A method for estimating clamping force in an electric vehicle EMB system without pressure sensors

Abstract

This invention discloses a method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless system, relating to the field of distributed electric drive drive-by-wire vehicle technology. Considering the viscous effect between the brake disc and friction pads during actual braking, a dynamic stiffness model of the clamping force is established to predict the clamping force. The descriptive equation of the dynamic stiffness model is discretized, and a discretized state-space equation for the clamping force is established. Based on the motor torque balance equation, a sliding mode observer for the motor load torque is designed to observe the motor load torque in real time during braking. According to the transmission model of the EMB system, the load torque observation value observed in real time by the sliding mode observer is converted into the clamping force observation value. Using a Kalman filter algorithm, combined with the discretized clamping force state-space equation and the clamping force observation value, and based on the optimal clamping force estimate obtained at the previous moment, the optimal clamping force estimate for the current moment is calculated, achieving real-time optimal estimation of the clamping force. This invention solves the problem that under pressure sensor control, the high temperature and high pressure during braking cause a decrease in the measurement accuracy of the pressure sensor or even its overall failure.

Description

Technical Field

[0001] This invention relates to the field of distributed electric drive drive-by-wire vehicle technology, and in particular to a method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless system. Background Technology

[0002] Electromechanical braking (EMB) is a relatively new braking system introduced in recent years for electric vehicles. Compared to traditional hydraulic braking, EMB uses a fully drive-by-wire approach, replacing the mechanical system with drive-by-wire, saving significant space and reducing weight. In terms of response speed, EMB employs intelligent drive-by-wire, eliminating the delay of hydraulic braking systems and greatly shortening braking response time. Finally, through relevant control algorithms, EMB can precisely achieve the target clamping force, resulting in precise braking.

[0003] However, during actual braking, the friction pads press against the brake disc, generating a large amount of heat. Pressure sensors operating in this high-temperature environment are highly susceptible to reduced measurement accuracy. Simultaneously, the pressure sensors installed within the brake disc are subjected to immense compressive force, making them prone to failure after prolonged use. When pressure sensors experience accuracy issues or even fail, the braking force may fall short of the required level, significantly impacting personal safety.

[0004] To avoid the pressure sensor in the EMB system from becoming inaccurate or even failing, thus failing to provide the required clamping force to the electric vehicle, it is an urgent problem to be solved by adopting a pressure sensorless control strategy to estimate the magnitude of the generated clamping force in real time and replace the pressure sensor. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides a method for estimating clamping force of electric vehicle EMB system based on pressure sensorless, which solves the serious safety problems caused by the high cost of pressure sensors in the prior art and their tendency to decrease in accuracy or even fail completely in harsh braking environments.

[0006] To achieve the above objectives, the present invention adopts the following technical solution, including:

[0007] A method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless system includes the following steps:

[0008] S1. Considering the viscous effect between the brake disc and the friction pad during actual braking, a dynamic stiffness model of clamping force is established to predict the clamping force.

[0009] S2, discretize the description equation of the clamping force dynamic stiffness model and establish the discretized clamping force state space equation;

[0010] S3. Based on the motor torque balance equation, a sliding mode observer for the motor load torque is designed to observe the motor load torque in real time during braking and obtain the observed load torque value.

[0011] S4. Based on the transmission model of the electromechanical braking system, i.e., the EMB system, the load torque observation value observed in real time by the sliding mode observer is converted into the clamping force observation value.

[0012] S5 uses the Kalman filter algorithm, combining the discretized clamping force state space equation and clamping force observations, and calculates the optimal clamping force estimate for the current moment based on the optimal clamping force estimate obtained at the previous moment, thus achieving real-time optimal estimation of the clamping force.

[0013] Preferably, in step S1, the dynamic stiffness model of the clamping force is:

[0014]

[0015] In the formula, F cl1 θ represents the predicted clamping force. m For the motor rotation angle, Λ3, Λ2, Λ1, and Λ0 represent the differential of the clamping force prediction with respect to time; Λ3, Λ2, Λ1, and Λ0 are the motor rotation angle constants; τ is the time constant.

[0016] Preferably, in step S2, the discretized state-space equation of the clamping force is:

[0017]

[0018] In the formula, F cl1 (k), F cl1 (k-1) represent the predicted clamping force values ​​at time k and k-1, respectively; α represents the clamping force prediction coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants; θ m (k) represents the motor rotation angle at time k.

[0019] Preferably, in step S3, the sliding mode observer for the load torque is:

[0020]

[0021] in,

[0022] In the formula, This refers to the observed angular velocity of the motor. Observed values ​​of motor angular velocity Differential with respect to time; This is the observed value of the load torque; For load torque observations The derivative with respect to time; g is the feedback gain of the sliding mode observer; U is the input control variable; ε is the gain of the sign function sgn(·) in U; ​​J represents the moment of inertia of the motor; p n ψ represents the number of pole pairs of the motor; f Indicates permanent magnet flux linkage; i q This represents the q-axis current.

[0023] Preferably, in step S3, the design of the sliding mode observer for the load torque includes the following steps:

[0024] S31, the motor torque balance equation is:

[0025]

[0026] In the formula, J represents the moment of inertia of the motor; T represents the derivative of the motor's angular velocity with respect to time. L B represents the load torque; ω represents the damping coefficient. m T represents the angular velocity of the motor. e Indicates electromagnetic torque;

[0027] in,

[0028]

[0029] In the formula, p n ψ represents the number of pole pairs of the motor; f Indicates permanent magnet flux linkage; i q Represents the q-axis current;

[0030] S32, with the motor angular velocity ω m With load torque T L As a state variable, since the actual sampling frequency of the controller is much higher than the frequency of load torque changes, its differential component can be considered as a constant of 0, resulting in the state equation:

[0031]

[0032] S33, the observed value of the motor angular velocity The actual value of the motor's angular velocity ω m The difference is taken as the sliding surface S, denoted as...

[0033] S34, the approach law of the sliding mode observer is chosen to be the constant-velocity approach law:

[0034]

[0035] In the formula, γ is the variable of the sliding surface; The gain of the sign function sgn(·); sgn(·) is the sign function;

[0036] S35, the motor angular velocity ω m With load torque T L As the object of observation, establish a sliding mode observer:

[0037]

[0038] in,

[0039] In the formula, Observed values ​​of motor angular velocity Differential with respect to time; This is the observed value of the load torque; For load torque observations The derivative with respect to time; g is the feedback gain of the sliding mode observer; U is the input control variable; ε is the gain of the sign function sgn(·) in U;

[0040] S36. Subtract the formulas from steps S35 and S32 to obtain the error equation for the sliding mode observer:

[0041]

[0042] In the formula, e1 represents the observation error of the motor angular velocity; e1 represents the derivative of the observation error of the motor angular velocity with respect to time; e2 represents the observation error of the load torque. The observation error e2 of the load torque is represented as a derivative with respect to time.

[0043] S37, based on the Lyapunov stability principle, the Lyapunov function is chosen as... Its derivative must satisfy Right now

[0044] The range of values ​​for the gain ε is obtained as follows:

[0045] In the formula, V is a Lyapunov function variable, For the Lyapunov function variable, differentiate it with respect to time;

[0046] S38, when the sliding surface variable S reaches the sliding surface S = 0, it satisfies Right now The expression in step S36 can then be simplified to:

[0047]

[0048] According to the stability criterion, if e2 is to approach 0, the range of the gain g is g < 0.

[0049] S39. Based on the value range of gain ε and g, configure appropriate gain ε and g, and based on the sliding mode observer established in step S35, observe the load torque value in real time under braking conditions.

[0050] Preferably, in step S4, the planetary gear transmission ratio i of the mechanical transmission structure is determined. g Ball screw lead L0, planetary gear transmission efficiency η p Ball screw transmission efficiency η s The load torque observation value is obtained by the sliding mode observer in real time. Converted to clamping force observation value F cl2 :

[0051]

[0052] Preferably, in step S5, the real-time optimal estimation of the clamping force includes the following steps:

[0053] S51, Establish the state-space equations:

[0054]

[0055] In the formula, X(k-1) and X(k) represent the state vectors at time k-1 and time k, respectively; U(k) represents the input vector at time k; A represents the state transition matrix; B represents the control input matrix; w(k-1) represents the process noise matrix at time k-1; Z(k) represents the observation vector at time k; H represents the observation matrix; and v(k) represents the observation noise matrix at time k.

[0056] S52, based on the discretized state-space equation of clamping force, we obtain:

[0057]

[0058] In the formula, F cl1 (k-1) represents the predicted clamping force at time k-1 obtained from the discretized clamping force state-space equation; α represents the clamping force coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants; θ m (k) represents the motor rotation angle at time k.

[0059]

[0060] In the formula, I represents the identity matrix, and F cl2 (k) represents the clamping force observation at time k, which is converted from the load torque observation value observed by the sliding mode observer;

[0061] S53, Based on the Kalman filter algorithm, the predicted clamping force value and the observed clamping force value are fused and estimated, as shown below:

[0062] S531, performing state prediction for the EMB system:

[0063] Calculate the prior estimate of the EMB system state at time k:

[0064] X(k|k-1)=AX(k-1|k-1)+BU(k);

[0065] In the formula, X(k|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1) represents the prior state estimate of the clamping force at time k; X(k-1|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1);

[0066] Calculate the prior estimate of the covariance of the EMB system at time k:

[0067] P(k|k-1)=AP(k-1|k-1)A T +Q;

[0068] In the formula, P(k|k-1) represents the prior state estimate of the covariance of the EMB system at time k using the optimal covariance estimate P(k-1|k-1) of the EMB system at time k-1; P(k-1|k-1) represents the optimal covariance estimate of the EMB system at time k-1; A T represents the transpose of the state transition matrix; Q represents the process noise covariance matrix;

[0069] S532, Measurement update for EMB system:

[0070] Calculate the Kalman gain of the EMB system:

[0071] K(k)=[P(k|k-1)H T ]·[HP(k|k-1)H T +R] -1 ;

[0072] In the formula, K(k) represents the Kalman gain of the EMB system at time k; H T R represents the transpose of the observation matrix; R represents the measurement noise covariance matrix.

[0073] Based on the clamping force observation Z(k), update the state estimate of the EMB system:

[0074] X(k|k)=X(k|k-1)+K(k)[Z(k)-HX(k|k-1)];

[0075] In the formula, X(k|k) represents the optimal clamping force estimate F at time k. cl (k);

[0076] Update the EMB system covariance matrix:

[0077] P(k|k)=(IK(k)H)·P(k|k-1);

[0078] In the formula, P(k|k) represents the estimated value of the covariance matrix of the EMB system at time k.

[0079] A readable storage medium having a computer program stored thereon, which, when executed, implements the above-described method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless design.

[0080] An electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless system.

[0081] A computer program product comprising a computer program / instructions that, when executed by a processor, implement the aforementioned method for estimating clamping force in an electric vehicle EMB system without pressure sensors.

[0082] The advantages of this invention are:

[0083] (1) Compared with the prior art, the present invention uses the Kalman filter algorithm, combining the clamping force prediction value obtained from the spatial state equation of the clamping force dynamic stiffness model and the clamping force observation value obtained from the sliding mode observer, and obtains the optimal clamping force estimate through real-time update iteration. The present invention solves the problem that under extreme braking conditions, the pressure sensor is in a high temperature and high pressure environment and the number of uses increases, resulting in a decrease in the accuracy of clamping force observation, and in severe cases, even the entire sensor fails, thus affecting safety.

[0084] (2) This invention takes into account the viscous effect between the friction plate and the brake disc, and establishes a dynamic stiffness model of the clamping force; the established dynamic stiffness model of the clamping force is transformed into a discretized state-space equation of the clamping force; a sliding mode observer for the motor torque is established, which receives the magnitude i of the q-axis current fed back from the sensor. q θ, the angular position of the motor rotor m and the angular velocity ω of the motor rotor m To obtain real-time observations of the external load torque of the motor. By using the transmission mechanism parameters of the EMB system, the load torque observation value is... The values ​​are converted into clamping force observations. Finally, using a Kalman filter algorithm, the clamping force state space equation and the clamping force observations are updated and iterated in real time to obtain the optimal clamping force estimate at different times. This invention, through the Kalman filter algorithm, takes into account external noise and updates and iterates in real time to obtain the optimal clamping force value, greatly improving the accuracy of clamping force estimation and effectively preventing the inaccuracy caused by a single model due to noise. It can replace pressure sensors. Attached Figure Description

[0085] Figure 1 This is a flowchart of the method of the present invention.

[0086] Figure 2 This is a schematic diagram of the experimental results of the present invention. Detailed Implementation

[0087] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0088] Depend on Figure 1 As shown, the present invention provides a method for estimating the clamping force of an electric vehicle EMB system based on a pressure sensorless system, comprising the following steps:

[0089] S1, Establishment of the dynamic stiffness model of clamping force: In the actual braking process, the clamping force of the caliper and the friction plate at the same motor rotation angle is not the same during the clamping-release process. Therefore, considering the viscous effect of the brake disc during actual braking, the traditional clamping force model based on the motor rotation angle polynomial is improved and a new dynamic stiffness model of clamping force is established.

[0090] During braking, the motor drives the planetary gears in the electromechanical braking (EMB) system to reduce speed and increase torque. The planetary gears drive the ball screw to convert the rotational motion into linear motion, pushing the friction pads to press against the brake disc for deceleration. When releasing the brake, the motor releases the friction pads by reversing its direction.

[0091] Due to the viscous effect between the friction pads and the brake disc during braking, the clamping force at the same motor rotation angle will vary during the clamping (or releasing) of the friction pads.

[0092] Traditional clamping force estimation involves measuring the relationship between clamping force and motor rotation angle, expressing the clamping force as a polynomial of the motor rotation angle, and finally estimating the clamping force based on the real-time motor rotation angle. However, this method ignores the error caused by the viscous effect, leading to a significant discrepancy between the estimated clamping force and the actual clamping force when clamping (or releasing) the friction pad.

[0093] Therefore, this invention establishes a more realistic dynamic stiffness model for the clamping force of electromechanical braking (EMB) systems, and estimates the clamping force as follows:

[0094] S11, considering the ideal situation where the clamping force is linearly related to the motor rotation angle, and taking into account the hysteresis of the clamping force caused by the viscosity effect, the transfer function of the EMB system can be expressed as:

[0095]

[0096] In the formula, G(s) is the transfer function, F cl1 (s), θ m (s) represent the predicted clamping force and the motor rotation angle in the complex domain, respectively, and K s τ is the linear relationship coefficient between clamping force and motor rotation angle, τ is the time constant, and s is a complex number.

[0097] S12, transforming the transfer function into its real-number representation, yields the expression for the clamping force as follows:

[0098]

[0099] In the formula, F cl1 θ represents the predicted clamping force. m For the motor rotation angle, This represents the derivative of the clamping force prediction with respect to time.

[0100] S13, In practice, clamping force is usually expressed as a polynomial of motor rotation angle, which yields a more accurate expression of clamping force, namely the clamping force dynamic stiffness model:

[0101]

[0102] In the formula, Λ3, Λ2, Λ1, and Λ0 are the motor rotation constants to be determined experimentally.

[0103] S2, Transformation of State-Space Equations: The describing equations of the clamping force dynamic stiffness model considering viscous effects are transformed into discretized clamping force state-space equations. Specifically, as shown below:

[0104] The discrete state-space equations are expressed as follows:

[0105] X(k) = AX(k-1) + BU(k);

[0106] In the formula, X(k-1) and X(k) represent the system state vectors at time k-1 and time k, respectively; U(k) represents the system input vector at time k; A represents the state transition matrix; and B represents the control input matrix.

[0107] Based on the discrete state-space equations described above, the clamping force expression obtained in step S13 is transformed into a discrete form. The discretized clamping force state-space equation is as follows:

[0108]

[0109] In the formula, F cl1 (k), F cl1 (k-1) represent the predicted clamping force values ​​at time k and k-1, respectively; α represents the clamping force prediction coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants to be determined experimentally; θ m (k) represents the motor rotation angle at time k.

[0110] Where X(k)=F cl1 (k); X(k-1)=F cl1 (k-1); A = α;

[0111]

[0112] S3, Construction of the sliding mode observer: Based on the motor torque balance equation, a sliding mode control algorithm is designed. Under braking conditions, the algorithm is based on the q-axis current i measured in real time by the sensor. q Motor rotation angle (motor rotor angular position) θ m Angular velocity ω of the motor (rotor) m The value is used to calculate the external load torque on the drive motor in real time. The details are as follows:

[0113] S31, the torque balance equation of the motor is expressed as:

[0114]

[0115] In the formula, J represents the moment of inertia of the motor; Represents the angular velocity ω of the motor m Differential with respect to time; T L B represents the load torque; ω represents the damping coefficient. m T represents the angular velocity of the motor. e Indicates electromagnetic torque;

[0116] The selected drive motor is a surface-mounted permanent magnet synchronous motor, using a d-axis current i d =0 control mode, its electromagnetic torque T e for:

[0117]

[0118] In the formula, p n ψ represents the number of pole pairs of the motor; f Indicates permanent magnet flux linkage; i q This represents the q-axis current.

[0119] S32, with the motor angular velocity ω m With load torque T L As a state variable, since the actual sampling frequency of the controller is much higher than the frequency of load torque changes, its differential component can be considered as a constant of 0, resulting in the state equation:

[0120]

[0121] In the formula, since the actual sampling frequency of the controller is much higher than the load torque T L The frequency of change, and therefore its differential components It can be considered as a constant 0.

[0122] S33, the observed value of the motor angular velocity The actual value of the motor's angular velocity ω m The difference is taken as the sliding surface S, which is expressed as:

[0123] S34, the approach law of the sliding mode observer is chosen to be the constant-velocity approach law:

[0124]

[0125] In the formula, γ is the variable of the sliding surface; The gain of the sign function sgn(·); sgn(·) is the sign function.

[0126] S35, the motor angular velocity ω m With load torque T L As the object of observation, establish a sliding mode observer:

[0127]

[0128] in,

[0129] In the formula, Observed values ​​of motor angular velocity Differential with respect to time; This is the observed value of the load torque; For load torque observations The derivative with respect to time; g is the feedback gain of the observer; U is the input control variable; ε is the gain of the sign function sgn(·) in U.

[0130] S36. Subtract the two equations from steps S35 and S32 to obtain the error equation for the sliding mode observer:

[0131]

[0132] In the formula, e1 represents the observation error of the motor angular velocity; e2 represents the derivative of the motor angular velocity observation error with respect to time; e2 represents the load torque observation error. This represents the derivative of the load torque observation error with respect to time.

[0133] S37, based on the Lyapunov stability principle, the Lyapunov function is chosen as... Its derivative must satisfy Right now:

[0134] The range of values ​​for the sliding mode gain ε is obtained as follows:

[0135] In the formula, V is a Lyapunov function variable, Let be the derivative of the Lyapunov function variable with respect to time.

[0136] S38, when the observed quantity follows the sliding surface, satisfies Right now The expression in step S36 can then be simplified to:

[0137]

[0138] Substituting into the first equation of step S36, we obtain the relationship between U and e2. Then, by representing U in the second equation of step S36 with e2, we can obtain the above equation.

[0139] According to the stability criterion, if e2 is to approach 0, the range of the gain g is g < 0.

[0140] S39. Based on the value range of gain ε and g, configure appropriate gain ε and g. Based on the sliding mode observer established in step S35, observe the observed value of motor load torque in real time under braking conditions.

[0141] S4, Conversion between load torque observations and clamping force observations: Based on the transmission model of the EMB system, the conversion is made according to the reduction ratio and transmission efficiency of the planetary gears and ball screws connected to the external rotating shaft of the motor, as well as the load torque observations obtained from the sliding mode observer. Real-time calculation of clamping force observation value F cl2 The details are as follows:

[0142] The mechanical transmission structure adopted is a motor shaft-planetary gear-ball screw structure. Based on the planetary gear transmission ratio i of the mechanical transmission structure... g Ball screw lead L0, planetary gear transmission efficiency η p Ball screw transmission efficiency η s The load torque observation value observed in real time by the sliding mode observer can be used. Converted to clamping force observation value F cl2 :

[0143]

[0144] S5. Optimal clamping force is derived through data fusion based on Kalman filtering: The Kalman filtering algorithm is used to combine the predicted and observed clamping forces, and based on the optimal clamping force value obtained at the previous moment, to calculate the estimated optimal clamping force for the current moment. Details are as follows:

[0145] S51, Establish the state-space equations:

[0146]

[0147] In the formula, X(k-1) and X(k) represent the state vectors at time k-1 and time k, respectively; U(k) represents the input vector at time k; A represents the state transition matrix; B represents the control input matrix; w(k-1) represents the process noise matrix at time k-1; Z(k) represents the observation vector at time k; H represents the observation matrix; and v(k) represents the observation noise matrix at time k.

[0148] S52, In order to achieve the fusion of clamping forces, the state-space equations are rewritten to obtain the following state-space equations:

[0149]

[0150] In the formula, F cl1 (k-1) represents the predicted clamping force at time k-1 obtained from the discretized clamping force state-space equation; α represents the clamping force coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants to be determined experimentally; θ m (k) represents the motor rotation angle at time k.

[0151]

[0152] In the formula, I represents the identity matrix, and F cl2 (k) represents the clamping force observation at time k, which is converted from the load torque observation value observed by the sliding mode observer.

[0153] S53. After obtaining the predicted and observed values ​​of the clamping force, the two are fused and estimated according to the Kalman filter theory, as shown below:

[0154] S531, performing state prediction for the EMB system:

[0155] Calculate the prior estimate of the EMB system state at time k:

[0156] X(k|k-1)=AX(k-1|k-1)+BU(k);

[0157] In the formula, X(k|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1) represents the prior state estimate of the clamping force at time k; X(k-1|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1).

[0158] Calculate the prior estimate of the covariance of the EMB system at time k:

[0159] P(k|k-1)=AP(k-1|k-1)A T +Q;

[0160] In the formula, P(k|k-1) represents the prior state estimate of the covariance of the EMB system at time k using the optimal covariance estimate P(k-1|k-1) of the EMB system at time k-1; P(k-1|k-1) represents the optimal covariance estimate of the EMB system at time k-1; A T represents the transpose of the state transition matrix; Q represents the process noise covariance matrix, whose parameters need to be configured according to the actual situation.

[0161] S532, Measurement update for EMB system:

[0162] Calculate the Kalman gain of the EMB system:

[0163] K(k)=[P(k|k-1)H T ]·[HP(k|k-1)H T +R] -1 ;

[0164] In the formula, K(k) represents the Kalman gain of the EMB system at time k; H T R represents the transpose of the observation matrix; R represents the measurement noise covariance matrix, whose parameters need to be configured according to the actual situation.

[0165] Based on the observations, update the state estimate of the EMB system:

[0166] X(k|k)=X(k|k-1)+K(k)[Z(k)-HX(k|k-1)];

[0167] In the formula, X(k|k) represents the optimal clamping force estimate F at time k. cl (k).

[0168] Update the covariance estimate of the EMB system:

[0169] P(k|k)=(IK(k)H)·P(k|k-1);

[0170] In the formula, P(k|k) represents the optimal covariance estimate of the EMB system at time k.

[0171] S54, When the electric vehicle enters braking mode, the ECU (Electronic Control Unit) receives a signal that the brake pedal has been depressed, and calculates the reference value F of the clamping force required for braking in real time based on the degree to which the brake pedal is depressed. ref Furthermore, it sends a command to the drive motor of the EMB system to enter closed-loop operating mode. When entering the pressure tracking phase, the ECU uses pre-written code to calculate the clamping force F at time k, which is derived from the load torque observation value observed by the sliding mode observer. cl2 (k) and by incorporating the Kalman filter algorithm code into the ECU, the optimal clamping force fusion estimate at different times is obtained through continuous iteration. This estimate is then used as the feedback of the force closed loop to replace the clamping force measured by the pressure sensor, thus realizing a pressure sensorless control strategy.

[0172] Figure 2 The invention demonstrates its practical effectiveness. Under different clamping force requirements, the invention reduces the influence of external noise on clamping force estimation by configuring the noise figure, and can replace the pressure sensor to achieve accurate estimation of clamping force.

[0173] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for estimating clamping force in an electric vehicle EMB system based on a pressure sensorless system, characterized in that, Includes the following steps: S1. Considering the viscous effect between the brake disc and the friction pad during actual braking, a dynamic stiffness model of clamping force is established to predict the clamping force. S2, discretize the description equation of the clamping force dynamic stiffness model and establish the discretized clamping force state space equation; S3. Based on the motor torque balance equation, a sliding mode observer for the motor load torque is designed to observe the motor load torque in real time during braking and obtain the observed load torque value. S4. Based on the transmission model of the electromechanical braking system, i.e., the EMB system, the load torque observation value observed in real time by the sliding mode observer is converted into the clamping force observation value. S5 uses the Kalman filter algorithm, combining the discretized clamping force state space equation and clamping force observations, and calculates the optimal clamping force estimate for the current moment based on the optimal clamping force estimate obtained at the previous moment, thus achieving real-time optimal estimation of the clamping force.

2. The method for estimating clamping force of an electric vehicle EMB system based on a pressure sensorless system according to claim 1, characterized in that, In step S1, the dynamic stiffness model of the clamping force is: In the formula, F cl1 θ represents the predicted clamping force. m For the motor rotation angle, Λ3, Λ2, Λ1, and Λ0 represent the differential of the clamping force prediction with respect to time; Λ3, Λ2, Λ1, and Λ0 are the motor rotation angle constants; τ is the time constant.

3. The method for estimating clamping force of an electric vehicle EMB system based on a pressure sensorless system according to claim 1, characterized in that, In step S2, the discretized state-space equation for the clamping force is: In the formula, F cl1 (k), F cl1 (k-1) represent the predicted clamping force values ​​at time k and k-1, respectively; α represents the clamping force prediction coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants; θ m (k) represents the motor rotation angle at time k.

4. In the method for estimating clamping force of an electric vehicle EMB system based on a pressure sensorless system according to claim 1, in step S3, the sliding mode observer for the load torque is: in, In the formula, This refers to the observed angular velocity of the motor. Observed values ​​of motor angular velocity Differential with respect to time; This is the observed value of the load torque; For load torque observations The derivative with respect to time; g is the feedback gain of the sliding mode observer; U is the input control variable; ε is the gain of the sign function sgn(·) in U; ​​J represents the moment of inertia of the motor; p n ψ represents the number of pole pairs of the motor; f Indicates permanent magnet flux linkage; i q This represents the q-axis current.

5. In the method for estimating clamping force of an electric vehicle EMB system based on a pressure sensor-free method according to claim 1, step S3, the design of the sliding mode observer for the load torque includes the following steps: S31, the motor torque balance equation is: In the formula, J represents the moment of inertia of the motor; T represents the derivative of the motor's angular velocity with respect to time. L B represents the load torque; ω represents the damping coefficient. m T represents the angular velocity of the motor. e Indicates electromagnetic torque; in, In the formula, p n ψ represents the number of pole pairs of the motor; f Indicates permanent magnet flux linkage; i q Represents the q-axis current; S32, with the motor angular velocity ω m With load torque T L As state variables, we obtain the state equation: S33, the observed value of the motor angular velocity The actual value of the motor's angular velocity ω m The difference is taken as the sliding surface S, denoted as... S34, the approach law of the sliding mode observer is chosen to be the constant-velocity approach law: In the formula, γ is the variable of the sliding surface; The gain of the sign function sgn(·); sgn(·) is the sign function; S35, the motor angular velocity ω m With load torque T L As the object of observation, establish a sliding mode observer: in, In the formula, Observed values ​​of motor angular velocity Differential with respect to time; This is the observed value of the load torque; For load torque observations The derivative with respect to time; g is the feedback gain of the sliding mode observer; U is the input control variable; ε is the gain of the sign function sgn(·) in U; S36. Subtract the formulas from steps S35 and S32 to obtain the error equation for the sliding mode observer: In the formula, e1 represents the observation error of the motor angular velocity; e1 represents the derivative of the observation error of the motor angular velocity with respect to time; e2 represents the observation error of the load torque. The observation error e2 of the load torque is represented as a derivative with respect to time. S37, based on the Lyapunov stability principle, the Lyapunov function is chosen as... Its derivative must satisfy Right now The range of values ​​for the gain ε is obtained as follows: In the formula, V is a Lyapunov function variable, For the Lyapunov function variable, differentiate it with respect to time; S38, when the sliding surface variable S reaches the sliding surface S = 0, it satisfies Right now Then the expression in step S36 can be simplified to: According to the stability criterion, if e2 is to approach 0, the range of the gain g is g < 0. S39. Based on the value range of gain ε and g, configure appropriate gain ε and g, and based on the sliding mode observer established in step S35, observe the load torque value in real time under braking conditions.

6. The method for estimating clamping force of an electric vehicle EMB system based on a pressure sensorless system according to claim 1, characterized in that, In step S4, based on the planetary gear transmission ratio i of the mechanical transmission structure... g Ball screw lead L0, planetary gear transmission efficiency η p Ball screw transmission efficiency η s The load torque observation value is obtained by the sliding mode observer in real time. Converted to clamping force observation value F cl2 :

7. The method for estimating clamping force of an electric vehicle EMB system based on a pressure sensorless system according to claim 1, characterized in that, In step S5, the real-time optimal estimation of the clamping force includes the following steps: S51, Establish the state-space equations: In the formula, X(k-1) and X(k) represent the state vectors at time k-1 and time k, respectively; U(k) represents the input vector at time k; A represents the state transition matrix; B represents the control input matrix; w(k-1) represents the process noise matrix at time k-1; Z(k) represents the observation vector at time k; H represents the observation matrix; and v(k) represents the observation noise matrix at time k. S52, based on the discretized state-space equation of clamping force, we obtain: In the formula, F cl1 (k-1) represents the predicted clamping force at time k-1 obtained from the discretized clamping force state-space equation; α represents the clamping force coefficient; α3, α2, α1, and α0 represent the discretized motor rotation constants; θ m (k) represents the motor rotation angle at time k; In the formula, I represents the identity matrix, and F cl2 (k) represents the clamping force observation at time k, which is converted from the load torque observation value observed by the sliding mode observer; S53, Based on the Kalman filter algorithm, the predicted clamping force value and the observed clamping force value are fused and estimated, as shown below: S531, performing state prediction for the EMB system: Calculate the prior estimate of the EMB system state at time k: X(k|k-1)=AX(k-1|k-1)+BU(k); In the formula, X(k|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1) represents the prior state estimate of the clamping force at time k; X(k-1|k-1) represents the optimal clamping force estimate F at time k-1. cl (k-1); Calculate the prior estimate of the covariance of the EMB system at time k: P(k|k-1)=AP(k-1|k-1)A T +Q; In the formula, P(k|k-1) represents the prior state estimate of the covariance of the EMB system at time k using the optimal covariance estimate P(k-1|k-1) of the EMB system at time k-1; P(k-1|k-1) represents the optimal covariance estimate of the EMB system at time k-1; A T represents the transpose of the state transition matrix; Q represents the process noise covariance matrix; S532, Measurement update for EMB system: Calculate the Kalman gain of the EMB system: K(k)=[P(k|k-1)H T ]·[HP(k|k-1)H T +R] -1 ; In the formula, K(k) represents the Kalman gain of the EMB system at time k; H T R represents the transpose of the observation matrix; R represents the measurement noise covariance matrix. Based on the clamping force observation Z(k), update the state estimate of the EMB system: X(k|k)=X(k|k-1)+K(k)[Z(k)-HX(k|k-1)]; In the formula, X(k|k) represents the optimal clamping force estimate F at time k. cl (k); Update the EMB system covariance matrix: P(k|k)=(IK(k)H)·P(k|k-1); In the formula, P(k|k) represents the estimated value of the covariance matrix of the EMB system at time k.

8. A readable storage medium, characterized in that, It stores a computer program, which, when executed, implements the clamping force estimation method for an electric vehicle EMB system based on a pressure sensorless system as described in any one of claims 1 to 6.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the clamping force estimation method for an electric vehicle EMB system based on any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes a computer program / instruction that, when executed by a processor, implements the clamping force estimation method for an electric vehicle EMB system based on any one of claims 1 to 6.

Images