CN112429005B - Pure electric vehicle personalized gear shifting rule optimization method considering transmission efficiency and application - Google Patents

Abstract

According to the pure electric vehicle personalized gear shifting rule optimization method considering the transmission efficiency, the efficiency model (comprising a motor efficiency model and a transmission efficiency model) of the power transmission system is established by using the motor efficiency and transmission efficiency data under different working conditions; and then, a comprehensive evaluation function is constructed by combining the individual requirements, the dynamic subobjective function and the economic subobjective function, the strength of an accelerator pedal and the speed are used as gear shift control parameters, and the individual gear shift rule of the pure electric vehicle is optimized by minimizing the comprehensive evaluation function. The obtained personalized gear shifting rule of the pure electric vehicle is applied to gear shifting control in the driving process of the vehicle, so that the pure electric automatic speed changing vehicle can more accurately realize the optimal comprehensive performance including the dynamic property and the economical property of the vehicle.

Description

Pure electric vehicle personalized gear shifting rule optimization method considering transmission efficiency and application

Technical Field

The invention belongs to the technical field of new energy automobile gear shifting, and relates to an optimization method of a pure electric automobile personalized gear shifting rule.

Background

The gear shifting rule is a rule that the automatic gear shifting time between two adjacent gears changes along with control parameters, is a core control technology of the automobile stepped automatic transmission, and directly influences the performances of the whole automobile such as dynamic property, economical efficiency and the like. The general method for making the gear shifting rule is to determine control parameters, construct an objective function with optimal economy or optimal dynamic performance, and then realize the calculation of the gear shifting rule with a certain algorithm.

The patent publication No. CN110550034A discloses a two-gear AMT comprehensive gear shifting method for a pure electric vehicle, and firstly, an optimal dynamic gear shifting rule and an optimal economic gear shifting rule are provided according to the requirement of the whole vehicle performance of the pure electric vehicle; on the basis of the two optimal gear shifting rules, the vehicle speed and the gear shifting delay amount of a gear shifting point are used as optimization variables, and the difference value of the energy consumption of the whole vehicle and the acceleration of the gear shifting point is used as a target function, so that a comprehensive performance gear shifting rule optimization model considering both economy and dynamic performance is established; and finally, solving the optimization model by using an NSGA-II genetic algorithm to obtain the comprehensive performance gear shifting rule. However, in this shifting method, the efficiency of the transmission system is regarded as a constant value when the shifting schedule is prepared, and the influence of the transmission efficiency change of each gear of the transmission on the shifting schedule is not considered. In fact, the efficiency of the transmission changes under different driving conditions, and the efficiency of the transmission system changes along with the change of the working conditions. Therefore, in the shifting rule making process, the transmission efficiency is regarded as a constant value, so that the made shifting rule cannot achieve the ideal single performance or the optimal comprehensive performance.

Disclosure of Invention

The invention aims to provide a pure electric vehicle personalized gear shifting rule optimization method aiming at the defects in the prior art, which considers the influence of the power transmission system efficiency (including motor efficiency and transmission efficiency) change of a power transmission system under different working conditions on the gear shifting rule in the actual running process of a vehicle, considers the requirements of a driver on the gear shifting performance, and determines the optimal target gear of the comprehensive performance of a transmission, so that the power performance and the economy of the pure electric vehicle are comprehensively optimal on the premise of embodying the driving intention.

According to the pure electric vehicle personalized gear shifting rule optimization method considering the transmission efficiency, firstly, a power transmission system efficiency model (comprising a motor efficiency model and a transmission efficiency model) is established by utilizing motor efficiency and transmission efficiency data under different working conditions; then, a comprehensive evaluation function is constructed by using the dynamic sub-objective function and the economic sub-objective function, the strength of an accelerator pedal and the speed are used as gear shifting control parameters, and the personalized gear shifting rule of the pure electric vehicle is optimized by minimizing the comprehensive evaluation function, wherein the comprehensive evaluation function is as follows:

wherein u represents the vehicle speed; w is a_d、w_eRespectively given dynamic weight and economic weight to reflect the performance expectation of the driver, w_d+w_e＝1；f_d'(u)、f_e' (u) are respectively a dynamic subobjective function and an economic subobjective function after normalization; f. of_d ^*、f_e ^*Respectively, the normalized optimal dynamic performance subobjective function value and the normalized optimal economic performance subobjective function value.

The parameters of the whole vehicle and the transmission system comprise the whole vehicle mass, the transmission ratio of each gear of the transmission, the transmission ratio of the main reducer, the wheel radius, the conversion coefficient of the rotating mass, the windward area, the wind resistance coefficient and the rolling resistance coefficient.

The invention carries out normalization processing on data according to the following formula:

where x' is the normalized value, x is the sample value, x_maxIs the maximum value of the sample, x_minIs the sample minimum.

The inverse normalization process is the inverse process of the normalization process, and x is obtained by calculation according to the formula and the normalized value x'.

The invention relates to a pure electric vehicle personalized gear shifting rule optimization method considering transmission efficiency, which comprises the following steps of:

s1, respectively establishing a motor efficiency model and a transmission efficiency model;

s2, dividing the strength of an accelerator pedal into a plurality of equal parts from 0-100%, and constructing a gear shifting vehicle speed range of two adjacent gears under the same strength of the accelerator pedal by using the minimum vehicle speed of the high gear and the maximum vehicle speed of the low gear of the two adjacent gears under the same strength of the accelerator pedal;

the method for determining the minimum vehicle speed of the high gear and the maximum vehicle speed of the low gear of two adjacent gears under the same accelerator pedal strength comprises the following steps: firstly, determining the motor rotating speed and the automobile acceleration corresponding to different driving speeds of a high gear and a low gear of two adjacent gears under the same accelerator pedal strength; then according to the motor rotating speed and the automobile acceleration corresponding to different driving speeds of the high gear, finding out the minimum speed of the automobile in the high gear, wherein the automobile acceleration is greater than 0 and the current rotating speed of the motor is greater than the minimum rotating speed of the motor, and the minimum speed is the minimum speed of the high gear; then according to the motor rotating speed and the automobile acceleration corresponding to different driving speeds of the low gear, finding out the maximum automobile speed in which the automobile acceleration in the low gear is greater than 0 and the current rotating speed of the motor is less than the maximum rotating speed of the motor, namely the maximum automobile speed of the low gear;

the motor rotating speed and the automobile acceleration corresponding to different driving speeds of a high gear or a low gear of two adjacent gears under the same accelerator pedal strength are obtained according to the following steps:

s21, importing parameters of the whole vehicle and a transmission system;

s22, obtaining motor rotating speeds corresponding to different driving speeds;

s23, motor torques corresponding to different driving speeds are obtained through the motor torque model;

s24, taking the motor rotating speed and the motor torque as input data, and obtaining motor efficiencies corresponding to different driving speeds through a motor efficiency model;

s25, acquiring transmission efficiency corresponding to different driving speeds through a transmission efficiency model by taking the transmission input speed, the transmission input torque and the given transmission oil temperature as input data;

s26, determining automobile acceleration corresponding to different driving speeds according to the motor torque, the motor efficiency and the transmission efficiency;

s3, establishing gear shifting vehicle speed constraint conditions of two adjacent gears under different accelerator pedal strengths;

s4, different dynamic weight values and economic weight values are given, the strength of each accelerator pedal is traversed, the gear shifting speed range and the constraint conditions of two adjacent gears are combined, the gear shifting speed with the minimum comprehensive evaluation function is obtained through the solution of an optimization algorithm, and then the gear shifting curve and the gear shifting curve under the different dynamic weight values and the economic weight values are obtained, namely the personalized gear shifting rule of the pure electric vehicle.

In the method for optimizing the personalized gear shifting rule of the pure electric vehicle, in step S1, the efficiency of the motor and the transmission under different working conditions is tested through efficiency experiments of the motor and the transmission rack respectively. In the motor bench efficiency experiment, the input end voltage, the current, the motor rotating speed and the motor torque corresponding to different working condition points in the motor full working condition range are collected, and the motor efficiency corresponding to each testing working condition point is calculated on the basis of data processing. In the transmission rack efficiency experiment, the input rotating speed, the input torque, the output rotating speed and the output torque of each gear of the transmission corresponding to different working condition points in the full working condition range are collected, and the transmission efficiency corresponding to each testing working condition point of each gear is calculated on the basis of data processing. And establishing a motor efficiency model and each gear efficiency model of the transmission by using the BP neural network.

The establishment process of the motor efficiency model comprises the following sub-steps:

s11, calculating input power and output power of the motor by using input end voltage, current, motor speed and motor torque data under different working conditions collected by a motor efficiency experiment, and taking the ratio of the output power and the input power of the motor as the motor efficiency value under each corresponding working condition;

s12, constructing a motor data collection by using the motor rotating speed, the motor torque and the motor efficiency value obtained in the step S11 under different working conditions, and selecting a motor efficiency model training set and a testing set from the motor data collection;

s13, training and testing the neural network model by using the motor efficiency model training set and the test set data constructed in the step S12 to obtain a motor efficiency model.

In step S11, the motor efficiency value is calculated according to the following formula:

wherein eta is_mThe motor efficiency; p is a radical of_omOutputting power for the motor; p is a radical of_imInputting power for the motor; t is_mIs the output torque of the motor; n is_mIs the output rotation speed of the motor.

Wherein u (N) is voltage signal data (instantaneous data) collected in the update period, i (N) is current signal data (instantaneous data) collected in the update period, and N_mTo sample the number of points, P_xIs the power of a certain phase (x is A, B, C).

In step S12, the training set data volume and the test set data volume account for 70% and 30% of the total motor data set, respectively; the training set and the test set should cover efficiency data of the motor under various working conditions. Therefore, the specific implementation mode adopted in the invention is as follows: dividing the motor speed into n_qDividing the motor torque into n_pIs equally divided into intervals forming n_q×n_pTraversing all motor data sets, wherein 10% of data in each set form an intersection set, 20% of data (of total data in the set) in the rest data are randomly selected as a non-intersection test set, non-selected data are selected as a non-intersection training set, the intersection set and the non-intersection training set form a training set, and the intersection set and the non-intersection test set form a test set.

In step S13, in the process of training and testing the BP neural network model by using the constructed motor efficiency model training set and test set data, the motor rotation speed and the motor torque are used as model input data P₁Output of the model Y₁Motor efficiency η obtained in step S11 as a motor efficiency prediction value_mIs T₁With T₁And Y₁Deviation e of₁(i.e. absolute error, e)₁＝|T₁-Y₁|) as the learning signal of the network, iteratively adjusting the weight and the threshold of the BP neural network model by using a Levenberg-Marquardt learning algorithm through error back propagation to enable the relative error (i.e. e) of training and testing₁′＝|T₁-Y₁|/T₁) The accuracy requirements are met so as to complete the establishment of the motor efficiency model.

The transmission efficiency model is corresponding to the transmission gear, namely different neural network models are adopted for different gears, but the transmission efficiency model establishing methods for different gears are completely the same. The process for establishing the efficiency model of the transmission with a certain gear comprises the following sub-steps:

s14, calculating the input power and the output power of a certain gear of the transmission by using the input rotating speed, the input torque, the output rotating speed and the output torque of the transmission under different working conditions of the certain gear, which are acquired by a transmission efficiency experiment, and taking the ratio of the output power and the input power of the certain gear of the transmission as the efficiency value of the certain gear of the transmission under the corresponding working condition;

s15, constructing a gear data collection of the transmission by using the transmission input rotating speed, the transmission input torque, the transmission oil temperature (namely the transmission lubricating oil temperature) and the transmission efficiency value obtained in the step S14 under different working conditions of a certain gear of the transmission, and selecting a transmission efficiency model training set and a test set from the gear data collection of the transmission;

s16, training and testing the neural network model by using the transmission efficiency model training set and the test set data constructed in the step S15 to obtain an efficiency model of a certain gear of the transmission.

In step S14, the pure electric vehicle transmission efficiency is calculated according to the following formula:

wherein eta is_tTo transmission efficiency; p is a radical of_otIs the power at the output of the transmission; p is a radical of_itIs the variator input power; t is_itIs transmission input torque; n is_itThe rotational speed of the input end of the transmission; t is_otIs the variator output torque; n is_otIs the speed of the output end of the speed changer.

In step S15, the training set data volume and the test set data volume respectively account for the transmission data total set70% and 30%; the training and test sets should cover efficiency data for the transmission under various operating conditions. Therefore, the specific implementation mode adopted in the invention is as follows: dividing input speed into n_sEqually dividing the input torque into n_tDivided equally into zones, oil temperature divided into n_oilIs equally divided into intervals forming n_s×n_t×n_oilAnd traversing all the transmission data sets, wherein in each set, 10% of data is randomly selected to form an intersection set, 20% of data (of the total data of the set) is randomly selected from the rest data to be used as a non-intersection test set, the non-selected data is used as a non-intersection training set, the intersection set and the non-intersection training set form a training set, and the intersection set and the non-intersection test set form a test set.

In step S16, in the process of training and testing the BP neural network model by using the constructed transmission efficiency model training set and test set data, the transmission input rotation speed, the transmission input torque and the transmission oil temperature are used as model input data P₂Output of the model Y₂The predicted transmission efficiency is the transmission efficiency η obtained in step S14_tIs T₂With T₂And Y₂Deviation e of₂(i.e. absolute error, e)₂＝|T₂-Y₂|) as the learning signal of the network, iteratively adjusting the weight and the threshold of the BP neural network model by using a Levenberg-Marquardt learning algorithm through error back propagation to enable the relative error (i.e. e) of training and testing₂′＝|T₂-Y₂|/T₂) The accuracy requirements are met so as to complete the establishment of the transmission efficiency model.

In the above steps S13 and S16, the network model used in the process of constructing the motor efficiency model or the transmission efficiency model is a BP neural network model. In a preferred implementation mode, the number of hidden layers of the used BP neural network is two, and the lower limit and the upper limit of the number of the neurons in each hidden layer are respectively M_minAnd M_maxThe training and testing method for the BP neural network model by using the training set and the test set data is carried out according to the following steps:

a1 data normalization processing, wherein the normalization processing is carried out on the input data of the model in the training set and the test set;

a2 neural network parameter initialization including hidden layer number, hidden layer neuron number initial value and hidden layer neuron number initial lower limit M_minAnd an initial upper limit M_maxSetting the number L of hidden layer as 1, and the initial value of the number of the hidden layer neurons as M_min；

A3 training a BP neural network model by using the data in the training set after normalization processing;

a4, after training in the step A3, testing the BP neural network model by using the test concentrated data after normalization processing;

a5 carrying out reverse normalization processing on the BP neural network model output values in the steps A3 and A4;

a6, judging whether the maximum value of the relative error between the BP neural network model output value obtained in the step A5 and the efficiency value corresponding to the data in the training set and the test set is less than or equal to 5%; if yes, go to step A16; otherwise, the number of the hidden layer neurons is increased by 1, and then the step A7 is carried out;

a7 determining whether the number of hidden neurons is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A3; otherwise, entering A8;

a8 sets the hidden layer number to L-2, and sets the neuron number of each layer to the initial value M_minThen proceed to step a 9;

a9 training a BP neural network model by using the data in the training set after normalization processing;

a10, after training in the step A9, testing the BP neural network model by using the test concentrated data after normalization processing;

a11 carrying out reverse normalization processing on the BP neural network model output values in the steps A9 and A10;

a12, judging whether the maximum value of the relative error between the BP neural network model output value obtained in the step A11 and the measured value corresponding to the data in the training set and the test set is less than or equal to 5%; if yes, go to step A16; otherwise, the number of the second hidden layer neurons is increased by 1, and then the step A13 is carried out;

a13 determination ofWhether the number of neurons in the hidden layer is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A9; otherwise, increasing the number of the first hidden layer neurons by 1, and setting the number of the second hidden layer neurons as an initial value M_minEntering A14;

a14 determining whether the number of first hidden layer neurons is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A9; otherwise, go to A15;

a15 upper limit M of number of hidden layer neurons_maxIncreasing by 10, keeping the lower limit unchanged, setting the hidden layer number to be L ═ 1, and setting the hidden layer neuron number to be an initial value M_minThen returns to step a 3;

a16 stores the trained BP neural network model parameters.

In the method for optimizing the personalized gear shifting rule of the pure electric vehicle, in step S2, the rotating speed n of the motor_mCalculated according to the following formula:

the motor torque model refers to a motor torque and rotating speed relation map or table established under the condition of given accelerator pedal strength. The relation curve of the motor torque and the rotating speed under the condition of 100% of the accelerator pedal strength in the general motor torque model can be obtained through conventional motor experiments in the field. The relationship between the motor torque and the rotating speed under other accelerator pedal strengths can be set according to the relationship curve between the motor torque and the rotating speed under 100% accelerator pedal strength. Through the motor torque model, the motor torque corresponding to the motor rotating speed can be found out.

Acceleration of a motor vehicle

Can be calculated from the following formula:

wherein, T_mIs the motor torque; i.e. i_gIs the transmission ratio of the transmission; i.e. i₀The transmission ratio of the main speed reducer is set; eta_gMain reducer efficiency; eta_tTo transmission efficiency; m is the mass of the whole vehicle; r is the wheel radius; f rolling resistance coefficient; c_dIs the wind resistance coefficient; a is the windward area; u is the vehicle speed; and delta is an automobile rotating mass conversion coefficient.

Transmission ratio i of the above-mentioned transmission_gMain reducer transmission ratio i₀Main reducer efficiency eta_gThe mass m of the whole vehicle, the radius r of the wheel, the rolling resistance coefficient f and the wind resistance coefficient C_dThe windward area A and the automobile rotating mass conversion coefficient delta belong to parameters of the whole automobile and a transmission system and can be obtained by conventional means in the field. Wherein the transmission has a transmission ratio i_gMain reducer transmission ratio i₀Main reducer efficiency eta_gThe mass m of the whole vehicle, the rolling resistance coefficient f and the wind resistance coefficient C_dAnd the frontal area a are given values. The automobile rotating mass conversion coefficient delta is calculated by the following formula:

wherein, I_WIs the rotational inertia of the wheel; i is_mIs the rotational inertia of the motor; η is the driveline efficiency, which is the product of the variator efficiency and the final drive efficiency; r is the tire radius. I is_WAnd I_mIs a given value.

Motor torque T_mMotor efficiency eta_mAnd transmission efficiency η_tObtained in accordance with the foregoing steps S21 to S25.

In the method for optimizing the personalized shift schedule of the pure electric vehicle, in step S3, the shift speed u of two adjacent gears_aThe constraint conditions of (1) are:

g₂(u_a)＝u_a-u_min≥0；

g₃(u_a)＝u_max-u_a≥0；

wherein u is_minMinimum vehicle speed in high gear for a given accelerator pedal intensity; u. of_maxFor the maximum vehicle speed in the low gear for a given accelerator pedal intensity.

In the method for optimizing the personalized gear shifting rule of the pure electric vehicle, in step S4, different dynamic weights and economic weights are given, the strength of each accelerator pedal is traversed, and a gear shifting vehicle speed which minimizes a comprehensive evaluation function is obtained by using a Particle Swarm Optimization (PSO) or a Genetic Algorithm (GA) in combination with gear shifting vehicle speed constraint conditions of two adjacent gears. The process of obtaining the shift speed with the minimum comprehensive evaluation function by adopting a Particle Swarm Optimization (PSO) algorithm comprises the following steps:

s41, determining a comprehensive evaluation function;

s42, under the condition of giving a dynamic weight, an economic weight and an accelerator pedal strength, taking a comprehensive evaluation function as a fitness function, and calling a particle swarm optimization algorithm to obtain a upshift speed under the given dynamic weight, economic weight and accelerator pedal strength;

s43 traversing each determined accelerator pedal strength, repeating the step S42, and obtaining the upshifting speed u with different accelerator pedal strengths under the given dynamic weight and economic weight_aupForming a gear-up curve under the given dynamic weight and economic weight;

s44 calculating the downshift vehicle speed u under different accelerator pedal strengths according to the following formula_adownAnd forming a downshift curve under the given dynamic weight and economic weight:

u_adown＝u_aup*(1-A)；

wherein A is a given coefficient; when the accelerator pedal strength is less than 30%, a is 0.4; when the strength of the accelerator pedal is more than or equal to 30%, A is 0.15;

s45, repeating the steps S42-S44 for given different given power weight values and economic weight values to obtain an upshift curve and a downshift curve under different given power weight values and economic weight values.

In the above step S41, according to the comprehensive evaluation function given above, f is determined_d'(u)、f_e'(u)、f_d ^*And f_e ^*A1 is to f_d'(u)、f_e'(u)、f_d ^*And f_e ^*Substituting, the comprehensive evaluation function can be determined.

Normalized dynamic partial objective function f_d' (u) and an economic objective function f_e' (u) acquisition mode: the absolute value of the acceleration difference under the same accelerator pedal strength of two adjacent gears of the pure electric vehicle transmission is used as a dynamic sub-target function; taking an absolute value of a unit mileage energy consumption difference under the same accelerator pedal strength of each adjacent gear of the pure electric vehicle transmission as an economic subobjective function; respectively obtaining the maximum value and the minimum value of a dynamic sub-objective function and an economic sub-objective function according to the upper limit and the lower limit of the gear shifting speed range of two adjacent gears under different accelerator pedal strengths; then, the dynamic subobjective function and the economic subobjective function are normalized, and the normalized dynamic subobjective function f is obtained_d' (u) and an economic objective function f_e'(u)。

The respective accelerations of two adjacent gears are calculated according to the automobile acceleration formula given above.

The energy consumption E of each unit mileage of two adjacent gears is calculated by the following formula:

wherein eta is_mThe motor efficiency; eta_cTo the motor controller efficiency. The motor controller efficiency is obtained by experiments.

Normalized optimal dynamic performance subobjective function value f_d ^*And the best economic objective function value f_e ^*The acquisition mode is as follows: f in the comprehensive evaluation function_d ^*The value is 0, f_e ^*The value is taken to be 0, let w_d＝1，w_eUsing an optimization algorithm to solve the minimum value f of the dynamic subobjective function under each accelerator pedal strength_dNormalizing the minimum value of the dynamic subobjective function under each accelerator pedal strength to obtain the normalized optimal dynamic subobjective function value f under each pedal strength_d ^*(ii) a Let w_d＝0，w_eSolving the minimum value f of the economic subobjective function under each accelerator pedal strength by using an optimization algorithm as 1_eNormalizing the economic subobjective function value under each accelerator pedal strength to obtain the normalized optimal economic subobjective function value f under each pedal strength_e ^*。

The invention further provides application of the pure electric vehicle individual gear shifting rule obtained by the optimization method in gear shifting in the driving process of the vehicle, and the pure electric vehicle individual gear shifting rule obtained by the optimization method can be stored in a gear shifting rule database in a data table form. When the method is applied, the expected power value (namely, the weight value of the power) and the expected economic value (namely, the weight value of the economic efficiency) are determined according to the prior art (see the application document disclosed by the application number 201911247447.8), then the gear shifting rule matched with the current expected power value and the current expected economic value (namely, the individualized gear shifting rule with the optimal comprehensive performance) is selected from the gear shifting rule database, and then the target gear is determined according to the vehicle speed and the strength of the accelerator pedal and the selected gear shifting rule. And then, the prior art is used for realizing the optimal gear shifting operation of the automobile with comprehensive performance, which is consistent with the performance expectation of a driver and takes the working condition efficiency of the power transmission system into consideration, through the gear shifting control system.

Compared with the prior art, the pure electric vehicle personalized gear shifting rule optimization method considering the transmission efficiency and the application have the following beneficial effects:

(1) according to the invention, the transmission efficiency is taken into consideration along with the change of working conditions, the accurate motor efficiency and transmission efficiency are determined according to the efficiency model of the power transmission system, the whole vehicle dynamic performance and economic performance indexes of the pure electric step-by-step automatic variable speed vehicle are calculated based on the accurate motor efficiency and transmission efficiency, the gear shifting speed range and constraint conditions are determined, and finally the accurate gear shifting speed is obtained by optimizing a comprehensive evaluation function, so that the established gear shifting rule enables the pure electric automatic variable speed vehicle to more accurately realize the optimal comprehensive performance including the vehicle dynamic performance and economic performance.

(2) According to the invention, the BP neural network is utilized to learn the motor efficiency and transmission efficiency experimental data, and the established motor efficiency model and transmission efficiency model can obtain accurate motor efficiency and transmission efficiency along with the change of working conditions, so that a foundation is laid for formulating a gear shifting rule with optimal comprehensive performance.

(3) The comprehensive evaluation function is constructed on the basis of the dynamic subobjective function and the economic subobjective function, the dynamic performance and the economic performance of the automobile are considered, the performance requirements of a driver are reflected, the gear shifting point which enables the comprehensive evaluation function to be minimum is solved through optimization methods such as a particle swarm optimization algorithm or a genetic algorithm, and therefore the obtained gear shifting rule enables the performance of the whole automobile to achieve personalized comprehensive optimization.

Drawings

FIG. 1 is a schematic flow chart of a pure electric vehicle personalized gear shifting rule optimization method considering transmission efficiency according to the invention;

FIG. 2 is a schematic diagram of a BP neural network modeling principle;

FIG. 3 is a schematic flow chart of a training and testing method of a BP neural network model;

FIG. 4 is a motor efficiency curved surface drawn by BP neural network model training and test results in an embodiment of the present invention;

FIG. 5 is a transmission 1-gear efficiency curved surface drawn by BP neural network model training and testing results in an embodiment of the present invention;

FIG. 6 is a transmission 2-gear efficiency curved surface drawn by BP neural network model training and test results in an embodiment of the present invention;

FIG. 7 is a flowchart illustrating a method for determining a maximum speed for gear 1 and a minimum speed for gear 2 for a given accelerator pedal intensity;

FIG. 8 is a diagram illustrating a relationship between a motor torque and a rotational speed corresponding to a motor torque model according to an embodiment of the present invention;

FIG. 9 is a schematic flow chart of a Particle Swarm Optimization (PSO) algorithm for obtaining a shift speed;

FIG. 10 shows an upshift curve and a downshift curve (i.e., personalized shift schedule) for different dynamics weights and economy weights determined according to an embodiment of the present invention;

fig. 11 is a schematic block diagram of shifting implemented by using the personalized shift schedule determined according to the embodiment of the present invention.

Detailed Description

The technical solutions of the present invention will be described in detail and fully with reference to the accompanying drawings, which are used for describing the embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Examples

The present embodiment is directed to an electric vehicle equipped with an automatic transmission (AMT) having two gears (1 gear and 2 gears). The method for optimizing the personalized gear shifting rule of the pure electric vehicle considering the transmission efficiency is also suitable for the pure electric vehicle with more than two gears, and the gear shifting rule given by the method refers to the gear shifting rule between two adjacent gears.

According to the pure electric vehicle personalized gear shifting rule optimization method considering the transmission efficiency, the used comprehensive evaluation function is constructed by a dynamic objective function and an economic objective function, and the two objective functions are defined as follows:

1) dynamic subobjective function: absolute value of difference between acceleration in 1-gear and 2-gear under the same pedal intensity, i.e.

Wherein the content of the first and second substances,

1 gear acceleration when driving at a vehicle speed u;

the acceleration is 2 nd gear acceleration when the vehicle is running at the vehicle speed u. Acceleration of a vehicle

Can be calculated by

Wherein, T_mIs the motor torque; i.e. i_gIs the transmission ratio of the transmission; i.e. i₀The transmission ratio of the main speed reducer is set; eta_gMain reducer efficiency; eta_tTo transmission efficiency; m is the mass of the whole vehicle; r is the tire radius; f rolling resistance coefficient; c_dIs the wind resistance coefficient; a is the windward area; u is the vehicle speed; and delta is an automobile rotating mass conversion coefficient.

The automobile rotating mass conversion coefficient delta is calculated by the following formula:

2) economic objective function: absolute value of energy consumption difference of unit mileage of 1 gear and 2 gears under same pedal intensity, namely

f_e(u)＝|E₁-E₂| (5)；

Wherein E is₁The energy consumption is 1-gear unit mileage energy consumption when the vehicle runs at a vehicle speed u; e₂The energy consumption is 2-gear unit mileage energy consumption when the vehicle runs at the vehicle speed u. The unit mileage energy consumption E is calculated by the following formula

Wherein eta is_mThe motor efficiency; eta_cThe motor controller efficiency (0.95 in this embodiment).

This example uses square additionAnd constructing a comprehensive evaluation function by a weight and ideal point method. Firstly, respectively taking a dynamic subobjective function and an economic subobjective function as optimization targets, and calculating an optimal dynamic shift law and an optimal economic shift law; then, respectively calculating the dynamic sub-objective function value and the economic sub-objective function value corresponding to the gear shifting point under each accelerator pedal strength, and respectively carrying out normalization processing to obtain the optimal dynamic sub-objective function value f after normalization under each accelerator pedal strength_d ^*And the best economic objective function value f_e ^*. Then, a comprehensive evaluation function is constructed according to the following formula

Wherein u represents the vehicle speed; w is a_d、w_eRespectively given dynamic weight and economic weight to reflect the performance expectation of the driver, w_d+w_e＝1；f_d'(u)、f_e' (u) are respectively a dynamic subobjective function and an economic subobjective function after normalization; f. of_d ^*、f_e ^*Respectively, the normalized optimal dynamic performance subobjective function value and the normalized optimal economic performance subobjective function value. And then forming an optimization problem by taking the minimum comprehensive evaluation function as a target.

Based on the above analysis, the pure electric vehicle personalized gear shifting law optimization method considering the transmission efficiency provided by the embodiment includes, as shown in fig. 1, the following steps:

s1 respectively establishing a motor efficiency model and a transmission efficiency model

In the step, the efficiency of the motor and the speed changer under different working conditions is tested through efficiency experiments of the motor and the speed changer rack respectively. In the motor bench efficiency experiment, the input end voltage, the current, the motor rotating speed and the motor torque corresponding to different working condition points in the motor full working condition range are collected, and the motor efficiency corresponding to each testing working condition point is calculated on the basis of data processing. In the transmission rack efficiency experiment, the input rotating speed, the input torque, the output rotating speed and the output torque of each gear of the transmission corresponding to different working condition points in the full working condition range are collected, and the transmission efficiency corresponding to each testing working condition point of each gear is calculated on the basis of data processing. And establishing a motor efficiency model and each gear efficiency model of the transmission by using the BP neural network.

As shown in fig. 2, the motor efficiency model is input with the motor rotation speed and the motor torque as model input data P₁Output of the model Y₁The motor efficiency is calculated as T₁With T₁And Y₁Deviation e of₁(i.e. absolute error, e)₁＝|T₁-Y₁|) as the learning signal of the network, iteratively adjusting the weight and the threshold of the BP neural network model by using a Levenberg-Marquardt learning algorithm through error back propagation to enable the relative error (i.e. e) of training and testing₁′＝|T₁-Y₁|/T₁) The accuracy requirements are met so as to complete the establishment of the motor efficiency model.

Therefore, the process of establishing the motor efficiency model comprises the following sub-steps:

s11, acquiring input power and output power of the motor by using the acquired input end voltage, current, motor rotating speed and motor torque sample data corresponding to the motor under different working conditions, and then taking the ratio of the output power and the input power of the motor as an actual measured value of the efficiency of the motor;

in this step, the motor efficiency is calculated according to the following formula:

wherein eta is_mThe motor efficiency; p is a radical of_omOutputting power for the motor; p is a radical of_imInputting power for the motor; t is_mIs the output torque of the motor; n is_mIs the output rotation speed of the motor.

Wherein u (N) is voltage signal data (instantaneous data) collected in the update period, i (N) is current signal data (instantaneous data) collected in the update period, and N_mTo sample the number of points, P_xIs the power of a certain phase (x is A, B, C).

S12, constructing a motor data collection set by using the motor rotating speed and the motor torque under different working conditions and the sample data of the motor efficiency measured value obtained in the step S11, and selecting a motor efficiency model training set and a test set from the motor data collection set; the training set data and the test set data account for 70% and 30% of the total motor data set, respectively. Specifically, the motor rotation speed is divided into 20 equal intervals, the motor torque is divided into 10 equal intervals to form 20 × 10 motor data sets, all the motor data sets are traversed, 10% of data in each set is randomly selected to form an intersection set, 20% of data (of total data in the set) in the rest data is randomly selected to serve as a non-intersection test set, non-selected data serves as a non-intersection training set, the intersection set and the non-intersection training set form a training set, and the intersection set and the non-intersection test set form a test set.

S13, training and testing the BP neural network model by using the obtained motor efficiency model training set and test set data to obtain a motor efficiency model.

As shown in FIG. 2, the transmission efficiency model is corresponding to the transmission gear, and for a certain gear transmission efficiency model, the transmission input rotation speed, the transmission input torque and the transmission oil temperature are taken as model input data P₂Output of the model Y₂For the predicted value of the transmission efficiency, the calculated transmission efficiency is T₂With T₂And Y₂Deviation e of₂(i.e. absolute error, e)₂＝|T₂-Y₂|) as the learning signal of the network, iteratively adjusting the weight and the threshold of the BP neural network model by using a Levenberg-Marquardt learning algorithm through error back propagation to enable the relative error (i.e. e) of training and testing₂′＝|T₂-Y₂|/T₂) The accuracy requirements are met so as to complete the establishment of the transmission efficiency model.

Therefore, the process of establishing an efficiency model of a certain gear transmission comprises the following sub-steps:

s14, calculating the input power and the output power of a certain gear of the transmission by using the input rotating speed, the input torque, the output rotating speed and the output torque of the transmission under different working conditions of the certain gear, which are acquired by a transmission efficiency experiment, and taking the ratio of the output power and the input power of the certain gear of the transmission as the efficiency value of the certain gear of the transmission under the corresponding working condition;

in the step, the efficiency of the pure electric automobile transmission is calculated according to the following formula:

wherein eta is_tTo transmission efficiency; p is a radical of_otIs the power at the output of the transmission; p is a radical of_itIs the variator input power; t is_itIs transmission input torque; n is_itThe rotational speed of the input end of the transmission; t is_otIs the variator output torque; n is_otIs the speed of the output end of the speed changer.

S15, constructing a gear data collection of the transmission by using the transmission input rotating speed, the transmission input torque, the transmission oil temperature (namely the transmission lubricating oil temperature) and the transmission efficiency value obtained in the step S14 under different working conditions of a certain gear of the transmission, and selecting a transmission efficiency model training set and a test set from the gear data collection of the transmission; the training set data and the test set data account for 70% and 30% of the total transmission data set, respectively. Specifically, the rotation speed of an input shaft is divided into 20 equal intervals, the input torque is divided into 10 equal intervals, the oil temperature is divided into 10 equal intervals to form 20 × 10 × 10 transmission data sets, all the transmission data sets are traversed, 10% of data are randomly selected to form an intersection set in each set, 20% of data (of total data of the set) in the rest of data are randomly selected to serve as a non-intersection test set, non-selected data serve as a non-intersection training set, the intersection set and the non-intersection training set form a training set, and the intersection set and the non-intersection test set form the test set.

S16, training and testing the neural network model by using the transmission efficiency model training set and the test set data constructed in the step S15 to obtain an efficiency model of a certain gear of the transmission.

In the process of constructing the motor efficiency and the transmission efficiency, the training process of the BP neural network is similar. The number of hidden layers of the BP neural network is two, and the lower limit and the upper limit of the number of the neurons in each hidden layer are respectively M_minAnd M_maxThe learning function adopts a Levenberg-Marquardt algorithm; mean square error (here, the average of the squares of the differences of all network outputs from the target value) training precision is set to 0.001; the maximum number of training times is set to 1000.

The method for training and testing the BP neural network model by using the training set and the test set data is carried out according to the following steps as shown in FIG. 3:

a1 data normalization processing, wherein the normalization processing is carried out on the input data of the model in the training set and the test set;

a2 neural network parameter initialization including hidden layer number, hidden layer neuron number n_H1Initial value and initial lower limit M of number of hidden layer neurons _min3 and an initial upper limit M _max15, setting the number L of hidden layer as 1, the number n of hidden layer neurons_H1Is M_min＝3；

A3 training a BP neural network model by using the data in the training set after normalization processing;

a4, after training in the step A3, testing the BP neural network model by using the test concentrated data after normalization processing;

a5 carrying out reverse normalization processing on the BP neural network model output values in the steps A3 and A4;

a6, judging whether the maximum value of the relative error between the BP neural network model output value obtained in the step A5 and the efficiency value corresponding to the data in the training set and the test set is less than or equal to 5%; if yes, go to step A16; otherwise, the number of the hidden layer neurons is increased by 1, and then the step A7 is carried out;

a7 judging the number n of hidden layer neurons_H1Whether or not it is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A3; otherwise enter A8；

A8 sets the hidden layer number to L-2, and sets the neuron number of each layer to the initial value M_minThen proceed to step a 9;

a9 training a BP neural network model by using the data in the training set after normalization processing;

a10, after training in the step A9, testing the BP neural network model by using the test concentrated data after normalization processing;

a11 carrying out reverse normalization processing on the BP neural network model output values in the steps A9 and A10;

a12, judging whether the maximum value of the relative error between the BP neural network model output value obtained in the step A11 and the measured value corresponding to the data in the training set and the test set is less than or equal to 5%; if yes, go to step A16; otherwise, the number of second hidden layer neurons n_H2Step A13 is entered after increasing by 1;

a13 judging the number n of second hidden layer neurons_H2Whether or not it is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A9; otherwise, the number n of the first hidden layer neurons is counted_H1Increasing the number of 1, second hidden layer neurons n_H2Set to an initial value M_minEntering A14;

a14 judging the number n of first hidden layer neurons_H1Whether or not it is less than or equal to the upper limit M of the number of neurons_max(ii) a If yes, returning to the step A9; otherwise, go to A15;

a15 upper limit M of number of hidden layer neurons_maxIncreasing by 10, keeping the lower limit unchanged, setting the hidden layer number to be L ═ 1, and setting the hidden layer neuron number to be an initial value M_minThen returns to step a 3;

a16 stores the trained BP neural network model parameters.

For a pure electric vehicle equipped with a 2-gear mechanical automatic transmission (AMT), an efficiency model of a power transmission system is established according to step S1: the topological structure of the motor efficiency model established according to the steps S11-S13 is 2 multiplied by 6 multiplied by 1, the number of hidden layer layers is 1, and the number of hidden layer neurons is 6; the topological structure of the 1-gear transmission efficiency model established in the steps S14-S16 is 3 multiplied by 6 multiplied by 4 multiplied by 1, the number of hidden layer layers is 2, the number of first hidden layer neurons is 6, and the number of second hidden layer neurons is 4; the topological structure of the 2-gear transmission efficiency model is 3 multiplied by 1, the number of hidden layer layers is 2, the number of first hidden layer neurons is 3, and the number of second hidden layer neurons is 1.

With the motor efficiency model of the power transmission system established, a motor efficiency curve is plotted as shown in fig. 4.

Using the established 1-speed transmission efficiency model and 2-speed transmission efficiency model of the powertrain, the 1-speed transmission efficiency and 2-speed transmission efficiency curves are plotted as shown in fig. 5 and 6.

As can be seen from fig. 4, 5 and 6, the motor efficiency and the transmission efficiency both change greatly under different operating conditions, which affects the shift speed of the pure electric vehicle, and thus has a great effect on the dynamic performance and the economic performance of the whole vehicle.

S2, dividing the strength of the accelerator pedal into a plurality of equal parts from 0-100%, and constructing the gear shifting vehicle speed range of two adjacent gears under the same strength of the accelerator pedal by using the minimum vehicle speed of a high gear and the maximum vehicle speed of a low gear under the same strength of the accelerator pedal.

Here, the high gear refers to 2, the low gear refers to 1, and i denotes a gear, which is 1 or 2. The method comprises the following steps of dividing the strength of an accelerator pedal from 0-100% into h equal to 20 equal parts, and constructing the shifting vehicle speed ranges of the 1 gear and the 2 gear with the minimum vehicle speed of the 2 gear and the maximum vehicle speed of the 1 gear under the strength of the jth accelerator pedal, namely:

u_min(2,j)≤u_a(j)≤u_max(1,j) (11)

in the formula u_a(j) Represents the shift speed at the jth accelerator pedal intensity; u. of_min(2, j) represents a2 nd gear minimum vehicle speed at the jth accelerator pedal intensity; u. of_max(1, j) represents the 1 st gear maximum vehicle speed at the jth accelerator pedal intensity.

As shown in fig. 7, the method for determining the maximum vehicle speed in gear 1 and the minimum vehicle speed in gear 2 at each accelerator pedal intensity includes the following steps:

and S21, importing parameters of the whole vehicle and the transmission system, and enabling i to be 1.

The whole experimental vehicle is adoptedVehicle and transmission system parameters comprise the mass m of the whole vehicle 1250kg and the transmission ratio i of 1 gear ₁1, 2-gear transmission ratio i₂Main speed reducer transmission ratio i of 2.27_o5.29, the wheel radius r is 0.316m, and the windward area A is 2m²Wind resistance coefficient C_d0.32 and 0.015 as the rolling resistance coefficient f.

S22 obtains the motor speed corresponding to the i-th gear different driving vehicle speed, and sets j equal to 1.

The motor rotating speed is calculated according to the following formula:

and S23, obtaining motor torques corresponding to different running vehicle speeds through the motor torque model.

The motor torque model used in the present embodiment is shown in fig. 8, and a motor torque and rotation speed relation curve under 100% accelerator pedal strength is obtained through a motor experiment; by equally dividing the accelerator pedal strength 20, the motor torque at each accelerator pedal strength is taken as the product of the accelerator pedal strength and the motor torque at 100% accelerator pedal strength, and the motor torque vs. rotational speed curves at other accelerator pedal strengths can be obtained.

As can be seen from fig. 8, at a given accelerator pedal intensity, the motor torque corresponding to the motor rotational speed can be found from the torque-rotational speed relationship map in accordance with the motor rotational speed calculated in step S22.

And S24, taking the motor rotating speed and the motor torque as input data, and obtaining the motor efficiency corresponding to different running speeds through the motor efficiency model obtained in the previous step.

The motor rotation speed and the motor torque are input into the motor efficiency model established in step S1, so that the motor efficiencies corresponding to different driving speeds at the current accelerator pedal intensity can be obtained.

And S25, acquiring the transmission efficiency corresponding to different running speeds by using the transmission efficiency model obtained in the previous step by taking the transmission input speed, the transmission input torque and the given transmission oil temperature as input data.

The transmission input rotation speed, the transmission input torque and the given transmission oil temperature are input into the transmission efficiency model established in step S1, and the transmission efficiency corresponding to the current accelerator pedal intensity and the different driving speeds can be obtained.

S26, according to the motor torque, the motor efficiency and the transmission efficiency, the automobile acceleration corresponding to different running speeds is determined.

The motor torque obtained in step S23, the motor efficiency obtained in step S24, and the transmission efficiency obtained in step S25 are substituted into formula (3), so that the vehicle acceleration corresponding to different driving speeds at the given current accelerator pedal intensity can be obtained.

S27, finding out the maximum speed with the acceleration greater than 0 and the current speed less than the maximum motor speed in the 1 gear according to the motor speeds and the automobile accelerations corresponding to different driving speeds of the 1 gear, namely the maximum speed of the 1 gear under the jth accelerator pedal strength;

or finding out the minimum speed with the acceleration greater than 0 and the current speed greater than the minimum motor speed in the 2-gear according to the motor speeds and the automobile accelerations corresponding to different driving speeds of the 2-gear, namely the minimum speed of the 2-gear under the jth accelerator pedal strength;

s28 determines whether j is h, if so, step S28 is performed, otherwise, 1 is added to j, and step S22 is performed;

s29 judges whether or not i is 2, if so, the determination of the 1 st gear maximum vehicle speed and the 2 nd gear minimum vehicle speed at each accelerator pedal intensity is completed, and the process ends, otherwise, 1 is added to i, and the process returns to step S21.

The shift speed ranges of the two- speed gears 1 and 2 at each accelerator pedal strength (h is 1 to 20) can be obtained through the steps S21 to S29.

S3 establishes gear shifting vehicle speed constraint conditions of two gears under different accelerator pedal strengths.

In the present embodiment, the shift speeds u of 1-gear and 2-gear are given the pedal strength_aSubject to the following conditions:

g₂(u_a)＝u_a-u_min(2)≥0 (14)；

g₃(u_a)＝u_max(1)-u_a≥0 (15)；

wherein u is_min(2) The lowest speed of 2 gears under the given pedal strength is set; u. of_max(1) The highest speed in gear 1 for a given pedal intensity.

The resulting optimization problem is thus expressed as follows:

minU(u)＝min{w_d[f_d'(u)-f_d ^*]²+w_e[f_e'(u)-f_e ^*]²}

s4, different dynamic weight values and economic weight values are given, the strength of each accelerator pedal is traversed, the gear shifting speed which enables a comprehensive evaluation function to be minimum is obtained through an optimization algorithm by combining the gear shifting speed range and the constraint condition of two adjacent gears, and then a gear shifting curve and a gear shifting curve under the given transmission oil temperature, the different dynamic weight values and the different economic weight values are obtained, namely the personalized gear shifting rule of the pure electric vehicle.

In this embodiment, a Particle Swarm Optimization (PSO) algorithm is used to obtain a shift speed that minimizes the comprehensive evaluation function.

Since the transmission oil temperature has a large influence on the transmission efficiency, the transmission oil temperature needs to be given first, and then the personalized shift schedule under the transmission oil temperature needs to be determined.

Because the oil temperature of the transmission used in the embodiment is basically maintained at about 90 ℃ after the cold vehicle is started and enters a normal working state, the embodiment optimizes the personalized gear shifting rule of the two-gear pure electric vehicle when the oil temperature of the transmission is 90 ℃, and sets the input temperature of the transmission efficiency model to 90 ℃ when the transmission efficiency calculation is related; for the gear shifting rules of other transmission oil temperatures, the input temperature of the transmission efficiency model is only required to be changed, and optimization is carried out according to the same method.

The process of obtaining the shift speed that minimizes the overall evaluation function using the Particle Swarm Optimization (PSO) algorithm, as shown in fig. 9, includes the following steps:

s41 determining a comprehensive evaluation function

Shift vehicle speed ranges of 1 st gear and 2 nd gear at different accelerator pedal strengths have been obtained according to step S2. According to the upper limit and the lower limit of the speed range of the 1-gear and 2-gear shifting under different accelerator pedal strengths, the maximum value and the minimum value of the dynamic subobjective function and the economic subobjective function can be obtained through the formulas (2), (3), (5) and (6), and then the normalized dynamic subobjective function f can be obtained by normalizing the dynamic objective function and the economic objective function through the formula (1) respectively_d' (u) and an economic objective function f_e'(u)。

Normalized optimal dynamic performance subobjective function value f_d ^*And the best economic objective function value f_e ^*The acquisition mode is as follows: f in the comprehensive evaluation function_d ^*The value is 0, f_e ^*The value is taken to be 0, let w_d＝1，w_eSolving the minimum value f of the dynamic performance sub-objective function under each accelerator pedal strength by using a Particle Swarm Optimization (PSO) algorithm as 0_dNormalizing the minimum value of the dynamic performance subobjective function under each accelerator pedal strength to obtain the normalized optimal dynamic performance subobjective function value f_d ^*(ii) a Let w_d＝0，w_eSolving the minimum value f of the economic subobjective function under each accelerator pedal strength by using an optimization algorithm as 1_eNormalizing the economic subobjective function value under each accelerator pedal strength to obtain the normalized optimal economic subobjective function value f_e ^*。

And then combine above f_d'(u)、f_e'(u)、f_d ^*And f_e ^*And substituting the comprehensive evaluation function to obtain the comprehensive evaluation function only containing the speed variable.

And S42, for the given dynamic weight, economic weight and accelerator pedal strength, calling a particle swarm optimization algorithm by taking the comprehensive evaluation function as a fitness function to obtain the upshift speed under the given dynamic weight, economic weight and accelerator pedal strength.

In this embodiment, the dynamic weight and the economic weight are shown in table 1.

TABLE 1 weight combination table

The method comprises the following steps of calling a particle swarm optimization algorithm under the conditions of given dynamic weight, economic weight and accelerator pedal strength to obtain the upshift speeds of the 1-gear and the 2-gear.

In the step, a comprehensive evaluation function is used as a fitness function, the vehicle speed is used as a particle position, and a gear shifting vehicle speed range [ u [ u ] ]_min(2,j),u_max(1,j)](wherein u_min(2, j) is the minimum vehicle speed of the 2 nd gear; u. of_max(1, j) the maximum vehicle speed of 1 st gear) as a particle position search range; the specific acquisition process of the upshift speed given the dynamic weight, the economic weight and the accelerator pedal strength comprises the following sub-steps:

s421 particle population initialization, setting acceleration constant c₁＝2、c ₂2; setting the maximum iteration number M to be 100; setting the population size N to 80; setting the maximum value w of the inertial weight_max0.9, minimum value w_min＝0.4。

S422 assigning initial values to the positions and velocities of all particles, i.e. the initial position p of the ith particle_l ⁰In [ u ]_min(2,j),u_max(1,j)]An internal random value (l ═ 1, …, N); initial velocity v of the first particle_l ⁰In [0,0.9 ]]The value of internal random (l ═ 1, …, N). Calculating the fitness corresponding to each particle initial position according to the comprehensive evaluation function (the vehicle speed in the function is the particle position) obtained in the step S41; respectively taking the current particle position and the corresponding fitness as the local optimal particle position pbest_lAnd local optimumpbestFitness_l(ii) a And respectively taking the minimum value of the fitness of all the particles and the positions of the particles corresponding to the minimum value as a global optimal value gbestFitness and a global optimal particle position gbest.

S423 starts with k ═ 1 and ends with k ═ M for each particle. And for the kth iteration, the method comprises the following steps:

s4231 calculating inertia weight w

S4232, for each particle in the population, repeating the steps of:

s4232a at [0, 1 ]]Taking random values r within the range₁、r₂。

S4232b updating the speed and position of the particle

v_l ^k＝wv_l ^k-1+c₁r₁(pbest_l-p_l ^k-1)+c₂r₂(gbest-p_l ^k-1) (18)；

Wherein v is_l ^kThe velocity of the particle l for the kth iteration; v. of_l ^k-1The velocity of the (k-1) th iteration particle l; p is a radical of_l ^k-1The position of the particle/for the (k-1) th iteration.

p_l ^k＝p_l ^k-1+v_l ^k-1 (19)；

Wherein p is_l ^kThe position of the particle/for the kth iteration.

S4232c combines the gear shifting vehicle speed constraint condition to process the particle position boundary crossing: if p is_l ^k＜u_min(2, j), let p_l ^k＝u_min(2, j); if p is_l ^k＞u_max(1, j), let p_l ^k＝u_max(1,j)。

S4232d calculates fitness pbestFitness of particle l according to the comprehensive evaluation function obtained in step S41_l ^k。

S4232e A pbestFitness_l ^k＜pbestFitness_lThen pbestFitness_l＝pbestFitness_l ^k，pbest_l＝p_l ^kElse pbestFitness_lAnd pbest_lRemain unchanged.

S4232f A pbestFitness_lIf < gbestFitness, then gbestFitness ═ pbestFitness_l，gbest＝pbest_l。

S424 determines the upshift speed, and after all the particles are optimized according to the step S423, the global optimal particle position gbest corresponding to the global optimal value gbestFitness is used as the upshift speed under the given dynamic weight, economic weight and accelerator pedal strength, namely u is the upshift speed under the given dynamic weight, economic weight and accelerator pedal strength_aup(j)＝gbest。

S43 steps through the determined accelerator pedal strengths (j is 1 to h), repeats step S42, and determines upshift vehicle speeds u of different accelerator pedal strengths given the power weight and the economy weight_aupAnd forming an upshift curve under the given dynamic weight and economic weight.

The strength of the accelerator pedal is divided into equal parts h equal to 20 from 0-100% (corresponding to positions where the accelerator pedal is released and the accelerator pedal is stepped to the bottom respectively), the strength of the accelerator pedal is taken from small to large, the step S42 is repeated until j equal to h ends the cycle, and the upshift vehicle speed u with different strengths of the accelerator pedal is obtained_aupAnd forming an upshift curve under the given dynamic weight and economic weight.

S44 calculating the downshift vehicle speed u under different accelerator pedal strengths according to the following formula_adownAnd forming a downshift curve under the given dynamic weight and economic weight:

u_adown＝u_aup*(1-A) (20)；

wherein A is a given coefficient; when the accelerator pedal strength is less than 30%, a is 0.4; when the strength of the accelerator pedal is more than or equal to 30%, A is 0.15.

S45 repeats the above steps S42-S44 for different given power weights and economic weights according to table 1 to obtain upshift curves and downshift curves for different given power weights and economic weights, as shown in fig. 10. In fig. 10, each curve corresponds to an upshift curve or a downshift curve of a combination of the power weighting value and the economy weighting value.

The personalized gear shifting rule of the pure electric vehicle obtained according to the optimization method considering the transmission efficiency provided by the embodiment can be stored in a gear shifting rule database in a form of a data table.

Application example

The application example adopts the obtained pure electric vehicle individual gear shifting rule of the embodiment, namely the gear shifting curve and the gear shifting curve under different given power weight values and economic weight values shown in fig. 10 to realize gear shifting control in the driving process of the vehicle. As shown in fig. 11, when in use, a Transmission Control Unit (TCU) collects the vehicle speed and the accelerator pedal strength signal in real time, and then obtains a power weight (i.e., a desired power value) and an economy weight (i.e., an expected economy value) according to the collected vehicle speed and accelerator pedal strength signal. And selecting a gear shifting rule (namely the individualized gear shifting rule with the optimal comprehensive performance) matched with the current dynamic weight and the economic weight from a gear shifting rule candidate library stored in the TCU according to the dynamic weight and the economic weight, and determining a target gear according to the selected gear shifting rule according to the vehicle speed and the strength of an accelerator pedal. And then, the prior art is used for realizing the optimal gear shifting operation of the automobile with comprehensive performance, which is consistent with the performance expectation of a driver and takes the working condition efficiency of the power transmission system into consideration, through the gear shifting control system.

For example, when the power weighting value is determined to be 0.8 (the economy weighting value is 0.2), an upshift curve with the power weighting value of 0.8 (the economy weighting value is 0.2) and a downshift curve with the power weighting value of 0.8 (the economy weighting value is 0.2) can be determined from fig. 10. When a point corresponding to the speed and the strength of an accelerator pedal passes through an upshift curve from the left side of the upshift curve in the driving process of the automobile, the automobile should be shifted into a high gear; when the point corresponding to the vehicle speed and the strength of the accelerator pedal crosses the downshift curve from the right side of the downshift curve, the low gear should be shifted down. Corresponding to the embodiment, if the strength of the accelerator pedal is not changed, if the vehicle speed is increased and passes through the upshift curve from the left side of the upshift curve, the gear 2 is shifted up; if the vehicle speed decreases and goes from the right side of the downshift curve to the left side of the downshift curve, the 1 st gear is shifted down.

The above-mentioned manner of obtaining the expected value of the driver's dynamic performance and the expected value of the economic performance can be realized by the prior art, for example, in "a method for quantifying the expected value of the driver's dynamic performance and the economic performance" (application No. 201911247447.8), which can continuously quantify the expected value of the driver's dynamic performance and the economic performance of the vehicle.

It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.

Claims (5)

1. A pure electric vehicle personalized gear shifting rule optimization method considering transmission efficiency is characterized by comprising the following steps:

s1, respectively establishing a motor efficiency model and a transmission efficiency model;

the establishment process of the motor efficiency model comprises the following sub-steps:

s11, calculating input power and output power of the motor by using input end voltage, current, motor speed and motor torque data under different working conditions collected by a motor efficiency experiment, and taking the ratio of the output power and the input power of the motor as the motor efficiency value under each corresponding working condition;

s12, constructing a motor data collection by using the motor rotating speed, the motor torque and the motor efficiency value obtained in the step S11 under different working conditions, and selecting a motor efficiency model training set and a testing set from the motor data collection;

s13, training and testing the neural network model by using the motor efficiency model training set and the test set data constructed in the step S12 to obtain a motor efficiency model;

the transmission efficiency model corresponds to a transmission gear, and the process for establishing the transmission efficiency model in a certain gear comprises the following sub-steps:

s14, calculating the input power and the output power of a certain gear of the transmission by using the input rotating speed, the input torque, the output rotating speed and the output torque of the transmission under different working conditions of the certain gear, which are acquired by a transmission efficiency experiment, and taking the ratio of the output power and the input power of the certain gear of the transmission as the efficiency value of the certain gear of the transmission under the corresponding working condition;

s15, constructing a gear data collection of the transmission by using the transmission input rotating speed, the transmission input torque and the transmission oil temperature of the transmission under different working conditions of a certain gear of the transmission and the efficiency value of the certain gear of the transmission obtained in the step S14, and selecting a transmission efficiency model training set and a test set from the gear data collection of the transmission;

s16, training and testing the neural network model by using the transmission efficiency model training set and the test set data constructed in the step S15 to obtain an efficiency model of a certain gear of the transmission;

s2, dividing the strength of an accelerator pedal into a plurality of equal parts from 0-100%, and constructing a gear shifting vehicle speed range of two adjacent gears under the same strength of the accelerator pedal by using the minimum vehicle speed of the high gear and the maximum vehicle speed of the low gear of the two adjacent gears under the same strength of the accelerator pedal;

the method for determining the minimum vehicle speed of the high gear and the maximum vehicle speed of the low gear of two adjacent gears under the same accelerator pedal strength comprises the following steps: firstly, determining the motor rotating speed and the automobile acceleration corresponding to different driving speeds of a high gear and a low gear of two adjacent gears under the same accelerator pedal strength; then according to the motor rotating speed and the automobile acceleration corresponding to different driving speeds of the high gear, finding out the minimum speed of the automobile in the high gear, wherein the automobile acceleration is greater than 0 and the current rotating speed of the motor is greater than the minimum rotating speed of the motor, and the minimum speed is the minimum speed of the high gear; then according to the motor rotating speed and the automobile acceleration corresponding to different driving speeds of the low gear, finding out the maximum automobile speed in which the automobile acceleration in the low gear is greater than 0 and the current rotating speed of the motor is less than the maximum rotating speed of the motor, namely the maximum automobile speed of the low gear;

the motor rotating speed and the automobile acceleration corresponding to different driving speeds of a high gear or a low gear of two adjacent gears under the same accelerator pedal strength are obtained according to the following steps:

s21, importing parameters of the whole vehicle and a transmission system;

s22, obtaining motor rotating speeds corresponding to different driving speeds;

s23, motor torques corresponding to different driving speeds are obtained through the motor torque model;

s24, taking the motor rotating speed and the motor torque as input data, and obtaining motor efficiencies corresponding to different driving speeds through a motor efficiency model;

s25, acquiring transmission efficiency corresponding to different driving speeds through a transmission efficiency model by taking the transmission input speed, the transmission input torque and the given transmission oil temperature as input data;

s26, determining automobile acceleration corresponding to different driving speeds according to the motor torque, the motor efficiency and the transmission efficiency;

s3, establishing gear shifting vehicle speed constraint conditions of two adjacent gears under different accelerator pedal strengths;

s4, different dynamic weight values and economic weight values are given, the strength of each accelerator pedal is traversed, the gear shifting speed which enables a comprehensive evaluation function to be minimum is obtained through a particle swarm optimization algorithm by combining the gear shifting speed range and constraint conditions of two adjacent gears, and then a gear shifting curve and a gear shifting curve under the different dynamic weight values and the different economic weight values are obtained, namely the personalized gear shifting rule of the pure electric vehicle;

the process of obtaining the gear shifting vehicle speed with the minimum comprehensive evaluation function by adopting the particle swarm optimization algorithm comprises the following steps:

s41, determining a comprehensive evaluation function;

the comprehensive evaluation function is as follows:

wherein u represents the vehicle speed; w is a_d、w_eRespectively given dynamic weight and economic weight to reflect the performance expectation of the driver, w_d+w_e＝1；f_d'(u)、f_e' (u) are each independentlyThe dynamic subobjective function and the economic subobjective function are normalized; f. of_d ^*、f_e ^*Respectively obtaining an optimal dynamic performance objective function value and an optimal economic performance objective function value after normalization;

s42, under the condition of giving a dynamic weight, an economic weight and an accelerator pedal strength, taking a comprehensive evaluation function as a fitness function, and calling a particle swarm optimization algorithm to obtain a upshift speed under the given dynamic weight, economic weight and accelerator pedal strength;

s43 traversing each determined accelerator pedal strength, repeating the step S42, and obtaining the upshifting speed u with different accelerator pedal strengths under the given dynamic weight and economic weight_aupForming a gear-up curve under the given dynamic weight and economic weight;

s44 calculating the downshift vehicle speed u under different accelerator pedal strengths according to the following formula_adownAnd forming a downshift curve under the given dynamic weight and economic weight:

u_adown＝u_aup*(1-A)；

wherein A is a given coefficient; when the accelerator pedal strength is less than 30%, a is 0.4; when the strength of the accelerator pedal is more than or equal to 30%, A is 0.15;

s45, repeating the steps S42-S44 for given different given power weight values and economic weight values to obtain an upshift curve and a downshift curve under different given power weight values and economic weight values.

2. The pure electric vehicle personalized gear shift schedule optimization method considering transmission efficiency according to claim 1, characterized in that in step S3, the gear shift speeds u of two adjacent gears under different accelerator pedal strengths_aThe constraint conditions of (1) are:

g₂(u_a)＝u_a-u_min≥0；

g₃(u_a)＝u_max-u_a≥0；

wherein u is_minMinimum vehicle speed in high gear for a given accelerator pedal intensity; u. of_maxFor the maximum vehicle speed in the low gear for a given accelerator pedal intensity.

3. The pure electric vehicle personalized gear shifting law optimization method considering transmission efficiency according to claim 1 or 2, characterized in that the normalized dynamic performance is a partial objective function f_d' (u) and an economic objective function f_e' (u) acquisition mode: the absolute value of the acceleration difference under the same accelerator pedal strength of two adjacent gears of the pure electric vehicle transmission is used as a dynamic sub-target function; taking an absolute value of a unit mileage energy consumption difference under the same accelerator pedal strength of each adjacent gear of the pure electric vehicle transmission as an economic subobjective function; respectively obtaining the maximum value and the minimum value of a dynamic sub-objective function and an economic sub-objective function according to the upper limit and the lower limit of the gear shifting speed range of two adjacent gears under different accelerator pedal strengths; then, the dynamic subobjective function and the economic subobjective function are normalized, and the normalized dynamic subobjective function f is obtained_d' (u) and an economic objective function f_e'(u)。

4. The pure electric vehicle personalized gear shift law optimization method considering transmission efficiency according to claim 1 or 2, characterized in that the normalized optimal dynamic performance is a target function value f_d ^*And the best economic objective function value f_e ^*The acquisition mode is as follows: f in the comprehensive evaluation function_d ^*The value is 0, f_e ^*The value is taken to be 0, let w_d＝1，w_eUsing an optimization algorithm to solve the minimum value f of the dynamic subobjective function under each accelerator pedal strength_dNormalizing the minimum value of the dynamic subobjective function under each accelerator pedal strength to obtain the normalized optimal dynamic subobjective function value f under each pedal strength_d ^*(ii) a Let w_d＝0，w_eSolving the minimum value f of the economic subobjective function under each accelerator pedal strength by using an optimization algorithm as 1_eNormalizing the economic subobjective function value under each accelerator pedal strength to obtain the normalized optimal economic subobjective function value f under each pedal strength_e ^*。

5. The application of the personalized gear shifting rule of the pure electric vehicle obtained by the optimization method in any one of claims 1 to 4 in gear shifting control during driving of the vehicle.

Images