the corresponding braking torque. 2. The vehicle braking energy recovery control method according to claim 1, further comprising: Using the control architecture of the model predictive control, at each moment based on the current state of the vehicle and previous historical information or expected vehicle speed, the genetic algorithm is used to calculate the optimal control sequence in the limited prediction time domain and the control time domain, Obtain the optimal control quantity at the current moment; The three control variables are placed in different subpopulations, and when the prediction calculation is performed, the individuals of the different subpopulations are combined, and then the maximum fitness of each individual in all its combinations is used as its fitness value, and finally the Individuals of the different subpopulations are iteratively updated; Then, using the method of uniform distribution of the initial population, for each of the different subpopulations, the available area that satisfies the constraints is divided into several average parts, and the boundary points of the several average parts are selected as the Individual values for the different subpopulations. 3. The vehicle braking energy recovery control method according to claim 1, further comprising: generating the different subpopulations of the three control variables within the constraints of the mean distribution; permuting and combining individuals in the different subpopulations of the three control variables; Use the prediction model to predict the result of the combined control variable sequence, and calculate the fitness of each individual based on the fitness function; When the end condition is reached, the calculation ends and the value of the first control period of the three control variables in the optimal control combination is output; When the end condition is not reached, the different subpopulations are selected according to the selection process of the genetic algorithm, and the selected individuals are crossed and mutated under the constraints to generate the next generation population individuals. 4. The vehicle braking energy recovery control method according to any one of claims 1-3, wherein the basic operator of the genetic algorithm comprises a selection operator, and the selection operator is: The individuals in the different subpopulations are evenly divided into the first level, the second level, the third level and the fourth level according to their fitness values. The probability of selection is 0.4, the probability of selection of the individual in the second level is 0.3, the probability of selection of the individual in the third level is 0.2, and the probability of selection of the individual in the fourth level is 0.1, Each of the selections is used to favor the selection of the individual that is superior in the selection of paternal and maternal species. 5. The vehicle braking energy recovery control method according to claim 4, wherein the basic operator of the genetic algorithm further comprises a crossover operator, and the crossover operator is: After the parent body and the mother body are determined, the next generation individual needs to be generated according to its genes, and the crossover operator is determined according to the first formula and the second formula; The first formula is: u _i,j (t+1)=P ₁ u _ik (t)+P ₂ u _ih (t); The second formula is: u _i,j+1 (t+1)=P ₂ u _ik (t)+P ₁ u _ih (t); Among them, P ₁ is a randomly generated value between 0 and 1, P ₂ is the difference between 1 and P ₁ , u _ik (t) and u _ih (t) are the parents selected in the t generation, u _{i, j} (t+1) and ui _,j+1 (t+1) are the offspring after cross-inheritance in the t+1 generation, i represents the number of subpopulations, and j represents the jth individual in the i population. 6. The vehicle braking energy recovery control method according to claim 5, wherein the basic operator of the genetic algorithm further comprises a mutation operator, and the mutation operator is: In the process of generating the next generation of new individuals, a random number between 0 and 10 is randomly generated at the same time. If the random number is less than 8, the individual is not mutated; if the random number is greater than or equal to 8 , then the individual mutates, and the value it carries is randomly generated within the constraints. 7. The vehicle braking energy recovery control method according to claim 6, further comprising: In the iterative calculation process, the method of keeping the optimal is adopted, that is, the optimal individual in the previous generation population is retained in the next generation population, the individual has the greatest fitness, and its genes are retained for faster and more efficient calculation. convergence. 8. A vehicle braking energy recovery control device, comprising: The calculation unit is used for calculating the required braking force of the vehicle according to the current state of the vehicle when the vehicle controller receives the braking signal, the required braking force of the vehicle is distributed to three control variables of the front and rear axles, the three The three control variables are the front wheel friction braking torque, the rear wheel friction braking torque and the motor regenerative braking torque, and the three control variables are calculated by using an improved genetic algorithm based on the prediction model; an execution unit, configured to execute the genetic algorithm under the framework of the model predictive control, that is, to obtain the values of the three control variables of the optimal control sequence by solving the optimal problem in the finite time domain at the current moment; An optimization unit that uses a multi-group combination iterative and average distribution method to improve computational efficiency and prevent it from converging on a local optimum; an output unit, configured to recalculate the values of the three control variables of the optimal control sequence at the next moment according to the vehicle state after outputting the optimal control sequence, so as to realize rolling optimization in the entire braking process; A sending unit, configured to send a control signal to the motor and its controller according to the motor regenerative braking torque of the optimal control sequence obtained at each moment by the vehicle controller, so that the motor Its controller controls the motor to output the corresponding braking torque. 9. A vehicle, comprising a wire-controlled pneumatic mechanical brake system composed of an air compressor, a cylinder and a brake valve, and a vehicle controller, a motor and its controller, a gearbox, a battery and its management unit, and the position of an accelerator pedal A braking energy recovery control system consisting of a sensor, a brake pedal position sensor, and a vehicle speed sensor, the wire-controlled pneumatic mechanical braking system is equipped with an air pressure regulating valve for each wheel, and the wire-controlled pneumatic mechanical braking system uses For individually adjusting and controlling the wheel cylinder pressure of each wheel, it is characterized in that it also includes a computer-readable storage medium and a processor storing a computer program, when the computer program is read and executed by the processor, the The vehicle braking energy recovery control method according to any one of claims 1-7. 10. A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is read and executed by a processor, any one of claims 1-7 is implemented. The described vehicle braking energy recovery control method.

Images