G y 0nx (n. G y d = I nu N (5) (6) J uu 2 = and. J ud H1 H

Size: px
Start display at page:

Download "G y 0nx (n. G y d = I nu N (5) (6) J uu 2 = and. J ud H1 H"

Transcription

1 29 American Control Conference Hatt Regenc Riverfront, St Lois, MO, USA Jne 1-12, 29 WeC112 Conve initialization of the H 2 -optimal static otpt feeback problem Henrik Manm, Sigr Skogesta, an Johannes Jäschke Abstract Recentl we have establishe a link between invariants for qaratic optimization problems an linearqaratic (LQ) optimal control [1 The link is that for LQ control one invariant is c k = k K k, which iels zero loss from optimalit when controlle to a constant setpoint c = c s = In general there eists infinitel man sch invariants to a qaratic programming (QP) problem In [2 we show how the link can be se to generate otpt feeback control b sing crrent an ol measrements In this paper we eten this approach b consiering in more etail some interesting eamples, an the se of aitional (ol) measrements In particlar, we show that if the nmber of measrements is less than the nmber of istrbances (initial states) pls inepenent inpts, we can not with this metho fin a polic k = K k that minimizes the original problem, becase K is not optimall constant However, this metho ma be se to fin initial vales for H 2 -optimal static otpt feeback snthesis Ine Terms linear qaratic control, fie-strctre control I INTRODUCTION Consier a finite horizon LQ problem of the form min J(,()) = E{ T, 1,, N 1 NP N + + N 1 k= sbject to = () [ T k Q k + T kr k } k+1 = A k + B k, k k = C k + n k, where k R n are the states, k R n are the inpts an k R n are the measrements Frther P = P T >, Q, an R > are matrices of appropriate imensions, an E{ } is the epectation operator It is well-known that if C = I an n =, sch that k = k, the soltion to (1) is state feeback k = K k, where the gain matri K can be fon b solving an iterative Riccati eqation For the case with white noise assmption on an (n ), the optimal soltion is k = Kˆ k, where ˆ k is a state estimate from a Kalman filter [3, which in effect gives a namic compensator K lqg (from to ) of same orer n as the plant In this paper we consier the static otpt feeback problem, k = K k, where K is a static gain matri Note that the case with a fie-orer controller of orer less than n ma also be broght back to the static otpt Department of Chemical Engineering, Norwegian Universit of Science an Technolog, N-7491 Tronheim, Norwa Athor to whom corresponence shol be aresse skoge@ntnno (1) Fig 1 G G c = H( + n ) n Measrement combination (H) Notation for self-optimizing control feeback problem A particlar controller consiere in this paper is the mlti-inpt mlti-otpt proportional-integralerivative controller (MIMO-PID) where we have as man controlle otpts c as there are inpts The allowe measrements k in the formlation in (1) are the present vale of the controlle otpt k c (P), the integrate vale k i= c i (I) an the erivative c k δt (D) This optimal soltion to this problem is nsolve [4 so one cannot epect to fin an analtic or conve nmerical soltion The contribtion of this paper is therefore to propose a conve approach to fin a goo initial estimate for K, as a starting point for a nmerical search A Notation Notation aopte from self-optimizing control is smmarize in figre 1 Tpicall, = (, 1,, N 1 ), = an = (,) or = (, t,,), bt also other variables will be consiere II MAIN RESULTS A Reslts from self-optimizing control 1) Nllspace metho: From [5 we have the following theorem: Theorem 1: (Nllspace theorem = Linear invariants for qaratic optimization problem) Consier an nconstraine qaratic optimization problem in the variables (inpt vector of length n ) an (istrbance vector of length n ) min J(,) = [ T J J (2) J T J In aition, there are measrement variables = G + G If there eists n n + n inepenent measrements [ (where inepenent means that the matri G = G G has fll rank), then the optimal soltion to (2) has the propert that there eists n c = n linear variable combinations (constraints) c = H that are invariant to /9/$25 29 AACC 1724

2 the istrbances The optimal measrement combination matri H is fon b selecting H sch that HF =, (3) where F = opt is the optimal sensitivit matri which can be obtaine from F = (G J 1 J G ), (4) (That is, H is in the left nllspace of F ) 2) Generalization: Eact local metho: A generalization of Theorem 1 is the following: Theorem 2: (Eact local metho = Loss b introcing linear constraint for nois qaratic optimization problem [5) Consier the nconstraine optimization problem in Theorem 1, (2), an a set of nois measrements m = + n, where = G +G Assme that n c = n constraints c = H m = c s are ae to the problem, which will reslt in a non-optimal soltion with a loss L = J(,) J opt () Consier istrbances an noise n with magnites [ = W ; n = W n n ; 1 n Then for a given H, the worst-case loss introce b aing the constraint c = H is L wc = σ 2 (M)/2, where M is M [ M M n M = J 1/2 (HG ) 1 HFW M n = J 1/2 (HG ) 1 HW n, an σ( ) is the maimm singlar vale The optimal H that minimizes the loss can be fon b solving the conve optimization problem min H F F H sbject to HG = J 1/2 Here F = [FW W n The reason for sing the Frobenis norm is that minimization of this norm also minimizes σ(m) [6 Remark 1: From [5 we have that an optimal H premltiplie b a non-singlar matri n c n c D, ie H 1 = DH is still optimal One implication of this is that for a sqare plant, n c = n, we can write c = H 1 = H m 1 m +I To see this, assme = ( m,), so H = [H m H, where H is a nonsinglar n n matri Now, H 1 = (H ) 1 [H m H = [(H ) 1 H m I Remark 2: More generall, for the case when F F T is singlar, we can solve the conve problem (6) sing for eample CVX, a package for specifing an solving conve programs [7, with the following coe: cv_begin variable H(N*n,n+n*N); minimize norm(h*ftile, fro ) sbject to H*G == sqrtm(j); cv_en Remark 3: Noise will not be frther iscsse in this paper, bt is covere in [8 2 (5) (6) B Some special cases Some special cases will now be consiere where eplicit epressions can be fon 1) Fll information: No new reslts are represente here, bt we show the matrices G, G, J, an J for LQoptimal control Assme that noise-free measrements of all the states are available It is well known that the LQ problem (1) can be rewritten on the form in (2) (see for eample [9) b treating as the istrbance, an letting = (, 1,, N 1 ) Ths, from Theorem 1 we know that for the LQ problem there eists infinitel man invariants (bt onl one of these involves onl present states) Withot loss of generalit consier the case when the moel in (1) is stable Let = (,, 1,, N 1 ) = (,) Note that this incles also ftre inpts, bt we will se the normal trick in MPC of implementing onl the present (first) inpt change Since we have n = n + n an no noise, we can se Theorem 1 The open loop moel becomes: = G + G [ G n (n = N) R (n+nn) (nn) I nn G = I n N R (n+nn) n (nn) n Here I m is an m m ientit matri an m n is a m n matri of zeros The matrices J an J are the erivatives of the linear qaratic objective fnctionfor the objective an process moel in (1) we show in [1 that an J 2 = J 2 = B T PB+R B T A T PB B T (A N 1 ) T PB B T PAB B T PB+R B T (A N 2 ) T PB B T PA N 1 B B T PA N 2 B B T B T B T B T PB+R P PA PA N 1 (7) (8) A (9) The sensitivit matri (optimal change in when is pertrbe) becomes: F = opt T = (G J 1 J G ) = I n J 1 (1) J We can se Theorem 1 to get the combination matri H, ie fin an H sch that HF = : I H1 H n 2 J 1 = H J 1 H 2 (J 1 J ) = (11) To ensre a non-trivial soltion we can choose H 2 = I nn an get the following optimal combination of an : c = H = J 1 J +, (12) 1725

3 which can be interprete as: Invariant 1: = K Invariant 2: 1 = K 1 Invariant N: N 1 = K N 1 (13) From Theorem 1 implementation of (13) give zero loss from optimalit, ie the correspon to the optimal inpt trajector, 1,, N 1 from the soltion of (1) Moreover, since the states captre all information, we mst have that 1 = K 1 = K 1 (A + BK ) 1 1 (14) =K From this we ece that the soltion to (1) can be implemente as k = K k, k =,1, In [1 we prove that this gives the same reslt as conventional linear qaratic control, b conventional meaning for eample eqation (3) in Rawlings an Mske [9 2) Otpt feeback: In this section we will show how Theorem 2 can be se for the special (bt common) case when k = C k + k, k =,1,,N an we look for controllers on the form k = K k If C is fll colmn rank, then we have fll information (state feeback), bt we here consier the general case where C has fll row rank (inepenent measrements), bt not fll colmn rank Let = (,) an as before = (, 1,, N 1 ) The istrbance = The open loop moel is now = [ C = +, (15) I G G an the sensitivit matri F is F = (G J 1 J G ) = C J 1 (16) J Since we now have that n = nỹ + n < n + n, where nỹ are the nmber of measrements from the plant, nỹ < n, we cannot simpl set HF =, bt we nee to solve (6) Let s analze this problem For G T = [ I, HG = J 1/2 is eqivalent to H 2 = J 1/2, where H nc (nỹ+n ) = H 1nc nỹ H 2nc n (17) With this partitioning we get that H F = H[FW W n = [HFW HW n, (18) an for W n =, ie the noise-free case, H F = [HFW (19) We want to minimize the Frobenis-norm of this matri an we have that [HFW F = HFW F + F (2) Assme withot loss of generalit that W = I, an let J = J 1 J With F T = [C T J T we have that HF = H 1 C + H 2 J = H 1 C J 1/2 J (21) H2=J 1/2 We want to minimize H 1 C J 1/2 J, hence we look for a H 1 sch that Using the pseo-inverse, we fin that an we get that the optimal H is H 1 C = J 1/2 J (22) H 1 = J 1/2 J C, (23) H = [J 1/2 J C J 1/2 (24) In the final implementation we can ecople the invariants in the inpts b H = J 1/2 H = [ J 1 J C I (25) This means that the open-loop optimal otpt feeback is k = J 1 J C k = K k, (26) K state feeback that is, for an optimal state feeback K, the optimal otpt feeback is KC This means that for this case we have Invariant 1: = K C Invariant 2: 1 = K 1 C Invariant N: N 1 = K N 1 C (27) We have calle these variable combinations invariants in qotation marks becase the are not invariant to the soltion of the original problem, bt rather the variable combinations that minimize the (open-loop) loss Inee, the non-negative loss is HF = J 1/2 J C C J 1/2 J = J 1/2 J (C C I) J 1/2 J C C I For otpt feeback we have in the least sqares sense (28) 1 = K 1 C C(A + BK C C) 1 C 1 (29) K 1 Unfortnatel, in general K 1 K an hence the open loop soltion (27) cannot be implemente as a constant feeback k = K C k, as was the case for state feeback, see (14) III MAIN ALGORITHM We now propose an algorithm for fining otpt feeback controllers This is a two-step procere where we first fin initial vales sing Theorem 2 These initial vales correspon to a controller that in the open-loop sense is closest to the optimal state feeback LQ controller Thereafter we improve this controller b solving a close-loop optimization problem where the controller parameters are the egrees of freeom In the previos section we showe that if = (C,,, N 1 ) Theorem 1 gives = K = K state feeback C The algorithm presente here is more 1726

4 n 1/2 W W 1/2 n R 1/2 P z k = [1; B c (si A c ) 1 K C Q 1/2 m Fig 2 Interconnection strctre for close-loop optimization of K k 1 1: k = K k : k = K k = [;1 3: k = K 1, k K 2, k 1 4: k = K 1 k K 2 k 1 5: k = K lqr k general in the sense that it hanles measrements sch as = (, 1,, M,,, N 1 ) (In the latter case a casal controller is M = K [ T T M T ) Fig Time Simlation reslts for istrbances in initial conitions, eample Algorithm 1 Low-orer controller snthesis 1: Define a finite-horizon qaratic objective J(,) = T N P N + N 1 i= T i Q i T i R i + 2 T i N i 2: Calclate J an J as in (8), (9) 3: Define caniate variables = G + G, = (, 1,, N 1 ) 4: Decie weights W an Wn (Defalt: W = I, Wn = ) 5: Fin H b solving the conve optimization problem (6) 6: Optional: Improve control b close-loop optimization (Section III-A) A Relationship between LQ-control an H 2 optimal control It is well-known that the LQG problem ma be cast into the H 2 framework an that a class of H 2 optimal controllers ma be implemente in an LQG-scheme with a Kalman estimator an a constant feeback gain from the estimate states [11 In this paper we propose to improve the soltion from the open-loop control b minimizing the H 2 norm min K F l(p,k) 2 (3) In this contet min K means minimizing over the parameters in K The lower-fractional transform F l (P,K) = P 11 + P 12 (I P 22 K) 1 P 21 for a P = [ P 11 P 12 P 21 P 22 [12 The interconnection strctre we se for P is shown in figre 2 The last row of Algorithm 1 consists of solving (3) with initial vales as H on step 5 in algorithm 1, which is the soltion to (6) IV EXAMPLES In this section two eamples will be consiere First we iscss P-control of a secon orer plant, then MIMO-PID control of a moel of a istillation colmn Eample 1: (P-control of secon orer plant) Consier the 2 plant g(s) = s 2 +3s+2 The plant is sample with T s = 1 to get k+1 = k + 91 k k = [ 1 (31) k The objective is to erive two LQ-optimal controllers for this process, one P-controller on the form k = K k an a PD-controller k = (K 1 k + K 2 k 1) In the snthesis of the controllers we se algorithm 1 The open loop objective to be minimize is J(,) = i= T i Q i + T i R i with Q = [ 1 an R = 1 The infinite horizon objective can be approimate b the following objective: N 1 J(,) = T NP N + T i Q i + T i R i, (32) i= with P = [ an N = 1 (P is a soltion to the iscrete Lapnov eqation P = A T PA+Q) The objective is now on the form of step 1 in the algorithm P-control: For the P-controller, the variables to combine are 1 = (,,, N 1 ) The matrices J an J are the same as those reporte in eqations (8) an (9) Since we o not consier noise W = I an W n = We can now fin H either b solving the conve problem in (6), or we can simpl se the eplicit formla in (23), ie H 1 = J 1 J C As shown in section II-B2 (see eqation 27) we now get N = 1 invariants to the soltion to the original optimization problem in (32) The first one of these invariants is reporte as controller 1 in table I We observe that H F >, which is epecte from Theorem 1, as n < n + n in this case (n = 1 + 1,n = 1,n = 2) Using this P-controller ( k = 44 k ) as the initial estimate, the H 2 -optimal close-loop controller K in Figre 2 is obtaine nmericall Note from row 2 in Table I that the H 2 -norm is onl rece slightl (from 2993 to 2981), althogh K changes from -44 to -313 PD-control: For the snthesis of a PD controller we again se algorithm 1 The variables to combine are now 2 = (, 1,,, N 1 ) The objective fnction an the matrices J an J remain the same The open-loop moel 1727

5 TABLE I CONTROLLERS FOR EXAMPLE 1 Controller H F F F l (P, K) 2 1: k = 44 k First invariant sing Theorem 2 for 1 = (,,, N 1 ) 2: k = 313 k 2981 Close-loop optimal P-controller 3: k = (149 k 111 k 1 ) 3176 Secon invariant sing Theorem 2 for 2 = (, 1,,, N 1 ) 4: k = (416 k 19 k 1 ) 2979 Close-loop optimal controller PD-controller 5: k = [ k 2972 LQR = G + G is now: 2 = 1 = CB C I + CA (33) I For this particlar variable combination (23) cannot be se, as the variables occr on ifferent instances in time We therefore solve the optimization problem (6) sing cv, as shown in remark 2 The soltion is again on the form of (27), for which the secon invariant is reporte as controller 3 in table I The soltion (all the invariants) gives H F =, which is epecte from Theorem 1, as n = n + n an no noise is present Frther nmerical optimization reces the H 2 norm from 3176 to 2979 It can be verifie that the variable combination is inee optimal after one step with the following calclations: = K, 1 = K 1 1 K 2 1 = (K 1 C + K 2 C(A BK) 1 ) 1, (34) =K where K is the LQR controller For implementation some sb-optimalit mst be epecte since we are not starting the control with LQR, rather we se the PD controller at all time instances Simlations: From the close-loop norms reporte in table I the controllers are epecte to perform similarl in close loop This is confirme in the close loop simlations of istrbances in initial states, see figre 3 Eample 2: (Linear namic moel of istillation colmn) In this eample we consier MIMO-PI an -PID control of colmn A in [13 The moel is se as an eample for offset-free control in [14 The moel is base on the following assmptions: binar separation, 41 stages, incling reboiler an total conenser, each stage is at eqilibrim, with constant relative volatilit α = 15, linearize liqi flow namics, negligible vapor holp, constant pressre The fee enters on stage 21 = [ V L D an = [ B We here consier the LV-configration, where D an B are se to control the levels With level controllers implemente (P-control with K c = 1) the rest of the colmn is stable Balance rection is se to rece the nmber of states to 16 Then integrate otpts are ae to the moel, reslting in a moel with 18 states If we let the otpts of top bottom Inpts L,V LQR 5 PI 1 3 PID pper: V lower: L Time [min Fig 4 Simlation reslts for eample 2 At t = a step-change of 1 in F occrs, an at t = 7 z F is change from 5 to 6 the moel be P, I, an D, we get a moel with the following strctre: [ẋ a b = + σ c σ P c (35) I = I + D σ ca cb This moel is sample with T s = 1 to get a iscrete time moel Again[ we set p an infinite time objective fnction, with Q = C T I C, an R = 1 I, an for intermeiate calclations we approimate this b a finite horizon objective with N = 15 an P = Q We now look for controllers on the form k = ( K P k P + K I k I + K D k D ) (36) an we assme measrements of the compositions with a sample time of 1 minte is available Table II shows first-move PI (= first move invariant realize as feeback), close loop PI an PID controllers, 1728

6 TABLE II CONTROLLERS FOR EXAMPLE 2 Description Control eqation H F F F l (P, K) ( ) First-move PI k = k P k I ( ) Close-loop optimal PI k = k P k I 365 ( First-move PID k = k P k I ) k D ( Close-loop optimal PID k = k P k I ) k D LQR k = k (1 : 6) k (7 : 12) k (13 : 18) TABLE III ITERATION COUNT USING FMINUNC (MATLAB c R28A) WITH DIFFERENT INITIALIZATIONS Algorithm 1 K = SIMC-tne PI controllers PI PID an the LQR controller for reference In aition to the initialziation propose in this paper we trie to initialize the nmerical search with K = an two SIMC-tne [15 PI controllers with τ c = 1 mintes, leaing to KSIMC = As reporte in table III i K = not converge, whereas intializing with two SIMC-tne PI controllers converge in both cases (both for PI an PID esign), thogh with some more iterations than the metho propose in this paper Figre 4 shows simlation reslts where we at t = introce a step in the fee rate an at t = 7 a step in the fee composition PI an PID refers to the closeloop optimal controllers As one observes is the MIMO-PID controller qite close in performance to the LQR controller V CONCLUSIONS In this paper we have iscsse snthesis of H 2 -optimal static otpt feeback, an in particlar the MIMO-PID We have shown that initial conitions for close loop optimization can be fon b solving a conve program, an that the reslting close loop optimization problem converges for some interesting cases VI ACKNOWLEDGMENTS The athors gratefll acknowlege the comments from Bjarne Grimsta [2, Eplicit MPC with otpt feeback sing self-optimizing control, in Proceeings of IFAC Worl Conference, Seol, Korea, 28 [3 K J Åström an B Wittenmark, Compter Controlle Sstems, T Kailath, E Prentice-Hall, 1984 [4 V Srmos, C Aballah, an P Dorato, Static otpt feeback: a srve, Decision an Control, 1994, Proceeings of the 33r IEEE Conference on, vol 1, pp vol1, Dec 1994 [5 V Alsta, S Skogesta, an E Hori, Optimal measrement combinations as controlle variables, 28, in Press: Jornal of Process Control, oi:1116/jjprocont2812 [6 V Kariwala, Y Cao, an S Janarhanan, Local self-optimizing control with average loss minimization, In Eng Chem Res, vol 46, pp , 27 [7 M Grant an S Bo, CVX: Matlab software for iscipline conve programming (web page an software), Agst 28 [Online Available: bo/cv [8 H Manm an S Skogesta, Conve initialization of the H 2 -optimal static feeback problem with noise, in In preparation for Proceeings of CDC, 29 [9 J B Rawlings an K R Mske, The stabilit of constraine receing-horizon control, in IEEE Transactions on Atomatic Control, vol 38, 1993, pp [1 H Manm, S Narasimhan, an S Skogesta, A new approach to eplicit MPC sing self-optimizing control, 27, avaliable at: [11 J Dole, K Glover, P Khargonekar, an B Francis, State-space soltions to stanar H 2 an H control problems, IEEE Transactions on Atomatic Control, vol 34, no 8, pp , 1989 [12 S Skogesta an I Postlethwaite, Mltivariable Feeback Control Wile, 25 [13 S Skogesta, Dnamics an control of istillation colmns - a ttorial introction, Trans IChemE, Part A, vol 75, pp , September 1997 [14 K R Mske an T A Bagwell, Distrbance moeling for offsetfree linear moel preictive control, Jornal of Process Control, vol 12, pp , 22 [15 S Skogesta, Simple rles for moel rection an pi controller tning, Jornal of Process Control, 23 REFERENCES [1 H Manm, S Narasimhan, an S Skogesta, A new approach to eplicit MPC sing self-optimizing control, in Proceeings of American Control Conference, Seattle, USA,

Optimal Operation by Controlling the Gradient to Zero

Optimal Operation by Controlling the Gradient to Zero Optimal Operation by Controlling the Graient to Zero Johannes Jäschke Sigr Skogesta Department of Chemical Engineering, Norwegian University of Science an Technology, NTNU, Tronheim, Norway (e-mail: {jaschke}{skoge}@chemeng.ntn.no)

More information

Model Predictive Control Lecture VIa: Impulse Response Models

Model Predictive Control Lecture VIa: Impulse Response Models Moel Preictive Control Lectre VIa: Implse Response Moels Niet S. Kaisare Department of Chemical Engineering Inian Institte of Technolog Maras Ingreients of Moel Preictive Control Dnamic Moel Ftre preictions

More information

NEURAL CONTROL OF NONLINEAR SYSTEMS: A REFERENCE GOVERNOR APPROACH

NEURAL CONTROL OF NONLINEAR SYSTEMS: A REFERENCE GOVERNOR APPROACH NEURAL CONTROL OF NONLINEAR SYSTEMS: A REFERENCE GOVERNOR APPROACH L. Schnitman Institto Tecnológico e Aeronática.8-900 - S.J. os Campos, SP Brazil leizer@ele.ita.cta.br J.A.M. Felippe e Soza Universiae

More information

Linear and Nonlinear Model Predictive Control of Quadruple Tank Process

Linear and Nonlinear Model Predictive Control of Quadruple Tank Process Linear and Nonlinear Model Predictive Control of Qadrple Tank Process P.Srinivasarao Research scholar Dr.M.G.R.University Chennai, India P.Sbbaiah, PhD. Prof of Dhanalaxmi college of Engineering Thambaram

More information

A new approach to explicit MPC using self-optimizing control

A new approach to explicit MPC using self-optimizing control 28 American Control Conference Westin Seattle Hotel, Seattle, Washington, USA June -3, 28 WeA3.2 A new approach to explicit MPC using self-optimizing control Henrik Manum, Sriharakumar Narasimhan an Sigur

More information

DISCRETE-TIME ANTI-WINDUP: PART 1 - STABILITY AND PERFORMANCE

DISCRETE-TIME ANTI-WINDUP: PART 1 - STABILITY AND PERFORMANCE DISCRETE-TIME ANTI-WINDUP PART - STABILITY AND PERFORMANCE Matthew C. Trner, Gio Herrmann an Ian Postlethwaite Control an Instrmentation Grop, Universit of Leicester, Universit Roa, Leicester, LE 7RH,

More information

Fault Tolerant Control - A Residual based Set-up

Fault Tolerant Control - A Residual based Set-up Joint 48th IEEE Conference on Decision an Control an 28th Chinese Control Conference Shanghai, P.R. China, December 16-18, 2009 FrC14.2 Falt Tolerant Control - A Resial base Set-p Henrik iemann Dept. of

More information

FREQUENCY WEIGHTED L 2 GAIN OPTIMISATION FOR IMPROVED ANTI-WINDUP PERFORMANCE. Phil March,1 Matthew C. Turner Guido Herrmann Ian Postlethwaite

FREQUENCY WEIGHTED L 2 GAIN OPTIMISATION FOR IMPROVED ANTI-WINDUP PERFORMANCE. Phil March,1 Matthew C. Turner Guido Herrmann Ian Postlethwaite FREQUENCY WEIGHTED L GAIN OPTIMISATION FOR IMPROVED ANTI-WINDUP PERFORMANCE Phil March, Matthew C. Trner Gio Herrmann Ian Postlethwaite Department of Engineering, Universit of Leicester, UK {pm7,mct6,gh7,ixp}@le.ac.k

More information

Move Blocking Strategies in Receding Horizon Control

Move Blocking Strategies in Receding Horizon Control Move Blocking Strategies in Receding Horizon Control Raphael Cagienard, Pascal Grieder, Eric C. Kerrigan and Manfred Morari Abstract In order to deal with the comptational brden of optimal control, it

More information

Math 273b: Calculus of Variations

Math 273b: Calculus of Variations Math 273b: Calcls of Variations Yacob Kreh Homework #3 [1] Consier the 1D length fnctional minimization problem min F 1 1 L, or min 1 + 2, for twice ifferentiable fnctions : [, 1] R with bonary conitions,

More information

1. State-Space Linear Systems 2. Block Diagrams 3. Exercises

1. State-Space Linear Systems 2. Block Diagrams 3. Exercises LECTURE 1 State-Space Linear Sstems This lectre introdces state-space linear sstems, which are the main focs of this book. Contents 1. State-Space Linear Sstems 2. Block Diagrams 3. Exercises 1.1 State-Space

More information

Modal Transient Analysis of a Beam with Enforced Motion via a Ramp Invariant Digital Recursive Filtering Relationship

Modal Transient Analysis of a Beam with Enforced Motion via a Ramp Invariant Digital Recursive Filtering Relationship oal ransient Analysis of a Beam ith Enforce otion via a Ramp nvariant Digital Recrsive iltering Relationship By om rvine Email: tomirvine@aol.com December, Variables f f ass matrix Stiness matrix Applie

More information

Proposal of Surface Topography Observer considering Z-scanner for High-speed AFM

Proposal of Surface Topography Observer considering Z-scanner for High-speed AFM 2 American Control Conference Marriott Waterfront, Baltimore, MD, USA Jne 3-Jly 2, 2 ThA.5 Proposal of Srface Topography Observer consiering Z-scanner for High-spee AFM Takayki Shiraishi an Hiroshi Fjimoto

More information

Numerical simulation on wind pressure transients in high speed train tunnels

Numerical simulation on wind pressure transients in high speed train tunnels Compters in ailways XI 905 Nmerical simlation on win pressre transients in high spee train tnnels S.-W. Nam Department of High Spee Train, Korea ailroa esearch Institte, Korea Abstract When a train passes

More information

Adaptive partial state feedback control of the DC-to-DC Ćuk converter

Adaptive partial state feedback control of the DC-to-DC Ćuk converter 5 American Control Conference Jne 8-, 5. Portlan, OR, USA FrC7.4 Aaptive partial state feeback control of the DC-to-DC Ćk converter Hgo Rorígez, Romeo Ortega an Alessanro Astolfi Abstract The problem of

More information

The Real Stabilizability Radius of the Multi-Link Inverted Pendulum

The Real Stabilizability Radius of the Multi-Link Inverted Pendulum Proceedings of the 26 American Control Conference Minneapolis, Minnesota, USA, Jne 14-16, 26 WeC123 The Real Stabilizability Radis of the Mlti-Link Inerted Pendlm Simon Lam and Edward J Daison Abstract

More information

Simple robustness measures for control of MISO and SIMO plants

Simple robustness measures for control of MISO and SIMO plants Preprints of the 8th IFAC World Congress Milano Ital) Agst 28 - September 2, 2 Simple robstness measres for control of MISO and SIMO plants W. P. Heath Sandira Gaadeen Control Sstems Centre, School of

More information

A Note on Irreducible Polynomials and Identity Testing

A Note on Irreducible Polynomials and Identity Testing A Note on Irrecible Polynomials an Ientity Testing Chanan Saha Department of Compter Science an Engineering Inian Institte of Technology Kanpr Abstract We show that, given a finite fiel F q an an integer

More information

SEG Houston 2009 International Exposition and Annual Meeting

SEG Houston 2009 International Exposition and Annual Meeting Fonations of the metho of M fiel separation into pgoing an ongoing parts an its application to MCSM ata Michael S. Zhanov an Shming Wang*, Universit of Utah Smmar The renee interest in the methos of electromagnetic

More information

Chapter 3 MATHEMATICAL MODELING OF DYNAMIC SYSTEMS

Chapter 3 MATHEMATICAL MODELING OF DYNAMIC SYSTEMS Chapter 3 MATHEMATICAL MODELING OF DYNAMIC SYSTEMS 3. System Modeling Mathematical Modeling In designing control systems we mst be able to model engineered system dynamics. The model of a dynamic system

More information

Mehmet Pakdemirli* Precession of a Planet with the Multiple Scales Lindstedt Poincare Technique (2)

Mehmet Pakdemirli* Precession of a Planet with the Multiple Scales Lindstedt Poincare Technique (2) Z. Natrforsch. 05; aop Mehmet Pakemirli* Precession of a Planet with the Mltiple Scales Linstet Poincare Techniqe DOI 0.55/zna-05-03 Receive May, 05; accepte Jly 5, 05 Abstract: The recently evelope mltiple

More information

Nonlinear predictive control of dynamic systems represented by Wiener Hammerstein models

Nonlinear predictive control of dynamic systems represented by Wiener Hammerstein models Nonlinear Dn (26) 86:93 24 DOI.7/s7-6-2957- ORIGINAL PAPER Nonlinear predictive control of dnamic sstems represented b Wiener Hammerstein models Maciej Ławrńcz Received: 7 December 25 / Accepted: 2 Jl

More information

Linear System Theory (Fall 2011): Homework 1. Solutions

Linear System Theory (Fall 2011): Homework 1. Solutions Linear System Theory (Fall 20): Homework Soltions De Sep. 29, 20 Exercise (C.T. Chen: Ex.3-8). Consider a linear system with inpt and otpt y. Three experiments are performed on this system sing the inpts

More information

Introdction In the three papers [NS97], [SG96], [SGN97], the combined setp or both eedback and alt detection lter design problem has been considered.

Introdction In the three papers [NS97], [SG96], [SGN97], the combined setp or both eedback and alt detection lter design problem has been considered. Robst Falt Detection in Open Loop s. losed Loop Henrik Niemann Jakob Stostrp z Version: Robst_FDI4.tex { Printed 5h 47m, Febrar 9, 998 Abstract The robstness aspects o alt detection and isolation (FDI)

More information

Outline. Model Predictive Control: Current Status and Future Challenges. Separation of the control problem. Separation of the control problem

Outline. Model Predictive Control: Current Status and Future Challenges. Separation of the control problem. Separation of the control problem Otline Model Predictive Control: Crrent Stats and Ftre Challenges James B. Rawlings Department of Chemical and Biological Engineering University of Wisconsin Madison UCLA Control Symposim May, 6 Overview

More information

Theorem (Change of Variables Theorem):

Theorem (Change of Variables Theorem): Avance Higher Notes (Unit ) Prereqisites: Integrating (a + b) n, sin (a + b) an cos (a + b); erivatives of tan, sec, cosec, cot, e an ln ; sm/ifference rles; areas ner an between crves. Maths Applications:

More information

FRTN10 Exercise 12. Synthesis by Convex Optimization

FRTN10 Exercise 12. Synthesis by Convex Optimization FRTN Exercise 2. 2. We want to design a controller C for the stable SISO process P as shown in Figre 2. sing the Yola parametrization and convex optimization. To do this, the control loop mst first be

More information

Fuzzy Control of a Nonlinear Deterministic System for Different Operating Points

Fuzzy Control of a Nonlinear Deterministic System for Different Operating Points International Jornal of Electrical & Compter Sciences IJECS-IJE Vol: No: 9 Fzz Control of a Nonlinear Deterministic Sstem for Different Operating Points Gonca Ozmen Koca, Cafer Bal, Firat Universit, Technical

More information

Robust H 2 Synthesis for Dual-stage Multi-sensing Track-following Servo Systems in HDDs

Robust H 2 Synthesis for Dual-stage Multi-sensing Track-following Servo Systems in HDDs Proceedings of the 6 American Control Conference Minneapolis Minnesota USA Jne 14-16 6 WeB18.1 Robst H Snthesis for al-stage Mlti-sensing Track-following Servo Sstems in Hs Rozo Nagamne Xinghi Hang and

More information

sa/3.0/

sa/3.0/ Feeback Control Sstems Introction This work is license ner the Creative Commons Attribtion- NonCommercial-ShareAlike 3.0 Unporte License. To view a cop of this license, visit http://creativecommons.org/licenses/b-ncsa/3.0/

More information

Chapter 2 Difficulties associated with corners

Chapter 2 Difficulties associated with corners Chapter Difficlties associated with corners This chapter is aimed at resolving the problems revealed in Chapter, which are cased b corners and/or discontinos bondar conditions. The first section introdces

More information

Classify by number of ports and examine the possible structures that result. Using only one-port elements, no more than two elements can be assembled.

Classify by number of ports and examine the possible structures that result. Using only one-port elements, no more than two elements can be assembled. Jnction elements in network models. Classify by nmber of ports and examine the possible strctres that reslt. Using only one-port elements, no more than two elements can be assembled. Combining two two-ports

More information

3 2D Elastostatic Problems in Cartesian Coordinates

3 2D Elastostatic Problems in Cartesian Coordinates D lastostatic Problems in Cartesian Coordinates Two dimensional elastostatic problems are discssed in this Chapter, that is, static problems of either plane stress or plane strain. Cartesian coordinates

More information

Control Using Logic & Switching: Part III Supervisory Control

Control Using Logic & Switching: Part III Supervisory Control Control Using Logic & Switching: Part III Spervisor Control Ttorial for the 40th CDC João P. Hespanha Universit of Sothern California Universit of California at Santa Barbara Otline Spervisor control overview

More information

On the Total Duration of Negative Surplus of a Risk Process with Two-step Premium Function

On the Total Duration of Negative Surplus of a Risk Process with Two-step Premium Function Aailable at http://pame/pages/398asp ISSN: 93-9466 Vol, Isse (December 7), pp 7 (Preiosly, Vol, No ) Applications an Applie Mathematics (AAM): An International Jornal Abstract On the Total Dration of Negatie

More information

A Model-Free Adaptive Control of Pulsed GTAW

A Model-Free Adaptive Control of Pulsed GTAW A Model-Free Adaptive Control of Plsed GTAW F.L. Lv 1, S.B. Chen 1, and S.W. Dai 1 Institte of Welding Technology, Shanghai Jiao Tong University, Shanghai 00030, P.R. China Department of Atomatic Control,

More information

Ramp Metering Control on the Junction of Freeway and Separated Connecting Collector-Distributor Roadway

Ramp Metering Control on the Junction of Freeway and Separated Connecting Collector-Distributor Roadway Proceeings of the 8th WSEAS International Conference on APPLIED MATHEMATICS, Tenerife, Spain, December 16-18, 5 (pp45-5) Ramp Metering Control on the Jnction of Freeway an Separate Connecting Collector-Distribtor

More information

Electromagnet 1 Electromagnet 2. Rotor. i 2 + e 2 - V 2. - m R 2. x -x. zero bias. Force. low bias. Control flux

Electromagnet 1 Electromagnet 2. Rotor. i 2 + e 2 - V 2. - m R 2. x -x. zero bias. Force. low bias. Control flux Low-Bias Control of AMB's Sbject to Satration Constraints anagiotis Tsiotras an Efstathios Velenis School of Aerospace Engineering Georgia Institte of Technology, Atlanta, GA 333-5, USA p.tsiotras@ae.gatech.e,

More information

Online Solution of State Dependent Riccati Equation for Nonlinear System Stabilization

Online Solution of State Dependent Riccati Equation for Nonlinear System Stabilization > REPLACE American HIS Control LINE Conference WIH YOUR PAPER IDENIFICAION NUMBER (DOUBLE-CLICK HERE O EDI) FrC3. Marriott Waterfront, Baltimore, MD, USA Jne 3-Jly, Online Soltion of State Dependent Riccati

More information

COMPARATIVE STUDY OF ROBUST CONTROL TECHNIQUES FOR OTTO-CYCLE MOTOR CONTROL

COMPARATIVE STUDY OF ROBUST CONTROL TECHNIQUES FOR OTTO-CYCLE MOTOR CONTROL ACM Symposim Series in Mechatronics - Vol. - pp.76-85 Copyright c 28 by ACM COMPARATIVE STUDY OF ROUST CONTROL TECHNIQUES FOR OTTO-CYCLE MOTOR CONTROL Marcos Salazar Francisco, marcos.salazar@member.isa.org

More information

STABILIZATIO ON OF LONGITUDINAL AIRCRAFT MOTION USING MODEL PREDICTIVE CONTROL AND EXACT LINEARIZATION

STABILIZATIO ON OF LONGITUDINAL AIRCRAFT MOTION USING MODEL PREDICTIVE CONTROL AND EXACT LINEARIZATION 8 TH INTERNATIONAL CONGRESS OF THE AERONAUTICAL SCIENCES STABILIZATIO ON OF LONGITUDINAL AIRCRAFT MOTION USING MODEL PREDICTIVE CONTROL AND EXACT LINEARIZATION Čeliovsý S.*, Hospodář P.** *CTU Prage, Faclty

More information

BOND-GRAPH BASED CONTROLLER DESIGN OF A TWO-INPUT TWO-OUTPUT FOUR-TANK SYSTEM

BOND-GRAPH BASED CONTROLLER DESIGN OF A TWO-INPUT TWO-OUTPUT FOUR-TANK SYSTEM BOND-GAPH BASED CONTOLLE DESIGN OF A TWO-INPUT TWO-OUTPUT FOU-TANK SYSTEM Nacsse, Matías A. (a) and Jnco, Sergio J. (b) LAC, Laboratorio de Atomatización y Control, Departamento de Control, Facltad de

More information

Modeling of a Self-Oscillating Cantilever

Modeling of a Self-Oscillating Cantilever Moeling of a Self-Oscillating Cantilever James Blanchar, Hi Li, Amit Lal, an Doglass Henerson University of Wisconsin-Maison 15 Engineering Drive Maison, Wisconsin 576 Abstract A raioisotope-powere, self-oscillating

More information

Nonparametric Identification and Robust H Controller Synthesis for a Rotational/Translational Actuator

Nonparametric Identification and Robust H Controller Synthesis for a Rotational/Translational Actuator Proceedings of the 6 IEEE International Conference on Control Applications Mnich, Germany, October 4-6, 6 WeB16 Nonparametric Identification and Robst H Controller Synthesis for a Rotational/Translational

More information

n s n Z 0 on complex-valued functions on the circle. Then sgn n + 1 ) n + 1 2) s

n s n Z 0 on complex-valued functions on the circle. Then sgn n + 1 ) n + 1 2) s . What is the eta invariant? The eta invariant was introce in the famos paper of Atiyah, Patoi, an Singer see [], in orer to proce an inex theorem for manifols with bonary. The eta invariant of a linear

More information

Gradient Projection Anti-windup Scheme on Constrained Planar LTI Systems. Justin Teo and Jonathan P. How

Gradient Projection Anti-windup Scheme on Constrained Planar LTI Systems. Justin Teo and Jonathan P. How 1 Gradient Projection Anti-windp Scheme on Constrained Planar LTI Systems Jstin Teo and Jonathan P. How Technical Report ACL1 1 Aerospace Controls Laboratory Department of Aeronatics and Astronatics Massachsetts

More information

Desert Mountain H. S. Math Department Summer Work Packet

Desert Mountain H. S. Math Department Summer Work Packet Corse #50-51 Desert Montain H. S. Math Department Smmer Work Packet Honors/AP/IB level math corses at Desert Montain are for stents who are enthsiastic learners of mathematics an whose work ethic is of

More information

A Macroscopic Traffic Data Assimilation Framework Based on Fourier-Galerkin Method and Minimax Estimation

A Macroscopic Traffic Data Assimilation Framework Based on Fourier-Galerkin Method and Minimax Estimation A Macroscopic Traffic Data Assimilation Framework Based on Forier-Galerkin Method and Minima Estimation Tigran T. Tchrakian and Sergiy Zhk Abstract In this paper, we propose a new framework for macroscopic

More information

Throughput Maximization for Tandem Lines with Two Stations and Flexible Servers

Throughput Maximization for Tandem Lines with Two Stations and Flexible Servers Throghpt Maximization for Tanem Lines with Two Stations an Flexible Servers Sigrún Anraóttir an Hayriye Ayhan School of Instrial an Systems Engineering Georgia Institte of Technology Atlanta GA 30332-0205

More information

The Oscillatory Stable Regime of Nonlinear Systems, with two time constants

The Oscillatory Stable Regime of Nonlinear Systems, with two time constants 6th WSES International Conference on CIRCUITS SYSTEMS ELECTRONICSCONTROL & SIGNL PROCESSING Cairo Egpt Dec 9-3 7 5 The Oscillator Stable Regime of Nonlinear Sstems with two time constants NUŢU VSILE *

More information

AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS

AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS Fang-Ming Y, Hng-Yan Chng* and Chen-Ning Hang Department of Electrical Engineering National Central University, Chngli,

More information

AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS

AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS AN ALTERNATIVE DECOUPLED SINGLE-INPUT FUZZY SLIDING MODE CONTROL WITH APPLICATIONS Fang-Ming Y, Hng-Yan Chng* and Chen-Ning Hang Department of Electrical Engineering National Central University, Chngli,

More information

Solution to Tutorial 8

Solution to Tutorial 8 Soltion to Ttorial 8 01/013 Semester I MA464 Game Theory Ttor: Xiang Sn October, 01 1 eview A perfect Bayesian eqilibrim consists of strategies an beliefs satisfying eqirements 1 throgh 4. eqirement 1:

More information

The Linear Quadratic Regulator

The Linear Quadratic Regulator 10 The Linear Qadratic Reglator 10.1 Problem formlation This chapter concerns optimal control of dynamical systems. Most of this development concerns linear models with a particlarly simple notion of optimality.

More information

1 Introduction. r + _

1 Introduction. r + _ A method and an algorithm for obtaining the Stable Oscillator Regimes Parameters of the Nonlinear Sstems, with two time constants and Rela with Dela and Hsteresis NUŢU VASILE, MOLDOVEANU CRISTIAN-EMIL,

More information

Lecture Notes: Finite Element Analysis, J.E. Akin, Rice University

Lecture Notes: Finite Element Analysis, J.E. Akin, Rice University 9. TRUSS ANALYSIS... 1 9.1 PLANAR TRUSS... 1 9. SPACE TRUSS... 11 9.3 SUMMARY... 1 9.4 EXERCISES... 15 9. Trss analysis 9.1 Planar trss: The differential eqation for the eqilibrim of an elastic bar (above)

More information

Design Method for RC Building Structure Controlled by Elasto-Plastic Dampers Using Performance Curve

Design Method for RC Building Structure Controlled by Elasto-Plastic Dampers Using Performance Curve Design Metho or RC Biling Strctre Controlle by Elasto-Plastic Dampers Using Perormance Crve W. P Whan University o Technology, China K. Kasai Tokyo Institte o Technology, Japan SUMMARY: This paper proposes

More information

Nonlinear parametric optimization using cylindrical algebraic decomposition

Nonlinear parametric optimization using cylindrical algebraic decomposition Proceedings of the 44th IEEE Conference on Decision and Control, and the Eropean Control Conference 2005 Seville, Spain, December 12-15, 2005 TC08.5 Nonlinear parametric optimization sing cylindrical algebraic

More information

Assignment Fall 2014

Assignment Fall 2014 Assignment 5.086 Fall 04 De: Wednesday, 0 December at 5 PM. Upload yor soltion to corse website as a zip file YOURNAME_ASSIGNMENT_5 which incldes the script for each qestion as well as all Matlab fnctions

More information

Hybrid modelling and model reduction for control & optimisation

Hybrid modelling and model reduction for control & optimisation Hybrid modelling and model redction for control & optimisation based on research done by RWTH-Aachen and TU Delft presented by Johan Grievink Models for control and optimiation market and environmental

More information

Optimal Contract for Machine Repair and Maintenance

Optimal Contract for Machine Repair and Maintenance Optimal Contract for Machine Repair an Maintenance Feng Tian University of Michigan, ftor@mich.e Peng Sn Dke University, psn@ke.e Izak Denyas University of Michigan, enyas@mich.e A principal hires an agent

More information

Development of Second Order Plus Time Delay (SOPTD) Model from Orthonormal Basis Filter (OBF) Model

Development of Second Order Plus Time Delay (SOPTD) Model from Orthonormal Basis Filter (OBF) Model Development of Second Order Pls Time Delay (SOPTD) Model from Orthonormal Basis Filter (OBF) Model Lemma D. Tfa*, M. Ramasamy*, Sachin C. Patwardhan **, M. Shhaimi* *Chemical Engineering Department, Universiti

More information

System identification of buildings equipped with closed-loop control devices

System identification of buildings equipped with closed-loop control devices System identification of bildings eqipped with closed-loop control devices Akira Mita a, Masako Kamibayashi b a Keio University, 3-14-1 Hiyoshi, Kohok-k, Yokohama 223-8522, Japan b East Japan Railway Company

More information

DESIGN AND SIMULATION OF SELF-TUNING PREDICTIVE CONTROL OF TIME-DELAY PROCESSES

DESIGN AND SIMULATION OF SELF-TUNING PREDICTIVE CONTROL OF TIME-DELAY PROCESSES DESIGN AND SIMULAION OF SELF-UNING PREDICIVE CONROL OF IME-DELAY PROCESSES Vladimír Bobál,, Marek Kbalčík and Petr Dostál, omas Bata University in Zlín Centre of Polymer Systems, University Institte Department

More information

2.13 Variation and Linearisation of Kinematic Tensors

2.13 Variation and Linearisation of Kinematic Tensors Section.3.3 Variation an Linearisation of Kinematic ensors.3. he Variation of Kinematic ensors he Variation In this section is reviewe the concept of the variation, introce in Part I, 8.5. he variation

More information

INPUT-OUTPUT APPROACH NUMERICAL EXAMPLES

INPUT-OUTPUT APPROACH NUMERICAL EXAMPLES INPUT-OUTPUT APPROACH NUMERICAL EXAMPLES EXERCISE s consider the linear dnamical sstem of order 2 with transfer fnction with Determine the gain 2 (H) of the inpt-otpt operator H associated with this sstem.

More information

Port-Hamiltonian descriptor systems

Port-Hamiltonian descriptor systems Port-Hamiltonian escriptor systems Christopher Beattie an Volker Mehrmann an Honggo X an Hans Zwart May 29, 2017 Abstract The moeling framework of port-hamiltonian systems is systematically extene to constraine

More information

State Space Models Basic Concepts

State Space Models Basic Concepts Chapter 2 State Space Models Basic Concepts Related reading in Bay: Chapter Section Sbsection 1 (Models of Linear Systems) 1.1 1.1.1 1.1.2 1.1.3 1.1.5 1.2 1.2.1 1.2.2 1.3 In this Chapter we provide some

More information

FEA Solution Procedure

FEA Solution Procedure EA Soltion Procedre (demonstrated with a -D bar element problem) EA Procedre for Static Analysis. Prepare the E model a. discretize (mesh) the strctre b. prescribe loads c. prescribe spports. Perform calclations

More information

Formal Methods for Deriving Element Equations

Formal Methods for Deriving Element Equations Formal Methods for Deriving Element Eqations And the importance of Shape Fnctions Formal Methods In previos lectres we obtained a bar element s stiffness eqations sing the Direct Method to obtain eact

More information

Lecture Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2

Lecture Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2 BIJU PATNAIK UNIVERSITY OF TECHNOLOGY, ODISHA Lectre Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2 Prepared by, Dr. Sbhend Kmar Rath, BPUT, Odisha. Tring Machine- Miscellany UNIT 2 TURING MACHINE

More information

Safe Manual Control of the Furuta Pendulum

Safe Manual Control of the Furuta Pendulum Safe Manal Control of the Frta Pendlm Johan Åkesson, Karl Johan Åström Department of Atomatic Control, Lnd Institte of Technology (LTH) Box 8, Lnd, Sweden PSfrag {jakesson,kja}@control.lth.se replacements

More information

1 Differential Equations for Solid Mechanics

1 Differential Equations for Solid Mechanics 1 Differential Eqations for Solid Mechanics Simple problems involving homogeneos stress states have been considered so far, wherein the stress is the same throghot the component nder std. An eception to

More information

HOMEWORK 2 SOLUTIONS

HOMEWORK 2 SOLUTIONS HOMEWORK 2 SOLUTIONS PHIL SAAD 1. Carroll 1.4 1.1. A qasar, a istance D from an observer on Earth, emits a jet of gas at a spee v an an angle θ from the line of sight of the observer. The apparent spee

More information

Online identification of continuous bimodal and trimodal piecewise affine systems

Online identification of continuous bimodal and trimodal piecewise affine systems Delft University of Technology Online ientification of continos bimoal an trimoal piecewise affine systems Le, Than; van en Boom, Ton; Bali, Simone DOI 9/ECC2678432 Pblication ate 26 Docment Version Peer

More information

1 Undiscounted Problem (Deterministic)

1 Undiscounted Problem (Deterministic) Lectre 9: Linear Qadratic Control Problems 1 Undisconted Problem (Deterministic) Choose ( t ) 0 to Minimize (x trx t + tq t ) t=0 sbject to x t+1 = Ax t + B t, x 0 given. x t is an n-vector state, t a

More information

NEURAL CONTROLLERS FOR NONLINEAR SYSTEMS IN MATLAB

NEURAL CONTROLLERS FOR NONLINEAR SYSTEMS IN MATLAB NEURAL CONTROLLERS FOR NONLINEAR SYSTEMS IN MATLAB S.Kajan Institte of Control and Indstrial Informatics, Faclt of Electrical Engineering and Information Technolog, Slovak Universit of Technolog in Bratislava,

More information

Material. Lecture 8 Backlash and Quantization. Linear and Angular Backlash. Example: Parallel Kinematic Robot. Backlash.

Material. Lecture 8 Backlash and Quantization. Linear and Angular Backlash. Example: Parallel Kinematic Robot. Backlash. Lectre 8 Backlash and Qantization Material Toda s Goal: To know models and compensation methods for backlash Lectre slides Be able to analze the effect of qantization errors Note: We are sing analsis methods

More information

Logarithmic, Exponential and Other Transcendental Functions

Logarithmic, Exponential and Other Transcendental Functions Logarithmic, Eponential an Other Transcenental Fnctions 5: The Natral Logarithmic Fnction: Differentiation The Definition First, yo mst know the real efinition of the natral logarithm: ln= t (where > 0)

More information

Constraints on fourth generation Majorana neutrinos

Constraints on fourth generation Majorana neutrinos Jornal of Physics: Conference Series Constraints on forth generation Majorana netrinos To cite this article: Alexaner Lenz et al 2010 J. Phys.: Conf. Ser. 259 012096 Relate content - Lepton nmber, black

More information

Active Flux Schemes for Advection Diffusion

Active Flux Schemes for Advection Diffusion AIAA Aviation - Jne, Dallas, TX nd AIAA Comptational Flid Dynamics Conference AIAA - Active Fl Schemes for Advection Diffsion Hiroaki Nishikawa National Institte of Aerospace, Hampton, VA 3, USA Downloaded

More information

Theoretical and Experimental Implementation of DC Motor Nonlinear Controllers

Theoretical and Experimental Implementation of DC Motor Nonlinear Controllers Theoretical and Experimental Implementation of DC Motor Nonlinear Controllers D.R. Espinoza-Trejo and D.U. Campos-Delgado Facltad de Ingeniería, CIEP, UASLP, espinoza trejo dr@aslp.mx Facltad de Ciencias,

More information

Decentralized Control with Moving-Horizon Linear Switched Systems: Synthesis and Testbed Implementation

Decentralized Control with Moving-Horizon Linear Switched Systems: Synthesis and Testbed Implementation 17 American Control Conference Sheraton Seattle Hotel May 4 6, 17, Seattle, USA Decentralized Control with Moving-Horizon Linear Switched Systems: Synthesis and Testbed Implementation Joao P Jansch-Porto

More information

MEG 741 Energy and Variational Methods in Mechanics I

MEG 741 Energy and Variational Methods in Mechanics I MEG 74 Energ and Variational Methods in Mechanics I Brendan J. O Toole, Ph.D. Associate Professor of Mechanical Engineering Howard R. Hghes College of Engineering Universit of Nevada Las Vegas TBE B- (7)

More information

Decoder Error Probability of MRD Codes

Decoder Error Probability of MRD Codes Decoder Error Probability of MRD Codes Maximilien Gadolea Department of Electrical and Compter Engineering Lehigh University Bethlehem, PA 18015 USA E-mail: magc@lehighed Zhiyan Yan Department of Electrical

More information

18-660: Numerical Methods for Engineering Design and Optimization

18-660: Numerical Methods for Engineering Design and Optimization 8-66: Nmerical Methos for Engineering Design an Optimization in Li Department of ECE Carnegie Mellon University Pittsbrgh, PA 53 Slie Overview Geometric Problems Maximm inscribe ellipsoi Minimm circmscribe

More information

Lecture 6 : Linear Fractional Transformations (LFTs) Dr.-Ing. Sudchai Boonto

Lecture 6 : Linear Fractional Transformations (LFTs) Dr.-Ing. Sudchai Boonto Lectre 6 : (LFTs) Dr-Ing Sdchai Boonto Department of Control System and Instrmentation Engineering King Mongkts Unniversity of Technology Thonbri Thailand Feedback Strctre d i d r e y z K G g n The standard

More information

FRÉCHET KERNELS AND THE ADJOINT METHOD

FRÉCHET KERNELS AND THE ADJOINT METHOD PART II FRÉCHET KERNES AND THE ADJOINT METHOD 1. Setp of the tomographic problem: Why gradients? 2. The adjoint method 3. Practical 4. Special topics (sorce imaging and time reversal) Setp of the tomographic

More information

Click to edit Master title style

Click to edit Master title style Click to edit Master title stle APMonitor Modeling Langage John Hedengren Brigham Yong Universit Advanced Process Soltions LLC http://apmonitorcom Overview of APM Software as a service accessible throgh:

More information

Motivations and Historical Perspective

Motivations and Historical Perspective Motivations and Historical Perspective Giovanni Marro DEIS, Universit of Bologna, Ital MTNS - Jl 5-9, 2010 G. Marro (Bologna, Ital) MTNS - Jl 5-9, 2010 1 / 32 Geometric Control Theor for Linear Sstems

More information

A Decomposition Method for Volume Flux. and Average Velocity of Thin Film Flow. of a Third Grade Fluid Down an Inclined Plane

A Decomposition Method for Volume Flux. and Average Velocity of Thin Film Flow. of a Third Grade Fluid Down an Inclined Plane Adv. Theor. Appl. Mech., Vol. 1, 8, no. 1, 9 A Decomposition Method for Volme Flx and Average Velocit of Thin Film Flow of a Third Grade Flid Down an Inclined Plane A. Sadighi, D.D. Ganji,. Sabzehmeidani

More information

Solving the Lienard equation by differential transform method

Solving the Lienard equation by differential transform method ISSN 1 746-7233, England, U World Jornal of Modelling and Simlation Vol. 8 (2012) No. 2, pp. 142-146 Solving the Lienard eqation by differential transform method Mashallah Matinfar, Saber Rakhshan Bahar,

More information

Integral Control via Bias Estimation

Integral Control via Bias Estimation 1 Integal Contol via Bias stimation Consie the sstem ẋ = A + B +, R n, R p, R m = C +, R q whee is an nknown constant vecto. It is possible to view as a step istbance: (t) = 0 1(t). (If in fact (t) vaies

More information

Lecture: Corporate Income Tax - Unlevered firms

Lecture: Corporate Income Tax - Unlevered firms Lectre: Corporate Income Tax - Unlevered firms Ltz Krschwitz & Andreas Löffler Disconted Cash Flow, Section 2.1, Otline 2.1 Unlevered firms Similar companies Notation 2.1.1 Valation eqation 2.1.2 Weak

More information

An extremum seeking approach to parameterised loop-shaping control design

An extremum seeking approach to parameterised loop-shaping control design Preprints of the 19th World Congress The International Federation of Atomatic Control An extremm seeking approach to parameterised loop-shaping control design Chih Feng Lee Sei Zhen Khong Erik Frisk Mattias

More information

CS 450: COMPUTER GRAPHICS VECTORS SPRING 2016 DR. MICHAEL J. REALE

CS 450: COMPUTER GRAPHICS VECTORS SPRING 2016 DR. MICHAEL J. REALE CS 45: COMPUTER GRPHICS VECTORS SPRING 216 DR. MICHEL J. RELE INTRODUCTION In graphics, we are going to represent objects and shapes in some form or other. First, thogh, we need to figre ot how to represent

More information

Linear Strain Triangle and other types of 2D elements. By S. Ziaei Rad

Linear Strain Triangle and other types of 2D elements. By S. Ziaei Rad Linear Strain Triangle and other tpes o D elements B S. Ziaei Rad Linear Strain Triangle (LST or T6 This element is also called qadratic trianglar element. Qadratic Trianglar Element Linear Strain Triangle

More information

Light flavor asymmetry of polarized quark distributions in thermodynamical bag model

Light flavor asymmetry of polarized quark distributions in thermodynamical bag model Inian Jornal of Pre & Applie Physics Vol. 5, April 014, pp. 19-3 Light flavor asymmetry of polarize qark istribtions in thermoynamical bag moel K Ganesamrthy a & S Mrganantham b* a Department of Physics,

More information

Simplified Identification Scheme for Structures on a Flexible Base

Simplified Identification Scheme for Structures on a Flexible Base Simplified Identification Scheme for Strctres on a Flexible Base L.M. Star California State University, Long Beach G. Mylonais University of Patras, Greece J.P. Stewart University of California, Los Angeles

More information

Hongliang Yang and Michael Pollitt. September CWPE 0741 and EPRG 0717

Hongliang Yang and Michael Pollitt. September CWPE 0741 and EPRG 0717 Distingishing Weak an Strong Disposability among Unesirable Otpts in DEA: The Example of the Environmental Efficiency of Chinese Coal-Fire Power Plants Hongliang Yang an Michael Pollitt September 2007

More information