Զѧ
ACTA AUTOMATICA SINICA
1998 24 4 Vol.24 No.4 1998




Դϵݴֵⷨ1

ժҪ״̬ջϵͳеĴݴ ⣬Ȼȶʽֽ⵼ݴ״̬ջϵͳȶĳֱҪ. ڴ˻ϣݷԷֵⷨдݴƵ״̬ Ʒ.ֵŻⷽջϵͳõһ״̬ݴ Ʒ.˷Ч.
ؼʡݴƣԣ״̬ʽȶԣֵ㷨
NUMERICAL METHODS FOR FAULTTOLERANCE
CONTROL AGAINST SENSOR FAULTS

ZHAO MINGWANG
(Dept. of Automation, Wuhan Yejin Univ. of Sci.& Tech., Wuhan 430081)
AbstractFor the faulttolerance control against sensor faults in linear state feedback systems, sufficient and necessary conditions that make the closedloop systems stable are proposed based on the fractionization metho d of the stable polynomials. Then, a designing method of the faulttolerance st ate feedback law is presented via the numerical solutions to the nonlinear equat ions. Finally, based on the nonlinear optimization technique, another designing method of the faulttolerance state feedback law for pole placement of the clos edloop systems is proposed. Computer simulation examples show the effectivenes s of these faulttolerance control methods.
Key wordsFaulttolerance control, integrity, state feedback, poly nomial stability, numerical algorithm.
1
ڹϵͳУڸشڴִ.Щʹָ½ صʹϵͳȶ.ˣݴнǿʵʱһֱܹע1-5.
ִϿɷΪ.һźͨͨ(ƿ)ģʽֻ޵ һźŷ仯Ϸ仯ռ.һϵݴϵͳͬʱһ.ڣݴйϷڿƻڣ ڿƻУͬʱĶڱضģ.ڶݴ  ΪһཨģȷԵ³⣬ڹģʽܲΪ΢С㶯.Ricatti̺LyapunovȶҪԵǲռ仯Ŀ⣬˶ԹΪɢĵһݴԣģ1-5Ricatti̺LyapunovȶԵõķǳֵĵǱҪģнϴı.
۶Եһ״̬ݴϵͳ.ڶԸݴ⣬ȱ гҪԵĽƷľгҪԵݴֵǷǳ .ñķٸ֮CADʵݴϵͳƾʵ.

2
״̬ܿصSISOԶϵͳģ
=Ax+Bu.(1)
xuֱΪnά״̬룬ABֱάϵͳ.
ϵͳ(1)Ƿͬһ״̬ʹջϵͳڴ¾ȶ
u=-KLx.(2)
K=k1k2knΪL=diagl1l2lnΪϾliΪ10״̬xiĴл޹.ʱ״̬(2)£ϵͳջ״̬Ϊ
=(A-BKL)x.(3)
ǵϴͬʱڹϵĸʽСһʱֻһ ϵݴ.޹ʱϾLΪλ󣬲ΪL0kʱLΪLk.ϵͳۿ֪ϵͳ(1)ĿʽΪsn+1sn-1++nһڱ任Pʹϵͳ(1)ܿر׼ͱʾ
(4)


ʱiֹģʽ״̬ɿɼΪu=-KLiPջʽΪ
(5)

ei,j=ai+KLiPn-j+1.(6)
ʵn+1ֹģʽL0L1LnƷKʹn1ջʽfc,i(s)ȶ.ɶʽֽ֪ȶıջʽfc,i(s)һԱʾΪ
(7)
Цi,jͦi,jΪʵֱΪnnΪn/2(n+1)2֣֮ǡΪn .nΪżʱn=n+1=n.ʽ(7)еĶʽ˻չɵ
(8)
gi,kΪi,jͦi,jĺ.Ƚʽ(5)ʽ(8)ɹɳ¶.
1.״̬ܿصSISOϵͳ(1)ݴ״̬(2)£ջϵͳȶĳҪΪԷ
F(x)=F0(K,A0B0)F1(K,A1B1)Fn(K,AnBn)ݦ=0(9)
ʵ.



ʽ(9)֪ÿֹģʽ·ΪnΪn(n+1)xάΪn(n+2) ䷽Сڱ.ݴƵĽǷȡڷ(9)Ƿн(ݵ)ݴɵҲתΪԷ(9).
3ݴݷԷֵ
ԷСڱݷԷ̣ģ6һЧţٷ.㷨֤һԷţٷһΪ׾ֲ.ø㷨ⷽ(9)Ӧݴ.ø㷨ʱݶȷdF(x)/dxΪ
(10)


pi,jΪ任PĵiеjԪأli,jΪϾLiĵjԽԪ .dFidKɺܷʽ(6)(9)Ƶ.dFidAidFidBiģ6.ݶȷdF(x)dxıģ6ݵݷ ֵⷨɱݴ״̬.

4ݴķŻֵ
ϽڵݴƷõֻһջȶѣδǱջӦ.֪ ϵͳļַܳӳϵͳĶ̬ܣȶԡԵ.ˣȶĻͬʱʹݴϵͳջ㾡ָǼ.
4.1򼫵õŻֵ
ԵiֹģʽıջΪsi,1,si,2,si,nıջʽΪ
fe,i(s)=(s-si,1)(s-si,2)(s-si,n)=sn+e,i1sn-1++e,i,n.
ˣݴнĳҪ(9)ڼõݴŻɶΪ
(11) 
ЦiΪԵiϵͳŻļȨϵ.Ż(11)ĽⲻʹݴƱջϵͳȶʹջʽܽӽʽ.뷣Ż(11)ȼΪ
(12)
ЦΪ.ǵŻ(12)ڷԶȲߵĶʽŻ⣬ŻĶƫѣĲ½ȡֵ.Ȼ½ΪһӲչѸ٣ٶȺʱһŻѲӦõĹؼ.Ͻڵĸֵʽ½ݶȷ

С(13)
(14)
(15)
4.2Żλõݴֵ
ȶ֪ϵͳȶмиʵԽԶϵͳȶԣԽ.ʵ ϵͳڼʱͨʵΪRmaxСΪmaxͦminSƽ.ڴˣŻջϵͳλõŻʾ

(16)

Re(i,j)ͦi,jֱΪiֹ£ջʽs2+2i,js2i,jʵ.Ż(16)΢Ż ⣬Ż㷨ԴΪ.Powellڲ΢ѵŻ .Ż϶ʱPowellЧʽ½.Ż(16)Ż Ϊ״̬nĶκn(n+2)Ż϶.ˣͳPowellԴЧ ϵ.Ż(16)Ƶľп΢ԼĲ΢Ż⣬ģ7һн ЧʵPowell.ø÷ɱŻ(16).

5㡡
¿ȶϵͳݴ
1.
2.
ñķĽ1ʾ.ɼ֪ĵĻݷԷ ͷŻݴֵⷨнϸߵЧʣʧΪЧݴʵ㷨.
1ֵ

12
ʼֵx02,2,1,2,1,2,1,22,2, 2,1,2,3,1,2,3,
ԷⷽK-1.41953.7491ݣ0.75571.18330.8047
ջʽL0S27.5882s+7.0077
L1:S2+4.7491s+2.7491
L2S2+3.8390s+3.2585L0S3+5.3770s2+8.7035s+4.4591
L1S34.6217s2+5.6808s+0.6808
L2S36.5604s2+13.4368s+9.1925
L3S34.5723s2+6.2893s+2.0450
ڼ
Żıջʽ
K-0.67951.7770ݣ0.59850.20810.6413
ջʽL0s2+4.1360s+2.8155
L1s2+2.7770s+0.7770
L2s2+2.3590s+1.0386L0s3+6.4479s2+13.1505s+8.7490
L1s3+5.8494s2+10.7564s+5.7564
L2s3+6.2398s2+12.3179s+7.9164
L3s3+5.8067s2+11.2267s+6.8252
ŻλõķRmax,max,min0.5,3,1/30.5,3,1/3 
K-0.62642.5061ݣ1.02251.20972.6580
ջʽL0s2+4.7589s+3.3853
L1s2+3.5061s+1.5061
L2s2+2.2528s+0.8792L0s3+6.4708s2+8.7768s+6.2187
L1s3+5.4483s2+4.6869s+1.1064
L2s3+7.6805s2+12.4060s+7.4284
L3s3+3.8127s2+3.4607s+0.9026


1)ʡȻѧĿ.
ߵλ人ұƼѧԶϵ430081
ο
1Shimemura E, Fujta M. A design method for linear state feedback system processing intergrity based on a solution of a Ricatti type equation.Int. J . Control,1985,42(4)887-899
2Veillette RJ. Reliable linear quadratic statefeedback control. Automatica, 1995,31(1)137-143
3,ۻݺ.ϵͳԵһƷ.Ӧ,1993,11(2)161-167
4ï,.Եſϵͳ.Ӧ,1992,9(3)245-249
5,,ڸյ.³ݴϵͳ.Ӧ,1994,11(3)376-380
6.״̬ͬʱ򶨵ķԷֵ.Ӧ,1996,13 (5)376-380
7.һ಻΢ŻPowell.,1998,13
ոڡ19961125
