CN104401232B - Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL - Google Patents
Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL Download PDFInfo
- Publication number
- CN104401232B CN104401232B CN201410793965.0A CN201410793965A CN104401232B CN 104401232 B CN104401232 B CN 104401232B CN 201410793965 A CN201410793965 A CN 201410793965A CN 104401232 B CN104401232 B CN 104401232B
- Authority
- CN
- China
- Prior art keywords
- data
- output
- prediction
- torque
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000005284 excitation Effects 0.000 claims abstract description 31
- 238000004088 simulation Methods 0.000 claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000013461 design Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000005540 biological transmission Effects 0.000 claims description 8
- 230000004044 response Effects 0.000 claims description 7
- 230000014509 gene expression Effects 0.000 claims description 6
- 238000011161 development Methods 0.000 abstract description 4
- 238000012360 testing method Methods 0.000 abstract description 3
- 230000007246 mechanism Effects 0.000 description 10
- 230000000694 effects Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000011160 research Methods 0.000 description 5
- 230000009194 climbing Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 3
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 229910052744 lithium Inorganic materials 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000003750 conditioning effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/72—Electric energy management in electromobility
Landscapes
- Feedback Control In General (AREA)
Abstract
A kind of electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL, belongs to electric vehicle engineering field。It is an object of the invention to can effectively carry out electric automobile drive and braking moment optimized distribution, to realize the electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL of vehicular longitudinal velocity tracking and controlling method。The present invention is first against the suitable excited data of the dynamic properties design of system, to ensure the abundant excitation to system；Secondly, the inputoutput data that excitation obtains is constructed the prediction output equation of system；Then, it is considered to executor's firm constraints of motor, set of cells and brake, the cost function of Model Predictive Control Algorithm structure moment of torsion optimal control is utilized；Finally, corresponding by solving cost function optimal problem, it is thus achieved that control input and act on system, thus realizing the control to system。Having only to when obtaining excited data in control algolithm of the present invention needs some off-line simulations to test, and development cost is relatively low。
Description
Technical Field
The invention belongs to the technical field of electric automobiles.
Background
With the increasing environmental pollution and energy crisis, pure electric vehicles become the major direction of the current automobile industry development due to the advantages of high energy efficiency, low emission, low noise, energy recovery and the like, a large number of preferential policies are provided by the country to support enterprises to research electric vehicles, and during the period of fifteen, the country 863 plans the major science and technology project of electric vehicles to be continuously supported by the country in two periods. The problem of torque optimization is involved in both traditional automobiles and electric automobiles, and how to design a controller to reasonably generate driving and braking torque signals to realize good performance of the vehicle is one of important problems of automobile industry research. For the electric automobile, the structure is simple, the control problem can be properly simplified, meanwhile, the response speed of the motor is high, and the torque and the rotating speed of the motor are easy to obtain, so that good basic conditions are provided for the torque optimization control of the electric automobile. Aiming at the torque optimization control of the centralized electric vehicle, the following problems are mainly solved:
1. the research object is a centralized pure electric vehicle, only one motor is provided, and the vehicle is driven by a front wheel, so that the problem of distribution optimization does not exist, the torque distribution optimization design can not be carried out like a four-wheel hub motor driven electric vehicle, and the torque distribution optimization design can only be considered from the aspect of generation of driving/braking torque in a vehicle controller.
2. The torque optimization control system of the electric automobile has the advantages of complex mechanism model, strong nonlinearity and difficult modeling. Meanwhile, the output driving and braking torque signals of the controller are required to meet the hard constraints of the motor and the brake of the actuator, namely the output driving and braking torque signals of the controller cannot exceed the actual maximum output torque of the motor and the maximum braking torque of the brake.
3. The electric automobile needs a power supply, a lithium battery pack is commonly used at present to supply power to the motor, and the supply voltage of the motor also influences the maximum output torque of the motor, so the influence of the output voltage of the battery pack must be considered when the maximum output torque of the motor is considered.
Disclosure of Invention
The invention aims to provide an electric vehicle torque optimization method based on data drive prediction control, which can effectively perform electric vehicle drive and brake torque optimization distribution to realize a vehicle longitudinal speed tracking control method.
The method comprises the following steps:
a. building a centralized electric automobile simulation model: a modularization method is adopted, a whole vehicle model of a centralized electric vehicle is built in AMESim, the whole electric vehicle simulation model mainly comprises an electric drive module, a transmission module, a tire module and vehicle longitudinal dynamics, and parameters of the whole vehicle model are as shown in table I:
electric vehicle parameter meter
；
b. Torque optimization data driven predictive controller design: the adopted data-driven predictive control algorithm is formed by combining subspace identification and model predictive control, and comprises the following specific steps:
selection of excitation data: selecting a driving torque command and a braking torque command as excitation input signals, and in an AMESim whole vehicle model, giving designed driving torque and braking torque data to a motor and a brake, and performing open-loop operation on the model to obtain a measured actual vehicle speed, wherein the driving torque command, the braking torque command and the measured vehicle speed are finally required system excitation data; and obtaining input Hankel matrixes Up and Uf of the system according to the data, and outputting Hankel matrixes Yp and Yf:
；
the prediction equation:
the prediction output equation of the system constructed by the input and output Hankel matrix is as follows:
，(1)
namely, the system future output value estimated by the subspace identification method is used for obtaining two prediction matrix coefficients Lw and Lu in the prediction equation by solving the least square problem
(2)
The solution for the two prediction matrix coefficients Lw and Lu is found by:
(3)
after a Hankel matrix of input and output information of a known system is obtained and Lw and Lu are obtained, a subspace prediction equation (1) is utilized to obtain a future output value of the system;
the final output prediction equation is as in equation (4),is a Hankel matrixFirst column of
，(4)
Designing a data-driven predictive controller: and solving a target cost function to obtain optimized driving torque and braking torque signals, wherein the mathematical expression form of the cost function is as follows:
(5)
the prediction equation (4) of the system output is rewritten as an incremental version:
；(6)
accumulating the increased term prediction output to obtain the following model prediction output equation:
(7)
wherein,
；
the F item is the system free response, and the S _ uf item is called the control response of the system correspondingly;
constructing a cost function of the following form by using the incremental prediction output model:
(8)
wherein,
，； (9)
where u and y are the weighting matrices for the input and output items, respectively, the control time domain is Nu and the prediction time domain is Np, the weighting factor Λ = u/y, the selected Λ value is:
the maximum torque constraint of the motor, the maximum braking torque constraint of the brake and the constraint of the actual vehicle speed in the torque optimization system are described by the following mathematical expressions:
(10)。
the invention mainly aims at the torque optimization problem of a centralized front wheel drive electric automobile, and realizes the optimization of the driving and braking torque of the electric automobile by designing a controller based on a data driving predictive control algorithm. The data-driven predictive control algorithm is a perfect combination of data-driven subspace identification and model predictive control, and a controller is designed directly through input and output data acquired offline and characteristics obtained through data processing, so that complex mechanism modeling is effectively avoided. A torque prediction equation can be obtained through the collected input and output data, then a cost function is constructed by using a model prediction control method, constraint conditions are considered, and optimized driving and braking torque command signals are obtained through optimization solution. The torque optimization controller can well solve three problems in the prior art, can effectively avoid complex system mechanism modeling based on a data-driven predictive control algorithm, can effectively avoid complex system mechanism modeling, and considers the actuator hard constraints of a motor, a battery pack and a brake. The beneficial effects are that:
1. most of traditional control algorithms are based on mechanism models of control systems, but in the actual torque optimization process, the mechanism modeling is complex and the workload is large. The data-driven predictive control algorithm directly designs the controller through input and output data acquired offline and characteristics obtained through data processing, so that complex mechanism modeling is effectively avoided, intermediate solution to a state space equation of the model is omitted, the calculated amount is reduced on one hand, and the problem of model errors introduced when the state space model is identified from data is avoided on the other hand.
2. The torque optimization controller system designed in the invention is a nonlinear system, and in consideration of the actuator hard constraints of a motor, a battery pack and a brake, the traditional control algorithm cannot effectively process the constraints of the system, while the data-driven predictive control algorithm can effectively process the control problem of multi-input multi-output belt constraints, and directly compiles the constraints into the S _ function in the simulink for on-line solution during the solution.
3. The development process of the traditional automobile electric control unit needs a great deal of experience and calibration tests. The control algorithm of the invention only needs some off-line simulation experiments when obtaining the excitation data, and the development cost is lower.
Drawings
FIG. 1 is a block diagram of an electric vehicle torque optimization control based on data-driven predictive control according to the present invention;
FIG. 2 is a centralized AMESim vehicle model of an electric vehicle according to the invention;
FIG. 3 is a graph of desired vehicle speed in m/s plotted on the abscissa for time in s, under the NEDC condition in accordance with the present invention;
FIG. 4 is a block diagram of a data-driven predictive control algorithm used in the present invention;
FIG. 5 is excitation data selected by the design torque optimization controller of the present invention with excitation inputs being drive and brake torque commands in Nm; the output data obtained off line is the actual vehicle speed, and the unit is m/s;
FIG. 6 is a diagram of the present invention for verifying the recognition result of a section of excitation data arbitrarily extracted from the excitation data by the prediction model, where the abscissa is time and the unit is s;
FIG. 7 is a block diagram of the recognition results obtained by redesigning a set of excitations for verifying the prediction model, where the abscissa is time and the unit is s;
FIG. 8 is a torque optimized data driven predictive controller built in simulink according to the present invention, which implements a data driven predictive control algorithm mainly through an M file compiler in Matlab;
FIG. 9 is a comparison of actual vehicle speed after the controller action and desired vehicle speed under the initial parameters described herein, wherein the dashed line represents actual vehicle speed, the solid line represents desired vehicle speed in m/s, and the abscissa is time in s;
FIG. 10 is a comparison of actual vehicle speed after being acted upon by the controller and desired vehicle speed with only grade change and other parameters held constant in accordance with the present invention, wherein the dashed line represents actual vehicle speed, the solid line represents desired vehicle speed, both in m/s, and the abscissa is time, in s;
FIG. 11 is a comparison of the actual vehicle speed after being acted on by the controller and the desired vehicle speed according to the present invention under the condition that only the load of the entire vehicle is changed and other parameters are kept unchanged, wherein the dotted line represents the actual vehicle speed, the solid line represents the desired vehicle speed, the units are m/s, the abscissa is time, and the unit is s;
FIG. 12 is a comparison of the actual vehicle speed after the controller action and the desired vehicle speed with only the wind speed changed and other parameters held constant, according to the present invention, where the dashed line represents the actual vehicle speed, the solid line represents the desired vehicle speed, both in m/s, and the abscissa is time, in s.
Detailed Description
The research method is based on data-driven model predictive control and comprises the following steps:
firstly, designing proper excitation data aiming at the dynamic characteristics of a system to ensure the sufficient excitation of the system; secondly, constructing a prediction output equation of the system by using input and output data obtained by excitation; secondly, considering actuator hard constraints of a motor, a battery pack and a brake, and constructing a cost function of torque optimization control by using a model predictive control algorithm; and finally, obtaining control input and acting on the system by solving the optimal problem corresponding to the cost function, thereby realizing the control of the system.
In order to solve the technical problems, the invention is realized by adopting the following technical scheme:
the torque optimization of the electric vehicle based on the data-driven predictive control is realized by the joint simulation of a software system. The software system consists of Matlab/Simulink and AMESim advanced simulation software. The AMESim software is a commercial complex system modeling simulation platform, and the AMESim software mainly plays a role in providing a high-fidelity concentrated motor type pure electric vehicle model which replaces a real vehicle as an implementation object of a control method in a simulation experiment; MATLAB/Simulink software is used for building a simulation model of the controller, namely the building of the torque optimization controller based on data-driven predictive control is completed through Simulink programming.
Functionally, the present invention may include the following: the controller simulation module and the centralized pure electric vehicle model. The function of each part is explained in detail as follows:
the centralized pure electric vehicle model mainly plays a role in designing the controller, namely input and output excitation data capable of reflecting the dynamic characteristics of the system are obtained through off-line simulation, and the controller constructs a prediction output equation of the system according to the data.
The controller module is mainly used for collecting various state information of the electric automobile model, such as automobile speed, battery output voltage, motor output torque and the like, then performing table lookup and operation, producing control signals, namely a driving torque signal and a mechanical braking torque signal of the motor, and sending the signals to an actuating mechanism, namely the motor and a traditional mechanical braking mechanism of the electric automobile;
the electric automobile model comprises an electric drive system model, a transmission shaft model, a vehicle dynamics model, a tire model and the like. The function of the device is to generate corresponding driving or braking torque according to the control signal given by the torque optimization controller, change the dynamic state of the vehicle, namely motor output torque, mechanical braking torque, vehicle speed and the like, and realize good performance of the vehicle.
A control block diagram implemented by the electric vehicle torque optimization method based on data drive prediction control is shown in a figure I, wherein a torque optimization controller is built in Simulink, the input of the controller is the expected vehicle speed, the actual vehicle speed is fed back to the controller in real time, Tmax is the maximum driving torque of a motor and is determined by the mechanical characteristics of the motor and the output voltage of a battery, the hard constraint condition of an actuator of the motor is embodied, and the influence of voltage reduction along with the increase of the discharge time of the battery on the performance of the whole vehicle is embodied. The drive torque obtained by the vehicle control unit must be equal to or less than Tmax, which is therefore given to the controller as a constraint. The centralized electric automobile model is built in AMESim and is used for simulating the operation of a real automobile. The driving torque signal and the braking torque signal which are obtained by optimizing the controller are respectively sent to the motor and the braking module to control the running of the vehicle, and the actual speed of the vehicle is fed back to the controller as a feedback signal.
The control objective of the invention is that the torque optimization controller compares the actual vehicle speed fed back in real time with the expected vehicle speed signal, optimizes and obtains driving and braking torque signals Tm and Tb on the premise of meeting the constraint condition, and sends the driving and braking torque signals Tm and Tb to a motor and a braking module in a whole vehicle model to control the running of the vehicle, and finally enables the actual vehicle speed to track the expected vehicle speed.
The invention provides a set of devices based on the operation principle and the operation process. Namely an offline electric vehicle torque optimization design test platform based on a PC. The construction and operation processes are as follows:
1. software selection
A controlled object of the control system and a simulation model of the controller are respectively built through software Matlab/Simulink and AMESim, the software versions are Matlab R2009a and AMESim R10, and solvers are selected to be ode3 and Euler respectively. The simulation step length is a fixed step length, and the step length is selected to be 0.1 s.
2. Joint simulation setup
To realize the joint simulation of Matlab/Simulink and AMESim, the environment variables of the PC computer must be set according to requirements, and the two are related to each other. Then, an interface module for communicating with Simulink is added in an AMESim interface, and variables needing communication between Matlab/Simulink and AMESim are connected to the module; and finally, after the system is compiled, the model information in the AMESim is retained in Simulink in the form of S-function, so that joint simulation and communication of the AMESim and the Simulink are realized. When the Simulink simulation model is run, the AMESim model is also calculated and solved at the same time. And data exchange is continuously carried out between the two in the simulation process. Recompilation is required if the model structure or parameter settings in the AMESim are modified. It is noted that the simulation steps for both must be identical.
The detailed steps of the invention are as follows:
a. building a centralized electric automobile simulation model: and (3) building a whole vehicle model of the centralized electric vehicle in AMESim by adopting a modularization method, as shown in figure two. The whole electric automobile simulation model mainly comprises an electric drive module, a transmission module, a tire module and vehicle longitudinal dynamics, and the parameters of the whole automobile model are as shown in the table I:
electric vehicle parameter meter
；
An electric drive system powers a vehicle and includes a battery portion and a motor portion. The battery is mainly used for supplying power to the motor, the battery part of the pure electric vehicle is formed by connecting a plurality of single batteries in series and in parallel, and the common battery pack is a lithium battery pack. The terminal voltage output by the battery pack is the sum of the output voltage of a single battery, the output terminal voltage of the battery system is the voltage provided by the battery pack to the motor, and the rated terminal voltage of the battery system in the model is 320V; the invention adopts a permanent magnet synchronous motor, the control method of the motor is a vector control method, and the maximum output torque of the motor is 300 Nm.
The transmission module mainly comprises a transmission and a drive shaft. The torque output by the motor is transmitted through the transmission to generate different speed ratios through different gear radiuses so as to reduce the rotating speed and increase the output torque, wherein the transmission ratio is 5:1, namely the rotating speed is changed into the original 1/5, and the torque is increased by 5 times, so that the problems of high rotating speed and insufficient torque of the motor are exactly solved. The invention mainly researches the linear running of the vehicle, and the output torque and the rotating speed of the left wheel and the right wheel are the same without difference, so that the differential mechanism is not arranged.
There are four tire models in the figure, and the tire uses Magic Formula model.
The vehicle longitudinal dynamics part, wherein the effects of driving force, braking force and driving resistance on the vehicle during driving are taken into consideration, wherein the driving resistance comprises air resistance, rolling resistance and friction resistance. In this module, parameters such as the overall mass of the vehicle, grade, wind speed, etc. can be set.
The selected cycle working condition of the model is a cycle working condition NEDC (New European driving cycle) of a light vehicle in Europe, the working condition comprises a plurality of times of long-time parking and uniform motion, the starting and braking curves are smooth, and the steady-state performance of the vehicle can be better reflected. The working condition simulates the driving state of the automobile under urban road conditions, and has strong practicability. The EUDC driving cycle takes 1200 seconds and the desired speed of the vehicle under this condition is shown in fig. three, which is the desired speed of the torque optimization controller.
b. The control objective of the controller in the present invention is to perform torque optimization while meeting the speed requirements for NEDC conditions, while the driving torque and braking torque should meet the actual motor torque capacity and battery constraints embodied in the constraint conditioning process when designing a data-driven predictive controller. The torque optimization data drive prediction controller of the invention is designed as follows: the adopted data-driven predictive control algorithm is formed by combining subspace identification and model predictive control, and the forming schematic diagram is shown in figure four. The method is mainly different from other control methods in that a controller is not required to be designed by using a traditional transfer function or state space model, a subspace linear prediction equation is applied to predict a future output value of the system, coefficients of the subspace prediction equation are obtained by directly calculating offline input and output data of the system, and the constrained data-driven prediction control method is obtained based on the prediction equation and by combining a basic principle of a model prediction control algorithm. The method comprises the following specific steps:
selection of excitation data: before designing a controller, appropriate excitation data needs to be designed according to the dynamic characteristics of the built whole vehicle model, and a system prediction output equation is obtained according to the excitation data. The driving torque command and the braking torque command are selected as excitation input signals, in an AMESim whole vehicle model, designed driving torque and braking torque data are sent to a motor and a brake, the model is operated in an open loop mode, the measured actual vehicle speed is obtained, and the simulation step length is 0.1 s. The drive torque command, the brake torque command, and the measured vehicle speed are the final required system excitation data as shown in fig. five. The excitation is selected as a result of the range of the drive torque: -300Nm, brake torque signal range: 0-1, vehicle speed range: 0-40m/s, and in order to ensure the accuracy of the excitation model, the excitation data is normalized, the normalization coefficient of the driving torque is 1/100, the normalization coefficient of the braking torque is 1, and the normalization coefficient of the vehicle speed is 1/40. It is noted that before the controller gives the drive and brake torque commands obtained by the optimization solution to the entire vehicle model, inverse normalization is required. After the excitation data are obtained, the excitation data are stored as data files, input Hankel matrixes Up and Uf of the system are obtained according to the data, and Hankel matrixes Yp and Yf are output:
the subscripts p and f of the matrix represent "past" and "future", respectively, and the number of columns of the Hankel matrix often needs to be sufficiently large in order to ensure that the data can contain sufficient information about the system status. For a single input single output system, the element in the Hankel matrix is a numerical value, and for a multiple input multiple output system, the element in the Hankel matrix is a column matrix of corresponding dimensions. The input of the invention is a driving torque command and a braking torque command, and the output is a vehicle speed, so that the elements input into the Hankel matrixes Up and Uf are 2-dimensional column matrixes, and the elements output into the Hankel matrixes Yp and Yf are a numerical value;
。
the prediction equation and the verification thereof are as follows:
the predicted output equation of the system constructed by the input and output Hankel matrix obtained in the previous step is as follows:
，(1)
namely, the system future output value estimated by the subspace identification method is used for obtaining two prediction matrix coefficients Lw and Lu in the prediction equation by solving the least square problem
(2)
The solution for the two prediction matrix coefficients Lw and Lu is found by:
(3)
after a Hankel matrix of input and output information of a known system is obtained and Lw and Lu are obtained, a subspace prediction equation (1) is utilized to obtain a future output value of the system; in practice, only the Hankel matrix is output when the controller is appliedThe first column of (a) is used as an estimation value of the system in the future, so that the calculation amount is greatly reduced on one hand, and the application of model predictive control is prepared on the other hand. The final output prediction equation is thus as in equation (4),is a Hankel matrixFirst column of
，(4)
And obtaining a prediction output model of the system according to the excitation data, and determining whether the prediction model can accurately predict the future output of the system for the data-driven prediction control algorithm so as to determine the quality of the control performance. In the following, the present invention performs certain verification on the obtained prediction equation through several sets of simulation experiments.
Fig. six is a recognition result of a piece of excitation data arbitrarily extracted from the excitation data. It can be seen that the predicted system output is very close to the actual system output. In addition, we redesign a set of stimuli as shown in fig. seven, and the predicted system output also tracks well with the actual system output. This indicates that the accuracy of the designed predictive model is satisfactory.
Designing a data-driven predictive controller: and after a prediction output equation output by the system is obtained, predicting a control algorithm according to the model. And obtaining optimized driving torque and braking torque signals by solving a target cost function, wherein the mathematical expression form of the cost function is as follows:
(5)
according to the prediction control theory, the incremental prediction controller is adopted, so that the controller has an integral function, the steady-state error of the system is reduced, and the control effect is improved. The prediction equation (4) for the system output is therefore rewritten as an incremental version:
；(6)
accumulating the increased term prediction output to obtain the following model prediction output equation:
(7)
wherein,
；
the F item is the system free response, and the S _ uf item is called the control response of the system correspondingly;
constructing a cost function of the following form by using the incremental prediction output model:
(8)
wherein,
，； (9)
in the invention, after a large number of experimental simulations are carried out in the later period, the selected prediction time domain is 20, and the selected control time domain is 4. the reference value Ref is the expected vehicle speed v _ q of the NEDC cycle working condition, and is updated to the controller in real time after being normalized. Λ = u/y is a weighting coefficient Λ value is selected after being adjusted by a large number of simulation experiments:
the maximum torque constraint of the motor, the maximum braking torque constraint of the brake and the constraint of the actual vehicle speed in the torque optimization system are described by the following mathematical expressions:
(10)。
the final torque-optimized data-driven predictive controller model is shown in fig. eight, primarily by Matlab
M file compiler implementation in (1). A vehicle speed simulation result graph obtained under initial parameters after debugging is shown in figure nine through an AMESim and Simulink combined simulation technology, and the graph shows that the actual vehicle speed is basically coincident with the expected vehicle speed, so that the control effect is good.
Experimental verification
In order to further verify the control performance of the torque optimization data-driven predictive controller, the performance of the torque optimization controller is verified under three working conditions that the gradient, the load and the wind speed of the whole vehicle are respectively changed and other parameters are unchanged in a centralized electric vehicle model.
1) Different slope
The requirement of the climbing capability of the electric automobile is generally more than or equal to 20 percent (namely 9 degrees), the climbing performance is directly influenced by the maximum torque of the motor through experiments (the climbing performance refers to that ENDC expected speed tracking can be realized under the slope), when the maximum torque is 100Nm, the climbing capability is 10 percent, and when the maximum torque exceeds 10 percent, the speed tracking cannot be realized; the maximum torque is 15% when 200Nm is reached, and the speed tracking cannot be realized after the maximum torque exceeds 15%; the maximum slope can reach 20% of the requirement at the maximum torque of 300 Nm. In the invention, the maximum torque of the motor is set to be 300Nm, other quantities are kept unchanged, the wind speed is 0m/s, the mass of the whole vehicle is 1000kg, only the gradient is changed, the maximum torque is respectively set to be 5%, 10% and 20%, and the obtained simulation result is shown in a tenth graph.
It can be seen from the figure that, since the vehicle has a tendency to slide down at the time of starting hill start due to the slope, the slope has a tendency to slide down more greatly, and the brake needs to be stepped on for actual hill start, which is reasonable in the model. The integral speed tracking effect is good along with the increase of the gradient, but because of the rigid condition constraint of the actuator of the maximum torque of the motor, when the gradient is large, the speed cannot be continuously increased after reaching the upper limit value, so that when the speed is high, the speed can only be maintained at the upper limit value, the tracking cannot be continuously carried out, but the speed tracking effect below the upper limit value of the speed is good. Where the upper speed limit is about 18m/s (64.8 km/h) at 20% grade, which is also reasonable in reality. The torque optimization data drive prediction controller still has good control performance under different gradient conditions.
2) Different load of the whole vehicle
The vehicle speed curve obtained by changing the entire vehicle mass m only, assuming that the average passenger weight is 75kg, setting the entire vehicle mass to 1500kg, (1500+2 × 75) kg, and (1500+4 × 75) kg, respectively, while keeping the other quantities constant, setting the gradient to 0 and the wind speed to 0m/s, is shown in fig. eleven.
It can be seen from the figure that the control effect is good overall under the quality, but the deviation increases along with the increase of the quality, and some overshoot occurs when the quality is (1500+4 × 75) kg, but the overall tracking effect is good, which indicates that the data-driven rolling optimization whole vehicle controller still has good control performance under different quality conditions.
3) Different wind speeds
The slope is set to 0 and the total vehicle mass is 1000kg, the ambient wind speed is only changed and set to 10m/s, 20m/s and 30m/s respectively, and the obtained vehicle speed comparison curve is shown in FIG. 11. As can be seen from the figure, after the ambient wind speed is changed to 30m/s, the tracking effect of the vehicle speed is still good, which shows that the torque optimization data drive prediction controller still has good control performance under the condition of different wind speeds.
The torque optimization controller based on data-driven predictive control is designed for a centralized electric automobile, the method skillfully combines a data-driven subspace identification method and a model predictive control method together, a prediction equation of a system is obtained according to input and output data of the system, control input of the system is obtained through model predictive control optimization, and meanwhile constraint of the system is explicitly considered. In order to verify the effectiveness of the torque optimization controller, a centralized front wheel drive electric automobile model is built in AMESim advanced simulation software, and the performance of the controller is verified under the working conditions of different gradients, different vehicle loads and different wind speeds. Simulation results show that the data-driven predictive controller vehicle control unit still has good control performance under the conditions of different qualities, different gradients and no ventilation speed.
Claims (1)
1. A torque optimization method of an electric vehicle based on data-driven predictive control is characterized by comprising the following steps: the method comprises the following steps:
a. building a centralized electric automobile simulation model: a modularization method is adopted, a whole vehicle model of a centralized electric vehicle is built in AMESim, the whole electric vehicle simulation model mainly comprises an electric drive module, a transmission module, a tire module and vehicle longitudinal dynamics, and parameters of the whole vehicle model are as shown in table I:
electric vehicle parameter meter
；
b. Torque optimization data driven predictive controller design: the adopted data-driven predictive control algorithm is formed by combining subspace identification and model predictive control, and comprises the following specific steps:
selection of excitation data: selecting a driving torque command and a braking torque command as excitation input signals, and in an AMESim whole vehicle model, giving designed driving torque and braking torque data to a motor and a brake, and performing open-loop operation on the model to obtain a measured actual vehicle speed, wherein the driving torque command, the braking torque command and the measured vehicle speed are finally required system excitation data; and obtaining input Hankel matrixes Up and Uf of the system according to the data, and outputting Hankel matrixes Yp and Yf:
；
the prediction equation:
the prediction output equation of the system constructed by the input and output Hankel matrix is as follows:
，(1)
namely, the system future output value estimated by the subspace identification method is used for obtaining two prediction matrix coefficients Lw and Lu in the prediction equation by solving the least square problem
(2)
The solution for the two prediction matrix coefficients Lw and Lu is found by:
(3)
after a Hankel matrix of input and output information of a known system is obtained and Lw and Lu are obtained, a subspace prediction equation (1) is utilized to obtain a future output value of the system;
the final output prediction equation is as in equation (4),is a Hankel matrixFirst column of
，(4)
Designing a data-driven predictive controller: and solving a target cost function to obtain optimized driving torque and braking torque signals, wherein the mathematical expression form of the cost function is as follows:
(5)
the prediction equation (4) of the system output is rewritten as an incremental version:
；(6)
accumulating the increased term prediction output to obtain the following model prediction output equation:
(7)
wherein,
；
the F item is the system free response, and the S _ uf item is called the control response of the system correspondingly;
constructing a cost function of the following form by using the incremental prediction output model:
(8)
wherein,
，； (9)
where u and y are the weighting matrices for the input and output items, respectively, the control time domain is Nu and the prediction time domain is Np, the weighting factor Λ = u/y, the selected Λ value is:
the maximum torque constraint of the motor, the maximum braking torque constraint of the brake and the constraint of the actual vehicle speed in the torque optimization system are described by the following mathematical expressions:
(10)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410793965.0A CN104401232B (en) | 2014-12-21 | 2014-12-21 | Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410793965.0A CN104401232B (en) | 2014-12-21 | 2014-12-21 | Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104401232A CN104401232A (en) | 2015-03-11 |
| CN104401232B true CN104401232B (en) | 2016-06-22 |
Family
ID=52638925
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410793965.0A Active CN104401232B (en) | 2014-12-21 | 2014-12-21 | Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104401232B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106647280B (en) * | 2017-01-16 | 2019-05-24 | 山东工商学院 | The data-driven Design of Predictive method and system of HEV mode switching |
| CN107269408B (en) * | 2017-05-15 | 2022-08-05 | 吉林大学 | Diesel engine optimized combustion controller and simulation model control method |
| CN107168104B (en) * | 2017-06-23 | 2020-06-16 | 吉林大学 | Observer-based longitudinal speed control method for pure electric intelligent automobile |
| CN107323308B (en) * | 2017-06-26 | 2019-06-25 | 杭州氢途科技有限公司 | A kind of fuel cell car tractive torque calculation method with predictive ability |
| CN107719187B (en) * | 2017-09-30 | 2020-06-23 | 合肥巨一动力系统有限公司 | Method for improving torque precision through torque control of electric automobile electric drive system |
| CN109035477B (en) * | 2018-07-04 | 2021-06-29 | 浙江中控技术股份有限公司 | Forklift equipment state comprehensive evaluation method, device and system |
| CN109039168A (en) * | 2018-07-13 | 2018-12-18 | 浙江工业大学之江学院 | The multiple dimensioned approximate explicit model forecast Control Algorithm of brshless DC motor |
| CN110877533B (en) * | 2018-09-05 | 2023-10-10 | 联合汽车电子有限公司 | Electric automobile power control system and method |
| CN110362871B (en) * | 2019-06-15 | 2023-04-18 | 江苏开沃汽车有限公司 | Simulation method of program for controlling tire slip |
| WO2021078390A1 (en) * | 2019-10-25 | 2021-04-29 | Zf Friedrichshafen Ag | Model-based predictive control of a drive machine of the powertrain of a motor vehicle and at least one vehicle component which influences the energy efficiency of the motor vehicle |
| CN115158037B (en) * | 2022-09-06 | 2022-12-09 | 中国重汽集团济南动力有限公司 | Torque distribution method and system for four-wheel drive motor of electric vehicle |
| CN115675121A (en) * | 2022-11-21 | 2023-02-03 | 智新控制系统有限公司 | Vehicle multidimensional torque control method and system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2964508B1 (en) * | 2010-09-06 | 2014-12-12 | Renault Sa | METHOD FOR CONTROLLING AN ELECTRIC MOTOR DRIVING A MOTOR VEHICLE |
| JP5830449B2 (en) * | 2012-08-30 | 2015-12-09 | 日立オートモティブシステムズ株式会社 | Electric vehicle drive system |
| CN103192737B (en) * | 2013-03-25 | 2015-10-07 | 吉林大学 | Drive control method for all-electric car |
| CN103303159B (en) * | 2013-05-09 | 2015-07-29 | 常州迈控智能科技有限公司 | Electric vehicle controller and Dual Drive control method thereof |
| CN104210383B (en) * | 2014-09-18 | 2016-05-25 | 上海工程技术大学 | A kind of four-wheel drive electric automobile torque distribution control method and system |
-
2014
- 2014-12-21 CN CN201410793965.0A patent/CN104401232B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN104401232A (en) | 2015-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104401232B (en) | Electric automobile torque optimization method based on data-driven PREDICTIVE CONTROL | |
| Wang et al. | Cyber-physical control for energy management of off-road vehicles with hybrid energy storage systems | |
| Huang et al. | Model predictive control power management strategies for HEVs: A review | |
| Khayyam et al. | Adaptive intelligent energy management system of plug-in hybrid electric vehicle | |
| Simmons et al. | Modeling and energy management control design for a fuel cell hybrid passenger bus | |
| CN107168104B (en) | Observer-based longitudinal speed control method for pure electric intelligent automobile | |
| Chen et al. | Adaptive energy-efficient control allocation for planar motion control of over-actuated electric ground vehicles | |
| Arce et al. | MPC for battery/fuel cell hybrid vehicles including fuel cell dynamics and battery performance improvement | |
| Zhao et al. | Parametric optimization of novel electric–hydraulic hybrid steering system based on a shuffled particle swarm optimization algorithm | |
| CN109933021A (en) | Consider the probabilistic vehicle platoon stability control method of Vehicle dynamic parameters | |
| CN102841315B (en) | Test platform for hardware-in-the-loop simulation of power storage battery system | |
| Wu et al. | Fast velocity trajectory planning and control algorithm of intelligent 4WD electric vehicle for energy saving using time‐based MPC | |
| Masoudi et al. | MPC-based battery thermal management controller for plug-in hybrid electric vehicles | |
| Sun et al. | Hybrid modeling and predictive control of intelligent vehicle longitudinal velocity considering nonlinear tire dynamics | |
| CN106080584A (en) | A kind of hybrid vehicle pattern based on Model Predictive Control Algorithm switching control method for coordinating | |
| Zhou et al. | A Rule‐Based Energy Management Strategy Based on Dynamic Programming for Hydraulic Hybrid Vehicles | |
| Sun et al. | A novel torque coordination control strategy of a single‐shaft parallel hybrid electric vehicle based on model predictive control | |
| Kim et al. | Thermal model development and validation for 2010 Toyota Prius | |
| Kong et al. | A data-driven energy management method for parallel PHEVs based on action dependent heuristic dynamic programming (ADHDP) model | |
| Kwon et al. | Motor and transmission multiobjective optimum design for tracked hybrid electric vehicles considering equivalent inertia of track system | |
| Sampathnarayanan et al. | Model predictive control as an energy management strategy for hybrid electric vehicles | |
| Hajidavalloo et al. | NMPC-based integrated thermal management of battery and cabin for electric vehicles in cold weather conditions | |
| Wang et al. | An Online Predictive Energy Management Strategy for Multi-Mode Plug-in Hybrid Electric Vehicle With Mode Transition Schedule Optimization | |
| Moore et al. | Fuel cell hardware-in-loop | |
| Zhang et al. | Research on coordinated control of braking mode switching for enhancing switching quality considering communication delay and time-varying model parameters |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |