From study faintness slippery model control reachs his to be in electro-hydraulic the application in servo

Using The Learning Control, the Renewal Of Fuzzy Control Rules And The Output Of Fuzzy Control Approaching The Equivalent Control Of Sliding Mode Are Completed.

Finally, theasymptotically Stable Sliding Mode Control Is Accomplished.

Taking Aelectrohydraulic Servo System As An Example, the Simulation Result Shows That The Method Is Effective.

Keywords:SControl of slippery model of foreword of Electrohydraulic Servo System1 of 　 of Nonlinear System of 　 of Elf-learning 　 Fuzzy Sliding Mode Control got quite rapid development nearly a few years to application from theory, because slippery model control is method of a kind of more effective rash club control,this basically is, and of the system move character to be able to be designed through slippery model beforehand set, no matter be right,linear system returns dispute linear system, slippery model control shows good control character, but, in be controlled actually, slippery model pilot shakes and tall gain control still is restricting slippery model pilot to apply actually. If groovy faintness controls the experience that relies on an expert greatly and experiment, the stability analysis that lacks a system and rash club sex analyse a method. In addition, to a few more complex control objects, ambiguous control obtains satisfactory control precision very hard. For this, document [1] , [2] offerred pilot of faintness slippery model basic idea, ambiguous pilot introduces to devise a method in control of tradition slippery model, achieved result of excel faintness pilot, weakened tradition slippery model to control what what exist to shake greatly phenomenon. Document [2] still make clear, the slippery model that includes attrib border layer controls equivalence at faintness. The article is based on document [3] ambiguous system is all-purpose is approached implement this one result, in the light of a kind of uncertainty nonlinear of the system dog control problem, apply ambiguous and logistic system to approach the equal in value control of slippery model, the faintness slippery standard that raised a kind to have capacity of self-study be used to controls a method, pass online study, replace ambiguous pilot regulation dynamicly, approach the equal in value control of slippery model in order to make the direction with stable asymptotic of model of slide of edge of ambiguous pilot output progressively, use this kind of method at electro-hydraulic of servo dog control problem, the emulation that gained satisfaction controls an outcome, this kind of method is had dog goodly precision, also maintained slippery model to control the characteristic with rash club strong sex at the same time, theory and emulate proving self-study reviews faintness slippery model controls a system is in Li Ya the asymptotic below Pu Nuofu meaning is stabilized. The description of 2 problems considers a kind of following copy to shoot nonlinear system (1) 　 among them: It is condition vector of the system, function of linear of dispute of F(x) , B(x) , and its are worth accurately sealed, u, Y is the output of control input of the system and system respectively, dominating an end is when the system (type (the F(x) of 1)) , B(x) , D(x)(has a bound to disturb) when existence uncertainty, output Y of the system can asymptotic dogs the output Xd of expectation. Controller of slippery model of faintness of 3 self-study be used to designs 　 (control of 1) slippery model takes tracking error vector to be: E=xd-x= [E, ... , e(n-1) ] T, acting E enters (the error condition equation that 1) type gets a system is: 　 of =-(f(x)+b(x)u+d(t))+x(n)d(2) of 　 of 　 of E(n)=x(n)d-x(n) 　 　 chooses function of slippery model switch to be: S=cTe=c1e+c2+ ... + E(n-1) , (Ci>0)(3) 　 can get control of slippery model equivalence to be by =0: (4) 　 is become when B, F, D is accurate and foregone, by type (4) can get Ueq directly, pilot of basis slippery model is basic and academic [4] , controlling a law to take here Us of 　 of structural 　 U=ueq+us(5) as follows is one nonlinear control, its action is generation slippery model, at the back, will give out the definition of Us. But, in be controlled actually, b, F, D often is existing element of a few uncertainty, be like parameter uncertainty, nonlinear the uncertainty that cause, make equal in value control (the average control that produces need of place of slippery model motion) cannot obtain directly, below this kind of circumstance, be based on document [3] , use ambiguous system to approach slippery model equivalence to control Ueq, such, type (5) repairs instead 　 U=uf+us(6) among them, uf is the output of ambiguous controller, by the design below the process is decided. (Design of 2) ambiguous controller is controlled with respect to slippery model and character, s and the relative speed when be apart from relatively and arriving at slip face S=0 what stated aleatoric description chooses slip side respectively, so, choice S, input variable as ambiguous pilot, ambiguous pilot outputs Uf by undermentioned M faintness controls regular R1, r2, ... , rm will decide, among them J regulation can state the Rj that it is 　 : If S Is A1j(m1j, 2 J(m of σ 1j) And Is A 2 J, 2 J) of σ 　 of 　 　 Then Ujf=Pj (J=1, 2, ... , m)(7) among them: A1j(m1j, 2 J of σ 1j) and A2 J(m, 2 J) of σ is obscure part A1(m1 respectively, σ 1) and A2 (m2, σ 2) ambiguous subclass, a1(m1, σ 1) and A2 (m2, σ 2) respectively correspondence inputs variable at faintness S and, the article uses A of μ of 　 of gauss subject function (X) ＝ Exp [- ((X-m) ／ σ ) 2] (8) so, obscure part A can express to be A(m, σ ) . μ Jf expresses to become J the output of ambiguous controller when regulation is contented. The action intensity that defines every regulation is: Wj=min(μ A1j(s) , μ A2 J) , (J=1... M)(9) among them μ A1j(s) and μ A2 J are ambiguous input respectively variable S and in its correspondence blurs subclass A1j(m1j, 2 J of σ 1j) and A2 J(m, subject function is worth calculative of the place on 2 J) of σ . Change the final output that can get ambiguous controller to be 　 according to adding authority to turn over faintness on average (10) among them: W= [W1, ... , wm] P= of T 　 　 [P1, ... , pm] here, p is sealed result parameter vector, can control a method to obtain through the study below. (The purpose of the design study of 3) study law and Us is to come true to be revised automatically (or update) after ambiguous pilot outputs result parameter Pj(parameter) , the way that exports stability of asymptotic of model of Uf edge slide finally in order to cause ambiguous pilot approachs slippery model equivalence to control Ueq stage by stage, make systematic condition maintains thereby go up in slip face and move along slip face. Study a law for derivation, do following hypothesis above all: Suppose 1: In (the B(x) in 1) type and its estimation are worth (X) satisfies next type: (11) seizes document [5] , can choose commonly for: (12) 　 assumes 2 [5] [6] [7] : Vector collect T= [S, ] T is one tight market and T ∈ Rn, put in P=p*= of vector of parameter of a group of results [P*1, p*2, ... , p*m] T, make next type establish 　 (13) among them: Your criterion (14) 　 is right (3) type both sides is begged guide, recycle type (2) , type (4) , type (5) can get: 　 of S=sb(ueq-uf-us)(15) of 　 of ∴ of =b(ueq-uf-us) 　 　 chooses Li Ya as follows 　 of 　 of 　 of Pu Nuofu function (16) among them: By (12) type is chosen. Right (16) type both sides is begged guide, choose Us following 　 Us=K0s+K1sgn(s)(17) among them: K0 ＞ 0, 0 　 of K1 ＞ by type (13) , (14) , (17) can get: (18) chooses: (19) has: (Type of 20) 　 (the result of 20) makes clear, should output Uf to approach slippery model equivalence to control Ueq with ambiguous pilot, can assure the stability of origin, and, slip mode also is stability of the asymptotic below meaning of husband of Nuo of Li Ya general, thereby, when T → ∞ , next type establish 　 (21) 　 type (18) also can express to be: (The place on 22) 　 put together is narrated, can reach as follows theorematic theorem to by (the system that 1) type describes, should use formula (6) , (10) , (the control keep under control that 17) place gives and type (the study rule that 18) place defines, should controlling system and state of slippery movable mould is Li Ya the asymptotic below Pu Nuofu meaning is stable, when T → ∞ , systematic tracking error comes along slippery model astringent 0. Attention: Take β to be ＜ of 0 ＜ β commonly 1, when S=0, the output Pj of ambiguous controller stops to update, ambiguous control brings to bear on directly control at controlled object. The initial value of β takes: 4 example emulate a consideration some is electro-hydraulic positional servo 　 (23) among them: Yp is output displacement of the system; α = [α 1, α 2, α 3] the parameter vector that T is a system, Δ α is the uncertainty of parameter vector; B(x) ＞ 0, be one nonlinear function, f is to the bound disturbs. To afore-mentioned systems (type (design stage of controller of slippery model of faintness of 23)) self-study be used to is as follows: (Variable of language of burnt pilot of 1) definite pattern and its are mixed by region choice S as language variable, its all are by region [- 1, 1] 　 takes (24) 　 among them: C1 ＞ 0, c2 ＞ 0, and C1, c2 designs a method to choose by convention slippery model. E=xd-x1, the system that Xd is expectation outputs signal. (Every language variable takes value of 2) choice language and corresponding subject function respectively 5 languages are worth () of 5 ambiguous subclass, its corresponding gauss subject function is: (I=1, 2 respectively correspondence mixes at S two languages are variable) among them: σ =0.

01, corresponding language value is respectively: (3) builds ambiguous and regular warehouse the article chooses 25 ambiguous regulation to make ambiguous and regular warehouse: (4) presses type (22) study law begs Pj(k)=pj(k-1) 　 　 (5) presses type (10) begs Uf, attention: The initial value of general Pj all is taken for 0. (6) presses type (17) gets Us. (7) is final by type (6) gets modular pilot controls an input to measure U teaching oneself reviewing blurring slipping. Attention: Shake to be eliminated further, type (the Sgn(s)(symbol function in 17) ) can use the saturated function Sat(s/ below φ ) replace [2] : Among them:  takes ＜ of 0 ＜  1. Graph 1 with the graph 2 it is faintness of afore-mentioned self-study be used to respectively program of slippery model control is right sine signal (Xd=sin10 π T) and square wave signal (frequency is 2.

5Hz) dogs answer a result. Emulation result makes clear: The faintness of self-study be used to that article place puts forward is slippery modular controller is had dog goodly character can gain satisfaction dog precision. Graph 1 dog the output of sine signal answers graph 2 dog the output of square wave signal answers a curve 5 conclusion are nonlinear to a kind of existence uncertainty system, use the control method that article place formulates, the slippery standard that can win asymptotic to stabilize is controlled and dog approvingly precision, this method does not need the knowledge of mathematical model of the system and uncertainty bound. Study a method introduce it is not easy to overcame groovy faintness to control regulation the inadequacy with affirmatory, subjective strong sex. CNC Milling CNC Machining