Distributed hybrid tractor driving force reconstruction and wheel-soil model online identification method

Abstract

The present invention relates to the field of hybrid tractor integrated control technology, and in particular to a distributed hybrid tractor driving force reconstruction and wheel-soil model online identification method. A tractor wheel-soil model online identification method is applied to a distributed hybrid tractor driving system, wherein the distributed hybrid tractor driving system comprises: a left motor (1), a left transmission (2), a left front wheel (3), a right motor (4), a right transmission (5), a right front wheel (6), a left motor controller (7), a right motor controller (8), a power battery pack (9), a battery management system (10), a vehicle controller (11), a diesel engine (12), a generator (13), a rear transmission (14), a differential (15), a left rear wheel (16), and a right rear wheel (17); the driving force reconstruction method of the present invention has a strategy that can control the front and rear axle tire driving forces to vary within a wide range in a relatively short period of time, so as to stimulate the tire slip rate as much as possible, expand the online identification sample range, and do not affect normal operation efficiency and operation quality.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated control of hybrid tractors, and in particular to a method for online identification of driving force reconstruction and wheel-soil model of a distributed hybrid tractor. Background Art

[0002] The front two wheels of a distributed hybrid tractor are driven by an electric motor system, while the rear two wheels are driven by a traditional internal combustion engine system. The vehicle's driving force is directly coupled between the four wheels and the ground, which means that the power coupling path requires multiple interactions between the tires and the ground. Therefore, in the process of precise energy management and power performance optimization of tractors, a relatively accurate tire-ground interaction model is crucial for real-time analysis of front and rear axle power output.

[0003] Previous researchers have often used empirical, semi-empirical, and numerical modeling methods to establish wheel-soil interaction models for off-road vehicles. However, these modeling methods are limited to specific tire and soil types and require parameter calibration based on tire specifications and soil properties. Due to the wide variety of tractor tires (including high-pressure, low-pressure, and ultra-low-pressure tires) and the complex and diverse operating surfaces (including wheat stubble, soft surfaces, sloped surfaces, and potholes), these modeling methods struggle to accurately represent the quantitative relationship between the real-time motion state of the tractor tires and the driving force when operating on the current plot. This makes it difficult to control the transient dynamics of each tire, which in turn affects the overall performance of the tractor.

[0004] Distributed hybrid tractors can achieve independent drive of the front and rear axles. A deliberately designed rule for front and rear drive force complementation allows for drive force regulation and reconfiguration in a fluctuating manner without affecting normal tillage, facilitating real-time identification of tire-soil dynamic models. Therefore, it is necessary to develop a distributed hybrid tractor front and rear axle drive force reconstruction strategy and a precise online identification method for the wheel-soil interaction model based on this strategy. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention proposes a driving force reconstruction method for a distributed hybrid tractor and an online wheel-soil model identification method. The driving force reconstruction method of the distributed hybrid tractor can control the driving force of the front and rear axle tires to vary within a large range in a relatively short period of time, so as to stimulate the tire slip rate as much as possible, expand the online identification sample range, and do not affect the normal operation efficiency and operation quality; the online identification method of the tractor wheel-soil model has simple calculations and fast convergence speed, and can improve the identification accuracy, reliability and real-time performance of the wheel-soil interaction model.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A distributed hybrid tractor driving force reconstruction method is applied to a distributed hybrid tractor drive system, wherein the distributed hybrid tractor drive system includes: a left motor 1, a left transmission 2, a left front wheel 3, a right motor 4, a right transmission 5, a right front wheel 6, a left motor controller 7, a right motor controller 8, a power battery pack 9, a battery management system 10, a vehicle controller 11, a diesel engine 12, a generator 13, a rear transmission 14, a differential 15, a left rear wheel 16, and a right rear wheel 17;

[0008] The output end of the left motor 1 is connected to the left transmission 2, and the output end of the left transmission 2 is connected to the left front wheel 3;

[0009] The output end of the right motor 4 is connected to the right transmission 5, and the output end of the right transmission 5 is connected to the right front wheel 6;

[0010] The output end of the diesel engine 12 is connected to the rear transmission 14, and the output end of the rear transmission 14 is respectively connected to the differential 15 and the generator 13, and the output end of the differential 15 is respectively connected to the left rear wheel 16 and the right rear wheel 17;

[0011] The battery management system 10 is electrically connected to the power battery pack 9, the generator 13, the vehicle controller 11, the left motor controller 7 and the right motor controller 8 respectively;

[0012] The left motor controller 7 is electrically connected to the left motor 1, and the right motor controller 8 is electrically connected to the right motor 4; the battery management system 10, the vehicle controller 11, the left motor controller 7, and the right motor controller 8 are signal-connected to each other and generate control signals for the left motor 1, the right motor 4, the diesel engine 12, the left transmission 2, the right transmission 5, and the rear transmission 14;

[0013] Four wheel speed and torque sensors 18 are coaxially fixed to the left front wheel 3, the right front wheel 6, the left rear wheel 16, and the right rear wheel 17, respectively;

[0014] The GNSS receiver 19 and the inertial measurement unit 20 are both fixed to the vehicle body;

[0015] Three pin-type tension sensors 21 are respectively installed at the three hinge points between the suspension lifting mechanism and the vehicle body;

[0016] The method comprises the following steps:

[0017] S1. Obtain the signals of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor 18 j and torque T j ; Get the current position of the tractor from the GNSS receiver 19 and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit 20 Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor 21 i In addition, the driver's expected speed v expect Obtained by analyzing the accelerator pedal and brake pedal signals;

[0018] S2, longitudinal speed and current acceleration Fusion is performed to obtain an estimate of the longitudinal vehicle speed

[0019]

[0020] In formula 1, is the estimated longitudinal vehicle speed in m / s; is the longitudinal velocity in m / s; is the current acceleration, in m / s 2 ;τ represents the sampling period of the inertial measurement unit, in seconds; N=(t k -t k-1 ) / τ-1;

[0021] S3, according to the load force F of each hinge point of the suspension lifting mechanism i and the angle θ between it and the horizontal plane i Calculate the total horizontal load force F total

[0022]

[0023] In formula 2, F total is the total horizontal load force, in N; F i is the load force of each hinge point of the suspension lifting mechanism, in N; θ i is the load force F at each hinge point of the suspension lifting mechanism i The angle with the horizontal plane, in degrees;

[0024] S4, according to the driver's expected speed v expect , longitudinal vehicle speed estimate and the total horizontal load force F total , calculate the required torque T req

[0025]

[0026] In formula 3, e is the speed error, in m / s; v expect is the expected vehicle speed, in m / s; is the estimated longitudinal speed, in m / s; T reqis the required torque, in Nm; F total is the total horizontal load force, in N; K P ,K I and K D are the coefficients of the proportional term, integral term, and differential term respectively; τ represents the sampling period of the inertial measurement unit, in seconds; t represents the current time, in seconds;

[0027] S5. Design a variable-amplitude, variable-frequency front axle target drive torque curve and distribute it evenly to the left and right front wheels. The rear axle compensates for the drive torque to maintain a roughly constant longitudinal vehicle speed. The set values ​​for the front and rear axle drive torques are as follows:

[0028]

[0029] T rear =T req -T front

[0030] In formula 4, T front and T rear represents the target driving torque of the front axle and rear axle respectively, in Nm; A and ω are parameters to be calibrated; t represents the current time, in seconds; T front_m Indicates the maximum output torque that the drive system can provide to the front axle, in Nm; T req Indicates the required torque in Nm;

[0031] S6, according to the front and rear axle drive torque setting value T front and T rear Match the output torque and working gear of each power source.

[0032] A tractor wheel-soil model online identification method is applied to the drive system of a distributed hybrid tractor. The distributed hybrid tractor drive system includes: a left motor 1, a left transmission 2, a left front wheel 3, a right motor 4, a right transmission 5, a right front wheel 6, a left motor controller 7, a right motor controller 8, a power battery pack 9, a battery management system 10, a vehicle controller 11, a diesel engine 12, a generator 13, a rear transmission 14, a differential 15, a left rear wheel 16, and a right rear wheel 17.

[0033] The output end of the left motor 1 is connected to the left transmission 2, and the output end of the left transmission 2 is connected to the left front wheel 3;

[0034] The output end of the right motor 4 is connected to the right transmission 5, and the output end of the right transmission 5 is connected to the right front wheel 6;

[0035] The output end of the diesel engine 12 is connected to the rear transmission 14, and the output end of the rear transmission 14 is respectively connected to the differential 15 and the generator 13, and the output end of the differential 15 is respectively connected to the left rear wheel 16 and the right rear wheel 17;

[0036] The battery management system 10 is electrically connected to the power battery pack 9, the generator 13, the vehicle controller 11, the left motor controller 7 and the right motor controller 8 respectively;

[0037] The left motor controller 7 is electrically connected to the left motor 1, and the right motor controller 8 is electrically connected to the right motor 4; the battery management system 10, the vehicle controller 11, the left motor controller 7, and the right motor controller 8 are signal-connected to each other and generate control signals for the left motor 1, the right motor 4, the diesel engine 12, the left transmission 2, the right transmission 5, and the rear transmission 14;

[0038] Four wheel speed and torque sensors 18 are coaxially fixed to the left front wheel 3, the right front wheel 6, the left rear wheel 16, and the right rear wheel 17, respectively;

[0039] The GNSS receiver 19 and the inertial measurement unit 20 are both fixed to the vehicle body;

[0040] Three pin-type tension sensors 21 are respectively installed at the three hinge points between the suspension lifting mechanism and the vehicle body;

[0041] Wherein, the method comprises the following steps:

[0042] S1. Obtain the signals of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor 18 j and torque T j ; Get the current position of the tractor from the GNSS receiver 19 and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit 20 Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor 21 i ;

[0043] S2, based on the estimated longitudinal vehicle speed The rotation speed of each wheel ω j and rolling radius r j , calculate the slip rate δ of each wheel j

[0044]

[0045] In formula 5, δ j is the slip rate of each wheel; r j is the rolling radius of each wheel, in m; ω jis the rotation speed of each wheel, in rad / s; is the estimated longitudinal vehicle speed in m / s; j = 1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively;

[0046] S3, based on the tractor mass m, wheelbase L, horizontal distance a from the center of gravity to the front axle, horizontal distance b from the center of gravity to the rear axle, and total horizontal load F total and its height H from the ground to calculate the vertical load W of each wheel j ; According to the wheel torque T j Calculate the longitudinal traction F of each wheel pull_j ; According to the vertical load W of each wheel j and traction force F pull_j Calculate the traction coefficient Φ j

[0047]

[0048] In formula 6, W j is the vertical load of each wheel, in N; m is the mass of the tractor, in kg; g is the acceleration due to gravity, in kg·m / s 2 ; L is the wheelbase, in meters; a is the horizontal distance from the center of gravity to the front axle, in meters; b is the horizontal distance from the center of gravity to the rear axle, in meters; F total is the total horizontal load force, in N; H is the height of the tractor's center of gravity from the ground, in m;

[0049] F pull_j is the longitudinal traction of each wheel, in N; T j is the torque of each wheel, in Nm; r j is the rolling radius of each wheel, in m;

[0050] Φ j is the traction coefficient;

[0051] j=1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel and right rear wheel respectively;

[0052] S4, according to the given slip rate-traction coefficient sample data set (δ k ,Φ k )(k=1,2,…,N) to establish a high-dimensional feature space

[0053]

[0054] In formula 7, δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; ρ is the deviation; φ(δ k ) means that δk Functions that map to high-dimensional feature spaces;

[0055] S5. According to the principle of structural risk minimization, the regression problem is transformed into a constrained optimization problem

[0056]

[0057]

[0058] In formula 8, J is the loss function; ρ is the deviation; e is the speed error, in m / s; δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; γ is the adjustable constant; e k is the error, e=[e1,…,e N ] T ∈R n is the error vector; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces;

[0059] S6. Define the Lagrangian function based on the optimization function

[0060]

[0061] In formula 9, J L is the Lagrangian function; ω is the weight vector; ρ is the deviation; e is the speed error, in m / s; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n Φ k is the sample data of traction coefficient; e k is the error; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces;

[0062] S7. Optimize and solve to obtain the dual matrix equation

[0063]

[0064] In formula 10, δ i and δ j is the slip rate sample data; σ is a constant; γ is an adjustable constant; ρ is the deviation; α k ∈R is the Lagrange multiplier; Φ k is the corresponding traction coefficient;

[0065] S8, obtaining the values ​​of parameters α and ρ, thereby obtaining an online identified wheel-soil model;

[0066]

[0067] In formula 11, Φ is the corresponding traction coefficient; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n ;δ is the measured slip rate;δ k is the sample data of slip rate; ρ is the deviation.

[0068] Among them, in step S1, the longitudinal vehicle speed estimation value Calculated by formula 1; the rolling radius r of each wheel j Obtained through measurement;

[0069]

[0070] In formula 1, is the estimated longitudinal vehicle speed in m / s; is the longitudinal velocity in m / s; is the current acceleration, in m / s 2 ;τ represents the sampling period of the inertial measurement unit, in seconds; N=(t k -t k-1 ) / τ-1.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] The present invention proposes a driving force reconstruction method for a distributed hybrid tractor. This strategy can control the driving force of the front and rear axle tires to vary within a large range in a relatively short period of time, so as to stimulate the tire slip rate as much as possible, expand the range of online identification samples, and not affect normal operating efficiency and quality.

[0073] The present invention proposes an online identification method for a tractor wheel-soil model. The method has simple calculation and fast convergence speed, and can improve the identification accuracy, reliability and real-time performance of the wheel-soil model. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 This is a schematic diagram of the structure of a distributed hybrid tractor drive system;

[0075] Figure 2 Schematic diagram of the installation location of each sensor;

[0076] Figure 3 Reconstruct the method control flow chart for the driving force;

[0077] Figure 4 This is the flow chart for online identification of tractor wheel-soil model.

[0078] The accompanying drawings are as follows:

[0079] 1. Left motor 2. Left transmission

[0080] 3. Left front wheel 4. Right motor

[0081] 5. Right transmission 6. Right front wheel

[0082] 7. Left motor controller 8. Right motor controller

[0083] 9. Power battery pack 10. Battery management system

[0084] 11. Vehicle controller 12. Diesel engine

[0085] 13. Generator 14. Rear transmission

[0086] 15. Differential 16. Left rear wheel

[0087] 17. Right rear wheel 18. Wheel encoder

[0088] 19. GNSS receiver 20. Inertial measurement unit

[0089] 21. Pin-type tension sensor DETAILED DESCRIPTION

[0090] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0091] like Figure 1 As shown, the drive system of the distributed hybrid tractor used in the present invention includes: a left motor 1, a left transmission 2, a left front wheel 3, a right motor 4, a right transmission 5, a right front wheel 6, a left motor controller 7, a right motor controller 8, a power battery pack 9, a battery management system 10, a vehicle controller 11, a diesel engine 12, a generator 13, a rear transmission 14, a differential 15, a left rear wheel 16 and a right rear wheel 17.

[0092] The output end of the left motor 1 is connected to the left transmission 2, and the output end of the left transmission 2 is connected to the left front wheel 3;

[0093] The output end of the right motor 4 is connected to the right transmission 5, and the output end of the right transmission 5 is connected to the right front wheel 6;

[0094] The output end of the diesel engine 12 is connected to the rear transmission 14, and the output end of the rear transmission 14 is respectively connected to the differential 15 and the generator 13, and the output end of the differential 15 is respectively connected to the left rear wheel 16 and the right rear wheel 17;

[0095] The battery management system 10 is electrically connected to the power battery pack 9, the generator 13, the vehicle controller 11, the left motor controller 7 and the right motor controller 8 respectively;

[0096] The left motor controller 7 is electrically connected to the left motor 1, and the right motor controller 8 is electrically connected to the right motor 4; the battery management system 10, the vehicle controller 11, the left motor controller 7, and the right motor controller 8 are signal-connected to each other and generate control signals for the left motor 1, the right motor 4, the diesel engine 12, the left transmission 2, the right transmission 5, and the rear transmission 14.

[0097] like Figure 2 As shown, the sensors used in the present invention include wheel speed and torque sensors 18, a GNSS receiver 19, an inertial measurement unit 20, and a pin-type tension sensor 21. There are four wheel speed and torque sensors 18, coaxially fixed to the left front wheel 3, right front wheel 6, left rear wheel 16, and right rear wheel 17, respectively; the GNSS receiver 19 and inertial measurement unit 20 are both fixed to the vehicle body; and there are three pin-type tension sensors 21, installed at the three hinge points between the suspension lifting mechanism and the vehicle body.

[0098] like Figure 3 As shown, the distributed hybrid tractor driving force reconstruction method of the present invention includes the following steps:

[0099] S1. Obtain the signals of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor 18 j and torque T j ; Get the current position of the tractor from the GNSS receiver 19 and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit 20 Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor 21 i In addition, the driver's desired speed v expect Obtained by analyzing the accelerator pedal and brake pedal signals.

[0100] S2, longitudinal speed and current acceleration Fusion is performed to obtain an estimate of the longitudinal vehicle speed

[0101]

[0102] In formula 1, is the estimated longitudinal vehicle speed in m / s; is the longitudinal velocity in m / s; is the current acceleration, in m / s2 ;τ represents the sampling period of the inertial measurement unit, in seconds; N=(t k -t k-1 ) / τ-1.

[0103] S3, according to the load force F of each hinge point of the suspension lifting mechanism i and the angle θ between it and the horizontal plane i Calculate the total horizontal load force F total

[0104]

[0105] In formula 2, F total is the total horizontal load force, in N; F i is the load force of each hinge point of the suspension lifting mechanism, in N; θ i is the load force F at each hinge point of the suspension lifting mechanism i The angle with the horizontal plane, in degrees.

[0106] S4, according to the driver's expected speed v expect , longitudinal vehicle speed estimate and the total horizontal load force F total , calculate the required torque T req

[0107]

[0108] In formula 3, e is the speed error, in m / s; v expect is the expected vehicle speed, in m / s; is the estimated longitudinal speed, in m / s; T req is the required torque, in Nm; F total is the total horizontal load force, in N; K P ,K I and K D are the coefficients of the proportional term, integral term, and differential term, respectively; τ represents the sampling period of the inertial measurement unit, in seconds; and t represents the current time, in seconds.

[0109] S5. Design a variable amplitude and variable frequency front axle target drive torque curve and distribute it evenly to the left and right front wheels. The rear axle compensates for the drive torque to maintain a roughly constant longitudinal vehicle speed. The set values ​​for the front and rear axle drive torques are as follows:

[0110]

[0111] T rear =T req -T front

[0112] In formula 4, T front and T rear represents the target driving torque of the front axle and rear axle respectively, in Nm; A and ω are parameters to be calibrated; t represents the current time, in seconds; T front_m Indicates the maximum output torque that the drive system can provide to the front axle, in Nm; T req Indicates the required torque in Nm.

[0113] S6, according to the front and rear axle drive torque setting value T front and T rear Match the output torque and working gear of each power source.

[0114] like Figure 4 As shown, the online identification method of a tractor wheel-soil model of the present invention includes the following steps:

[0115] S1. Obtain the signals of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor 18 j and torque T j ; Get the current position of the tractor from the GNSS receiver 19 and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit 20 Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor 21 i ;

[0116] S2, based on the estimated longitudinal vehicle speed The rotation speed of each wheel ω j and rolling radius r j , calculate the slip rate δ of each wheel j

[0117]

[0118] In formula 5, δ j is the slip rate of each wheel; r j is the rolling radius of each wheel, in m; ω j is the rotation speed of each wheel, in rad / s; is the estimated longitudinal vehicle speed in m / s; j=1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively.

[0119] Among them, the estimated longitudinal speed Calculated by formula 1. The rolling radius r of each wheel j Obtained through measurement.

[0120] S3, based on the tractor mass m, wheelbase L, horizontal distance a from the center of gravity to the front axle, horizontal distance b from the center of gravity to the rear axle, and total horizontal load Ftotal and its height H from the ground to calculate the vertical load W of each wheel j ; According to the wheel torque T j Calculate the longitudinal traction F of each wheel pull_j ; According to the vertical load W of each wheel j and traction force F pull_j Calculate the traction coefficient Φ j

[0121]

[0122] In formula 6, W j is the vertical load of each wheel, in N; m is the mass of the tractor, in kg; g is the acceleration due to gravity, in kg·m / s 2 ; L is the wheelbase, in meters; a is the horizontal distance from the center of gravity to the front axle, in meters; b is the horizontal distance from the center of gravity to the rear axle, in meters; F total is the total horizontal load force, in N; H is the height of the tractor's center of gravity from the ground, in m;

[0123] F pull_j is the longitudinal traction of each wheel, in N; T j is the torque of each wheel, in Nm; r j is the rolling radius of each wheel, in m;

[0124] Φ j is the traction coefficient; j=1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel and right rear wheel respectively.

[0125] S4, according to the given slip rate-traction coefficient sample data set (δ k ,Φ k )(k=1,2,…,N) to establish a high-dimensional feature space

[0126]

[0127] In formula 7, δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; ρ is the deviation; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces.

[0128] S5. According to the principle of structural risk minimization, the regression problem is transformed into a constrained optimization problem

[0129]

[0130]

[0131] In formula 8, J is the loss function; ρ is the deviation; e is the speed error, in m / s; δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; γ is the adjustable constant; e k is the error, e=[e1,…,e N ] T ∈R n is the error vector; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces.

[0132] S6. Define the Lagrangian function based on the optimization function

[0133]

[0134] In formula 9, J L is the Lagrangian function; ω is the weight vector; ρ is the deviation; e is the speed error, in m / s; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n Φ k is the sample data of traction coefficient; e k is the error; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces.

[0135] S7. Optimize and solve to obtain the dual matrix equation

[0136]

[0137] In formula 10, δ i and δ j is the slip rate sample data; σ is a constant; γ is an adjustable constant; ρ is the deviation; α k ∈R is the Lagrange multiplier; Φ k is the corresponding traction coefficient.

[0138] S8. Obtain the values ​​of parameters α and ρ, thereby obtaining an online identified wheel-soil model.

[0139]

[0140] In formula 11, Φ is the corresponding traction coefficient; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n;δ is the measured slip rate;δ k is the sample data of slip rate; ρ is the deviation.

Claims

1. A distributed hybrid tractor driving force reconstruction method, applied to a distributed hybrid tractor driving system, the distributed hybrid tractor driving system comprising: Left motor (1), left transmission (2), left front wheel (3), right motor (4), right transmission (5), right front wheel (6), left motor controller (7), right motor controller (8), power battery pack (9), battery management system (10), vehicle controller (11), diesel engine (12), generator (13), rear transmission (14), differential (15), left rear wheel (16) and right rear wheel (17); The output end of the left motor (1) is connected to the left transmission (2), and the output end of the left transmission (2) is connected to the left front wheel (3); The output end of the right motor (4) is connected to the right transmission (5), and the output end of the right transmission (5) is connected to the right front wheel (6); The output end of the diesel engine (12) is connected to the rear transmission (14), the output end of the rear transmission (14) is respectively connected to the differential (15) and the generator (13), and the output end of the differential (15) is respectively connected to the left rear wheel (16) and the right rear wheel (17); The battery management system (10) is electrically connected to the power battery pack (9), the generator (13), the vehicle controller (11), the left motor controller (7) and the right motor controller (8); The left motor controller (7) is electrically connected to the left motor (1), and the right motor controller (8) is electrically connected to the right motor (4); the battery management system (10), the vehicle controller (11), the left motor controller (7), and the right motor controller (8) are signal-connected to each other and generate control signals for the left motor (1), the right motor (4), the diesel engine (12), the left transmission (2), the right transmission (5), and the rear transmission (14); Four wheel speed torque sensors (18) are coaxially fixed to the left front wheel (3), the right front wheel (6), the left rear wheel (16) and the right rear wheel (17); The GNSS receiver (19) and the inertial measurement unit (20) are both fixed to the vehicle body; Three pin-type tension sensors (21) are respectively installed on three hinge points between the suspension lifting mechanism and the vehicle body; The method is characterized in that: S1. Obtain the signal of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor (18) j and torque T j ; Get the current position of the tractor from the GNSS receiver (19) and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit (20) Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor (21) i In addition, the driver's expected speed v expect Obtained by analyzing the accelerator pedal and brake pedal signals; S2, longitudinal speed and current acceleration Fusion is performed to obtain an estimate of the longitudinal vehicle speed In formula 1, is the estimated longitudinal vehicle speed in m / s; is the longitudinal velocity in m / s; Current acceleration in m / s 2 ;τ represents the sampling period of the inertial measurement unit, in seconds; N=(t k -t k-1 ) / τ-1; S3, according to the load force F of each hinge point of the suspension lifting mechanism i and the angle θ between it and the horizontal plane i Calculate the total horizontal load force F total In formula 2, F total is the total horizontal load force, in N; F i is the load force of each hinge point of the suspension lifting mechanism, in N; θ i is the load force F at each hinge point of the suspension lifting mechanism i The angle with the horizontal plane, in degrees; S4, according to the driver's expected speed v expect , longitudinal vehicle speed estimate and the total horizontal load force F total , calculate the required torque T req In formula 3, e is the speed error, in m / s; v expect is the expected vehicle speed, in m / s; is the estimated longitudinal speed, in m / s; T req is the required torque, in Nm; F total is the total horizontal load force, in N; K P ,K I and K D are the coefficients of proportional term, integral term and differential term respectively; τ represents the sampling period of the inertial measurement unit, in seconds; t represents the current time, in seconds; S5. Design a variable-amplitude, variable-frequency front axle target drive torque curve and distribute it evenly to the left and right front wheels. The rear axle compensates for the drive torque to maintain a roughly constant longitudinal vehicle speed. The set values ​​for the front and rear axle drive torques are as follows: T rear =T req -T front In formula 4, T front and T rear represents the target driving torque of the front axle and rear axle respectively, in Nm; A and ω are parameters to be calibrated; t represents the current time, in seconds; T front_m Indicates the maximum output torque that the drive system can provide to the front axle, in Nm; T req Indicates the required torque in Nm; S6, according to the front and rear axle drive torque setting value T front and T rear Match the output torque and working gear of each power source.

2. An online identification method for tractor wheel-soil models, applied to the drive system of a distributed hybrid tractor. The drive system of the distributed hybrid tractor includes: Left motor (1), left transmission (2), left front wheel (3), right motor (4), right transmission (5), right front wheel (6), left motor controller (7), right motor controller (8), power battery pack (9), battery management system (10), vehicle controller (11), diesel engine (12), generator (13), rear transmission (14), differential (15), left rear wheel (16) and right rear wheel (17); The output end of the left motor (1) is connected to the left transmission (2), and the output end of the left transmission (2) is connected to the left front wheel (3); The output end of the right motor (4) is connected to the right transmission (5), and the output end of the right transmission (5) is connected to the right front wheel (6); The output end of the diesel engine (12) is connected to the rear transmission (14), the output end of the rear transmission (14) is respectively connected to the differential (15) and the generator (13), and the output end of the differential (15) is respectively connected to the left rear wheel (16) and the right rear wheel (17); The battery management system (10) is electrically connected to the power battery pack (9), the generator (13), the vehicle controller (11), the left motor controller (7) and the right motor controller (8); The left motor controller (7) is electrically connected to the left motor (1), and the right motor controller (8) is electrically connected to the right motor (4); the battery management system (10), the vehicle controller (11), the left motor controller (7), and the right motor controller (8) are signal-connected to each other and generate control signals for the left motor (1), the right motor (4), the diesel engine (12), the left transmission (2), the right transmission (5), and the rear transmission (14); Four wheel speed torque sensors (18) are coaxially fixed to the left front wheel (3), the right front wheel (6), the left rear wheel (16) and the right rear wheel (17); The GNSS receiver (19) and the inertial measurement unit (20) are both fixed to the vehicle body; Three pin-type tension sensors (21) are respectively installed on three hinge points between the suspension lifting mechanism and the vehicle body; The method is characterized in that: S1. Obtain the signal of each sensor: Obtain the speed ω of each wheel from the wheel speed torque sensor (18) j and torque T j ; Get the current position of the tractor from the GNSS receiver (19) and calculate the longitudinal speed Get the current acceleration of the tractor from the inertial measurement unit (20) Obtain the load force F of each hinge point of the tractor suspension lifting mechanism from the pin-type tension sensor (21) i ; S2, based on the estimated longitudinal vehicle speed The rotation speed of each wheel ω j and rolling radius r j , calculate the slip rate δ of each wheel j In formula 5, δ j is the slip rate of each wheel; r j is the rolling radius of each wheel, in m; ω j is the rotation speed of each wheel, in rad / s; is the estimated longitudinal vehicle speed in m / s; j = 1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively; S3, based on the tractor mass m, wheelbase L, horizontal distance a from the center of gravity to the front axle, horizontal distance b from the center of gravity to the rear axle, and total horizontal load F total and its height H from the ground to calculate the vertical load W of each wheel j ; According to the wheel torque T j Calculate the longitudinal traction F of each wheel pull_j ; According to the vertical load W of each wheel j and traction force F pull_j Calculate the traction coefficient Φ j In formula 6, W j is the vertical load of each wheel, in N; m is the mass of the tractor, in kg; g is the acceleration due to gravity, in kg·m / s 2 ; L is the wheelbase, in meters; a is the horizontal distance from the center of gravity to the front axle, in meters; b is the horizontal distance from the center of gravity to the rear axle, in meters; F total is the total horizontal load force, in N; H is the height of the tractor's center of gravity from the ground, in m; F pull_j is the longitudinal traction of each wheel, in N; T j is the torque of each wheel, in Nm; r j is the rolling radius of each wheel, in m; Φ j is the traction coefficient; j=1, 2, 3, 4 represent the left front wheel, right front wheel, left rear wheel and right rear wheel respectively; S4, according to the given slip rate-traction coefficient sample data set (δ k ,Φ k )(k=1,2,…,N) to establish a high-dimensional feature space In formula 7, δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; ρ is the deviation; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces; S5. According to the principle of structural risk minimization, the regression problem is transformed into a constrained optimization problem In formula 8, J is the loss function; ρ is the deviation; e is the speed error, in m / s; δ k and Φ k are the sample data of slip rate and traction coefficient respectively; ω is the weight vector; γ is the adjustable constant; e k is the error, e=[e1,…,e N ] T ∈R n is the error vector; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces; S6. Define the Lagrangian function based on the optimization function In formula 9, J L is the Lagrangian function; ω is the weight vector; ρ is the deviation; e is the speed error, in m / s; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n Φ k is the sample data of traction coefficient; e k is the error; φ(δ k ) means that δ k Functions that map to high-dimensional feature spaces; S7. Optimize and solve to obtain the dual matrix equation In formula 10, δ i and δ j is the slip rate sample data; σ is a constant; γ is an adjustable constant; ρ is the deviation; α k ∈R is the Lagrange multiplier; Φ k is the corresponding traction coefficient; S8, obtaining the values ​​of parameters α and ρ, thereby obtaining an online identified wheel-soil model; In formula 11, Φ is the corresponding traction coefficient; α k ∈R is the Lagrange multiplier, α=[α1,…,α N ] T ∈R n ;δ is the measured slip rate;δ k is the sample data of slip rate; ρ is the deviation.

3. The online identification method for a tractor wheel-soil model according to claim 2, characterized in that: In step S1, the rolling radius r of each wheel j Obtained by measurement; longitudinal vehicle speed estimate Calculated by formula 1; In formula 1, is the estimated longitudinal vehicle speed in m / s; is the longitudinal velocity in m / s; Current acceleration in m / s 2 ;τ represents the sampling period of the inertial measurement unit, in seconds; N=(t k -t k-1 ) / τ-1.

Images