Nonlinear robust fault-tolerant compensation control method for new energy vehicle air-conditioning compressor drive system

Abstract

The nonlinear robust fault-tolerant compensation control method for the air-conditioning compressor drive system of new energy vehicles belongs to the advanced control field of brushless DC motors and includes the following steps: Step 1: Establish a control system model for the air-conditioning compressor drive motor of new energy vehicles; Step 2: Establish a discrete state space model with partial actuator failure, parameter uncertainty and nonlinear characteristics; Step 3: Design a conventional robust fault-tolerant controller; Step 4: Design a nonlinear dynamic fault-tolerant compensation controller at the previous moment; Step 5: Design a nonlinear dynamic incremental fault-tolerant compensation controller. In the case of partial actuator failure, this method proposes a new solution for the design of the air-conditioning compressor drive motor control system of new energy vehicles with uncertainty and strong nonlinearity, so that the system output can better track the set value and has a good control effect.

Description

Technical Field

[0001] The method belongs to the field of advanced control of brushless DC motors and relates to a nonlinear robust fault-tolerant compensation control method for a new energy vehicle air-conditioning compressor drive system. Background Art

[0002] As the air-conditioning systems of new energy vehicles on the market become increasingly complex, the reliability and safety of the control system of brushless DC motors, as the most common drive motors for new energy vehicle air-conditioning systems, are receiving more and more attention. In actual driving, harsh environments and severe vibrations may cause some actuators in the motor control system to fail. Therefore, in order to meet people's needs and improve the comfort of riding, these minor faults will be hidden dangers that cannot be ignored. In order to ensure the stability and safety of the control system, in the case of partial actuator failure, how to safely and effectively control the control system of the new energy vehicle air-conditioning compressor drive motor with uncertainty and strong nonlinearity has become a difficult problem faced in actual production.

[0003] At present, there are three mainstream control methods for brushless DC motors: FOC (also known as vector frequency conversion, magnetic field vector oriented control), square wave control (also known as trapezoidal wave control, 120° control, 6-step commutation control) and sine wave control. Among them, square wave control uses Hall sensors or sensorless estimation algorithms to obtain the position of the motor rotor, and then performs 6 commutations (commutation every 60°) within a 360° electrical cycle according to the position of the rotor. Its advantages are simple control algorithms and low hardware costs. A controller with ordinary performance can obtain a higher motor speed, but the torque fluctuation is large, there is a certain current noise, and the efficiency cannot reach the maximum value. Therefore, using nonlinear compensation control to improve the control system of the air-conditioning compressor drive motor of new energy vehicles and integrating advanced control algorithms into it is of great significance to the research in this field.

[0004] This patent proposes a nonlinear robust fault-tolerant compensation control method for the air-conditioning compressor drive motor control system of new energy vehicles. On the one hand, the operation of brushless DC motors often has nonlinear characteristics, making it difficult to establish an accurate control system model. To solve this problem, this patent discretizes the controlled object model during the modeling process and expresses it as the sum of a linear model and dynamic nonlinear terms. When dealing with these dynamic nonlinear terms, if they are simply ignored, the designed controller often fails to achieve the expected effect. On the other hand, it is difficult to avoid errors or failures in the actuator during actual operation. At this time, the fault tolerance of the control system is particularly important. Fault-tolerant control can ensure the stable and safe operation of the system while still meeting certain control requirements when some actuators fail, but its strong conservatism seriously affects the control accuracy of the controller. Therefore, combining nonlinear compensation control with robust fault-tolerant control to design a control method for brushless DC motors of new energy vehicle air conditioners has become an inevitable requirement for actual production. Summary of the invention

[0005] In order to solve the above technical problems, this patent proposes a nonlinear robust fault-tolerant compensation control method for the air-conditioning compressor drive system of new energy vehicles. In the case of partial actuator failure, this method adopts a combination of robust fault-tolerant control and signal compensation to better control the motor speed within the required range for the uncertainty and strong nonlinearity of the control system.

[0006] First, based on the operating principle of the brushless DC motor, a control system model of the air-conditioning compressor drive motor of a new energy vehicle with partial actuator failure is established. The Euler discretization is introduced to establish a discrete state space model, and the controller is designed on this basis: the uncertainty of the system is taken into account at the beginning of the design, and a robust fault-tolerant controller is designed for the linear part of the model, so that the entire control system can still show the expected control performance when facing the uncertainty of the object parameters in actual control. Then, the signal compensation method is used to design a nonlinear dynamic fault-tolerant compensation controller for the previous moment to eliminate the influence of the data information of the dynamic nonlinear term on the closed-loop system. Finally, a nonlinear dynamic incremental fault-tolerant compensation controller is designed to eliminate the influence of the change rate of the dynamic nonlinear term on the system.

[0007] The method is implemented through the following technical solutions:

[0008] Step 1: Establish a new energy vehicle air conditioning compressor drive motor control system model;

[0009] Assuming that each winding is completely symmetrical, the resistance, self-inductance and mutual inductance of the three-phase stator winding are fixed values, then L A =L B =L C=L,M AB =M AC =M BA =M BC =M,M CA =M CB =M, L is the self-inductance of the motor phase winding, and M is the mutual inductance between the motor phase windings. Since the three-phase windings are connected in a star connection, the matrix form of the voltage balance equation can be expressed as:

[0010]

[0011] In the formula, u x is the phase voltage of the motor, i x is the phase current of the motor, is the motor phase induced electromotive force, ψ x is the permanent magnet flux of the motor, and the subscript x represents the three-phase winding A, B, and C. x is the phase resistance of the motor. In the three-phase symmetrical winding, R A =R B =R C =R.

[0012] When the motor is running, it absorbs electric power from the power supply, which is equal to the sum of the products of the reverse potential of the three-phase winding and the phase current. Ignoring the mechanical loss of the rotor, the electromagnetic power is completely converted into rotor motion, so P e =T e Ω,P e is the electromagnetic power, T e is the electromagnetic torque of the motor, Ω is the mechanical angular velocity of the motor, and another form of the torque equation can be obtained:

[0013] T e =p[ψ m f A (θ)i A +ψ m f B (θ)i B +ψ m f C (θ)i C ] (2)

[0014] In the formula, B(θ) is the radial air gap flux density distribution of the rotor permanent magnet, which is trapezoidal along the angle θ. p is the number of motor poles, ψ m It is the maximum value of the permanent magnet flux linkage of each phase winding turn.

[0015] The equation of motion for a brushless DC motor is:

[0016]

[0017] In the formula, K is the electromagnetic torque coefficient of the motor, J is the motor moment of inertia, and B v is the motor damping coefficient.

[0018]

[0019] Selecting three-phase current and motor speed as state variables, the fourth-order state space model equation is obtained as follows:

[0020]

[0021] In the formula, x=[i A i B i C Ω],u=u A =u B =u C , A, B are the motor model parameters.

[0022] Step 2: Establish a discrete state space model with partial actuator failure, parameter uncertainty and strong nonlinearity;

[0023] In the case of partial actuator failure, the discrete state space model with parameter uncertainty and strong nonlinear characteristics is as follows:

[0024]

[0025] In the formula, x(k+1) represents the system state at discrete time k+1, u(k) represents the control input at discrete time k, and v(k) represents the nonlinear dynamics between the discretized model and the original model at discrete time k. B is the input matrix to be identified, α is the actuator fault coefficient, and satisfies α ≤1 is the lower limit of the failure coefficient, is the upper bound of the fault coefficient, C is the output matrix, A(k)=A+Δ a (k) is the state matrix of the system at discrete time k, A is the state matrix to be identified, Δ a (k) is the uncertain perturbation of the state matrix at discrete time k, satisfying Δ a (k) = NΔ(k)H, N is the unified matrix of system uncertainty, H is the state matrix of system uncertainty, is the uncertain perturbation that depends on the discrete time k, where Δ1, Δ2, Δ3, Δ4 are random numbers between [-1, 1].

[0026] In order to simplify the design, the failure coefficient is defined as:

[0027] α=(I+α0)β (7)

[0028] Among them, α0 is the fault coefficient factor and a random number that satisfies |α0|≤β0≤I, β0 is the ratio of the sum and difference of the upper and lower bounds of the fault coefficient, β is the median of the upper and lower bounds of the fault coefficient, and I is a unit matrix of appropriate dimension.

[0029] By moving the first term of equation (6), we can know that v(k) = x(k+1)-A(k)x(k)-B(β+β0β)u(k). However, x(k+1) is unknown at time k, and it is not feasible to eliminate nonlinear dynamics based on this design. Therefore, this method expresses v(k) as v(k) = v(k-1) + Δv(k), where v(k-1) is the nonlinear dynamics between the discretized model and the original model of the system at discrete time k-1, and Δv(k) is the nonlinear dynamic increment of the system at discrete time k.

[0030] Step 3: Design a conventional robust fault-tolerant controller;

[0031] This method deeply integrates the advantages of feedback, feedforward, compensation control and conventional robust fault-tolerant control, and proposes a nonlinear robust fault-tolerant compensation controller.

[0032] First, the conventional robust fault-tolerant controller u1(k) is designed as follows:

[0033] u1(k)=K1(k)x(k) (8)

[0034] Where K1(k) is the control law gain of the robust fault-tolerant controller at discrete time k.

[0035] Therefore, if there exists a known scalar θ>0, is the unknown positive definite symmetric matrix for which the linear matrix inequality needs to be solved, is the unknown positive definite symmetric matrix for which the linear matrix inequality needs to be solved, is the unknown matrix that needs to be solved by linear matrix inequality, ε1>0 is the unknown scalar that needs to be solved by linear matrix inequality, ε2>0 is the unknown scalar that needs to be solved by linear matrix inequality, so that the following linear matrix inequality condition holds, which can make the controlled system asymptotically stable.

[0036]

[0037]

[0038] Where K1(k)=YL -1 , P = θL -1 , I is the identity matrix of appropriate dimension, Q and R are the control parameters of the conventional robust fault-tolerant controller.

[0039] Step 4: Design a nonlinear dynamic fault-tolerant compensation controller at the previous moment;

[0040] According to the data information of the nonlinear dynamics at the previous moment, the nonlinear dynamic fault-tolerant compensation controller u2(k) at the previous moment is designed, that is,

[0041] u2(k)=-K2(k)v(k-1) (11)

[0042] Where K2(k) is the controller control law gain at discrete time u2(k).

[0043] In order to perform dynamic and static compensation for the nonlinear dynamics of the previous moment and ensure the control performance of the controller u2(k), it is defined as:

[0044]

[0045] In the formula, A(z -1 ) is the value of A(k) with respect to z -1 The weighted polynomial of .

[0046] Define the generalized ideal output,

[0047] ξ(k+1)=P(z -1 )x(k+1)=B(β+β0β)u(k)+v(k-1)+Δv(k)-K1(k)x(k) (13)

[0048] On the basis of u1(k), considering that the nonlinear dynamics caused by the nonlinearity of the system has the least impact on the system state, that is, without considering u3(k), the difference between the generalized output and the generalized ideal output is the smallest, we can get:

[0049]

[0050] In order to ensure the stability of the system and to perform dynamic and static compensation for the nonlinear dynamic v(k-1) at the previous moment, we can get A(z -1 )-z -1 B(β+β0β)K1(k)≠0, the one-step optimal feedforward control law with the nonlinear dynamic v(k-1) compensation of the previous moment is:

[0051]

[0052] Step 5: Design a nonlinear dynamic incremental fault-tolerant compensation controller;

[0053] Finally, a nonlinear dynamic incremental compensation controller is designed, defining y r is the set value of the controlled object, and e(k) is the system output error at discrete time k.

[0054] At this time, the one-step optimal forecast of the tracking error e(k+1|k) is:

[0055]

[0056] In order to minimize the error, the nonlinear dynamic incremental fault-tolerant compensation controller can be obtained as:

[0057]

[0058] Where Δv(k-1) is the nonlinear dynamic increment of the system at discrete time k-1.

[0059] Compared with the prior art, the beneficial effects of this method are as follows: a nonlinear robust fault-tolerant compensation control method is proposed for the air-conditioning compressor drive motor control system of new energy vehicles. On the one hand, based on the conventional robust fault-tolerant control method, this method combines the operating data information of the air-conditioning compressor drive motor control system of new energy vehicles, and designs a feedforward compensator for the nonlinear dynamics in the controlled object model, thereby improving the control accuracy of conventional robust fault-tolerant control for systems with model uncertainty and strong nonlinearity. On the other hand, unlike traditional nonlinear control methods, considering the difference between nonlinear models and linear models, combined with the one-step optimal control idea, the dynamic error information of the system is effectively used to design a compensator that eliminates the error signal, thereby designing a nonlinear robust fault-tolerant compensation control method with weak conservatism and good tracking effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a random noise signal;

[0061] Figure 2 is the actuator failure coefficient;

[0062] Figure 3 This is the output response simulation effect diagram of this method;

[0063] Figure 4 This is the tracking performance effect diagram of this method;

[0064] Figure 5 A flow chart of the steps for designing a controller. DETAILED DESCRIPTION

[0065] The method is described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] Example:

[0067] This method studies the control system of the air-conditioning compressor drive motor of new energy vehicles. Through repeated tests, the state space model parameters of the control system of the air-conditioning compressor drive motor of new energy vehicles are: A(k) = A + Δ a (k), C = [0 00 1], Δ a (k) = NΔ(k)H, is the uncertain perturbation that depends on the discrete time k, where Δ1, Δ2, Δ3, Δ4 are random numbers between [-1, 1].

[0068] The estimated values ​​of the model parameters in equation (3) obtained by alternating recursive least squares and neural network identification are:

[0069]

[0070] The random noise signal noise(k) superimposed on the simulation model of the controlled object is as follows: Figure 1 As shown, the actuator failure coefficient is as follows Figure 2 shown.

[0071] The control objective can be expressed as:

[0072] |e(k)|=|y r (k)-y(k)|<5,0<k<∞ (19)

[0073] The set value y r (k)=3000.

[0074] Since the controller parameters designed by this method will be optimized in a rolling manner as the system runs, and the controller parameters will only fluctuate within a small range, only the controller parameters at the time are given here for reference, and the values ​​are as follows:

[0075]

[0076] Therefore, when k=500s, the conventional robust predictive controller outputs u1(k)=-0.0206, the nonlinear dynamic compensation controller outputs u2(k)=0.0044 at the previous moment, the nonlinear dynamic incremental compensation controller outputs u3(k)=-0.0168, and the nonlinear robust fault-tolerant compensation controller outputs u(k)=0.1019.

[0077] The introduced performance evaluation function is:

[0078]

[0079] Where E(k) is the performance evaluation index at discrete time k.

[0080] The simulation is conducted for the air-conditioning compressor drive motor control system of new energy vehicles. The simulation adopts a nonlinear robust fault-tolerant compensation control method. The output response and tracking performance of this method are shown in Figure 2. Figure 3 , 4 shown.

[0081] Depend on Figure 3It can be seen that under the influence of external noise, this method can well control the speed of the brushless DC motor of the new energy vehicle air conditioner to near the set value, avoiding large fluctuations while ensuring control accuracy. Figure 4 It can be seen that this method controls the motor speed within the control requirement range, and the time exceeding the upper limit accounts for 0% of the total simulation time, which improves the tracking performance of the system.

[0082] In summary, the nonlinear robust fault-tolerant compensation control method can be effectively applied to the control system of the air-conditioning compressor drive motor of new energy vehicles. Through the analysis of simulation results, in the case of partial actuator failure, this method can track the set value well and has a good control effect. Therefore, the nonlinear robust fault-tolerant compensation control method can well meet the actual control needs, and puts forward a new solution for the design of the air-conditioning compressor drive motor control system of new energy vehicles with uncertainty and strong nonlinearity. It has good application prospects and is of great significance to the development of fault-tolerant control of such brushless DC motors.

Claims

1. A nonlinear robust fault-tolerant compensation control method for a new energy vehicle air-conditioning compressor drive system, characterized in that: Step 1: Establish a new energy vehicle air conditioning compressor drive motor control system model; Assuming that each winding is completely symmetrical, the resistance, self-inductance and mutual inductance of the three-phase stator winding are fixed values, then L A =L B =L C =L,M AB =M AC =M BA =M BC =M,M CA =M CB =M, L is the self-inductance of the motor phase winding, and M is the mutual inductance between the motor phase windings. Since the three-phase windings are connected in a star connection, the matrix form of the voltage balance equation can be expressed as: In the formula, u x is the phase voltage of the motor, i x is the phase current of the motor, is the phase induced electromotive force of the motor, ψ x is the permanent magnet flux of the motor, subscript x represents the three-phase winding A, B, C; R x is the phase resistance of the motor. In the three-phase symmetrical winding, R A =R B =R C =R; When the motor is running, it absorbs electric power from the power supply, which is equal to the sum of the products of the reverse potential of the three-phase winding and the phase current. Ignoring the mechanical loss of the rotor, the electromagnetic power is completely converted into rotor motion, so P e =T e Ω,P e is the electromagnetic power, T e is the electromagnetic torque of the motor, Ω is the mechanical angular velocity of the motor, and another form of the torque equation can be obtained: T e =p[ψ m f A (i)i A +ψ m f B (i)i B +ψ m f C (i)i C ] (2) In the formula, f A (θ)=θ, is the radial air gap flux density distribution of the rotor permanent magnet, which is trapezoidal along the angle θ, p is the number of motor pole pairs, ψ m is the maximum value of the permanent magnet flux of each phase winding turn; The equation of motion for a brushless DC motor is: In the formula, K is the electromagnetic torque coefficient of the motor, J is the motor moment of inertia, and B v is the motor damping coefficient; Selecting three-phase current and motor speed as state variables, the fourth-order state space model equation is obtained as follows: In the formula, x=[i A i B i C Ω],u=u A =u B =u C ; A, B are the motor model parameters; Step 2: Establish a discrete state space model with partial actuator failure, parameter uncertainty and strong nonlinearity; In the case of partial actuator failure, the discrete state space model with parameter uncertainty and strong nonlinear characteristics is as follows: In the formula, x(k+1) represents the system state at discrete time k+1, u(k) represents the control input at discrete time k, v(k) represents the nonlinear dynamics between the discretized model and the original model at discrete time k; B is the input matrix to be identified, α is the actuator fault coefficient, and satisfies α ≤1 is the lower limit of the failure coefficient, is the upper bound of the fault coefficient, C is the output matrix, A(k)=A+Δ a (k) is the state matrix of the system at discrete time k, A is the state matrix to be identified, Δ a (k) is the uncertain perturbation of the state matrix at discrete time k, satisfying Δ a (k) = NΔ(k)H, N is the unified matrix of system uncertainty, H is the state matrix of system uncertainty, is the uncertain perturbation that depends on the discrete time k, where Δ1, Δ2, Δ3, Δ4 are random numbers between [-1, 1]; In order to simplify the design, the failure coefficient is defined as: α＝(I+α0)β (7) Among them, α0 is the fault coefficient factor and a random number that satisfies |α0|≤β0≤I, β0 is the ratio of the sum and difference of the upper and lower bounds of the fault coefficient, β is the median of the upper and lower bounds of the fault coefficient, and I is a unit matrix of appropriate dimension; By moving the first term of equation (6), we can know that v(k)=x(k+1)-A(k)x(k)-B(β+β0β)u(k). However, x(k+1) is unknown at time k, and it is not feasible to eliminate the nonlinear dynamics based on this design. Therefore, this method expresses v(k) as v(k)=v(k-1)+Δv(k), where v(k-1) is the nonlinear dynamics between the discretized model and the original model of the system at discrete time k-1, and Δv(k) is the nonlinear dynamic increment of the system at discrete time k. Step 3: Design a conventional robust fault-tolerant controller; This method deeply integrates the advantages of feedback, feedforward, compensation control and conventional robust fault-tolerant control, and proposes a nonlinear robust fault-tolerant compensation controller; First, the conventional robust fault-tolerant controller u1(k) is designed as follows: u1(k)=K1(k)x(k) (8) Where K1(k) is the control law gain of the robust fault-tolerant controller at discrete time k; Therefore, if there exists a known scalar θ>0, is the unknown positive definite symmetric matrix for which the linear matrix inequality needs to be solved, is the unknown positive definite symmetric matrix for which the linear matrix inequality needs to be solved, is the unknown matrix that needs to be solved by the linear matrix inequality, ε1>0 is the unknown scalar that needs to be solved by the linear matrix inequality, ε2>0 is the unknown scalar that needs to be solved by the linear matrix inequality, so that the following linear matrix inequality conditions are established, which can make the controlled system asymptotically stable; Where K1(k)=YL -1 , P = θL -1 , I is the identity matrix of appropriate dimension, Q and R are the control parameters of the conventional robust fault-tolerant controller; Step 4: Design a nonlinear dynamic fault-tolerant compensation controller at the previous moment; According to the data information of the nonlinear dynamics at the previous moment, the nonlinear dynamic fault-tolerant compensation controller u2(k) at the previous moment is designed, that is, u2(k)=-K2(k)v(k-1) (11) Where K2(k) is the controller control law gain of u2(k) at discrete time k; In order to perform dynamic and static compensation for the nonlinear dynamics of the previous moment and ensure the control performance of the controller u2(k), it is defined as: In the formula, A(z -1 ) is the value of A(k) with respect to z -1 A weighted polynomial of ; Define the generalized ideal output, ξ(k+1)=P(z -1 )x(k+1)=B(β+β0β)u(k)+v(k-1)+Δv(k)-K1(k)x(k) (13) On the basis of u1(k), considering that the nonlinear dynamics caused by the nonlinearity of the system has the least impact on the system state, that is, without considering u3(k), the difference between the generalized output and the generalized ideal output is the smallest, we can get: In order to ensure the stability of the system and to perform dynamic and static compensation for the nonlinear dynamic v(k-1) at the previous moment, we can get A(z -1 )-z -1 B(β+β0β)K1(k)≠0, the one-step optimal feedforward control law with the nonlinear dynamic v(k-1) compensation of the previous moment is: Step 5: Design a nonlinear dynamic incremental fault-tolerant compensation controller; Finally, design a nonlinear dynamic incremental compensation controller; define y r is the set value of the controlled object, e(k) is the system output error at discrete time k; At this time, the one-step optimal forecast of the tracking error e(k+1|k) is: In order to minimize the error, the nonlinear dynamic incremental fault-tolerant compensation controller can be obtained as: Where Δv(k-1) is the nonlinear dynamic increment of the system at discrete time k-1; The conventional robust fault-tolerant controller, the previous moment nonlinear dynamic fault-tolerant compensation controller and the nonlinear dynamic incremental fault-tolerant compensation controller are combined to finally form a nonlinear robust fault-tolerant compensation controller.

Images