CN105577058A - Novel fuzzy active disturbance rejection controller based five-phase fault-tolerant permanent magnet motor speed control method - Google Patents

Abstract

The invention discloses a novel fuzzy active disturbance rejection controller based five-phase fault-tolerant permanent magnet motor speed control method, and designs the novel fuzzy active disturbance rejection controller. A rotating speed differential value generated by a tracking differentiator and a system disturbance value observed by a linear expansion state observer are taken as the input of a fuzzy logic theory machine; and the output bandwidth value omega<c> of the fuzzy logic theory machine is taken as bandwidth input of a proportional controller. The fuzzy controller can change parameters of the controller according to the working conditions of the system in real time; the design difficulty of the controller is lowered, and the controller parameters can be adjusted in real time according to the running working conditions of the system; the novel tracking differentiator of the novel fuzzy active disturbance rejection controller ensures that the motor has no overshoot rapid response in the whole dynamic process; and compared with a conventional linear active disturbance rejection controller, the novel fuzzy active disturbance rejection controller designed in the invention has strong disturbance resisting capacity and adaptive capacity for complex working conditions, and the excellent dynamic performance.

Description

Based on five mutually fault-tolerant magneto method for control speed of novel fuzzy automatic disturbance rejection controller

Technical field

The present invention relates to a kind of fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller, be applicable to magneto high-precision servo control field.

Background technology

Along with the application of permagnetic synchronous motor in SERVO CONTROL field is more and more extensive, its requirement for control precision is also more and more higher, and the most important factor wherein affecting magneto control precision is the disturbance of system.The disturbance of system is normally perturbed by the inner parameter of system and its exterior interference forms, and the existence of these disturbances can make system become unstable.In addition, when control system is in different operating conditions, need there is corresponding change in the parameter of controller, and this just needs the parameter of carrying out manual adjustment controller.If cannot realize parameter to regulate in real time along with the change of operating mode, the unperturbed that just cannot realize system runs.Therefore, in order to improve the precision of magneto speed control system further, controller not only will have very strong interference rejection ability, and will have very strong adaptive capacity.

In traditional permagnetic synchronous motor Vector Speed-Control System, the double closed-loop control system of usual employing speed outer shroud and current inner loop, and controller all adopts conventional PID controllers.Although PID controller has simplicity of design, be easy to the advantages such as grasp, because conventional PID controllers exists overshoot greatly, the shortcomings such as the response time is long, and Ability of Resisting Disturbance is poor, can not reach satisfied control effects usually.

In order to solve the problem that automatic disturbance rejection controller parameter is difficult to regulate, Chinese invention patent application number is that the Digital Servo Sysem based on fuzzy automatic disturbance rejection controller of design in 201310129388.0 " a kind of PMSM servo system control methods based on fuzzy Active Disturbance Rejection Control " is using the input of the differential of speed error and speed error as obfuscation module, using the output of the controling parameters of three in nonlinearity erron Feedback Control Laws as obfuscation module, the yield value in extended state observer regulates in advance.Although decrease the regulating parameter of system like this, but still do not eliminate the adjustment problem of parameter completely, and due to speed error is had certain hysteresis quality as the input of fuzzy control logic for control system, cannot according to the parameter of the real-time conditioning controller of the disturbance of system.At document (J.Gai, S.Huang, Q.Huang, M.Li, H.Wang; D.Luo, X.WuandW.Liao. " Anewfuzzyactive-disturbancerejectioncontrollerappliedinP MSMpositionservosystem, " 17 ^thinternationalConferenceonElectricalMachinesandSystems (ICEMS2014), pp:2055-2059,2014.) in devise a kind of fuzzy automatic disturbance rejection controller of second order position ring, decrease the required parameter regulated of automatic disturbance rejection controller, can stable operation within the specific limits.The document uses the error of given position and feedback position and differential signal as the input of fuzzy controller, and the output of fuzzy controller is three input signals of nonlinear state Error Feedback, does not thoroughly solve automatic disturbance rejection controller parameter and regulates problem.

Summary of the invention

The object of the invention is for the deficiencies in the prior art, propose a kind of five mutually fault-tolerant permanent magnet motor system control methods based on novel fuzzy Auto Disturbances Rejection Control Technique.The method fully combines the advantage of linear active disturbance rejection (LADRC) technology and fuzzy control technology, controller parameter can be made along with the adjustment of motor operating condition real-time adaptive, ensure that in five mutually fault-tolerant magneto dynamic processes and respond rapidity and non-overshoot, enhance performance of noiseproof and the robust stability of system simultaneously.

The technical solution used in the present invention has following steps:

Based on a fault-tolerant magneto method for control speed for novel fuzzy automatic disturbance rejection controller, comprise the following steps:

S1, is obtained the angular position theta of five mutually fault-tolerant permanent magnet machine rotors, and is obtained the speed omega of rotor by differential calculation by photoelectric coded disk;

S2, detects five phase current i of five mutually fault-tolerant magnetoes _a, i _b, i _c, i _d, i _e, utilize rotor position information θ and obtain direct-axis current i through 5s/2r (Clark-Park) conversion _dwith quadrature axis current i _q;

S3, according to speed setting value ω ^*the set-point of quadrature axis current is obtained through novel fuzzy automatic disturbance rejection controller with speed feedback value ω

S4, direct-axis current is given as 0, and quadrature axis current is given as the output valve of der Geschwindigkeitkreis controller with current feedback values i _dand i _qpoor respectively, difference obtains direct-axis voltage U respectively through PI controller _dwith quadrature-axis voltage U _q;

S5, utilizes rotor position information, to direct-axis voltage U _dwith quadrature-axis voltage U _qcarry out the anti-Park conversion of 2r/2s, obtain alpha-beta shaft voltage U _αand U _β;

S6, U _αand U _βas the input of SVPWM module, produce 10 road pwm pulses, control the voltage that five phase voltage source inventers produce five phase pulse width variation, drive five mutually fault-tolerant magnetoes to rotate.

Further, in described step S3, the design procedure of fuzzy automatic disturbance rejection controller is as follows:

S3.1, given motor speed value ω ^*, adopt novel Nonlinear Tracking Differentiator 1, obtain the real-time set-point v of motor speed ₁;

S3.2, utilizes linear extended state observer 2 to observe load torque disturbed value d, adopts Nonlinear Tracking Differentiator 2 from the tach signal of reality, obtain the Differential of Speed signal ec of noise-less pollution;

S3.3, sends into the load torque disturbed value d that observes and Differential of Speed signal ec in fuzzy logic inference machine, precisely controlled device parameter value, they respectively: controller bandwidth value ω _c, linear extended state observer 1 two yield value β ₁₁and β ₁₂; Relation between them meets: β ₁₁=2* ω ₀, ω ₀=(4 ~ 5) ω _c;

S3.4, by tach signal ω and q shaft current set-point be multiplied by the value that b obtains to send in linear extended state observer 1, then by two yield value β of the linear extended state observer 1 obtained in step 3.3 ₁₁and β ₁₂, calculate the measured value z of rotational speed omega ₁with the disturbed value z of system ₂;

S3.5, by by the real-time set-point v of the motor speed obtained in step 3.1 ₁with the motor speed measured value z obtained in step 3.4 ₁making difference sends in proportional controller, according to the controller bandwidth value ω obtained in step 3.3 _cadjustment obtains controlled quentity controlled variable u ₀, the scale parameter of proportional controller is k _p=ω _c;

S3.6, by by the controlled quentity controlled variable u obtained in step 3.5 ₀deduct the system disturbance value z obtained in step S3.4 ₂with the business of b, obtain real controlled quentity controlled variable sent in the current loop control system of motor in synchrony rotating coordinate system, drive fault-tolerant magneto to run.

Further, in described step S3.1, novel Nonlinear Tracking Differentiator 1 comprises to rotating speed transition process arranging: rotary speed setting value ω ^*rotating speed transition signal v is obtained through designed novel Nonlinear Tracking Differentiator 1 ₁, this transient process function representation is

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_1=v_2\\ {\stackrel{\&CenterDot;}{v}}_2=k^2({\mathrm{\&omega;}}^{*}-v_1)-2{\mathrm{kv}}_2\end{array}\right.$

Wherein, ω ^*for rotary speed setting value, v ₁for rotating speed transition signal, v ₂for the differential signal of transition signal, k is Turbo Factor function, and its expression formula is

$k=\left\{\begin{array}{c}\frac{t}{T_1}{a}_t,0<t\&le;T_1\\ a_t,T_1<t\&le;T_2\\ a_t\left(\frac{T_3-t}{T_3-T_2}\right),T_2<t\&le;T_3\\ a_t,T_3<t\\ a_t=a_{\max }+h\hat{f}\end{array}\right.$

Wherein, t represents system operation time, T ₁, T ₂, T ₃represent three time points of system cloud gray model respectively, the relation between them is determined by motor electrical time constant and mechanical time constant, a _maxexpression system peak acceleration, a _texpression system real time acceleration, f represents the real-time Assumption torque disturbance of linear active disturbance rejection controller, and h is torque feedback coefficient.T ₁＝0.005、T ₂＝0.045、T ₃＝0.05、a _max＝330、h＝0.25。

T ₁＝0.005、T ₂＝0.045、T ₃＝0.05、a _max＝330、h＝0.25

Further, in described step S3.2, the expression formula of Nonlinear Tracking Differentiator 2 is:

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_3=v_4\\ {\stackrel{\&CenterDot;}{v}}_4=k_2^2(\mathrm{\&omega;}-v_3)-2k_2{v}_4\end{array}\right.$

Wherein, k ₂for Turbo Factor constant, ω represents motor speed actual value, v ₃for actual speed transition signal, v ₄for actual speed differential signal, i.e. required Differential of Speed signal ec=v in described step S3.3 ₄, wherein parameter k ₂be chosen for k ₂=320.

Further, the expression formula of linear extended state observer 2 is:

$\left\{\begin{array}{c}{e}_2=z_3-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_3=z_4-{\mathrm{\&beta;}}_{21}{e}_2+u\\ {\stackrel{\&CenterDot;}{z}}_4=-{\mathrm{\&beta;}}_{22}{e}_2\end{array}\right.$

Wherein, e ₂for the difference of measured value and actual value, z ₃speed observation value, z ₄the total disturbance of system, i.e. torque disturbed value d=z in step S3.2 ₄, u is automatic disturbance rejection controller output signal β ₂₁, β ₂₂for the yield value of linear extended state observer 2.Wherein observer gain value β ₂₁, β ₂₂be chosen for β ₂₁=3600, β ₂₂=1000000.

Further, the expression formula of described step S3.3 neutral line extended state observer 1 is:

$\left\{\begin{array}{c}{e}_1=z_1-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_1=z_2-{\mathrm{\&beta;}}_{11}{e}_1+bu\\ {\stackrel{\&CenterDot;}{z}}_2=-{\mathrm{\&beta;}}_{12}{e}_1\end{array}\right.$

Wherein, e ₁for the difference of measured value and actual value, z ₁speed observation value, z ₂system disturbance measured value, system disturbance value u is automatic disturbance rejection controller output signal β ₁₁, β ₁₂for the yield value of linear extended state observer 1.

Further, in described step S3.3, indistinct logic computer relevant arranges as follows:

Load torque disturbed value d is set to [0,20Nm], its scope is treated to [-10Nm, 10Nm], then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest];

The excursion of Differential of Speed signal parameter ec is [-0.5,3.5], its scope is treated to [-2,2] at input, then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest];

Controller bandwidth value ω _cthe i.e. output valve of fuzzy logic inference machine, its obfuscation language is [NB, NS, ZO, PS, PB], and corresponding de-fuzzy quantizing factor is [-2 ,-1,0,1,2], and corresponding accurate output area is ω _c∈ [60,140]; Differential of Speed signal parameter ec, load torque disturbed value d and controller bandwidth value ω _cmembership function be isosceles triangle membership function:, the expression formula of isosceles triangle membership function is:

$\mathrm{\&mu;}\left(x\right)=\left\{\begin{array}{c}\frac{x-a}{b-a},a<x<b\\ \frac{x-c}{b-c},b<x<c\end{array}\right.$

Wherein: parameter a, relation between b, c reflect shape and the distribution of membership function, condition that their meet is b-a=c-b.

The present invention has following beneficial effect:

1) the physics peak acceleration allowed according to motor vector control system arranges a fast and stably transient process, and utilize linear active disturbance rejection controller to carry out real-Time Compensation to acceleration, make motor realize non-overshoot in whole dynamic process to respond fast, solve the contradiction of electric machine control system rapidity and overshoot veritably, ensure that motor keeps stable state (error of given speed signal and feedback speed signal is very little) in dynamic process always, enhance the antijamming capability of system in dynamic (comprising startup) process simultaneously, there is good dynamic and static state performance.

2) utilize the load torque disturbing signal d of the system that observes that extended state observer 2 can be real-time, reflection system that simultaneously can be real-time is by the information of external disturbance; Because Nonlinear Tracking Differentiator 2 can suppress the noise of tach signal effectively, obtain reliable Differential of Speed signal, effective filtering external interference, rotating speed step and measurement noises etc., thus the situation avoiding differential signal super large and cannot use; The torque disturbance signal d that the Differential of Speed signal ec obtained by Nonlinear Tracking Differentiator 2 and linear expansion state observer 2 obtain gives fuzzy controller, ensure that the Static and dynamic information of the acquisition system change that fuzzy controller can be real-time.Nonlinear Tracking Differentiator, linear extended state observer and fuzzy controller combine and effectively can obtain system Static and dynamic performance information accurately; Abandon in traditional fuzzy controller the method directly utilizing the method for tach signal error to be used as fuzzy controller input.The conditioning controller parameter value utilizing fuzzy controller real-time, enhances system robustness, turn improves the adaptive capacity of system for different operating mode simultaneously.

3) owing to adopting linear active disturbance rejection controller, except Nonlinear Tracking Differentiator, the parameter that controller need regulate only has a controller bandwidth value ω _c, observer bandwidth omega ₀with controller bandwidth omega _cfollowing rule need be met: ω ₀=(4 ~ 5) ω _c.Fuzzy reasoning table only needs one, simplifies the design of controller, the function that the parameter achieving controller regulates in real time with operating mode.The disturbance information (uncertainty, moment of friction, load torque as system modelling) estimated in system cloud gray model that employing linear active disturbance rejection controller can be real-time simultaneously, by the feedforward of its compensation rate to controller output end, the interference free performance of system can be improved.

4) novel Nonlinear Tracking Differentiator is adopted, fuzzy controller, extended state observer and linear active disturbance rejection controller combine form novel fuzzy automatic disturbance rejection controller, the operating mode residing for system that can be real-time is different and change the parameter value of automatic disturbance rejection controller, the population parameter achieving system regulates, overcome the shortcoming of manual adjustments parameter, achieve the unperturbed operation of system and the rotational speed regulation of quick non-overshoot, the error of the system rotating speed of ensure that under any circumstance all and between given rotating speed is very little, effectively prevent the overshoot of system, improve the rapidity of system responses, enhance the robustness of system, improve dynamic property and the steady-state behaviour of system.

Accompanying drawing explanation

Fig. 1 is the five mutually fault-tolerant magneto control structure block diagrams adopting novel fuzzy automatic disturbance rejection controller;

The mutually fault-tolerant cross-sectional view of permanent magnet electric machine of Fig. 2 five;

Fig. 3 is five mutually fault-tolerant magneto quadrature axis equivalent circuit diagrams;

Fig. 4 is Turbo Factor function waveform figure in novel Nonlinear Tracking Differentiator;

Fig. 5 is conventional linear automatic disturbance rejection controller block diagram

Fig. 6 is novel fuzzy automatic disturbance rejection controller block diagram;

Fig. 7 is the load torque and observation torque profile figure that do not apply in the same time;

Fig. 8 is Differential of Speed oscillogram;

Fig. 9 is membership function figure;

Figure 10 is fuzzy control rule table;

Figure 11 is having the rotating speed response oscillogram adopting traditional automatic disturbance rejection controller in load torque step situation;

Figure 12 is having the rotating speed response oscillogram adopting novel fuzzy automatic disturbance rejection controller in load torque step situation;

Figure 13 is the Random Load torque disturbance figure applied;

Figure 14 is the rotating speed response comparison of wave shape figure when there being Random Load torque disturbance;

Figure 15 is the rotating speed step response waveform figure adopting traditional automatic disturbance rejection controller under no-load condition;

Figure 16 is the rotating speed step response waveform figure adopting novel fuzzy automatic disturbance rejection controller under no-load condition.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described.

Be described below in detail embodiments of the invention, the example of described embodiment is shown in the drawings, and wherein same or similar label represents same or similar element or has element that is identical or similar functions from start to finish.Being exemplary below by the embodiment be described with reference to the drawings, only for explaining the present invention, and can not limitation of the present invention being interpreted as.

As shown in Figure 1, a kind of fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller of the present invention, comprises following rate-determining steps:

S1, is obtained the angular position theta of five mutually fault-tolerant permanent magnet machine rotors, and is obtained the speed omega of rotor by differential calculation by photoelectric coded disk;

S2, detects five phase current i of five phase fault tolerant permanent magnet machines _a, i _b, i _c, i _d, i _e, utilize rotor position information θ and obtain direct-axis current i through 5s/2r (Clark-Park) conversion _dwith quadrature axis current i _q;

S3, according to speed setting value ω ^*the set-point of quadrature axis current is obtained by fuzzy automatic disturbance rejection controller with speed feedback value ω

S4, direct-axis current be given as 0 and quadrature axis current be given as the output valve of der Geschwindigkeitkreis controller respectively with the value of feedback i of electric current _dand i _qdiffer from, difference is obtained direct-axis voltage U through PI controller respectively _dwith quadrature-axis voltage U _q;

S5, utilizes rotor position information, to direct-axis voltage U _dwith quadrature-axis voltage U _qcarry out the anti-Park conversion of 2r/2s, obtain alpha-beta shaft voltage U _αand U _β;

S6, U _αand U _βas the input of SVPWM module, produce 10 road pwm pulses, control Five-phase inverter and produce five cross streams voltages, drive five mutually fault-tolerant magnetoes to rotate.

The present invention is based on rotor flux-orientation vector control technology, obtain A, B, C, D, E five phase current by current sensor, obtain direct-axis current i through 5s/2r (Clark-Park) conversion _dwith quadrature axis current i _q, as the value of feedback of two electric current loops; Calculated position and the rotating speed of rotor by photoelectric coded disk, utilize the positional information of rotor to carry out 5s/2r (Clark-Park) and 2r/2s (anti-Park) conversion; Rotating speed is as the feed back input of Fuzzy-LADRC (i.e. fuzzy automatic disturbance rejection controller) rotational speed governor; The output of der Geschwindigkeitkreis controller is as the reference input of q shaft current ring controller the output of current loop controller is as direct-axis voltage U _dwith quadrature-axis voltage U _qset-point, produce the voltage U under rest frame through 2r/2s (anti-Park) conversion _αand U _β; Two phase voltages produce pwm pulse through SVPWM module, and control inverter produces five cross streams voltages, drive five mutually fault-tolerant magnetoes to rotate.

Fig. 2 is the sectional view of five mutually fault-tolerant magnetoes.As shown in Figure 2, the V-shaped arrangement of select five mutually fault-tolerant permanent magnet of permanent magnet motor, permanent magnet is embedded in rotor.Stator winding adopts the centralized distribution of individual layer, can reduce copper loss, raise the efficiency.The alternately arrangement of the armature tooth of motor and fault-tolerant teeth, and the width of fault-tolerant teeth is less than armature tooth, optimizes back-emf with this, reduces cogging torque and Driving Torque pulsation simultaneously.The centralized winding of individual layer and fault-tolerant teeth structure can reduce being coupled between phase and phase, improve the fault freedom of motor.

As one embodiment of the present of invention, the present invention is on der Geschwindigkeitkreis linear active disturbance rejection controller basis, devise and adopt a kind of novel fuzzy automatic disturbance rejection controller for fault-tolerant magneto speed ring, adopt novel Nonlinear Tracking Differentiator to rotating speed transition process arranging, five mutually fault-tolerant magnetoes are made to be in stable state all the time, observe the Differential of Speed signal that load torque disturbance and Nonlinear Tracking Differentiator produce, the parameter that provide controller real-time by fuzzy logic inference machine according to linear extended state observer simultaneously.Specific embodiments comprises following steps:

1) rotary speed setting value ω ^*rotating speed transition signal v is obtained through designed novel Nonlinear Tracking Differentiator 1 ₁, this transient process function representation is

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_1=v_2\\ {\stackrel{\&CenterDot;}{v}}_2=k^2({\mathrm{\&omega;}}^{*}-v_1)-2{\mathrm{kv}}_2\end{array}\right.---\left(1\right)$

Wherein, ω ^*for rotary speed setting value, v ₁for rotating speed transition signal, v ₂for the differential signal of transition signal, k is Turbo Factor function, and its expression formula is

$k=\left\{\begin{array}{c}\frac{t}{T_1}{a}_t,0<t\&le;T_1\\ a_t,T_1<t\&le;T_2\\ a_t\left(\frac{T_3-t}{T_3-T_2}\right),T_2<t\&le;T_3\\ a_t,T_3<t\\ a_t=a_{\max }+h\hat{f}\end{array}\right.---\left(2\right)$

Wherein, t represents system operation time, T ₁, T ₂, T ₃represent three time points of system cloud gray model respectively, the relation between them is determined by motor electrical time constant and mechanical time constant, a _maxexpression system peak acceleration, a _texpression system real time acceleration, represent the real-time Assumption torque disturbance of linear active disturbance rejection controller, equal the z in linear extended state observer 1 ₂, h is torque feedback coefficient.

T ₁, T ₂, T ₃the reflection of three parameters be relation between current response and rotating speed response, to this point be considered when choosing this three parameters, therefore needing to find a kind of constant that can reflect electromechanics and electrical characteristic.

If mechanical time constant t _m, electrical time constant t _e, the concrete method of measurement of two constants is:

T _m: on winding, add step voltage U, measure when angular speed reaches 63.2% of maximum angular rate the time used;

T _e: rotor is fixed, winding adds step voltage U, measure the required time when electric current reaches 63.2% of maximum current.

Fig. 3 is five phase fault tolerant permanent magnet machine quadrature axis equivalent circuit diagrams, and known motor electrical time constant is mechanical time constant l _qfor the q axle inductance of magneto, R _sfor motor stator resistance, J is electric machine rotation inertia.K _eand K _tbe back electromotive-force constant and electromagnetic torque constant respectively, their computing formula is respectively with (wherein p represents the number of pole-pairs of motor, ψ _rrepresent the rotor permanent magnet magnetic flux of motor), K as calculated _t=23.8K _e, t _m≈ 0.032, t _e≈ 0.004, t _m/ t _e≈ 8.For convenience of calculation, ratio is taken as 10, therefore T ₁, T ₂, T ₃pass between three parameters is 8T ₁=8 (T ₃-T ₂)=T ₂-T ₁.

In the Nonlinear Tracking Differentiator function of the present invention's design, the parameter choose of transient process Turbo Factor function k is: T ₁=0.005, T ₂=0.045, T ₃=0.05, a _max=330, h=0.25.

Fig. 4 is the oscillogram of Turbo Factor function, therefrom can find out it and general direct current machine start or shock load time current waveform very similar.Utilize this Turbo Factor function, little (nothing) error that can realize in magneto dynamic process is run.

2) as shown in Figure 1, fault-tolerant permanent magnet machine rotor position θ is obtained by photoelectric encoder, 5s/2r coordinate transformation module changes phase current into d-q shaft current, adopts 2r/2s coordinate transformation module to change d-q shaft voltage into alpha-beta shaft voltage, and calculates rotational speed omega.

3) be illustrated in figure 6 the novel fuzzy automatic disturbance rejection controller structure chart that the present invention proposes, the design of novel fuzzy automatic disturbance rejection controller comprises following step:

S3.1, given motor speed value ω ^*, adopt novel Nonlinear Tracking Differentiator 1, obtain the real-time set-point v of motor speed ₁;

S3.2, utilizes linear extended state observer 2 to observe load torque disturbed value d, adopts Nonlinear Tracking Differentiator 2 from the tach signal of reality, obtain the Differential of Speed signal ec of noise-less pollution;

S3.3, sends into the load torque disturbed value d that observes and Differential of Speed signal ec in indistinct logic computer, precisely controlled device parameter value, they respectively: controller bandwidth value ω _c, linear extended state observer 1 two yield value β ₁₁and β ₁₂;

S3.4, by tach signal ω and q shaft current set-point be multiplied by the value that b obtains to send in linear extended state observer 1, then by two yield value β of the linear extended state observer 1 obtained in step 3.3 ₁₁and β ₁₂, calculate the measured value z of rotational speed omega ₁with the disturbed value z of system ₂;

S3.5, by by the real-time set-point v of the motor speed obtained in step 3.1 ₁with the motor speed measured value z obtained in step 3.4 ₁making difference sends in proportional controller, according to the controller bandwidth value ω obtained in step 3.3 _cadjustment obtains controlled quentity controlled variable u ₀;

S3.6, by by the controlled quentity controlled variable u obtained in step 3.5 ₀deduct the system disturbance value z obtained in step S3.4 ₂with the business of b, obtain real controlled quentity controlled variable sent in the control system of motor, drive fault-tolerant magneto to run.Detailed process is:

The first step, the determination of input variable and output variable

In novel fuzzy automatic disturbance rejection controller, the parameter value of proportional controller and second-order linearity extended state observer 1 is provided by fuzzy controller, the load torque disturbed value d that select linear extended state observer 2 observes and the Differential of Speed value ec that drawn by Nonlinear Tracking Differentiator 2 is as input variable, and output variable is the parameter value ω of controller _c, β ₁₁, β ₁₂, they are two yield values of proportional controller parameter value and linear extended state observer 1 respectively.

In order to observe the load torque disturbance of motor fast and accurately, what need the yield value of extended state observer 2 to regulate is comparatively large, and the expression formula of linear extended state observer 2 is:

$\left\{\begin{array}{c}{e}_2=z_3-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_3=z_4-{\mathrm{\&beta;}}_{21}{e}_2+u\\ {\stackrel{\&CenterDot;}{z}}_4=-{\mathrm{\&beta;}}_{22}{e}_2\end{array}\right.---\left(3\right)$

Wherein, e ₂for the difference of measured value and actual value, z ₃speed observation value, z ₄the total disturbance of system, i.e. disturbed value d=z in step S3.2 ₄, u is automatic disturbance rejection controller output signal β ₂₁, β ₂₂for the yield value of linear extended state observer 2.Wherein observer gain value β ₂₁, β ₂₂be chosen for β ₂₁=3600, β ₂₂=1000000.The expression formula of linear extended state observer 1 is:

$\left\{\begin{array}{c}{e}_1=z_1-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_1=z_2-{\mathrm{\&beta;}}_{11}{e}_1+bu\\ {\stackrel{\&CenterDot;}{z}}_2=-{\mathrm{\&beta;}}_{12}{e}_1\end{array}\right.$

Wherein, e ₁for the difference of measured value and actual value, z ₁speed observation value, z ₂system disturbance measured value, system disturbance value u is automatic disturbance rejection controller output signal β ₁₁, β ₁₂for the yield value of linear extended state observer 1.In order to obtain the differential signal of rotating speed accurately, adopt Nonlinear Tracking Differentiator 2, its expression formula is:

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_3=v_4\\ {\stackrel{\&CenterDot;}{v}}_4=k_2^2(\mathrm{\&omega;}-v_3)-2k_2{v}_4\end{array}\right.---\left(4\right)$

Wherein, k ₂for Turbo Factor constant, ω represents motor speed actual value, v ₃for actual speed transition signal, v ₄for actual speed differential signal, the Differential of Speed signal ec=v namely in step S3.3 ₄, parameter k ₂be chosen for k ₂=320.

Second step, the determination of input/output variable domain

The domain of torque disturbance value d is determined according to torque at rated load disturbance, and load torque scope is [0,20Nm], as shown in Figure 7, load torque can be estimated accurately at this scope internal linear extended state observer 2.As shown in Figure 8, because differential numerical value is comparatively large, the scope after processing Differential of Speed signal is [-0.5,3.5] to the oscillogram of Differential of Speed ec.

3rd step, the quantification of input/output argument

Load torque disturbed value d (Rated motor electromagnetic torque) is set to [0,20Nm], its scope is treated to [-10Nm, 10Nm], then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is exactly [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest].

The excursion of differential parameter ec is [-0.5,3.5], its scope is treated to [-2,2] at input, then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is exactly [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest].

In order to the precision that enhancement mode gelatinization controls, make it to combine with automatic disturbance rejection controller, selection input variable ec of the present invention, d and output variable ω _c, β ₁₁, β ₁₂membership function be isosceles triangle (trimf) function, as shown in Figure 9, the expression formula of isosceles triangle membership function is:

$\mathrm{\&mu;}\left(x\right)=\left\{\begin{array}{c}\frac{x-a}{b-a},a<x<b\\ \frac{x-c}{b-c},b<x<c\end{array}\right.---\left(5\right)$

Relation wherein between parameter a, b, c reflects shape and the distribution of membership function, and the condition that they meet is b-a=c-b.

The output parameter ω of controller _cwith ω _c∈ [60,140] parameter is example, is [NB, NS, ZO, PS, PB] in output obfuscation Language Processing, and correspond to quantizing factor for [-2 ,-1,0,1,2], the accurate output area of corresponding de-fuzzy is [60,140].

Due to two yield value β of linear extended state observer 1 ₁₁and β ₁₂with controller bandwidth value ω _cthere is the relation determined, that is, β ₁₁=2* ω ₀, ω ₀=(4 ~ 5) ω _c; So without the need to two yield value β to linear extended state observer 1 ₁₁and β ₁₂fuzzy processing, only need controller bandwidth value ω _c.Known through Matlab/Simulink simulation study, the parameter area that linear extended state observer 1 normally runs is β ₁₁∈ [600,1400], β ₁₂∈ [90000,490000], only needs a fuzzy controller.

4th step, fuzzy control rule

The parameter ω of controller is provided according to control law _cthe experience summed up is as follows:

A) the electric motor starting stage and rotating speed step sudden change dynamic process in, because needs actual speed can tracing preset rotating speed exactly, the parameter of observer is needed to tune up, owing to adopting novel Nonlinear Tracking Differentiator, the feature of the error of set-point and actual value need not be adopted judge, only need observe differential value change can (startup stage and rotating speed sudden change the stage larger).

B) when motor is subject to external torque disturbance, the parameter value of observer needs corresponding change greatly, can strengthen the performance of noiseproof of system like this.

Owing to determining that proportional controller parameter just can know the parameter of linear extended state controller 1, so only need two fuzzy rules above, concrete fuzzy logic control rule as shown in Figure 10.

4) according to speed preset value and speed feedback value, the speed ring based on novel fuzzy automatic disturbance rejection controller shown in Fig. 6 is analyzed.In Fig. 6, linear extended state observer 1 to export the rotating speed of fault-tolerant Control System of AC Permanent Magnet Synchronous and observation acquisition speed observation value z is carried out in the total disturbance of system ₁with disturbance estimated value z ₂

$\left\{\begin{array}{c}{e}_1=z_1-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_1=z_2-{\mathrm{\&beta;}}_{11}e+bu\\ {\stackrel{\&CenterDot;}{z}}_2=-{\mathrm{\&beta;}}_{12}e\end{array}\right.---\left(6\right)$

Wherein, e ₁for the difference of measured value and actual value, the differential signal of speed observation value, be the differential signal of the total disturbance of system, u is automatic disturbance rejection controller output signal, determined by system model, β ₁₁, β ₁₂for the yield value of linear extended state observer 1.

Above observer is write as following matrix form further:

$\left\{\begin{array}{c}\stackrel{\&CenterDot;}{z}=\left(\begin{array}{c}{\stackrel{\&CenterDot;}{z}}_1\\ {\stackrel{\&CenterDot;}{z}}_2\end{array}\right)=\left(\begin{array}{cc}-{\mathrm{\&beta;}}_{11}& 1\\ -{\mathrm{\&beta;}}_{12}& 0\end{array}\right)\left(\begin{array}{c}{z}_1\\ z_2\end{array}\right)+\left(\begin{array}{cc}b& {\mathrm{\&beta;}}_{11}\\ 0& {\mathrm{\&beta;}}_{12}\end{array}\right)\left(\begin{array}{c}u\\ n\end{array}\right)\\ z=\left(\begin{array}{c}{z}_1\\ z_2\end{array}\right)=\left(\begin{array}{cc}1& 0\\ 0& 1\end{array}\right)\left(\begin{array}{c}{z}_1\\ z_2\end{array}\right)+\left(\begin{array}{cc}0& 0\\ 0& 0\end{array}\right)\left(\begin{array}{c}u\\ n\end{array}\right)\end{array}\right.---\left(7\right)$

Wherein, $\begin{array}{cccc}A=\left(\begin{array}{cc}-{\mathrm{\&beta;}}_{11}& 1\\ -{\mathrm{\&beta;}}_{12}& 0\end{array}\right),& B=\left(\begin{array}{cc}b& {\mathrm{\&beta;}}_{11}\\ 0& {\mathrm{\&beta;}}_{12}\end{array}\right),& C=\left(\begin{array}{cc}1& 0\\ 0& 1\end{array}\right),& D=\left(\begin{array}{cc}0& 0\\ 0& 0\end{array}\right)\end{array}$

Observer proper polynomial is

$|\mathrm{\&lambda;}E-A|=\left|\begin{array}{cc}\mathrm{\&lambda;}+{\mathrm{\&beta;}}_{11}& -1\\ {\mathrm{\&beta;}}_{12}& \mathrm{\&lambda;}\end{array}\right|=\mathrm{\&lambda;}(\mathrm{\&lambda;}+{\mathrm{\&beta;}}_{11})+{\mathrm{\&beta;}}_{12}={\mathrm{\&lambda;}}^2+{\mathrm{\&beta;}}_{11}\mathrm{\&lambda;}+{\mathrm{\&beta;}}_{12}={(\mathrm{\&lambda;}+{\mathrm{\&omega;}}_0)}^2---\left(8\right)$

Observer is regarded as a second-order system, systematic function (as stability, response speed etc.) be made to meet some requirements, need adjustment feature polynomial parameters β ₁₁, β ₁₂.Because systematic function and characteristic root (i.e. the limit of ssystem transfer function) have much relations, observer energy reliable observation system mode be made, will make system features root entirely on the left of complex plane.The speed of system features root absolute value larger observer observer state amount is faster, but conference causes concussion, so two of system gain parameters are set to β excessively ₁₁=2 ω ₀, make two POLE PLACEMENT USING of ssystem transfer function at-ω ₀, ω ₀for observer bandwidth.

Linearity error Feedback Control Laws is by rotating speed transition signal v ₁measured value z is exported with rotating speed ₁error e ₂original control signal u is obtained through proportional controller ₀, then the controlled quentity controlled variable u of automatic disturbance rejection controller is obtained through disturbance compensation

$\left\{\begin{array}{c}{e}_2=v_1-z_1\\ u_0=k_p{e}_2\\ u=u_0-\frac{z_2}{b}\end{array}\right.---\left(9\right)$

Wherein, k _pfor proportional controller gain, usually get k _p=ω _c, ω _cfor controller bandwidth.

Observer bandwidth omega ₀with controller bandwidth omega _cfollowing rule need be met

ω ₀＝(4～5)ω _c(10)

In the present invention, the parameter of the fault-tolerant magneto chosen is: the number of phases is 5 phases, number of pole-pairs p=11, quadrature axis inductance L _q=0.48mH, stator resistance R _s=0.12 Ω, rotor flux ψ _f=0.06Wb, moment of inertia J=0.03kgm ²so, .The parameter ω that need regulate is only had in visible controller _c, all the other parameters are all relevant to it.

5) as control signal u (i.e. quadrature axis current set-point that automatic disturbance rejection controller exports by Fig. 1 ) as inner ring i _qthe input of controller is given obtains quadrature-axis voltage U _q, adopt i _d=0 control mode produces direct-axis voltage U _d.The transformation matrix being tied to rest frame through rotational coordinates obtains the voltage U of rest frame alpha-beta axle _α, U _β, sent into the PWM ripple signal that space voltage vector SVPWM modulation module produces each phase of motor, the rotating speed realizing motor controls.

6) in order to the advantage adopting this kind of novel fuzzy linear active disturbance rejection controller is described, in Matlab/Simulink, built simulation model, and itself and conventional linear automatic disturbance rejection controller have been carried out comparative analysis.As shown in Figure 5, wherein conventional linear automatic disturbance rejection controller parameter choose is ω to conventional linear automatic disturbance rejection controller schematic diagram _c1=100, ω ₀₁=500, get k _p1=100, β ₁=1000, β ₂=250000, all the other parameters are identical with adopting the parameter of fuzzy automatic disturbance rejection controller.

In Fig. 7, dotted line is the load torque disturbance quantity do not applied in the same time, as can be seen from the figure, applies the load torque of 20Nm, load torque removed again, observing rotation speed change in such cases when 150ms when 80ms.Figure 11 adopts conventional linear automatic disturbance rejection controller rotating speed response waveform, and can find out, when torque is risen, the maximum pulsation of rotating speed reaches 11r/min, and recovery time is 15ms; And after removing load, fluctuation of speed amount reaches 8r/min, and the time returning to given rotating speed reach 25ms; Figure 12 is the rotating speed response waveform of the novel fuzzy automatic disturbance rejection controller adopting this patent, can find out, significantly change does not occur rotating speed, and the moment fluctuation of speed of load torque impact only has 3r/min, and only has 5ms recovery time; Load torque is prominent when unloading, and the fluctuation of speed only has 3r/min, and rotating speed only has 12ms recovery time.Based on the above, relative to conventional linear controller, the novel fuzzy automatic disturbance rejection controller of the present invention's design has stronger Ability of Resisting Disturbance, and shortens the time of system responses disturbance.

In order to investigate the performance of system under complex working condition, be the Random Load torque profile figure applied as shown in figure 13, as can be seen from the figure, load torque change very fast, can the performance of test controller by this.If Figure 14 is the enlarged drawing that speed waveform that traditional automatic disturbance rejection controller and the present invention design novel fuzzy automatic disturbance rejection controller responds, represented by dotted arrows adopts the rotating speed response waveform of traditional automatic disturbance rejection controller, and solid line representative adopts the rotating speed response waveform of novel fuzzy automatic disturbance rejection controller.As can be seen from the figure, adopt the waveform of traditional automatic disturbance rejection controller to decline at most 4.5r/min, rise at most 2.5r/min, but recovery time is longer; And adopt the waveform of novel fuzzy automatic disturbance rejection controller to decline at most 2r/min, rise at most 1.6r/min, and recovery time is extremely short.Can be drawn by the above, novel fuzzy automatic disturbance rejection controller better can adapt to complex working condition.

As can be seen from Figure 15 and Figure 16, the time adopting traditional automatic disturbance rejection controller to start is 55ms, and after adopting novel fuzzy automatic disturbance rejection controller of the present invention, system start-up time is 50ms.Under no-load condition, motor speed is adjusted to 400r/min from 800r/min respectively, then is adjusted to from 400r/min time and the overshoot that 800/min observes its arrival stable state.It is 45ms that Figure 15 adopts the rotating speed of traditional automatic disturbance rejection controller to drop to 400r/min required time from 800r/min, and overshoot is 60r/min; Rising to 800r/min required time from 400r/min is 43ms, and overshoot is 20r/min; And Figure 16 adopts novel fuzzy automatic disturbance rejection controller rotating speed of the present invention to drop to 400r/min required time from 800r/min is 25ms, non-overshoot; Rising to 800r/min required time from 400r/min is 25ms, non-overshoot.Illustrate and adopt the system rotating speed response time of the novel fuzzy automatic disturbance rejection controller of the present invention's design to shorten relative to traditional automatic disturbance rejection controller, more crucially there is no overshoot, solve the contradiction between response rapidity and overshoot.

In sum, first the novel fuzzy active disturbance rejection rotational speed governor of the present invention's design adopts a kind of novel Nonlinear Tracking Differentiator to make the dynamic process of motor start-up procedure and velocity jump present little error state to given rotating speed signal, guarantees that whole dynamic process is close to stable state; Then the load torque disturbed value d observed according to the linear extended state observer 2 and Differential of Speed value ec produced by Nonlinear Tracking Differentiator 2, use practical engineering experience design fuzzy logic inference rule, set up the fuzzy if-then rules table of controller bandwidth value, after its de-fuzzy, precisely controlled parameter, realizes parameter online adaptive and regulates; Finally, the difference of the rotating speed given transient process rotating speed and linear extended state observer 1 observed is sent into proportional controller and is produced control signal, deduct the disturbing signal produced by linear extended state observer 1 again, form real control signal and send in electric current loop.This control method solves in electric system dynamic process the contradiction responded between rapidity and overshoot, the rotating speed of system under any operating mode is made all to have little error or error free, enhance system rejection to disturbance performance and robust performance, improve steady-state behaviour and the dynamic property of system, better can adapt to complex working condition.

Claims (7)

1., based on a fault-tolerant magneto method for control speed for novel fuzzy automatic disturbance rejection controller, it is characterized in that, comprise the following steps:

S1, is obtained the angular position theta of five mutually fault-tolerant permanent magnet machine rotors, and is obtained the speed omega of rotor by differential calculation by photoelectric coded disk;

S2, detects five phase current i of five mutually fault-tolerant magnetoes _a, i _b, i _c, i _d, i _e, utilize rotor position information θ and obtain direct-axis current i through 5s/2r (Clark-Park) conversion _dwith quadrature axis current i _q;

S3, according to speed setting value ω ^*the set-point of quadrature axis current is obtained by novel fuzzy automatic disturbance rejection controller with speed feedback value ω

S4, the output valve that direct-axis current is given as 0, quadrature axis current is given as der Geschwindigkeitkreis controller respectively with the value of feedback i of electric current _dand i _qdiffer from, difference obtains direct-axis voltage U respectively through PI controller _dwith quadrature-axis voltage U _q;

S5, utilizes rotor position information, to direct-axis voltage U _dwith quadrature-axis voltage U _dcarry out the anti-Park conversion of 2r/2s, obtain alpha-beta shaft voltage U _αand U _β;

S6, U _αand U _βas the input of SVPWM module, produce 10 road pwm pulses, control the voltage that Five-phase inverter produces five phase pulse width variation, drive five mutually fault-tolerant magnetoes to rotate.

2. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 1, it is characterized in that, in described step S3, the design procedure of novel fuzzy automatic disturbance rejection controller is as follows:

S3.1, given motor speed value ω ^*, adopt novel Nonlinear Tracking Differentiator 1, obtain the real-time set-point v of motor speed ₁;

S3.2, utilizes linear extended state observer 2 to observe load torque disturbed value d, adopts Nonlinear Tracking Differentiator 2 from the tach signal of reality, obtain the Differential of Speed signal ec of noise-less pollution;

S3.3, sends into the load torque disturbed value d observed and Differential of Speed signal ec in fuzzy logic inference machine, tries to achieve accurate controller parameter value: controller bandwidth value ω _c, linear extended state observer 1 two yield value β ₁₁and β ₁₂; Relation between them meets: β ₁₁=2* ω ₀, ω ₀=(4 ~ 5) ω _c;

S3.4, by tach signal ω, q shaft current set-point be multiplied by two the yield value β obtained in the value and step 3.3 that b obtains ₁₁and β ₁₂send into linear extended state observer 1, obtain measured value z ₁with the disturbed value z of system ₂;

S3.5, by by the real-time set-point v of the motor speed obtained in step 3.1 ₁with the motor speed measured value z obtained in step 3.4 ₁making difference sends in proportional controller, according to the controller bandwidth value ω obtained in step 3.3 _cadjustment obtains controlled quentity controlled variable u ₀, the scale parameter of proportional controller is k _p=ω _c;

S3.6, by by the controlled quentity controlled variable u obtained in step 3.5 ₀deduct the system disturbance value z obtained in step S3.4 ₂with the business of b, obtain real controlled quentity controlled variable sent in the current loop control system of motor in synchrony rotating coordinate system, drive fault-tolerant magneto to run.

3. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 2, it is characterized in that, in described step S3.1, novel Nonlinear Tracking Differentiator 1 makes rotary speed setting value ω ^*become rotating speed transition signal v ₁, its function is

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_1=v_2\\ {\stackrel{\&CenterDot;}{v}}_2=k^2({\mathrm{\&omega;}}^{*}-v_1)-2{\mathrm{kv}}_2\end{array}\right.$

Wherein, ω ^*for rotary speed setting value, v ₁for rotating speed transition signal, v ₂for the differential signal of transition signal, k is Turbo Factor function, and its expression formula is

$k=\left\{\begin{array}{c}\frac{t}{T_1}{a}_t,0<t\&le;T_1\\ a_t,T_1<t\&le;T_2\\ a_t\left(\frac{T_3-t}{T_3-T_2}\right),T_2<t\&le;T_3\\ a_t,T_3<t\\ a_t=a_{\max }+h\hat{f}\end{array}\right.$

Wherein, t represents system operation time, T ₁, T ₂, T ₃represent three time points of system cloud gray model respectively, the relation between them is determined by motor electrical time constant and mechanical time constant, a _maxexpression system peak acceleration, a _texpression system real time acceleration, represent the real-time Assumption torque disturbed value of linear active disturbance rejection controller, h is torque feedback coefficient.The parameter choose of described transient process Turbo Factor function k is: T ₁=0.005, T ₂=0.045, T ₃=0.05, a _max=330, h=0.25.

4. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 2, it is characterized in that, in described step S3.2, the expression formula of Nonlinear Tracking Differentiator 2 is:

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{v}}_3=v_4\\ {\stackrel{\&CenterDot;}{v}}_4=k_2^2(\mathrm{\&omega;}-v_3)-2k_2{v}_4\end{array}\right.$

Wherein, k ₂for Turbo Factor constant, ω represents motor speed actual value, v ₃for actual speed transition signal, v ₄for actual speed differential signal.

5. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 2, is characterized in that, the expression formula of described step S3.2 neutral line extended state observer 2 is:

$\left\{\begin{array}{c}{e}_2=z_3-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_3=z_4-{\mathrm{\&beta;}}_{21}{e}_2+u\\ {\stackrel{\&CenterDot;}{z}}_4=-{\mathrm{\&beta;}}_{22}{e}_2\end{array}\right.$

Wherein, e ₂for the difference of measured value and actual value, z ₃speed observation value, z ₄the total disturbance of system, torque disturbance value d=z ₄, u is automatic disturbance rejection controller output signal β ₂₁, β ₂₂for the yield value of linear extended state observer 2; Described observer gain value β ₂₁, β ₂₂be chosen for β ₂₁=3600, β ₂₂=1000000.

6. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 2, is characterized in that, the expression formula of described step S3.3 neutral line extended state observer 1 is:

$\left\{\begin{array}{c}{e}_1=z_1-\mathrm{\&omega;}\\ {\stackrel{\&CenterDot;}{z}}_1=z_2-{\mathrm{\&beta;}}_{11}{e}_1+bu\\ {\stackrel{\&CenterDot;}{z}}_2=-{\mathrm{\&beta;}}_{12}{e}_1\end{array}\right.$

Wherein, e ₁for the difference of measured value and actual value, z ₁speed observation value, z ₂system disturbance measured value, system disturbance value u is automatic disturbance rejection controller output signal β ₁₁, β ₁₂for the yield value of linear extended state observer 1.

7. the fault-tolerant magneto method for control speed based on novel fuzzy automatic disturbance rejection controller according to claim 2, is characterized in that, in described step S3.3, fuzzy logic inference machine relevant arranges as follows:

Load torque disturbed value d is set to [0,20Nm], its scope linearisation is adjusted to [-10Nm, 10Nm], then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest];

The excursion of Differential of Speed signal parameter ec is [-0.5,3.5], its scope linearisation is adjusted to [-2,2] at input, then quantizing factor is written as [-2 ,-1,0,1,2], corresponding in fuzzy language is [NB, NS, ZO, PS, PB], be expressed as [negative large, negative little, zero, just little, honest];

Controller bandwidth value ω _c, i.e. the output valve of fuzzy logic inference machine, its obfuscation language is [NB, NS, ZO, PS, PB], and corresponding quantizing factor is [-2 ,-1,0,1,2], and the corresponding accurate output area of de-fuzzy is ω _c∈ [60,140]; Differential of Speed signal ec, load torque disturbed value d and controller bandwidth value ω _cmembership function be isoceles triangle shape function, the expression formula of isosceles triangle membership function is:

$\mathrm{\&mu;}\left(x\right)=\left\{\begin{array}{c}\frac{x-a}{b-a},a<x<b\\ \frac{x-c}{b-c},b<x<c\end{array}\right.$

Wherein: parameter a, relation between b, c reflect shape and the distribution of membership function, condition that their meet is b-a=c-b.