CN103746630A - Active control method for low-frequency vibration of electric drive system - Google Patents

Abstract

The invention discloses an active control method for low-frequency vibration of an electric drive system. The active control method is characterized by comprising the following steps: determining a control scheme, measuring the rotating speed or acceleration of a motor according to the low-frequency vibration characteristic of the electric drive system, and designing an active control strategy; selecting a control algorithm, comparing the characteristics and current application statuses of various intelligent control algorithms, and determining a sliding mode variable structure control algorithm; defining a variable structure switching function, designing a variable structure control rate, exploring a key problem in the application of the sliding mode variable structure algorithm, namely, the root cause of vibration, and reasonably designing a switching function and a control rate; optimizing a switching function parameter according to a genetic algorithm, and applying a genetic algorithm thought to the optimization of the switching function parameter of a sliding mode variable structure in order to further eliminate vibration caused by sliding mode switching; determining a braking torque, and tracking a vibration signal through a controller to provide a brake torque for realizing active control on the low-frequency vibration of the electric drive system. The active control method has the advantages of good control effect, high adaptability and the like.

Description

A kind of Active Control Method for electric drive system low-frequency vibration

Technical field

Patent of the present invention relates to electric drive system low-frequency vibration control field, a kind of Active Control Method for electric drive system low-frequency vibration of invention application is realized the control to electric drive system low-frequency vibration, this method can be used for reducing electric drive system low-frequency vibration and noise, improves system works stability.

Background technology

Vibration is all electric equipments ubiquitous phenomenons in running, in electric drive system, motor is the same with other equipment, in operation process, can there is vibration in various degree, vibration is mainly manifested in the following aspects to the harm of electric drive system: increase energy consumption, dynamical system Efficiency Decreasing; Directly injury motor bearings, accelerates the wearing and tearing of motor bearings, has greatly shortened the useful life of bearing; Rotor magnetic pole is loosening, causes stator and rotor mutually to wipe and touch, thereby causes rotor bending, fracture; Motor end wiring is loosening, causes end winding phase mutual friction, and insulation resistance reduces, insulation is shortened useful life, causes insulator breakdown or be affected with the running of other equipment of electric drilling match when serious, causes some part loosening, even part of damage, causes the accident.Vibration and noise not only can make physical unit and equipment fatigue, lost efficacy or disturb sensation and the discriminating of other acoustical signal, cause mechanical driving device vibration; While surpassing certain limit, also can damage people's health, the noise of extra-heavy, even can cause building construction vibration.

The vibration control method adopting is at present mostly Passive Control mode, adopts the modes such as suspension system, vibration isolation material, spring, elastomeric material, and inconvenience is installed, easily aging.And it is better that dither is controlled to effect, but it is general that low-frequency vibration is controlled to effect.The passive supporting construction of original electric drive system mainly can only be isolated the low-frequency vibration that motor produces, and adopt electromagnetic actuator can significantly decay on this basis simple harmonic vibration that motor excitation produces.

Traditional control system, as: PI with PID controls and control and all can not respond fast the intrinsic vibration of mechanical driving part with the PI of resistance-trap filter, the hysteresis effect of its corrective action is equivalent to the excitation of this vibration, rather than makes it stable.Moreover traditional control system will be processed above-mentioned nonlinear problem, as: Backlash Nonlinearity problem is almost impossible.

Variable structure control method, as a kind of conventional nonlinear control method, for complicated strongly coupled system, there is good robustness, but its major defect is: slip plane adopts POLE PLACEMENT USING or quadratic performance optimum to design, rule of thumb provide boundary layer thickness and shiver to eliminate, therefore be difficult to obtain one group of preferably slip plane parameter and boundary layer thickness.

Summary of the invention

In view of the existing defect of above-mentioned prior art, the object of the invention is to, a kind of Active Control Method for electric drive system low-frequency vibration is proposed, its feature is, improve existing control algolithm, the synovial membrane of genetic algorithm optimization is become to structure algorithm to be applied in the low frequency ACTIVE CONTROL of electric drive system, and the anti-vibration control scheme of the active that provides low-frequency vibration, Active Control Method is according to the vibration signal detecting, apply certain control strategy, through calculating in real time, and then drive ram applies certain impact to controlling target, reach the object that suppresses or eliminate vibration, there is the effective of control, the advantages such as strong adaptability.By simulation result, show, the method is controlled successful to low-frequency vibration.

The object of the invention is to be realized by following technical scheme: a kind of Active Control Method for electric drive system low-frequency vibration, it is characterized in that, it comprises the following steps:

1) determine control program: electric drive system is mainly the passive support of motor, because the difference of motor model has different supporting forms and the strong point, on each strong point of motor, arrange an execution unit, by each execution unit of real-time control, adjust the acceleration that self exciting of decay motor produces, thereby the transmission of motor dynamic load to electric drive system that also just indirectly decayed; The main cause producing according to motor low-frequency vibration, described execution unit is selected electromagnetic actuator, while being connected with alternating current in electromagnetic actuator coil, electromagnetic action produces exciting force to oscillator, make it forced vibration, externally output is controlled and is used as power, this is used as power and directly acts on the vertical direction of motor, and equate with the vibration amplitude of motor self exciting force, during single spin-echo, these two power are cancelled out each other, and the vibration being caused by them also will be cancelled, the vibration of motor will be cancelled like this, and the power that is finally delivered to electric drive system is also zero;

2) select control algolithm: control algolithm is selected Sliding mode variable structure control algorithm, once electric drive system enters sliding mode, just interference and parameter perturbation have consistency to external world under certain condition, thereby have strong robustness, adaptivity,

3) definition becomes structure switching function, design becomes structure control rate:

Definition global sliding mode face is:

become structure control rate function,

u=k _d1·tanh(s)+k _d·s

K in formula _d1, k _dfor proportionality coefficient and k _d1, k _d>0;

4) according to genetic algorithm optimization switching function parameter: application genetic algorithm determines that switching function parameter and bound thickness slacken buffeting, and fitness function is selected: $J_{\mathrm{OBJ}}={\∫}_0^{\∞}(e{\left(t\right)}^{T}e\left(t\right)+{\mathrm{wy}}^2)\mathrm{dt},$

In formula: e (t)=y (t)-y _f, y _ffor desirable end-state value; W is weights, and performance index are minimized, and story hardness function f changes into:

by selecting, intersect, making a variation, select optimal solution, return to v=[c D], realize the optimization of sliding-mode surface parameter and bound thickness, wherein: c is that sliding formwork parameter and D are boundary layer thickness;

5) determine braking moment: according to frequency and the amplitude of motor oscillating, determine the braking moment that suppresses motor oscillating.

Described definite braking moment comprises, according to frequency and the amplitude of simplified model emulation vibration, determines the output of motor braking moment by the sliding formwork change algorithm of genetic algorithm optimization.

Described braking moment is the braking moment of vertical direction, and with the single spin-echo of motor oscillating, by electromagnetic actuator, apply a kind of reverse disturbing moment, to offset original disturbing moment.

The present invention compared with prior art has obvious advantage and beneficial effect.By technique scheme, to guarantee that sliding mode variable structure control method has disturbance to external world insensitive, when keeping good robustness, effectively slackened the system chatter that the continuous switching of structure brings, thereby avoided high frequency to switch the impact bringing to material resources system, and improved the precision of controlling.Simple general-purpose, strong robustness, be suitable for parallel processing, and because braking moment acts directly on motor shaft, effectiveness in vibration suppression is very obvious.By simulation result, show, the method is controlled successful to low-frequency vibration.

Accompanying drawing explanation

Fig. 1 is the correlation step that realizes electric drive system low-frequency vibration Active Control Method

Fig. 2 is reduced to second order single-degree-of-freedom inertia force vertical vibration model schematic diagram;

Fig. 3 is execution unit mechanical model schematic diagram;

Fig. 4 is genetic algorithm optimization sliding formwork handoff parameter and convenient thickness schematic diagram;

Fig. 5 is that the sliding formwork of optimizing based on GA becomes control flow chart;

Fig. 6 is Torque Control input signal schematic diagram;

Fig. 7 is that motor oscillating is controlled tracking emulation schematic diagram;

Fig. 8 is the vibration signal schematic diagram producing by simplified model;

Fig. 9 is the ACTIVE CONTROL effect emulation schematic diagram for electric drive system low-frequency vibration;

Figure 10 is the ACTIVE CONTROL effect emulation schematic diagram changing after vibration frequency for electric drive system low-frequency vibration.

Embodiment

Utilize drawings and Examples to be elaborated to a kind of Active Control Method for electric drive system low-frequency vibration of the present invention below.

With reference to Fig. 1, a kind of Active Control Method for electric drive system low-frequency vibration, comprises the following steps:

1) determine control program: electric drive system is mainly the passive support of motor, because the difference of motor model has different supporting forms and the strong point, on each strong point of motor, arrange an execution unit, by each execution unit of real-time control, adjust the acceleration that self exciting of decay motor produces, thereby the transmission of motor dynamic load to electric drive system that also just indirectly decayed; The main cause producing according to motor low-frequency vibration, described execution unit is selected electromagnetic actuator, while being connected with alternating current in electromagnetic actuator coil, electromagnetic action produces exciting force to oscillator, make it forced vibration, externally output is controlled and is used as power, this is used as power and directly acts on the vertical direction of motor, and equate with the vibration amplitude of motor self exciting force, during single spin-echo, these two power are cancelled out each other, and the vibration being caused by them also will be cancelled, the vibration of motor will be cancelled like this, and the power that is finally delivered to electric drive system is also zero;

2) select control algolithm: control algolithm is selected Sliding mode variable structure control algorithm, once electric drive system enters sliding mode, just interference and parameter perturbation have consistency to external world under certain condition, thereby have strong robustness, adaptivity,

3) definition becomes structure switching function, design becomes structure control rate:

Definition global sliding mode face is:

become structure control rate function,

u=k _d1·tanh(s)+k _d·s

K in formula _d1, k _dfor proportionality coefficient and k _d1, k _d>0;

4) according to genetic algorithm optimization switching function parameter: application genetic algorithm determines that switching function parameter and bound thickness slacken buffeting, and fitness function is selected: $J_{\mathrm{OBJ}}={\∫}_0^{\∞}(e{\left(t\right)}^{T}e\left(t\right)+{\mathrm{wy}}^2)\mathrm{dt},$

In formula: e (t)=y (t)-y _f, y _ffor desirable end-state value; W is weights, and

performance index are minimized, and story hardness function f changes into:

by selecting, intersect, making a variation, select optimal solution, return to v=[c D], realize the optimization of sliding-mode surface parameter and bound thickness, wherein: c is that sliding formwork parameter and D are boundary layer thickness;

5) determine braking moment: according to frequency and the amplitude of motor oscillating, determine the braking moment that suppresses motor oscillating.

Concrete grammar is: according to Fig. 2, set up simplified model:

In the vibration of bearing, rolling bearing is one of motor stronger vibration source in service, main exciting force is all to take the periodic function that speed of crankshaft is independent variable, therefore its vibration has stronger periodic feature, and main vibrational energy concentrates in one or several frequency.Conventionally, the vertical and rocking vibration being produced with tilting moment effect by uneven reciprocal inertia force is the principal mode of motor oscillating, especially more outstanding with vertical vibration.

Therefore the model of vibration of electric drive system is reduced to second order single-degree-of-freedom inertia force vertical vibration model, it is driving source that the sinusoidal signal of take replaces the second-order inertia masterpiece of motor, carries out simulation analysis.

A. the vibration maths model of motor

If external force F (t) is input variable, self vertical exciting force during machine operation, the out-of-balance force producing during motor rotation, the displacement y of mass ₁for output variable, F ₂(t) for electromagnetic actuator acts on being used as power of motor vertical direction.According to Newton's second law, the vibration maths model that obtains motor is:

$m_1{\stackrel{\·\·}{y}}_1+c_1{\stackrel{\·}{y}}_1+k_1{y}_1=F_2+F_1---\left(1\right)$

In formula: m ₁quality for the single support of motor; c ₁for the single passive supporting damping coefficient of motor; k ₁for the single passive support stiffness of motor.

Formula (1) is carried out to Laplace transformation, the impact being used as power if only consider, the transfer function that can obtain being used as power with motor vertical vibration acceleration is:

$\frac{F_2\left(S\right)}{Y_1\left(S\right)}=\frac{1}{m_1{s}^2+c_{1}s+k_1}---\left(2\right)$

The transfer function that in like manner can obtain motor self exciting force and motor vertical vibration acceleration is:

$\frac{F_1\left(S\right)}{Y_1\left(S\right)}=\frac{1}{m_1{s}^2+c_{1}s+k_1}---\left(3\right)$

Parameters:

If k=43.8N/m, m=18.2kg, f=1.49N/ (ms ^-1),

As calculated, $\frac{f}{m}=0.8187,\frac{k}{m}=2.40659,\frac{1}{m}=0.05495.$

Trying to achieve transfer function is:

$G\left(S\right)=\frac{438}{{4s}^2+3s+10}.---\left(4\right)$

Change spring vibration coefficient, can change its vibration frequency.

B. the mechanical model of execution unit:

The mechanical model of execution unit can be reduced to single-degree-of-freedom Forced Vibration System, and its mechanical model is as Fig. 3.

The Mathematical Modeling of electromagnetic actuator is:

$\stackrel{\·\·}{m_2{y}_2}+c_2{\stackrel{\·}{y}}_2+k_2{y}_2=F_2\left(t\right)---\left(5\right)$

In formula: m ₂for execution unit quality; y ₂vertical displacement for execution unit; c ₂for execution unit internal viscosity damping coefficient, k ₂for the spring rate of execution unit, F ₂for the electromagnetic force producing after execution unit energising.

By electromagnetic force F ₂(t) response is expressed as plural number:

y ₂=H(ω)F ₂(t) （6）

Thereby response be used as power into:

$f_2=c_2{\stackrel{\·}{y}}_2+k_2{z}_2=(k_2+{\mathrm{ic}}_2\mathrm{\ω})y_2=\frac{k_2+{\mathrm{ic}}_2\mathrm{\ω}}{k_2-{\mathrm{\ω}}^2{m}_2+{\mathrm{ic}}_2\mathrm{\ω}}{F}_2---\left(7\right)$

Make s=j ω, can obtain electromagnetic force and with the transfer function being used as power that is applied on machine shaft be:

$G\left(S\right)=\frac{f_2\left(s\right)}{F_2\left(s\right)}=\frac{c_{2}s+k_2}{m_2{s}^2+c_{2}s+k_2}---\left(8\right)$

Execution unit control coil can be regarded as the circuit of a resistance and inductance series connection, and the pass of its voltage and current is:

$\frac{i}{U}=\frac{1}{R+\mathrm{j\ωL}}---\left(9\right)$

In formula: R is the resistance of execution unit coil; L is coil inductance.

Make s=j ω, the complex frequency domain that can obtain the relation of voltage and current is expressed as:

$\frac{i\left(s\right)}{U\left(s\right)}=\frac{1}{R+\mathrm{Ls}}---\left(10\right)$

Electric current to be applied to execution unit coil on the electromagnetic force that produces be directly proportional:

F ₂=ki （11）

In formula: k is proportionality coefficient, relevant with structure and the magnetic field of execution unit.

(2) design of synovial membrane controller

A. choosing of sliding-mode surface:

A major issue of moding structural control system design is the sliding-mode surface that How to choose is suitable; make dynamic property and the steady-state behaviour optimization of system; for second-order system, the model of vibration of electrical equipment drive system can be similar to regards a second-order system as, and sliding-mode surface function is elected as conventionally:

$s\left(e\right)=\stackrel{\·}{e}+\mathrm{ce}$

The deviation that in formula, e is state variable; C is sliding-mode surface coefficient.

State is upper once operate between sliding area, state deviation is with exponential damping to balance point, and c value determines the rate of decay, and c value affects the speed that system slides into balance point. and little for state trajectory excursion, dynamic response process be take sliding process as main system, generally gets larger c value.But too high c value, will reduce between sliding area, make system get back to cycle of balance point elongated, the system of impact enters the process of steady operation.

Definition global sliding mode face is

$s=\stackrel{\·}{e}+\mathrm{ce}-F\left(t\right)---\left(12\right)$

In order to realize global sliding mode, function F (t) needs to meet following three conditions:

（1） $F\left(0\right)={\stackrel{\·}{e}}_0+{\mathrm{ce}}_0;$

（2）F(t)→0as t→∞；

(3) F (t) single order can be led.

E wherein ₀with

site error during for t=0 and derivative thereof.

Condition (1) makes system mode be positioned at sliding-mode surface,

Condition (2) has guaranteed that closed-loop system is stable,

Condition (3) is the requirement of sliding formwork existence condition.

According to above analysis, F (t) is defined as

F(t)=s(0)exp(-λt) （13）

λ >0 wherein, the s (t) that s (0) is initial time.

B. choosing of control rate function:

When state arrives sliding-mode surface, by switching controls rate function anancastia track, along sliding-mode surface, slide into balance point, the control rate function of for this reason selecting is:

u=k·sgn(s)

Sgn in formula (s) is sign function, i.e. (14)

Although insensitive for system parameter variations and external interference, there is very strong robustness, due to its discontinuous control characteristic in essence, be attended by the chattering phenomenon that is difficult to overcome, buffeting may damage execution unit, worsens control performance.So take measures, reduce to buffet.This control law must cause the intrinsic flutter of variable structure control system.Because the high frequency of controlling switches the high frequency mode that may encourage flexible appendage, therefore must revise to eliminate flutter to formula (14).After s converges to a less scope, switch, when | during s|≤D, make u=s * Mf _max/ D (Mf herein, _maxfor execution unit maximum output torque; D is boundary layer thickness), become being optimized for of structure control rule:

u=k _d1·tanh(s)+k _d·s

K in formula _d1, k _d> 0 (15)

K _d1, k _dfor proportionality coefficient, if k _d1value get less, k _dvalue larger, can guarantee that velocity of approach is large and be gradually to very little speed k near diverter surface time when away from diverter surface _d1thereby, have concurrently and buffet advantage little and that dynamic process is fast.

(3) genetic algorithm of application based on neural net

A. the selection of genetic coding

The problem that first coding will solve while being application genetic algorithm is one of committed step of genetic algorithm.It has also determined the individual phenotypic coding/decoding method that transforms to solution space from the genotype of search volume except determining individual Chromosomal arrangement form.If employing binary coding, the chromosomal length of degree of precision is long, can cause the solution space of genetic algorithm training excessive, may produce to overflow or to obtain computing time of desired result long.Real coding has advantages of that precision is high, is convenient to large space search, therefore adopt this coding method, sliding formwork parameter c is become to the chromosome of required problem with boundary layer thickness D direct coding, is designated as v=[c D].Algorithm flow as shown in Figure 4.

B. fitness function:

For obtaining satisfied transient process dynamic characteristic, the minimum target function that adopts state error quadratic term to select as parameter time integral performance index.Adopt punitive function to avoid overshoot, once execution unit span of control limit of control produces overshoot, just using overshoot quadratic term as optimum index.The optimum index that parameter is chosen is selected

$J_{\mathrm{OBJ}}={\∫}_0^{\∞}(e{\left(t\right)}^{T}e\left(t\right)+{\mathrm{wy}}^2)\mathrm{dt}---\left(16\right)$

In formula: e (t)=y (t)-y _f(y herein, _ffor desirable end-state value); W is weights, and

Performance index are minimized, and story hardness function f changes into:

$f=\frac{1}{J_{\mathrm{OBJ}}}---\left(18\right)$

C. select:

During selection, according to the size of ideal adaptation degree, determine to participate in the individuality of coupling.Adopt conventional roulette wheel selection method, the individual selected Probability p of i in colony _ifor:

$p_i=f_i/\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{f}_k---\left(19\right)$

In formula: the size that n is colony.Obviously, fitness is larger, and selected probability is just larger, but the little individuality of fitness also may be selected, can increase like this diversity of Xia Dai colony.Adopt the mechanism that retains optimized individual simultaneously, in order to avoid the larger individuality of fitness is lost in exchange and mutation operation, make the too fast globally optimal solution that converges to of algorithm.

D. intersect

Intersection is the core of genetic manipulation.Adopt the arithmetic cross method that is applicable to real coding, integration switch method is to two parent chromosome v=[c D].

$v_1=\left[\begin{array}{cc}{c}_1^{\left(1\right)}& D^{\left(1\right)}\end{array}\right],$ $v_{21}=\left[\begin{array}{cc}{c}_2^{\left(2\right)}& D^{\left(2\right)}\end{array}\right]$ Exchange, its offspring individual is:

$\left\{\begin{array}{c}{v}_1^{\′}=v_1-\mathrm{\λ}(v_1-v_2)\\ v_2^{\′}=v_2+\mathrm{\λ}(v_1-v_2)\end{array}\right.---\left(20\right)$

In formula: λ is random number, and λ ∈ [0,1].

E. variation

Variation is to change a certain gene in chromosome with certain probability.In the genetic algorithm application of real coding, variation is a very important genetic operator, directly affects the search performance of genetic algorithm.

The mutation algorithm adopting is described as:

For chromosome v=[c D]=[x ₁x ₂], if its element x _k(k=1,2) are selected and make a variation, and the element generating in offspring is:

$x_k^{\′}=x_k^L+(x_k^U-x_k^L)\·r---\left(21\right)$

In formula:

be respectively x _kbound; R is random number, and r ∈ [0,1].

(4) model following becomes structure controller

Make e ₁=y-y _f, e ₂=y '-y ' _f, y wherein _f, y ' _ffor tracked signal.So desirable sliding formwork equation is:

s=ce ₁+e ₂=0.c>0 (22)

According to sliding mode, arrive condition:

$s\stackrel{\&CenterDot;}{s}<0---\left(23\right)$

When s>0,

adequate condition be:

s=ce ₁+e ₂=c(y-y _f)+y′-y′ _f<0 （24）

Be c (y-y _f)-y ' _f<bu ₁-f

When s<0,

adequate condition be:

s=ce ₁+e ₂=c(y-y _f)+y′-y′ _f>0 （25）

Be c (y-y _f)-y ' _f>bu ₂-f

So, can be in the hope of meeting the inequality of arrival condition:

bu ₂-f<c(y-y _f)-y′ _f<bu ₁-f (26)

Wherein, when s>0, u=u ₁<0 (27)

When s<0, u=u ₂>0 (28)

Because sliding mode itself is stable, so as long as meet above formula, system is exactly stable.

(5) Simulink emulation

In general, low-frequency vibration and dither do not have clear and definite boundary, according to the GB/T14277-93 national standard of the IEC581 of International Power association standard and China, that to divide like this frequency range: 0-150Hz be low-frequency range, 150-500Hz is medium and low frequency section, 500-5KHz is medium-high frequency section, 5K-16KHz high band.

Interpretation of result:

(1) because we will carry out emulation with low-frequency vibration signal, the low-frequency vibration that is 4Hz by Simulink simulation frequency, when we change damping coefficient, the simulation result of Simulink shows as Figure 10.By having changed the damping coefficient of this model, can realize the simulation waveform of different vibration frequencies.

(2) as seen from Figure 6, when input is sinusoidal signal, the output of sliding mode controller is except slight buffeting, substantially reduced original sinusoidal waveform, guaranteed the accuracy of next step emulation.

In order to eliminate vibration, follow the tracks of signal, follow the tracks of good position is the prerequisite of eliminating vibration, the result that follow the tracks of position is as shown in Figure 7.

(3) application genetic algorithm is carried out the simulation result of position tracking in the method for electric drive system low-frequency vibration in conjunction with sliding moding structure algorithm application, as shown in Figure 7; By the result of emulation, can see that the method has good effect for position tracking, substantially mate with origin-location, simulated effect is good.In the good situation of Position Tracking Systems emulation, to adopting the method for subtracting each other between two signals, will eliminate vibration.

(4) for Fig. 8 frequency, control the control result of effect emulation, have simulation result to find out, the method has completed the elimination to vibration substantially, though there is slight fluctuation, whole structure is good.When changing the frequency of low-frequency vibration signal, it is carried out to simulation analysis, obtain result as shown in Figure 9 and Figure 10.

Claims (3)

1. for an Active Control Method for electric drive system low-frequency vibration, it is characterized in that, it comprises the following steps:

1) determine control program: electric drive system is mainly the passive support of motor, because the difference of motor model has different supporting forms and the strong point, on each strong point of motor, arrange an execution unit, by each execution unit of real-time control, adjust the acceleration that self exciting of decay motor produces, thereby the transmission of motor dynamic load to electric drive system that also just indirectly decayed; The main cause producing according to motor low-frequency vibration, described execution unit is selected electromagnetic actuator, while being connected with alternating current in electromagnetic actuator coil, electromagnetic action produces exciting force to oscillator, make it forced vibration, externally output is controlled and is used as power, this is used as power and directly acts on the vertical direction of motor, and equate with the vibration amplitude of motor self exciting force, during single spin-echo, these two power are cancelled out each other, and the vibration being caused by them also will be cancelled, the vibration of motor will be cancelled like this, and the power that is finally delivered to electric drive system is also zero;

2) select control algolithm: control algolithm is selected Sliding mode variable structure control algorithm, once electric drive system enters sliding mode, just interference and parameter perturbation have consistency to external world under certain condition, thereby have strong robustness, adaptivity,

3) definition becomes structure switching function, design becomes structure control rate:

Definition global sliding mode face is:

become structure control rate function,

u=k _d1·tanh(s)+k _d·s

K in formula _d1, k _dfor proportionality coefficient and k _d1, k _d>0;

4) according to genetic algorithm optimization switching function parameter: application genetic algorithm determines that switching function parameter and bound thickness slacken buffeting, and fitness function is selected: $J_{\mathrm{OBJ}}={\&Integral;}_0^{\&infin;}(e{\left(t\right)}^{T}e\left(t\right)+{\mathrm{wy}}^2)\mathrm{dt},$

In formula: e (t)=y (t)-y _f, y _ffor desirable end-state value; W is weights, and

performance index are minimized, and story hardness function f changes into:

by selecting, intersect, making a variation, select optimal solution, return to v=[c D], realize the optimization of sliding-mode surface parameter and bound thickness, wherein: c is that sliding formwork parameter and D are boundary layer thickness;

5) determine braking moment: according to frequency and the amplitude of motor oscillating, determine the braking moment that suppresses motor oscillating.

2. the Active Control Method for electric drive system low-frequency vibration according to claim 1, it is characterized in that: described definite braking moment comprises, according to frequency and the amplitude of simplified model emulation vibration, by the sliding formwork change algorithm of genetic algorithm optimization, determine the output of motor braking moment.

3. the Active Control Method for electric drive system low-frequency vibration according to claim 1 and 2, it is characterized in that: described braking moment is the braking moment of vertical direction, and with the single spin-echo of motor oscillating, by electromagnetic actuator, apply a kind of reverse disturbing moment, to offset original disturbing moment.

Images