CN108123648A

CN108123648A - Linear servo Position Tracking Control based on linear matrix inequality and sliding formwork control

Info

Publication number: CN108123648A
Application number: CN201711244933.5A
Authority: CN
Inventors: 孙宜标; 仲原; 刘春芳
Original assignee: Shenyang University of Technology
Current assignee: Dongneng (Shenyang) Energy Engineering Technology Co.,Ltd.
Priority date: 2017-11-30
Filing date: 2017-11-30
Publication date: 2018-06-05
Anticipated expiration: 2037-11-30
Also published as: CN108123648B

Abstract

Linear servo Position Tracking Control based on linear matrix inequality and sliding formwork control, designs mechanical subsystem and electrical subsystem carrying out control design case as a whole, it may be difficult to which the challenge of solution is converted into relatively simple linear matrix inequality（LMI）Problem using Lyapunov theorem-proving system stability, recycles MATLAB LMI tool boxes to solve its result.Off-line learning is carried out to end effect, friction quasi-periodic non-linear factor using neutral net Thrust Observer to ensure system real time, it reapplies and load disturbance is observed in system, finally achieve the object of the present invention the i.e. realization good robust tracking performance of system.

Description

Linear servo Position Tracking Control based on linear matrix inequality and sliding formwork control

Technical field

The present invention relates to fields of numeric control technique, and in particular to a kind of Position Tracking Control method of permanent magnet linear synchronous motor And device.

Technical background

Numeric Control Technology is the technology controlled with digital information the Machine Movement Process of controlled device, and numerical control equipment is The electromechanical integration that new technology using Numeric Control Technology as representative forms the infiltration of conventionally manufactured industry and emerging manufacturing industry is produced Product.Modern NC Machine Tool and the numerically-controlled machine tool of early stage it is most basic had occurred difference lies in its process velocity and machining accuracy it is huge Variation, has been respectively increased nearly thousand times.Since high speed, high-precision processing technology can greatly improve processing speed, raising product Quality and class, improve the market competitiveness, thus, using high-speed cutting, high speed feed and high manufacturing accuracy as main feature High-speed Machining Technology has become one of important development trend of modern number control processing technology.

Traditional line servo-drive system is made of current control (electrical) subsystem and speed control (machinery) subsystem.Two Electric current and the respective change procedure of speed differ greatly in time scale in a subsystem, at least more than an order of magnitude. It is simplify control in the case where control requirement is not relatively high, usually ignores the Nonlinear Dynamic coupling between speed and electric current, and Can be used Oriented Vector Control realize PMLSM Partial Variable between static decoupling, to obtain inearized model.Simultaneously Depending on independently of one another, mutually not coupled interference, then forms electric current _ speed two close cycles knot using linear control method between two subsystems Structure is independently to control electric current and speed.But for high speed, high-precision Position Tracking Control system, if above-mentioned two The time scale size of subsystem is relatively close to, and just be must take into consideration and is coupled between non-linear and electric current and the speed of model, because Non-linear coupling between electric current and speed will result directly in the tracking deterioration of system.

Scholars propose some control methods, such as using the anti-control method for pushing away control and variable-structure control is combined, It uses singular perturbation theory by system decomposition for fast, slow subsystem control method, is more to use feedback linearization Method, but the preferable control effect in the case of Parameter Perturbation of the method for feedback linearization control effect in nominal parameters is not It is good.Have scholar added on the basis of feedback linearization robust H∞ controller ensure system well tracking in the case of into One step enhances system robustness, but H ∞ Theoretical Designs controller is utilized to have certain conservative.

The content of the invention

In order to overcome the Nonlinear Dynamic coupled problem between conventional vector control medium velocity and electric current, the present invention provides one Kind suitable for the linear servo position that permanent magnet linear synchronous motor is combined based on linear matrix inequality with sliding formwork control with Track control method, electrical subsystem and mechanical subsystem are controlled as a whole, possess response speed using sliding formwork control Soon, controller is designed the advantages that strong robustness, it may be difficult to which the challenge of solution is converted into simple linear matrix inequality (LMI) problem utilizes MATLAB-LMI tool boxes.Periodical non-linear factor is carried out using neutral net Thrust Observer from Line learns, and reapplies in system to carrying out disturbance observation, finally achieves the object of the present invention the i.e. realization good robust of system Tracking performance.

A kind of linear servo position being combined based on linear matrix inequality with sliding formwork control designed by the present invention with Track control principle block diagram is as shown in Fig. 1.

It is a kind of to be included based on linear matrix inequality with the linear servo Position Tracking Control design that sliding formwork control is combined Following several steps：

1. by controlled quentity controlled variable u_qAnd u_dTwo parts are decomposed, sliding-mode surface and sliding formwork control ratio are designed using LMI.

Make q shaft voltages u_q=u_1f+u₁, d shaft voltages u_d=u_2f+u₂, wherein control law u_1f、u_2fBe designed to compensating line The Nonlinear Dynamic of servo-drive system, u₁、u₂As sliding formwork control ratio system mode track is forced to tend to sliding-mode surface.

Define error state vector：Wherein,For mover displacement error,It is missed for mover linear velocity Difference,For d, q shaft current error.

Systematic error state equation is：

Wherein, coefficient of combination matrixInput matrixUdd is disturbance quantity, and u is system control input.

DefinitionFor matrix B^TOrthonormal basis.MatrixFor two LMI's Solution.The sliding variable of system is：

σ=Sx=(B^TX^-1B)^-1B^TX^-1x (2)

u_eqFor Equivalent control law, expression formula is：

u_eq=-SAx (3)

u_swFor switching law, expression formula is

Wherein, adjustable parameter boundary layer δ ∈ (0,1), handoff gain k ∈ (0,1000), | | σ | | it is 2 norms of σ.U=u_eq +u_sw=(u₁u₂)^T。

2. design neutral net load thrust observer

This observer is a Neural Network Observer in parallel on the basis of linear observer, the two output and conduct Final observation.Schematic diagram is as shown in Fig. 2, NNTO is Neural Network Observer in parallel, and selecting tool here, there are one input sections Point, 8 hidden layer nodes, 3 layers of BP (error back propagation) neutral net of two output nodes.

Above-mentioned control method is embedded in the position control realized in DSP control circuits to PM linear servo system System, wherein realizing a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control designed by the present invention The hardware block diagram of tracking control system is as shown in Figure 4.Whole system includes main circuit, control circuit and controlled device three parts； Control circuit includes DSP, position and velocity checking circuits, current detection circuit, optical coupling isolation circuit, driving circuit and failure inspection Survey and protect circuit；Main circuit includes rectification filtering unit and IPM inversion units；Controlled device is three-phase permanent Linear Synchronous electricity Machine, fuselage are equipped with grating scale.

The method of the present invention is finally realized by the control program in embedded dsp processor, is as follows：

Step 1 control program starts；

Step 2 system initialization；

Step 3 allows INT1, INT2 to interrupt；

Step 4 starts the interruption of T1 underflows, EPWM is interrupted and QEP is interrupted；

Step 5 program data initializes；

Step 6 opens total interruption；

Step 7 interrupt latency；

Step 8 judges whether to terminate to run, if not return to step 7；

Step 9 terminates；

Wherein above-mentioned TN1 interrupt processings control program is according to the following steps：

Step 1 T1 Interrupt Subroutines；

Step 2 keeps the scene intact；

Step 3 current sample；

Step 4 CLARK is converted；

Step 5 PARK is converted；

Step 6 disturbance estimation programme-control；

Step 7 d q shaft voltages are adjusted；

Step 8 PARK inverse transformations；

Step 9 calculates CMPPx and PWM outputs；

Step 10 position samples；

Step 11 initial position program；

Step 12 restoring scene；

Step 13, which is interrupted, to be returned；

Wherein the disturbance of above-mentioned steps 6 estimation controls program according to the following steps：

Step 1 thrust disturbance observer program starts；

Step 2 reads initialization data；

Step 3 reads the sampled value (measured value) of initial time speed；

Step 4 sets the initial parameter of Systems with Linear Observation device；

Step 5 sets initial parameter, the structure of Parallel neural networks observer；

Step 6 calculating speed measured value and the difference of NNTO outputs；

Step 7 is by system discretization；

Step 8 calculates the state estimator of Systems with Linear Observation device；

Step 9 calculates the state estimator of Parallel neural networks observer；

The state estimator of Systems with Linear Observation device and Parallel neural networks observer is superimposed by step 10；

Step 11 is by system serialization；

Step 12 calculates the output quantity of NNTO；

Step 13 calculates and the estimator of output disturbance；

Step 14 returns to caller.

Wherein QEP interruption positions adjust processing son control program according to the following steps：

Step 1 QEP interrupt service subroutines start；

Step 2 keeps the scene intact；

Step 3 reads initialization data；

Step 4 reads initial time position, speed, d shaft currents, the sampled value (measured value) of q shaft currents；

Step 5 calculates the desired value of q shaft currents and speed；

Step 6 calculates state variable feedback error and its derivative by reference input and stater output；

Step 7 sets LMI- sliding formwork initial parameters；

Step 8 judges whether on sliding-mode surface is preset, if it is to carry out step 10, if otherwise carrying out step 9；

Step 9 calculates switching control u_sw；

Step 10 calculates equivalent control u_eq；

Step 11 calculates Linear Control rule u_1f、u_2f；

Step 12 Linear Control rule is added with LMI- sliding formwork control ratios；

Break in step 13 Central Shanxi Plain；

Step 14 restoring scene；

Step 15, which is interrupted, to be returned.

The advantages of present invention relates to is：For linear servo system, the present invention provides a kind of based on linear matrix inequality With the linear servo Position Tracking Control system and method that sliding formwork control is combined.It designs mechanical subsystem and electrical subsystem Control design case is carried out as a whole, it may be difficult to which the challenge of solution is converted into relatively simple linear matrix inequality (LMI) Problem using Lyapunov theorem-proving system stability, recycles MATLAB-LMI tool boxes to solve its result.Utilize god Off-line learning is carried out to end effect, friction quasi-periodic non-linear factor through network Thrust Observer to ensure that system is real-time Property, it reapplies and load disturbance is observed in system, finally achieve the object of the present invention the i.e. realization good robust of system Tracking performance.

Description of the drawings

A kind of linear servo position being combined based on linear matrix inequality with sliding formwork control designed by Fig. 1 present invention Tracking control system block diagram；

Neutral net Thrust Observer structure chart designed by Fig. 2 present invention；

The cut-away view of LMI- sliding mode controllers designed by Fig. 3 present invention；

Fig. 4 realizes the hardware system schematic diagram of the present invention；

Fig. 5 (a) electric machine control system main circuit schematic diagrams；

Fig. 5 (b) A, B phase current sampling circuit schematic diagrams；

Fig. 5 (c) grating scale signal sample circuit schematic diagrams；

Fig. 5 (d) driver circuit schematic diagrams；

Fig. 6 QEP interruption positions adjust processing son control program flow diagram；

Fig. 7 neutral nets Thrust Observer controls program flow diagram；

Fig. 8 master control system program flow diagrams；

Fig. 9 T1 interrupt processings son control program flow diagram；

Step response curve when Figure 10 impacts disturb；

The special aircraft pursuit courses of Figure 11；

Figure 12 particular curve tracking error curves；

Figure 13 sinusoidal position tracking error curves.

Specific embodiment

The embodiment of the present invention provides a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control Tracking control system and method can realize good Robust Speed tracing property to position command at a high speed, under high-precision occasion Energy.

Realize that the present invention is as follows：

Step 1：Permanent magnet linear synchronous motor nonlinear mathematical model is established.

Permanent magnet fundamental wave excitation field axis is taken as d axis, is rotated by 90 ° electrical angle counterclockwise as q axis so that mover electric current Vector is with stator field in orthogonal space.Ignore magnetic circuit saturation, nonlinear dynamic mathematical model can represent as follows：

ψ_d=L_di_d+ψ_f (7)

ψ_q=L_qi_q (8)

u_d、u_qFor straight, quadrature-axis voltage (V)；i_d、i_qFor straight, quadrature axis current (A)；ψ_fFor permanent magnet flux linkage (Wb)；ψ_d、ψ_qFor Directly, quadrature axis excitation flux linkage (Wb)；L_d、L_qFor straight, quadrature axis inductance (H)；Angular rate(v is mover speed (m/s))；τ For pole span (mm)；R_sFor stator resistance (Ω).

The electromagnetic push equation of PMLSM is：

Wherein, p is number of pole-pairs, in durface mounted permanent magnet linear synchronous generator, L_d=L_q=L_s(L_sFor stator inductance), it can

In formula (10),For electromagnetic push coefficient (N/A).

The mechanical motion equation of PMLSM is：

Wherein, f is broad sense load disturbance (N) (the sum of force oscillation caused by load, friction and disturbance etc.).

Fetch bit moves s, speed v, cross, straight shaft current i_q、i_dFor state variable, then PMLSM state space equations are：

Wherein, M is mover quality (kg)；S is mover mechanical displacement (m)；B_vFor viscous friction coefficient (Ns/m)；d₁(t) And d₂(t) it is equivalent exogenous disturbances caused by power supply each harmonic and other systems uncertainty；u_dAnd u_qInput quantity in order to control.

Step 2：Permanent magnet linear synchronous motor LMI- sliding mode controller designs

Define s_refFor given position signal,The desired value for making d shaft currents isWhen tracking Setting signal, Can desired q shaft currents be calculated according to the speed of Setting signal, acceleration and load disturbance, it can using formula (10), (11) The desired values of q shaft currents be：

Then displacement error：

Velocity error：

D, q shaft currents error：

The error state equation that PMLSM can be obtained by formula (12)~(19) is：

Wherein,

According to formula (20) design control law, by control input u_qAnd u_dIt decomposes：

Wherein, control law u is utilized_1f、u_2fDesign with the Nonlinear Dynamic of compensating line servo-drive system, u₁、u₂As cunning Mould control law forces the system mode track to tend to sliding-mode surface.

Define control law u_1fAnd u_2f：

So formula (20) described error state equation can be converted into：

Formula (23) can be written as：

In formula,

For the system design sliding-mode surface of formula (24).DefinitionFor the orthonormal basis of matrix B T.Consider matrixFor The solution of two LMI in formula (25).

Make S=(BTX-1B)-¹BTX-1, then sliding-mode surface can be defined as：

σ=Sx=(B^TX^-1B)^-1B^TX^-1X=(σ₁σ₂)^T (26)

Then sliding formwork control ratio can be designed as：

U=-SAx-ksgn (σ)=u_eq+u_sw (27)

Wherein, k is constant, SA=G=(G₁ G₂)^T。u_eqFor Equivalent control law, u_swFor switching law.

Switching law u_swMiddle introducing boundary layer δ, sliding formwork control ratio can be written as：

By formula (21), (22), (38), control input u_dAnd u_qIt can be written as：

Step 3：Write the subprogram part for realizing that LMI- sliding formwork control ratios are realized.

LMI- sliding formwork control ratios are realized by software programming, flow chart such as Fig. 6, and step is as follows：

Step 1 QEP interrupt service subroutines start；

Step 2 keeps the scene intact；

Step 3 reads initialization data；

Step 5 calculates the desired value of q shaft currents and speed；

Step 7 sets LMI- sliding formwork initial parameters；

Step 9 calculates switching control u_sw；

Step 10 calculates equivalent control u_eq；

Step 11 calculates Linear Control rule u_1f、u_2f；

Break in step 13 Central Shanxi Plain；

Step 14 restoring scene；

Step 15, which is interrupted, to be returned.

Step 4：Neutral net Thrust Observer

Above-mentioned LMI- sliding mode positionings tracing control needs the load disturbance of observation system, but when system parameter variations, Particularly when systematic parameter matrix perturbation amount is larger, conventional linear thrust observer (such as imperial Burger observer) it is difficult to ensure that Satisfied observation effect.Here a Neural Network Observer in parallel, output and linear push on the basis of traditional observer The output of force observer is added, the two and as final observation.

By the use of v and f as system state variables, PMLSM state equations can be written as：

Output equation is：

In formula (30), a=-B_v/ M, b=-K_f/ M, d=-1/M,

By equation (30), (31) discretization, can obtain：

Wherein, G, H, C are respectively state equation (30), the coefficient matrix corresponding to (31).The structure of observer such as Fig. 2 institutes Show, L=(L₁ L₂)^TFor the gain matrix of observer, the original of the characteristic value of system is must be considerably larger than according to the characteristic value of observer L gain battle arrays are then chosen, so that observation system rapid decay.NN is Neural Network Observer in parallel, is selected herein defeated with 1 Ingress, 8 hidden nodes, 3 layers of BP neural network of 2 output nodes.Define network target function be：

The weighed value adjusting formula of neutral net is：

W_j1(k+1)=W_j1(k)+ZW_ib_j(k), j=1~8, i=1~²。

W_ij(k+1)=W_ij(k)+ZW_ji_q(k)

W_j=b_j(k)[1-b_j(k)]W_iW_j1 (34)

In formula：Z --- learning rate；

W_j1--- the weights of hidden layer to output layer；

W_ij--- it inputs to the weights of hidden layer；

Δx_i(k+1) --- i-th of system mode increment size of network output；

b_j(k) --- the output of j-th of hidden node of network.

After state x is observed out,From observed quantityMiddle extraction is i.e. available.

Step 5：Write the subprogram part for realizing that disturbance estimation is realized

It disturbs estimation to be realized by software programming, flow chart such as Fig. 7, step is as follows：

Step 1 thrust disturbance observer program starts；

Step 2 reads initialization data；

Step 3 reads the sampled value (measured value) of initial time speed；

Step 4 sets the initial parameter of Systems with Linear Observation device；

Step 6 calculating speed measured value and the difference of NNTO outputs；

Step 7 is by system discretization；

Step 9 calculates the state estimator of Parallel neural networks observer；

Step 11 is by system serialization；

Step 12 calculates the output quantity of NNTO；

Step 13 calculates and the estimator of output disturbance；

Step 14 returns to caller.

The hardware system schematic diagrams of Fig. 4 to realize the present invention.The system includes main circuit, control circuit and controlled pair As three parts；Control circuit includes DSP, position and velocity checking circuits, current detection circuit, optical coupling isolation circuit, driving electricity Road and fault detect and protection circuit；The QEP ports link position and velocity checking circuits of DSP, the ADC ports connection electricity of DSP Current detection circuit, the PWM port of DSP connect optical coupling isolation circuit, optical coupling isolation circuit connection driving circuit with PDPINT ports With fault detect and protection circuit, driving circuit connection IPM inversion units；Main circuit includes rectification filtering unit and IPM inversions Unit；Controlled device is three-phase permanent linear synchronous generator, and fuselage is equipped with grating scale；Rectification filtering unit connects IPM inversion lists Member, IPM inversion units connection three-phase permanent linear synchronous generator.

The SCI ports connection host computer of DSP, the SPI port connection display circuit of DSP, the GPIO port connection I/O of DSP Interface circuit；Fault detect connects control power supply with protection circuit.

DSP uses the TMS320F28335 chips of TI companies

Fig. 8, Fig. 9 are to realize the master control system program flow diagram of the control system and T1 interrupt processings son control using DSP Processing procedure program flow diagram, is as follows：Master control system program.

Step 1 control program starts；

Step 2 system initialization；

Step 3 allows INT1, INT2 to interrupt；

Step 5 program data initializes；

Step 6 opens total interruption；

Step 7 interrupt latency；

Step 8 judges whether to terminate to run, if not return to step 7；

Step 9 terminates；

Step 1T1 Interrupt Subroutines；

Step 2 keeps the scene intact；

Step 3 current sample；

Step 4CLARK is converted；

Step 5PARK is converted；

Step 6 disturbance estimation programme-control；

Step 7 d q shaft voltages are adjusted；

Step 8 PARK inverse transformations；

Step 9 calculates CMPPx and PWM outputs；

Step 10 position samples；

Step 11 initial position program；

Step 12 restoring scene；

Step 13, which is interrupted, to be returned；

Disturbance estimation interrupt processing of wherein above-mentioned steps 6 controls program according to the following steps：

Step 1 thrust disturbance observer program starts；

Step 2 reads initialization data；

Step 3 reads the sampled value (measured value) of initial time speed；

Step 4 sets the initial parameter of Systems with Linear Observation device；

Step 6 calculating speed measured value and the difference of NNTO outputs；

Step 7 is by system discretization；

Step 9 calculates the state estimator of Parallel neural networks observer；

Step 11 is by system serialization；

Step 12 calculates the output quantity of NNTO；

Step 13 calculates and the estimator of output disturbance；

Step 14 returns to caller.

Step 1QEP interrupt service subroutines start；

Step 2 keeps the scene intact；

Step 3 reads initialization data；

Step 5 calculates the desired value of q shaft currents and speed；

Step 7 sets LMI- sliding formwork initial parameters；

Step 9 calculates switching control u_sw；

Step 10 calculates equivalent control u_eq；

Step 11 calculates Linear Control rule u_1f、u_2f；

Break in step 13 Central Shanxi Plain；

Step 14 restoring scene；

Step 15, which is interrupted, to be returned.

Shown in the control system main circuit such as Fig. 5 (a) for realizing the present invention, rectification filtering unit is uncontrollable whole using bridge-type Stream, bulky capacitor filtering, coordinates appropriate resistance capaciting absorpting circuit, can obtain the constant DC voltage needed for IPM work.IPM is adopted With company of Fuji 6MBP50RA060 intelligent power module, pressure-resistant 600V, maximum current 50A, maximum operating frequency 20kHz.IPM It is powered with the 15V driving powers of four groups of independence.Main power source input terminal (P, N), leading-out terminal (U, V, W), main terminal is with carrying Screw fix, it can be achieved that electric current transmit.P, N is the main power source input terminal after the rectifying conversion smothing filtering of frequency converter, and P is Anode, N are negative terminal, and the three-phase alternating current of inverter output is connected to motor by leading-out terminal U, V, W.

The core of the control circuit of the present invention is TMS320F28335 processors, and it is read-only that mating development board includes target Memory, analog interface, serial boot ROM, user lamp, reset circuit, can be configured to RS232/ at eCAN interfaces The outer 256K*16 RAM of asynchronous serial port, SPI synchronous serial interfaces and piece of RS422/RS485.

Current sample uses LEM companies Hall current sensor LT58-S7 in actual control system.By two Hall currents Sensor detects A, B phase current, obtains current signal, by current sampling circuit, is converted into the voltage signal of 0~3.3V, most The binary number of 12 precision is converted by the A/D modular converters of TMS320LF2812 afterwards, and is stored in numerical register. A, shown in the current sampling circuit of B phases such as Fig. 5 (b).Adjustable resistance VR2 Regulate signal amplitudes, adjustable resistance VR1 Regulate signals are inclined Signal by the adjusting to the two resistance, can be adjusted to 0~3.3V by shifting amount, then be sent to AD0, AD1 pipe of DSP Foot.Voltage-stabiliser tube in figure is to be sent into the signal of DSP in order to prevent more than 3.3V, and DSP is caused to be damaged by high pressure.Operational amplifier is adopted With OP27, power supply connects positive and negative 15V voltages, in the indirect decoupling capacitor of voltage and ground.Circuit input end connects capacitor filtering, to remove height Frequency signal interference improves sampling precision.

The A phases and B phase pulse signals of grating scale output will isolate signal by rapid light coupling 6N137, Ran Houjing Cross bleeder circuit and signal level be converted into 3.3V by 5V, be eventually connected to DSP two-way quadrature coding pulse interface QEP1 and QEP2.Shown in circuit theory such as Fig. 5 (c).Fig. 5 (d) is the driving circuit for the hardware system invented, and linear motor drives Circuit mainly includes an intelligent power module, and that the present invention selects is IRAMSl0UP60B, it is suitable for the electricity of relatively high power In machine, the power of motor scope that it can drive is 400W~750W；The three-phase bridge circuit being mainly made of 6 IGBT, control The pwm control signal that dsp chip generates on plate is input to power module, controls the shut-off of 3 bridge arms, generates appropriate drive electricity Pressure drives the control signal that the HIN1 in linear motor motion diagram and LIN1 is the upper and lower bridge arm of the first phase respectively, they are all Low level is effective.The operating voltage VDD of IRAMSl0UP60B is 15V, and VSS is ground terminal, in order to reach good decoupling effect, Two decoupling capacitors in parallel are added at this both ends.Since the PWM wave signal of input is digital signal, and IRAMS10UP60B is not Possess the function digital signal and power signal isolation, therefore need to add before the input control signal of IRAMS10UP60B Light-coupled isolation, TLPll3 is to realize the function that the pwm signal of input is converted into analog signal in figure, is then input to correspondence The control signal input of bridge arm.The bootstrap capacitor of a 2.2uF is added respectively in output terminal u, v, w of three-phase voltage. When Itrip ports are low level, chip normal work, when the control signal of the upper part of bridge arm input is low level, lower part This mutually has output voltage when being high level；When upper for high level, it is lower for it is low level when output voltage be zero；Both It is impermissible for occurring for low level situation, short circuit can be caused like that, burn chip.When Itrip ports are high level, core Piece does not work, and without voltage output, therefore a pull down resistor is added to make Itrip ports for low level in circuit, such power Module can work normally.Power chip itself had mild overcurrent protection, and self-protection can be played when circuit occurs abnormal Effect.

The case history of the present invention

It is controlled device to select Killmorgen IL06-050A1 non iron-cores linear motor, the parameter of electric machine：M=0.32kg, B_v =0.001Ns/m, k_f=28.5N/A, normal thrust F_e=49.7N, pole span τ=32mm, stator inductance L_s=0.003H, it is fixed Sub- resistance R_s=8.6 Ω, permanent magnet flux linkage ψ_f=0. 03226Wb.

Thus the parameter in formula (20) is extrapolated：k₁=89.1, k₂=0.003125, k₃=2866.7, k₄=1055.7, k₅ =333, k₆=89.2.Force of sliding friction Disturbance Model selects Stribeck models, and mathematic(al) representation is：Wherein Stribeck empirical parameters stiction F_s=3N, minimum Coulomb friction Power F_c=1N, critical Stribeck speed v_s=0.01m/s.Equivalent interference caused by supply harmonic and other systems uncertainty Input：d₁(t)=d₂(t)=0.6k₅·sin(100πt).LMI is solved using MATLAB LMI tool boxes：

Choose sliding formwork gain coefficient k=250, boundary layer δ=0.1.

Experiment one：Position command curve s_refThe step signal for being 0.01m for amplitude, linear motor starting under no load, motor ginseng Number is nominal parameter, impact 50N load resistances during time t=0.25s.Position response curve is as shown in Figure 10.By position tracking Curve (Figure 10) understands that the control method position tracking that this patent is provided responds fast, non-overshoot, when being disturbed in face of shock load, Recovery time is shorter, and the time of stabilizing for 0.02s, shows good robustness.

Experiment two：Position command curve s_refInput instruction is particular curve, as shown in figure 11, position tracking error curve As shown in figure 12, method provided by the invention makes system have good tracking performance when particular curve tracks.

Experiment three：Position command curve s_refIt is 1mm for amplitude, frequency is the sine curve of 80Hz, and T is disturbed outside time-varying_L= 50sin (2t) N, and systematic parameter changes, inductance variation delta L=20%, and mover quality becomes 1.5M, tracking error Curve is as shown in figure 13.Method provided by the present invention makes system have a good tracking accuracy, and for Parameter Perturbation and Good tracking performance can be still kept in the case of being disturbed outside time-varying.

Claims

1. a kind of linear servo Position Tracking Control method being combined based on linear matrix inequality with sliding formwork control, feature It is：The control method according to PM linear servo system by electrical subsystem and mechanical subsystem as a whole It accounts for, by speed, displacement, d, the input of q shaft currents and its error as LMI- sliding mode controllers, LMI- sliding mode controllers Output quantity u_qAnd u_dInput as controlled device；Neutral net Thrust Observer need to be utilized by wherein calculating q shaft currents desired value The disturbance quantity observed；Entire control system hardware includes main circuit, control circuit and controlled device three parts；Control circuit bag Include dsp processor, current sampling circuit, rotor position sample circuit, IPM isolated drive circuits and IPM protection circuits；Main circuit Including rectification filtering unit and IPM inversion units；Controlled device is three-phase permanent linear synchronous generator, and fuselage is equipped with grating scale.

2. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 1 Tracking control unit, it is characterised in that：By control input u_qAnd u_dTwo parts are decomposed into, sliding-mode surface and sliding formwork control are designed using LMI System rule；

Make q shaft voltages u_q=u_1f+u₁, d shaft voltages u_d=u_2f+u₂, wherein utilizing control law u_1f、u_2fDesign watched with compensating line The Nonlinear Dynamic of dress system, u₁、u₂As sliding formwork control ratio system mode track is forced to tend to sliding-mode surface；

Define error state vector：Wherein,For mover displacement error,For mover linear velocity error,For d, q shaft current error；

Systematic error state equation is：

<mrow> <mover> <mi>x</mi> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <mi>A</mi> <mi>x</mi> <mo>+</mo> <mi>B</mi> <mrow> <mo>(</mo> <mi>u</mi> <mo>+</mo> <msub> <mi>u</mi> <mrow> <mi>d</mi> <mi>d</mi> </mrow> </msub> <mo>)</mo> </mrow> </mrow>

Wherein coefficient matrixMatrixu_ddFor disturbance quantity, u is system control input；

DefinitionFor matrix B^TOrthonormal basis；MatrixFor twoSolution；

The sliding variable of system is：

σ=Sx=(B^TX^-1B)^-1B^TX^-1x

u_eqFor Equivalent control law, expression formula is：

u_ep=-SAx

u_swFor switching law, expression formula is：

<mrow> <msub> <mi>u</mi> <mrow> <mi>s</mi> <mi>w</mi> </mrow> </msub> <mo>=</mo> <mo>-</mo> <mi>k</mi> <mrow> <mo>(</mo> <mfrac> <mi>&sigma;</mi> <mrow> <mo>|</mo> <mo>|</mo> <mi>&sigma;</mi> <mo>|</mo> <mo>|</mo> <mo>+</mo> <mi>&delta;</mi> </mrow> </mfrac> <mo>)</mo> </mrow> </mrow>

Wherein, boundary layer δ ＞ 0, handoff gain k ＞ 0；| | σ | | it is 2 norms of σ,

U=u_eq+u_sw=(u₁ u₂)^T。

3. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 1 Tracking control unit, feature in：Design neutral net load thrust observer observation load disturbance；This observer is linearly to see A Neural Network Observer in parallel on the basis of survey device, the two the exports and final observation of conduct；NNTO is god in parallel Through network observations device, select here tool there are one input node, 8 hidden layer nodes, two output nodes 3 layers of BP nerve nets Network.

4. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 2 Tracking control unit, it is characterised in that：This method is as follows：

Step 1：Permanent magnet linear synchronous motor nonlinear mathematical model is established；

Step 2：Permanent magnet linear synchronous motor LMI- sliding mode controller designs；

Step 3：Realize LMI- sliding formwork control ratios part；

Step 4：Neutral net thrust observes part；

Step 5：Write the subprogram part for realizing that disturbance estimation is realized.

5. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 4 Tracking control unit, it is characterised in that：Step 1：Permanent magnet linear synchronous motor nonlinear mathematical model is established；

Permanent magnet fundamental wave excitation field axis is taken as d axis, is rotated by 90 ° electrical angle counterclockwise as q axis so that mover current phasor With stator field in orthogonal space；Ignore magnetic circuit saturation, nonlinear dynamic mathematical model can represent as follows：

<mrow> <msub> <mi>u</mi> <mi>d</mi> </msub> <mo>=</mo> <msub> <mi>R</mi> <mi>s</mi> </msub> <msub> <mi>i</mi> <mi>d</mi> </msub> <mo>+</mo> <mfrac> <mrow> <msub> <mi>d&psi;</mi> <mi>d</mi> </msub> </mrow> <mrow> <mi>d</mi> <mi>t</mi> </mrow> </mfrac> <mo>-</mo> <msub> <mi>&omega;&psi;</mi> <mi>q</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow>

<mrow> <msub> <mi>u</mi> <mi>q</mi> </msub> <mo>=</mo> <msub> <mi>R</mi> <mi>s</mi> </msub> <msub> <mi>i</mi> <mi>q</mi> </msub> <mo>+</mo> <mfrac> <mrow> <msub> <mi>d&psi;</mi> <mi>q</mi> </msub> </mrow> <mrow> <mi>d</mi> <mi>t</mi> </mrow> </mfrac> <mo>+</mo> <msub> <mi>&omega;&psi;</mi> <mi>d</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>

ψ_d=L_di_d+ψ_f (7)

ψ_q=L_qi_q (8)

u_d、u_qFor straight, quadrature-axis voltage (V)；i_d、i_qFor straight, quadrature axis current (A)；ψ_fFor permanent magnet flux linkage (Wb)；ψ_d、ψ_qFor straight, friendship Axis excitation flux linkage (Wb)；L_d、L_qFor straight, quadrature axis inductance (H)；Angular rateV is mover speed (m/s)；τ is pole span (mm)；R_sFor stator resistance (Ω)；

The electromagnetic push equation of PMLSM is：

Wherein, p is number of pole-pairs, in durface mounted permanent magnet linear synchronous generator, L_d=L_q=L_s(L_sFor stator inductance), it can obtain

In formula (10),For electromagnetic push coefficient (N/A)；

The mechanical motion equation of PMLSM is：

Wherein, f is broad sense load disturbance (N) (the sum of force oscillation caused by load, friction and disturbance)；

Wherein, M is mover quality (kg)；S is mover mechanical displacement (m)；B_vFor viscous friction coefficient (Ns/m)；d₁(t) and d₂ (t) it is equivalent exogenous disturbances caused by power supply each harmonic and other systems uncertainty；u_dAnd u_qInput quantity in order to control；

Define s_refFor given position signal,The desired value for making d shaft currents isIt, can basis when tracking Setting signal Speed, acceleration and the load disturbance of Setting signal calculate desired q shaft currents, utilize formula (10), (11) available q axis The desired value of electric current is：

<mrow> <msubsup> <mi>i</mi> <mi>q</mi> <mo>*</mo> </msubsup> <mo>=</mo> <mrow> <mo>(</mo> <mfrac> <msub> <mi>B</mi> <mi>v</mi> </msub> <mi>M</mi> </mfrac> <msub> <mi>v</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msub> <mo>+</mo> <msub> <mover> <mi>v</mi> <mo>&CenterDot;</mo> </mover> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msub> <mo>+</mo> <mfrac> <mn>1</mn> <mi>M</mi> </mfrac> <mi>f</mi> <mo>)</mo> </mrow> <mo>/</mo> <mfrac> <msub> <mi>k</mi> <mi>f</mi> </msub> <mi>M</mi> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>16</mn> <mo>)</mo> </mrow> </mrow>

Then displacement error：

Velocity error：

D, q shaft currents error：

The error state equation that PMLSM can be obtained by formula (12)~(19) is：

<mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mover> <mover> <mi>s</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mover> <mover> <mi>v</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <msub> <mi>k</mi> <mn>1</mn> </msub> <msub> <mover> <mi>i</mi> <mo>~</mo> </mover> <mi>q</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>2</mn> </msub> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mover> <mover> <mi>i</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mi>q</mi> </msub> <mo>=</mo> <msub> <mi>k</mi> <mn>3</mn> </msub> <msub> <mi>i</mi> <mi>q</mi> </msub> <mo>+</mo> <msub> <mi>k</mi> <mn>4</mn> </msub> <mi>v</mi> <mo>-</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <msub> <mi>u</mi> <mi>q</mi> </msub> <mo>+</mo> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>d</mi> </msub> <mo>+</mo> <msubsup> <mover> <mi>i</mi> <mo>&CenterDot;</mo> </mover> <mi>q</mi> <mo>*</mo> </msubsup> <mo>+</mo> <msub> <mi>d</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mover> <mover> <mi>i</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mi>d</mi> </msub> <mo>=</mo> <msub> <mi>k</mi> <mn>3</mn> </msub> <msub> <mover> <mi>i</mi> <mo>~</mo> </mover> <mi>d</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <msub> <mi>u</mi> <mi>d</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>q</mi> </msub> <mo>+</mo> <msubsup> <mover> <mi>i</mi> <mo>&CenterDot;</mo> </mover> <mi>d</mi> <mo>*</mo> </msubsup> <mo>+</mo> <msub> <mi>d</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>20</mn> <mo>)</mo> </mrow> </mrow>

Wherein,

Wherein, control law u is utilized_1f、u_2fDesign with the Nonlinear Dynamic of compensating line servo-drive system, u₁、u₂As sliding formwork control System restrains that system mode track is forced to tend to sliding-mode surface；

Define control law u_1fAnd u_2f：

<mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>u</mi> <mrow> <mn>1</mn> <mi>f</mi> </mrow> </msub> <mo>=</mo> <mrow> <mo>(</mo> <mrow> <mo>-</mo> <msub> <mi>k</mi> <mn>3</mn> </msub> <msub> <mi>i</mi> <mi>q</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>4</mn> </msub> <mi>v</mi> <mo>-</mo> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>d</mi> </msub> <mo>-</mo> <msubsup> <mover> <mi>i</mi> <mo>&CenterDot;</mo> </mover> <mi>q</mi> <mo>*</mo> </msubsup> </mrow> <mo>)</mo> </mrow> <mo>/</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>u</mi> <mrow> <mn>2</mn> <mi>f</mi> </mrow> </msub> <mo>=</mo> <mrow> <mo>(</mo> <mrow> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>q</mi> </msub> </mrow> <mo>)</mo> </mrow> <mo>/</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>22</mn> <mo>)</mo> </mrow> </mrow>

So formula (20) described error state equation can be converted into：

<mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mover> <mover> <mi>s</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mover> <mover> <mi>v</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <msub> <mi>k</mi> <mn>1</mn> </msub> <msub> <mover> <mi>i</mi> <mo>~</mo> </mover> <mi>q</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>2</mn> </msub> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mover> <mover> <mi>i</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mi>q</mi> </msub> <mo>=</mo> <mo>-</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>+</mo> <msub> <mi>d</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mover> <mover> <mi>i</mi> <mo>~</mo> </mover> <mo>&CenterDot;</mo> </mover> <mi>d</mi> </msub> <mo>=</mo> <msub> <mi>k</mi> <mn>3</mn> </msub> <msub> <mover> <mi>i</mi> <mo>~</mo> </mover> <mi>d</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <msub> <mi>u</mi> <mn>2</mn> </msub> <mo>+</mo> <msub> <mi>d</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>23</mn> <mo>)</mo> </mrow> </mrow>

Formula (23) can be written as：

<mrow> <mover> <mi>x</mi> <mo>&CenterDot;</mo> </mover> <mo>=</mo> <mi>A</mi> <mi>x</mi> <mo>+</mo> <mi>B</mi> <mrow> <mo>(</mo> <mi>u</mi> <mo>+</mo> <msub> <mi>u</mi> <mrow> <mi>d</mi> <mi>d</mi> </mrow> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>24</mn> <mo>)</mo> </mrow> </mrow>

In formula,

For the system design sliding-mode surface of formula (24)；DefinitionFor matrix B^TOrthonormal basis；Consider matrixFor formula (25) solution of two LMI in；

Make S=(B^TX^-1B)^-1B^TX^-1, then sliding-mode surface can be defined as：

σ=Sx=(B^TX^-1B)^-1B^TX^-1X=(σ₁ σ²)^T (26)

Then sliding formwork control ratio can be designed as：

U=-SAx-ksgn (σ)=u_eq+u_sw (27)

Wherein, k is constant, SA=G=(G₁ G₂)^T；u_eqFor Equivalent control law, u_swFor switching law；

<mrow> <mi>u</mi> <mo>=</mo> <msub> <mi>u</mi> <mrow> <mi>e</mi> <mi>q</mi> </mrow> </msub> <mo>+</mo> <msub> <mi>u</mi> <mrow> <mi>s</mi> <mi>w</mi> </mrow> </msub> <mo>=</mo> <mo>-</mo> <mi>S</mi> <mo>&CenterDot;</mo> <mi>A</mi> <mi>x</mi> <mo>-</mo> <mi>k</mi> <mrow> <mo>(</mo> <mfrac> <mi>&sigma;</mi> <mrow> <mo>|</mo> <mo>|</mo> <mi>&sigma;</mi> <mo>|</mo> <mo>|</mo> <mo>+</mo> <mi>&delta;</mi> </mrow> </mfrac> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>28</mn> <mo>)</mo> </mrow> </mrow>

By formula (21), (22), (38), control input u_dAnd u_qIt can be written as：

<mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>u</mi> <mi>q</mi> </msub> <mo>=</mo> <mrow> <mo>(</mo> <mo>-</mo> <msub> <mi>k</mi> <mn>3</mn> </msub> <msub> <mi>i</mi> <mi>q</mi> </msub> <mo>-</mo> <msub> <mi>k</mi> <mn>4</mn> </msub> <mi>v</mi> <mo>-</mo> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>d</mi> </msub> <mo>-</mo> <msubsup> <mover> <mi>i</mi> <mo>&CenterDot;</mo> </mover> <mi>q</mi> <mo>*</mo> </msubsup> <mo>)</mo> </mrow> <mo>/</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <mo>-</mo> <msub> <mi>G</mi> <mn>1</mn> </msub> <mi>x</mi> <mo>-</mo> <mi>k</mi> <mrow> <mo>(</mo> <mfrac> <msub> <mi>&sigma;</mi> <mn>1</mn> </msub> <mrow> <mo>|</mo> <mo>|</mo> <mi>&sigma;</mi> <mo>|</mo> <mo>|</mo> <mo>+</mo> <mi>&delta;</mi> </mrow> </mfrac> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>u</mi> <mi>d</mi> </msub> <mo>=</mo> <msub> <mi>k</mi> <mn>6</mn> </msub> <msub> <mi>vi</mi> <mi>q</mi> </msub> <mo>/</mo> <msub> <mi>k</mi> <mn>5</mn> </msub> <mo>-</mo> <msub> <mi>G</mi> <mn>2</mn> </msub> <mi>x</mi> <mo>-</mo> <mi>k</mi> <mrow> <mo>(</mo> <mfrac> <msub> <mi>&sigma;</mi> <mn>2</mn> </msub> <mrow> <mo>|</mo> <mo>|</mo> <mi>&sigma;</mi> <mo>|</mo> <mo>|</mo> <mo>+</mo> <mi>&delta;</mi> </mrow> </mfrac> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>.</mo> </mrow>

6. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 4 Tracking control unit, it is characterised in that：Step 4：

Here a Neural Network Observer in parallel on the basis of traditional observer, exports defeated with linear thrust observer Go out be added, the two and as final observation；

<mrow> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mover> <mover> <mi>v</mi> <mo>^</mo> </mover> <mo>&CenterDot;</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mover> <mover> <mi>f</mi> <mo>^</mo> </mover> <mo>&CenterDot;</mo> </mover> </mtd> </mtr> </mtable> </mfenced> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mi>d</mi> </mtd> <mtd> <mi>a</mi> </mtd> </mtr> <mtr> <mtd> <mn>0</mn> </mtd> <mtd> <mn>0</mn> </mtd> </mtr> </mtable> </mfenced> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mover> <mi>v</mi> <mo>^</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mover> <mi>f</mi> <mo>^</mo> </mover> </mtd> </mtr> </mtable> </mfenced> <mo>+</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mi>b</mi> </mtd> </mtr> <mtr> <mtd> <mn>0</mn> </mtd> </mtr> </mtable> </mfenced> <msub> <mi>i</mi> <mi>q</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>30</mn> <mo>)</mo> </mrow> </mrow>

Output equation is：

In formula (30), a=-B_v/ M, b=-K_f/ M, d=-1/M,

By equation (30), (31) discretization, can obtain：

Wherein, G, H, C are respectively state equation (30), the coefficient matrix corresponding to (31)；L=(L₁ L₂)^TFor the increasing of observer Beneficial matrix, the principle that the characteristic value of system is must be considerably larger than according to the characteristic value of observer chooses L gain battle arrays, so that observation system Rapid decay；NN is Neural Network Observer in parallel, is selected herein with 1 input node, 8 hidden nodes, 2 outputs 3 layers of BP neural network of node；Define network target function be：

The weighed value adjusting formula of neutral net is：

W_j1(k+1)=W_j1(k)+ZW_ib_j(k), j=1~8, i=1~2；

W_ij(k+1)=W_ij(k)+ZW_ji_q(k)

<mrow> <msub> <mi>W</mi> <mi>i</mi> </msub> <mo>=</mo> <mo>&lsqb;</mo> <mi>y</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mover> <mi>y</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>&rsqb;</mo> <msub> <mi>&Delta;x</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>&times;</mo> <mo>&lsqb;</mo> <mn>1</mn> <mo>-</mo> <msub> <mi>&Delta;x</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>&rsqb;</mo> </mrow>

W_j=b_j(k)[1-b_j(k)]W_iW_j1 (34)

In formula：Z --- learning rate；

W_j1--- the weights of hidden layer to output layer；

W_ij--- it inputs to the weights of hidden layer；

Δx_i(k+1) --- i-th of system mode increment size of network output；

b_j(k) --- the output of j-th of hidden node of network；

7. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 1 Tracking control unit, it is characterised in that：DSP uses model TMS320F28335 processors, and the control method main program is in DSP Middle realization step is as follows：

Step 1 control program starts；

Step 2 system initialization；

Step 3 allows INT1, INT2 to interrupt；

Step 5 program data initializes；

Step 6 opens total interruption；

Step 7 interrupt latency；

Step 8 judges whether to terminate to run, if not return to step 7；

Step 9 terminates.

8. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 7 Tracking control unit, it is characterised in that：INT1 interrupt processings son control program is according to the following steps：

Step 1T1 Interrupt Subroutines；

Step 2 keeps the scene intact；

Step 3 current sample；

Step 4CLARK is converted；

Step 5PARK is converted；

Step 6 disturbance estimation programme-control；

Step 7d q shaft voltages are adjusted；

Step 8PARK inverse transformations；

Step 9 calculates CMPPx and PWM outputs；

Step 10 position samples；

Step 11 initial position program；

Step 12 restoring scene；

Step 13, which is interrupted, to be returned.

9. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 8 Tracking control unit, it is characterised in that：The step 6 disturbance estimation control program is according to the following steps：

Step 1 thrust disturbance observer program starts；

Step 2 reads initialization data；

Step 3 reads sampled value, that is, measured value of initial time speed；

Step 4 sets the initial parameter of Systems with Linear Observation device；

Step 6 calculating speed measured value and the difference of NNTO outputs；

Step 7 is by system discretization；

Step 9 calculates the state estimator of Parallel neural networks observer；

Step 11 is by system serialization；

Step 12 calculates the output quantity of NNTO；

Step 13 calculates and the estimator of output disturbance；

Step 14 returns to caller.

10. a kind of linear servo position being combined based on linear matrix inequality with sliding formwork control according to claim 7 Put tracking control unit, it is characterised in that：

QEP interruption positions adjust processing son control program according to the following steps：

Step 1QEP interrupt service subroutines start；

Step 2 keeps the scene intact；

Step 3 reads initialization data；

Step 4 reads initial time position, speed, d shaft currents, the sampled value (measured value) of q shaft currents,

Step 5 calculates the desired value of q shaft currents and speed；

Step 7 sets LMI- sliding formwork initial parameters；

Step 9 calculates switching control u_sw；

Step 10 calculates equivalent control u_eq；

Step 11 calculates Linear Control rule u_1f、u_2f；

Break in step 13 Central Shanxi Plain；

Step 14 restoring scene；

Step 15, which is interrupted, to be returned.