Safety fault-tolerant control method and device for linear motor motion unit and medium

Abstract

The invention discloses a safety fault-tolerant control method and device for a linear motor motion unit and a medium, and relates to the technical field of motor control. The invention aims to solve the problem that a safe and reliable protection mechanism in a multi-constraint state of a system cannot be effectively established due to the fact that an existing fault-tolerant control method for the linear motor does not involve all working conditions. According to the invention, the controller parameters are rapidly evaluated and adjusted to adapt to the ill-conditioned environment. A disturbance observer with finite time convergence is considered to enhance system stability. A fault-tolerant controller is designed based on a specified performance control framework to ensure that the system performance meets an expected envelope. And asymptotically stable tracking of the system can be ensured. The method is suitable for the field of motion control of the linear motor unit in mechanical-electrical integration equipment, and is especially suitable for the field of safety fault-tolerant control under an unhealthy constraint working condition.

Description

Technical Field

[0001] The invention belongs to the technical field of motor control. Background Art

[0002] As a new generation of drive units, linear servo motors have the characteristics of high precision, small gap, and fast response compared to traditional screws, lead screws or gear transmissions. They are widely used in refined fields such as semiconductor assembly, CNC machining, and medical testing. In recent years, industry and academia have successively carried out in-depth research to meet the control needs of actual production. It is worth noting that the electromechanical physical system-linear motor will inevitably suffer from the risk of failure or constraint when it runs for a long time under various complex working conditions. If proper preparations are not made in advance, minor faults may also lead to mission failure and even cause serious safety accidents. Therefore, the study of safe fault-tolerant control of linear servo units has important practical significance and application value.

[0003] Linear motors achieve higher execution efficiency through direct drive, but they are easily disturbed by external disturbances during movement because there is no buffer transfer structure. In addition, hidden dangers such as actuator saturation or failure exist, which makes it difficult to guarantee the steady-state and transient performance indicators of the system when various crisis conditions occur. Existing research results on linear motors are mostly based on precision motion control under healthy conditions, while the related fault-tolerant control rarely involves various complex conditions such as actuator saturation, failure, performance preset, etc., resulting in the inability to effectively establish a safe and reliable protection mechanism under the multi-constraint state of the system. Summary of the invention

[0004] The present invention aims to solve the problem that the existing linear motor fault-tolerant control methods rarely involve operating conditions such as actuator saturation, failure, performance preset, etc., resulting in the inability to effectively establish a safe and reliable protection mechanism under the system's multi-constraint state. A safe fault-tolerant control method, equipment and medium for a linear motor motion unit are now provided to improve the processing quality and safety assurance factor of production and manufacturing.

[0005] A safe fault-tolerant control method for a linear motor motion unit, comprising:

[0006] Design adaptive control law υ a , anti-saturation control term υ c , robust feedback term υ r and error compensation term υ n , and a ,υ c ,υ r and n Add together to obtain the safe fault-tolerant control input of the linear motor,

[0007] The adaptive control law v a , anti-saturation control term υc , robust feedback term υ r and error compensation term υ n The expressions are:

[0008]

[0009] Among them, ψ is the regressor before filtering, is the estimated value of the parameterized vector θ of the linear motor mathematical model. The subscripts 2-4 represent the indexes from the 2nd to the 4th elements. d is the expected reference input of the linear motor mathematical model, c1 and c2 are both positive constants, ξ1 and ξ2 represent the state variables of the linear motor system and have initial values ​​ξ1(0)=ξ2(0)=0, k r is the robust gain constant, σ is a quasi-sliding mode variable, is the finite time neural network observation quantity, the intermediate quantity The expressions of and r are:

[0010]

[0011] Λ>0 is the gain constant, e is the state error of the linear motor system,

[0012]

[0013] ζ is a scalar that controls the error convergence speed, t represents time, and are the maximum overshoot and steady-state value respectively, and They are all positive numbers, e0 is the initial value of e.

[0014] Furthermore, the mathematical model of the linear motor under actuator failure and saturation is:

[0015]

[0016] Among them, x1, x2 and are the position, velocity and acceleration of the linear motor system respectively, h(υ) is a smooth function that approximates the standard saturation function sat(υ), S f (·) is a nonlinear smooth function related to velocity, M, K, B, and A represent the inertial mass, voltage-torque gain coefficient, viscous friction coefficient, and Coulomb friction coefficient of the motion unit, respectively. η is the effective factor. is the unknown constant parameter term decomposed from the lumped error d of the linear motor system, is the unknown uncertainty term decomposed from the total error d of the linear motor system g(υ) is the saturation approximation error, u b is the bias fault constant.

[0017] Furthermore, the above-mentioned design of adaptive control law υ a ,include:

[0018] Design of saturation-type projective adaptive law

[0019]

[0020] in, is the upper bound of the adaptive learning law, is the standard adaptive projection operator, φ and τ represent the update rate and adaptive function respectively;

[0021] Based on the above saturation projection adaptive law Design model-compensated adaptive control lawυ a .

[0022] Furthermore, a recursive least squares estimation algorithm with a forgetting factor is designed to calculate φ and τ:

[0023]

[0024] Among them, l and are the forgetting factor and the standardization factor, respectively, M is the upper bound of the update rate φ; λ max ( ) and tr{} represent the maximum eigenvalue and trace of the parameter, respectively. is the estimation error, ψ f is the regressor after filtering.

[0025] Furthermore, the anti-saturation control term υ c ,include:

[0026] Constructing anti-saturation auxiliary system:

[0027]

[0028] Among them, Δυ=sat(υ)-υ, υ is the safety fault-tolerant control input of the linear motor;

[0029] When saturation does not occur, then:

[0030]

[0031] When saturation occurs, there is a Lyapunov function V0(t):

[0032]

[0033] Differentiate V0(t):

[0034]

[0035] in,

[0036] design Make the auxiliary system bounded;

[0037] Then we can obtain the anti-saturation controller υ c .

[0038] Furthermore, the above design robust feedback term υ r ,include:

[0039] The state error of the linear motor system is defined as:

[0040] e=x1-x d -ξ1,

[0041] The state error satisfies the constraint:

[0042]

[0043] The state error is used to design a sliding mode variable:

[0044]

[0045] in,

[0046] Design of Lyapunov function based on quasi-sliding mode variables:

[0047]

[0048] Substitute the linear motor mathematical model, anti-saturation auxiliary system, the second-order derivative of ε(t) and the sliding mode-like variable into the first-order derivative of V1(t) and simplify to obtain:

[0049]

[0050] in, is the estimated value of the model parameterization vector θ, is the estimation error;

[0051] Then the robust feedback term υ is designed r .

[0052] Furthermore, the above-mentioned design error compensation term υ n ,include:

[0053] Introducing the RBF neural network unit makes the optimal output of the neural network Δ, then:

[0054]

[0055] in, is the true value of the network weight, is the error between the network weight and the optimal weight, is the network output error;

[0056] The finite-time neural network observer is:

[0057]

[0058] Among them, m, c and β1 are all positive numbers and are greater than or equal to The maximum value of Parameters p and q satisfy 0<p<q; is the observation error and

[0059] Then the error compensation term υ is designed n .

[0060] Furthermore, the output of the kth neuron node in the above RBF neural network unit is for:

[0061]

[0062] Among them, x is the input of network training, c k and b k are the center vector and width of the activation function respectively.

[0063] A safety fault-tolerant control device for a linear motor motion unit, the safety fault-tolerant control device comprising a processor and a memory, the memory storing at least one instruction, the at least one instruction being loaded and executed by the processor to implement the safety fault-tolerant control method for a linear motor motion unit as described above.

[0064] A computer storage medium stores at least one instruction, which is loaded and executed by a processor to implement the above-mentioned safe fault-tolerant control method for a linear motor motion unit.

[0065] The present invention proposes a safe fault-tolerant control method for linear motor motion units, which solves the control problems of poor accuracy and low safety under extreme working conditions such as actuator failure, saturation and system performance constraints. Compared with the prior art, it has the following effects:

[0066] (1) Use a smooth function instead to avoid the risk of chattering caused by the saturation inflection point;

[0067] (2) An adaptive fault-tolerant control method is designed, which can effectively handle problems in the system, including actuator failure, input saturation, performance preset and nonlinear disturbance, and has more universal application value.

[0068] The present invention discloses a safe fault-tolerant control method for a linear motor motion unit. Different from the complex design process of existing fault-tolerant strategies based on bounded functions, the controller parameters are adjusted through rapid evaluation to adapt to pathological environments. In addition, a disturbance observer with finite-time convergence is considered to enhance system stability. In order to prevent the risk of system jitter caused by saturated inflection point output, a smooth surrogate function is cleverly designed. At the same time, a fault-tolerant controller is designed based on the specified performance control framework to ensure that the system performance meets the expected envelope. It is worth noting that the proposed control strategy can ensure asymptotically stable tracking of the system under the single or multiple constraints mentioned. Finally, the effectiveness of the proposed method was evaluated and verified by conducting a linear motor comparison experiment.

[0069] The present invention is applicable to the field of motion control of linear motor units in mechatronic equipment, and is particularly applicable to the field of safe fault-tolerant control under non-healthy constraint conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a safety fault-tolerant control flow chart for a linear motor motion unit;

[0071] Figure 2 The actual tracking position curve diagram described in the example;

[0072] Figure 3 The tracking error variation curve diagram described in the example;

[0073] Figure 4 This is a control output variation curve diagram described in the example. DETAILED DESCRIPTION

[0074] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0075] Specific implementation method 1: For linear motor systems with multiple constraints such as faults, saturation and performance presets, existing research is mostly based on precision control under healthy conditions, and rarely involves the above-mentioned fully covered complex scenarios. This implementation method proposes a simple, safe and reliable fault-tolerant control method, which can effectively handle control problems under single or multiple extreme conditions, and at the same time has a positive role in promoting the steady-state accuracy and dynamic response of the closed-loop system.

[0076] Reference Figures 1 to 4 Specifically describing this embodiment, the safety fault-tolerant control method for a linear motor motion unit described in this embodiment includes:

[0077] Step 1: Construct a mathematical model of a linear motor with actuator failure and saturation, including:

[0078] Define the position, velocity and acceleration information of the linear motor system as x1, x2 and Considering the kinematic and dynamic characteristics of the motion unit, the second-order mathematical modeling of the linear motor system is constructed as follows:

[0079]

[0080] Among them, M, K, B, and A represent the inertial mass, voltage-torque gain coefficient, viscous friction coefficient, and Coulomb friction coefficient of the motion unit, respectively; and They are the unknown constant parameters decomposed from the total error d of the linear motor system. and unknown uncertainties S f (x2) = arctan (6000x2) is a nonlinear smooth function term related to speed; υ F represents the control input including actuator failure and saturation, and the specific form is as follows:

[0081] υ F =ηu(υ)+u b (1.2),

[0082] Among them, η∈(0,1] is the effective factor; u b is the bias fault constant; u(υ) represents the value of the control input υ when saturation occurs, and u(υ)=sat(υ)=h(υ)+g(υ); sat(υ) is the standard saturation function and is determined by the upper and lower bounds υ m and m h(υ) is used as a smooth function to approximate sat(υ), and h(υ) = υ M tanh(k1 / υ M ); M and k1 are the intermediate coefficient and saturation approximation gain respectively, and there is g(υ) is the saturated approximation error; tanh and sign are the hyperbolic tangent function and the standard sign function, respectively.

[0083] The mathematical model of the linear motor under actuator failure and saturation is obtained by combining formula (1.1) and formula (1.2):

[0084]

[0085] in, represents the new linear motor system lumped error; is the model parameterization vector containing the linear motor mathematical model information, The actual physical system has the characteristic θ∈[θ min ,θ max ] and |▽|≤▽ max , where θ min and θ max are boundary values ​​and ▽ max is a bounded function.

[0086] Step 2: Design the adaptive control law based on the model parameter vector, including:

[0087] definition is the estimated value of the model parameterization vector θ, Represents the estimation error.

[0088] Design of saturation-type projective adaptive law

[0089]

[0090] in, is the upper bound of the adaptive learning law, is the standard adaptive projection operator, φ and τ represent the update rate and adaptive function respectively.

[0091] In order to ensure the convergence of parameter estimates, a recursive least squares estimation algorithm with a forgetting factor is designed as follows:

[0092]

[0093] Among them, l and Represent the forgetting factor and the standardization factor respectively; M is the upper bound of the update rate φ; λ max ( ) and tr{} represent the maximum eigenvalue and trace of the parameter, respectively; is the estimation error and ψ f is the regressor after filtering, and:

[0094]

[0095] Among them, the subscript f indicates that the parameter has been filtered. Equation (2.3) can be obtained from the regression value ψ before filtering T =[h(υ),-x2,-S f (x2),1] is obtained after being processed by a low-pass filter.

[0096] Design model-compensated adaptive control lawυ a for:

[0097]

[0098] Among them, the subscripts 2-4 indicate the indexes of the 2nd to 4th elements; x d is the desired reference input of the linear motor mathematical model, Represents x d The second derivative of .

[0099] The purpose of this step is to design an adaptive control law for the linear motor mathematical model of the linear motor system in step 1 to achieve parameter estimation and model compensation.

[0100] Step 3: Introduce the intermediate auxiliary system and design an anti-saturation controller, including:

[0101] Constructing anti-saturation auxiliary system:

[0102]

[0103] Where c1 and c2 are both positive constants; ξ1 and ξ2 represent system state variables, and have initial values ​​ξ1(0)=ξ2(0)=0; Δυ=sat(υ)-υ is the deviation between saturation. is the estimated value of θ1.

[0104] When saturation does not occur (Δυ=0), then:

[0105]

[0106] Obviously, the auxiliary system is asymptotically stable, that is, as time t→∞, the system states ξ1→0 and ξ2→0.

[0107] When saturation occurs, the Lyapunov function is defined as:

[0108]

[0109] After taking the derivative of V0(t), we can get:

[0110]

[0111] in, Design parameters and It can ensure that the auxiliary system is bounded.

[0112] Design of anti-saturation controller υ c for:

[0113]

[0114] The purpose of this step is to design an anti-saturation auxiliary system for the linear motor mathematical model of the linear motor system in step 1, and to ensure that the auxiliary system is bounded and stable regardless of whether saturation occurs.

[0115] Step 4: Implement equivalent constraint transformation and design robust feedback items based on motion information, including:

[0116] The state error of the linear motor system is defined as:

[0117] e=x1-x d -ξ1(4.1).

[0118] Design error state constraints, that is, the state error satisfies the inequality:

[0119]

[0120] in, and are all positive numbers to be designed; e0 is the initial value of the state error e; ζ is the scalar of the control error convergence speed; represents the performance function and has:

[0121]

[0122] and are the maximum overshoot and steady-state value respectively, and there is a constant t represents time.

[0123] To achieve equivalent constraint transformation, the hyperbolic tangent function tanh is used as the error transformation function Then we have:

[0124]

[0125] in, ε is the conversion error.

[0126] For equation (4.3), we can find the first and second order derivatives with respect to time:

[0127]

[0128]

[0129] in, for The first derivative of .

[0130] The designed sliding mode variables are:

[0131]

[0132] Among them, Λ>0 represents the gain constant.

[0133] Further, define the Lyapunov function:

[0134]

[0135] Taking the derivative of V1(t) and substituting it into equations (1.3), (3.1), (4.5) and (4.6), we can simplify it to obtain:

[0136]

[0137] in, is the first-order derivative of r; is bounded.

[0138] The robust feedback term is designed as:

[0139]

[0140] Among them, k r >0 indicates a robust gain constant.

[0141] The control input υ consists of four parts, including υ in step 2 a , υ in step 3 c , the above formula r and the error compensation term to be designed is the estimated value to be designed.

[0142] Substituting equations (2.4), (3.4) and (4.7) into equation (4.8), further simplification yields:

[0143]

[0144] in, is the limit of the estimation error, which can be guaranteed by the neural network observer designed in step 5; ∈=(2k r -1) and design k r >1 / 2, From Lyapunov's theorem, we know that the system is bounded.

[0145] This step aims to design an equivalent constraint transformation and robust controller, while ensuring that the system is bounded under the control input υ.

[0146] Step 5: Develop a finite-time neural network observer for the influence of lumped disturbances. The design process includes:

[0147] Introducing the RBF neural network unit, the output of the kth neuron node for:

[0148]

[0149] Among them, x is the input of network training; c k and b k are the center vector and width of the activation function respectively.

[0150] The optimal output of the designed neural network is:

[0151]

[0152] in, is the true value of the network weight, is the error between the network weight and the optimal weight, is the error between the network outputs, and this error is bounded

[0153] The finite-time neural network observer is designed as:

[0154]

[0155] Among them, m, c and β1 are all positive numbers to be designed, and are greater than or equal to The parameters p and q satisfy 0<p<q; the observation error variable dynamics is designed to be

[0156] The adaptive law of the neural network is constructed as:

[0157]

[0158] Among them, Γ represents the network weight update rate, and the coefficient β2 is a constant.

[0159] Redefine the Lyapunov function V2(t):

[0160]

[0161] By taking the derivative and simplifying equation (5.5), we can obtain:

[0162]

[0163] Where ρ = min[λ min (β1)2 a ,λ min (β2)2 a ], and W * is the optimal weight of the neural network.

[0164] Design Γ≥diag(2,2,...,2), then we can obtain:

[0165]

[0166] therefore, is bounded, that is, The neural network observer (5.3) can ensure that the observation error In a finite time T it approaches zero.

[0167] This step aims to design a finite-time neural network observer that can eliminate the lumped error in the linear motor system to improve the tracking performance of the system.

[0168] Step 6: Integrate and optimize the control unit to achieve safe fault-tolerant control of the linear motor, including:

[0169] Combining the controllers designed in steps 2 to 5, the safety fault-tolerant controller of the linear motor motion unit can be expressed as:

[0170] υ=υ a +υ c +υ r +υ n (6.1),

[0171] Among them, a The adaptive control law designed based on the recursive least squares estimation algorithm is used to dynamically adjust the changes in model parameters caused by the external environment or faults;

[0172] Anti-saturation controller c Designed to compensate for the under-drive effect caused by saturation;

[0173] Robust controller r It is used to process and offset the influence of error dynamics and external disturbance after equivalent constraint conversion;

[0174] The compensation term υ constructed by the finite neural network observer n The purpose is to deal with nonlinear effects such as regression residuals and unmodeled dynamics, and further improve the tracking performance of the system.

[0175] An embodiment is given below:

[0176] Given the expected input x d (t) = 40-40cos(πt). The nominal value and upper and lower bounds of the system model obtained based on offline parameter identification are:

[0177] θ min =[5.0,0.4,0.2,-0.1],θ max =[8.0,3.2,1.8,0.1].

[0178] The remaining parameters of this controller are designed as follows:

[0179] S f (x2) = arctan(6000x2), simulate disturbance selection d(t) = 0.4sint;

[0180] The saturated smooth function gain matrix k1 = 1.5, and the upper and lower bounds are υ m =-0.3;

[0181] The parameter estimation algorithm sets l = 0.02, Ξ M =5000, φ0=diag[1000,1000,1000,1000];

[0182] The anti-saturation amplitude system parameters are c1=180, c2=180;

[0183] Sliding mode gain Λ=200;

[0184] The neural network observer parameters are m=1, c=1, β1=0.2, β2=0.5, p=3, q=5 and the update rate is Γ k =10;

[0185] The training input in the activation function is Width b k =5, center value c k =[-0.15,-0.1,-0.05,0,0.05,0.1,0.15] kx7 ;

[0186] The robust controller parameter is K r =1000;

[0187] choose Design performance function

[0188] Actuator faults and saturation conditions are simulated during operation, namely:

[0189]

[0190]

[0191] Figure 2 The real-time tracking trajectories under the two control strategies of this implementation and the traditional PID are demonstrated.

[0192] Figure 3 The error variation curve when following the desired trajectory under different working conditions (fault, saturation, performance constraint) is shown in Figure 1. Compared with the traditional proportional-integral-differential (PID) control, the safety fault-tolerant control error proposed in the present invention is kept within a smaller range while meeting the preset steady-state and transient performance.

[0193] Figure 4 The control output change curves when different working conditions occur are shown in FIG. It can be seen from the curve that after a fault occurs, the control input fluctuation of the present invention is significantly more stable than that of PID because the control parameters can be adjusted dynamically and quickly. After the saturation phenomenon occurs, the two control inputs are limited to the specified range together.

[0194] Specific implementation method 2: The safety fault-tolerant control device for a linear motor motion unit described in this implementation method includes a processor and a memory, wherein the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the safety fault-tolerant control method for a linear motor motion unit as described in specific implementation method 1.

[0195] Specific implementation method three: A computer storage medium described in this implementation method stores at least one instruction, and the at least one instruction is loaded and executed by a processor to implement the safe fault-tolerant control method for a linear motor motion unit as described in specific implementation method one.

[0196] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that the features described in conjunction with a single embodiment may be used in other described embodiments.

Claims

1. A safe fault-tolerant control method for a linear motor motion unit, characterized in that: include: Design adaptive control law υ a , anti-saturation control term υ c , robust feedback term υ r and error compensation term υ n , and a ,υ c ,υ r and n Add together to obtain the safe fault-tolerant control input of the linear motor; The adaptive control law v a , anti-saturation control term υ c , robust feedback term υ r and error compensation term υ n The expressions are: Among them, ψ is the regressor before filtering, is the estimated value of the parameterized vector θ of the linear motor mathematical model. The subscripts 2-4 represent the indexes from the 2nd to the 4th elements. d is the expected reference input of the linear motor mathematical model, c1 and c2 are both positive constants, ξ1 and ξ2 represent the state variables of the linear motor system and have initial values ​​ξ1(0)=ξ2(0)=0, k r is the robust gain constant, σ is a quasi-sliding mode variable, is the finite time neural network observation quantity, the intermediate quantity The expressions of and r are: Λ>0 is the gain constant, e is the state error of the linear motor system, ζ is a scalar that controls the error convergence speed, t represents time, and are the maximum overshoot and steady-state value respectively, and They are all positive numbers, e0 is the initial value of e.

2. The safe fault-tolerant control method for a linear motor motion unit according to claim 1, characterized in that: The mathematical model of the linear motor under actuator failure and saturation is: Among them, x1, x2 and are the position, velocity and acceleration of the linear motor system respectively, h(υ) is a smooth function that approximates the standard saturation function sat(υ), S f (·) is a nonlinear smooth function related to velocity, M, K, B, and A represent the inertial mass, voltage-torque gain coefficient, viscous friction coefficient, and Coulomb friction coefficient of the motion unit, respectively. η is the effective factor. is the unknown constant parameter term decomposed from the lumped error d of the linear motor system, is the unknown uncertainty term decomposed from the total error d of the linear motor system g(υ) is the saturation approximation error, u b is the bias fault constant.

3. The safe fault-tolerant control method for a linear motor motion unit according to claim 1 or 2, characterized in that: The designed adaptive control law υ a ,include: Design of saturation-type projective adaptive law in, is the upper bound of the adaptive learning law, is the standard adaptive projection operator, φ and τ represent the update rate and adaptive function respectively; Based on the above saturation projection adaptive law Design model-compensated adaptive control lawυ a .

4. The safe fault-tolerant control method for a linear motor motion unit according to claim 3, characterized in that: Design a recursive least squares estimation algorithm with a forgetting factor to calculate φ and τ: Among them, l and are the forgetting factor and the standardization factor, respectively, M is the upper bound of the update rate φ; λ max () and tr{} represent the maximum eigenvalue and trace of the parameter, respectively. is the estimation error, ψ f is the regressor after filtering.

5. The safe fault-tolerant control method for a linear motor motion unit according to claim 1 or 2, characterized in that: The design anti-saturation control term υ c ,include: Constructing anti-saturation auxiliary system: Among them, Δυ=sat(υ)-υ, υ is the safety fault-tolerant control input of the linear motor; When saturation does not occur, then: When saturation occurs, there is a Lyapunov function V0(t): Differentiate V0(t): in, design Make the auxiliary system bounded; Then the anti-saturation controller υ is designed c .

6. The safe fault-tolerant control method for a linear motor motion unit according to claim 5, characterized in that: The design robust feedback term υ r ,include: The state error of the linear motor system is defined as: e=x1-x d -ξ1; The state error satisfies the constraint: The state error is used to design a sliding mode variable: in, Design of Lyapunov function based on quasi-sliding mode variables: Substitute the linear motor mathematical model, anti-saturation auxiliary system, the second-order derivative of ε(t) and the sliding mode-like variable into the first-order derivative of V1(t) and simplify to obtain: in, is the estimated value of the model parameterization vector θ, is the estimation error; Then the robust feedback term υ is designed r .

7. The safe fault-tolerant control method for a linear motor motion unit according to claim 1 or 2, characterized in that: The design error compensation term υ n ,include: Introducing the RBF neural network unit makes the optimal output of the neural network Δ, then: in, is the true value of the network weight, is the error between the network weight and the optimal weight, is the network output error; The finite-time neural network observer is: Among them, m, c and β1 are all positive numbers and are greater than or equal to The maximum value of Parameters p and q satisfy 0<p<q; is the observation error and Then the error compensation term υ is designed n .

8. The safe fault-tolerant control method for a linear motor motion unit according to claim 7, characterized in that: The output of the kth neuron node in the RBF neural network unit for: Among them, x is the input of network training, c k and b k are the center vector and width of the activation function respectively.

9. A safe fault-tolerant control device for a linear motor motion unit, characterized in that: The safety fault-tolerant control device includes a processor and a memory, wherein the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the safety fault-tolerant control method for a linear motor motion unit as described in any one of claims 1 to 8.

10. A computer storage medium, characterized in that: The computer storage medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the safe fault-tolerant control method for a linear motor motion unit as claimed in any one of claims 1 to 8.

Images