Bifurcation & Chaos in Structural Dynamics: A Comparison of the Hilber-Hughes-Taylor-α and the Wang-Atluri (WA) Algorithms

Size: px
Start display at page:

Download "Bifurcation & Chaos in Structural Dynamics: A Comparison of the Hilber-Hughes-Taylor-α and the Wang-Atluri (WA) Algorithms"

Transcription

1 Bfurcaon & Chaos n Srucural Dynamcs: A Comparson of he Hlber-Hughes-Taylor-α and he Wang-Alur (WA) Algorhms Absrac Xuechuan Wang*, Wecheng Pe**, and Saya N. Alur* Texas Tech Unversy, Lubbock, TX, 7945 Cener for Advanced Research n he Engneerng Scences The WA algorhms (Compuaonal Mechancs, Vol. 59, No. 5, pp , 07) are rederved as beng based on opmal-feedback-acceleraed Pcard eraon, wheren he soluon vecors a any me n a fnely large me nerval + are correced by a weghed (wh a marx λ ) negral of he error from o. The 3 WA algorhms are rederved n deal, based on 3 dfferen approxmae soluons o he Euler-Lagrange equaons for λ. The nerval ( + ) n he 3 WA algorhms can be several hundred mes larger han he ncremen ( ) requred n he HHT-α, for he same sably and accuracy. Moreover, he WA algorhms-, 3 do no requre he nverson of he angen sffness marx, as s requred n HHTα. I s found ha WA algorhms-,, 3 (especally WA algorhm-) are far more superor o HHT-α and ode-45 n erms of compuaonal speed, accuracy, and convergence. Keywords: Nonlnear srucural dynamcs; bfurcaon and chaos; HHT-α algorhm; Wang-Alur algorhms * Vsng Scholar from Norhwesern Polyechncal Unversy, and Graduae Suden n Texas Tech Unversy, Correspondng Auhor, Emal: xc.wang@u.edu ** Vsng Scholar, from Behang Unversy * Drecor, Cener for Advanced Research n he Engneerng Scences

2 . Inroducon Nonlneares n he mechancs of srucures could arse from varous sources such as large deformaons, large roaons, nonlneares of he maeral, dampng, and boundary condons, ec. [] They are so common n real lfe ha he nonlnear behavors can be found everywhere rangng from cvl engneerng, auomoble manufacory, arcraf desgn, o roboc dynamcs, and on-orb maneuvers of spacecraf. Wh he developmen of lgher, more flexble and mulfunconal maerals and srucures, he effec of nonlneary wll become more sgnfcan n hese areas. Lnearzaon was once wdely used o smplfy nonlnear problems, bu n he meanme, was also realzed ha he basc prncples of lnear analyss are no vald even n some cases of weak nonlneary. The nonlnear dynamcal sysem s characerzed by some unque and complex phenomena, ncludng bfurcaon, jump phenomena, snap-hrough, and chaos, ha he lnear sysem canno exhb []. For weakly nonlnear problems, he perurbaon mehod has been successfully used o oban asympocally analycal approxmaons of he soluons. However, for srongly nonlnear dynamcal sysems wh complex forms, he soluon can only be obaned usng numercal negraon mehods, n he curren sae of mahemacs. The nvesgaon of srucural vbraon helps o enhance he relably and performance of srucures, as well as reveal poenal falures. Wh proper srucural models, one can use numercal smulaon as an alernave o expermen, and o save a lo of me and expense. In leraure, he mehods for numercally solvng a sysem of nonlnear ordnary dfferenal equaons n srucural dynamcs are roughly dvded no explc and mplc mehods. The fne dfference mehods are a broad class of explc mehods nvolvng cenral dfference mehod, Runge-Kua mehods and Taylor seres expanson schemes [3]. These mehods are very easy o mplemen, bu ofen lack sably n compuaons usng large sze seps. The mplemenaon of an mplc mehod resuls n nonlnear algebrac equaons of unknown saes, whch are usually solved wh Newon-Raphson mehods. For some ypcal mplc mehods, such as he Newmarkβ mehods [4] and he Hlber-Hughes-Taylor-α mehod [5], he sably s guaraneed only for lnear problems. However, for nonlnear problems, he sably s doubful and ofen depends on he specfc problem and he sep sze adoped n he mehod. Wh relavely large sze seps, some mplc mehods could become so unsable and naccurae ha hey are obvously unsued for numercal negraon o fnd he long-erm nonlnear dynamcal behavors [6]. Even hough, n he analyss of vbrang srucures by a fne elemen spaal dscrezaon mehod, for he negraon of he sem-dscree nonlnear me ODEs, schemes such as he cenral dfference, he Newmark and he Hlber-Hughes-Taylor-α (HHT-α ) mehods are sll wdely used n ndusral sofware whou relable proofs showng ha he nonlnear behavors ncludng bfurcaon and chaos can be accuraely predced by hese mehods. Ths paper concerns he phenomena of bfurcaon and chaos n srucural dynamcs, governed by a semdscree sysem of equaons n me, afer a spaal dscrezaon has been carred ou. Analyses of srucural dynamcs problems have been carred ou exensvely n he pas, by usng he Newmark- β and s slgh modfcaon wh added dampng, he Hlber-Hughes-Taylor- α mehods, wheren Taylor seres expansons of he dsplacemen and velocy vecors are used over a very small me sep, around he generc me. On he oher hand, he Wang-Alur algorhms, whch nvolve fnely large me nervals ( + ) were frs nroduced n [7], and appled o orbal mechancs problems n [8]. In orbal mechancs problems, he WA algorhms were proved o be he bes n leraure so far, n erms of compuaonal me, accuracy, and speed of convergence. These algorhms are renerpreed n he presen paper as beng derved from hghly opmal feedback acceleraed Pcard eraon mehods, wheren he soluon vecor conssng of dsplacemens ( x ) and veloces ( x ) a any me, +, s correced by an opmally

3 weghed (wh he weghng marx λ ) negral of he error n he soluon vecor from o. By m assumng orhogonal polynomals for he soluon vecor over ( + ), and usng collocaon pons ( m=,,..., M ) whn ( + ), we use he opmally acceleraed Pcard-eraon from m o, o reduce he sem-dscree nonlnear ODEs o algebrac equaons. We presen 3 approxmaons o solve he Euler-Lagrange equaons for he opmal wegh funcons λ. Thus we presen 3 algorhms denoed as Wang-Alur algorhms-,, 3. In hese algorhms, he me nerval ( + ) can be several hundred mes larger han he used n he HHT-α mehod, for he same sable and accurae performance. Even m m he dsance beween he collocaon pons + whn he nerval + can be several mes larger han he used n he HHT-α mehod. In addon, he WA algorhms- and 3 do no nvolve he nverson of a Jacoban (angen sffness) marx, as s he case n he HHT-α mehod. I s shown ha all he WA algorhms-,, 3 (especally he WA algorhm-) are far more superor o he HHT-α algorhm n compuaonal speed, accuracy, and speed of convergence, n he presenly consdered problems of bfurcaon and chaos n srucural dynamcs. The WA algorhms are also superor o he opmzed ODE45 algorhm n MATLAB, n erms of speed and accuracy. The WA algorhms- may be consdered as a sgnfcan mprovemen o he sae of scence n nonlnear srucural dynamcs. The nonlnear vbraon of a buckled beam s very represenave n nonlnear srucural dynamcs and has araced a number of researchers [, ]. Commonly he dynamcs of a buckled beam can be well modelled by fne elemen, boundary elemen, or meshless mehods. However, s shown n [3] ha he clamped-clamped buckled beam model can also be well dscrezed no a four-mode approxmaon usng he spaal Galerkn mehod. In hs paper, boh he Hlber-Hughes-Taylor-α and Wang-Alur algorhms are used o nvesgae he nonlnear vbraons of a buckled beam whch may exhb bfurcaon, jump phenomena, and chaos. The valdy of he four-mode buckled beam model s frs examned, and he nonlnear dynamcal behavors such as perod doublng process and chaoc moon are hen nvesgaed. The numercal resuls obaned by Wang-Alur mehod are compared wh hose by Hlber-Hughes-Taylorα mehod o evaluae he performance of hese wo mehods. I s shown ha he WA algorhms are far more superor o he HHT-α algorhms n erms of accuracy, compuaonal me, and convergence speed, n problems nvolvng bfurcaon and chaos n nonlnear srucural dynamcs.. An Elucdaon of HHT-α and WA Algorhms In srucural dynamcs modelng, he vbraon of a srucure s ofen descrbed by a sysem of second-order nonlnear ordnary dfferenal equaons afer spaal dscrezaon has already been carred ou, where x s he vecor of generalzed dsplacemens, and x are he acceleraons, M s he mass marx, C s he dampng marx, N s he vecor of nonlnear resorng forces whch may depend on x and he velocy vecor x, and F s he exernal force. Mx + Cx + N( x, x) = F (). () Eq. () can be furher rewren as a sysem of nonlnear frs order ODEs: x x =, () x M Cx M N( x, x) + M F() 3

4 where N denoes he nonlnear erms, x and x represen dsplacemen and velocy vecor of he moon. Eq. () may be consdered as he mxed-form equvalen of Eq. (), wheren x s he dsplacemen vecor and x s he velocy vecor.. HHT-α Algorhm Based on he heory of Taylor seres, he dscrezed dsplacemen x ( ) + (where s nfnesmally small) can be approxmaed [as done by Newmark (4)] by 3 x( + ) = x( ) + x ( ) + x ( ) + x ( ) + 3!, (3) 3 = x( ) + x ( ) + x ( ) + βx ( ) where β. Smlarly, usng Taylor seres expanson, x 6 ( ) + can be approxmaed [as done by Newmark (4)] by where γ. x ( ) x ( ) x ( ) x ( ) = x ( ) + x ( ) + γ x ( ) + = + + +, (5) Snce s very small, x ( ) n he las erm of Eq. (3) can be approxmaed as Subsung Eq. (6) no Eq. (3) leads o x ( ) = [ x ( + ) x ( )]. (6) x ( + ) = x ( ) + x ( ) + x ( ) + β[ x ( + ) x ( )] = x( ) + x ( ) + ( β) x ( ) + βx ( + ) Furher, x ( ) n he las erm of Eq. (5) s approxmaed by Subsung no Eq. (5) leads o. (7) x ( ) = [ x ( + ) x ( )]. (8) x( + ) = x( ) + x ( ) + γ[ x ( + ) x ( )]. (9) = x ( ) + ( γ) x ( ) + γx ( + ) 4

5 From Eq. (), we have x ( + ) = x( + ). (0) x ( + ) = x ( + ) = M Cx( + ) M N[ x( + ), x( + )] + M F( + ) By subsung Eq. (0) no Eqs. (7, 9), a sysem of nonlnear algebrac equaons abou x ( ) + and x ( ) + are obaned as M Cx( + ) M F( + ) x( + ) x( ) x ( ) ( β) x ( ) + β 0 = + M Nx [ ( + ), x( + ) ] M Cx( + ) M F( + ) ( ) ( ) ( γ) x + x x ( ) + γ + = 0 M Nx [ ( + ), x( + )] In HHT- α mehod, he proceedng algorhms of Newmark- β mehod are slghly modfed by replacng x ( ) + and x ( ) + n Eq. (0) wh x ( ) + α and x ( ) + α respecvely, where Therefore, Eq. (0) s modfed as () x( + α ) = x( + ) + α[ x( + ) x( )]. () x( + α ) = x( + ) + α[ x( + ) x( )] x ( + ) = x( + α ) ( α ) x +. (3) ( ) x + = x ( + ) = M Cx( + α ) M N + ( + α ) ( + α ) M F x Subsung Eq. (3) no Eqs. (7, 9) leads o M Cx( + α ) M F( + α ) x( + ) x( ) x ( ) ( β) x ( ) + β 0 = + M Nx [ ( + α ), x( + α ) ] (4) ( + α ) ( + α ) ( ) ( ) ( γ) M Cx M F x + x x ( ) + γ = 0 + M Nx [ ( + α ), x( + α )] α The parameers n HHT-α mehod are oponal. Normally hey are seleced as γ =, and α,0 3. α β = 5

6 Eqs. (, 4) provde a sysem of nonlnear algebrac equaons abou x ( ) + and x ( ) + ha can be solved for usng Newon-Raphson mehod, where he Jacoban marx and s nverson has o be calculaed n each eraon sep. The Newon-Raphson eraon formula for solvng Eq. (4) s wren as where R j HHT j+ j x ( + ) x( + ) j ( ) j+ = j JHHT x ( + ) x( + ) x x x x = M Cx x x x M Cx R j HHT, (5) j j ( + ) ( ) ( ) ( β) ( ) + β j j + M N x( + α ), x( + α ) j j ( ) ( ) ( γ) + ( ) + γ j j + M Nx [ ( α x ( + α ) M F( + α ) ( + α ) M F( + α ) + ), ( + α )] (6) The Jacoban marx of j R wh respec o ( ) j x + and x ( ) + s derved as j HHT J + βm ( + α)[ N x ( + α )] β( + α){ M C+ M [ N x ( + α )]} =. (7) γ ( + α)[ ( + α )] + γ( + α){ + [ ( + α )]} j j j j j HHT j j j j M N x M C M N x. Wang-Alur Algorhms Based on Opmal Error-Feedback Acceleraed Pcard Ieraon Conceps Le u and u be he ral funcons for x and x n a fnely large local me nerval [, + ]. By subsung hem no Eq. (), he error resdual funcon s obaned as R() u u = +, u + M Cu + M N( u, u) M F() [, ]. (8) To opmally correc he approxmae soluon a, a smple mechansm of an opmally weghed feedback of he error s adoped heren, whch has he expresson: u() u() u u ( τ) dτ () = () + λ + u + (, ) () c u u M Cu M N u u M F [, + ],, where he subscrp c on he lef-hand sde ndcaes he correcon. In Eq. (9), λ ( τ ), a marx, s he se of opma weghng funcons for he feedback of he error resdual R (). Eq. (9) ndcaes ha he soluon vecor [ u( ); u ( )] a any me n he nerval + s correced by an opmally weghed error resdual from me o. The eraon formula of he orgnal Pcard eraon mehod [4, 5] can be regarded as a specal case of Eq. (9), where λ ( τ ) s smply seleced as he negave un marx λ( τ ) = I. Alhough he orgnal (9) 6

7 Pcard eraon mehod may converge, s no he mos effcen approach, snce λ( τ ) = I s seleced oo roughly. The dervaon of he opmal λ ( τ ) s, however, as follows. rue rue For u and u n he neghborhood of he rue soluons,.e. u = u +δu, u = u +δu, λ s expeced o be opmally deermned so ha rue u() u () rue () =. (0) u c u () rue rue I means ha he varaon of Eq. (9) should equals o be zero f u = u, u = u, whch leads o: u u u u u u δ = δ + λδ dτ δ = () + λ 0 u + () c u u M Cu M N M F u M Cu M N M F, () Snce u and u n Eq. () are supposed o be he rue soluons, we have Therefore, Eq. () leads o: where u u + = u M Cu + M N M F() 0. () u u u u δ = δ + λδ dτ + u () c u u M Cu + M N M F, (3) u u = ( I + λ) δ + { λ + λj} δ dτ = 0 u u 0 I J = ( ) + ( ). (4) M N u M C M N u In Eq. (3), here are varaons boh nsde and ousde he negral over me, hus he varaons are made o be zero separaely by enforcng he weghng funcon marx λ o sasfy he followng condons: λ() + I = 0, and λ (,) τ + λ(,) τ J() τ = 0, τ [,] (5) rue Obvously, λ s relaed o u, u, whch are supposed o be u and u rue. However, he rue soluons are no known n advance. Consderng ha, we calculaed λ n he presen paper usng approxmaed soluons nsead of he rue soluons n he mplemenaon of he algorhms, whch are labeled as WA algorhms for convenence..3 Large Tme Inerval [, + ] Orhogonal Polynomal Collocaon 7

8 Furher, he followng weak formulaon of Eq. (9) can be esablshed, usng a marx of es funcons v (). u () u () u u v() d () dτ d () = v + + () λ. u u u M Cu + M N( u, u ) M F() + + c In hs formula, v () are es funcons, he same as hose n he classcal weghed resdual mehods. Le v () be a dagonal marx v = dag[ v, v,...], and v be Drac Dela funcon for a group of collocaon m pons, n he fne large me nerval o +. For smplcy, he collocaon pons m are smply denoed as m n he followng pars. hen Eq. (6) becomes (6) v= δ ( m ), m [, + ], m=,,..., M, (7) u( m) u( m) m u u dτ ( ) = + + ( ) m λ u c u m u M Cu + M N M F [, + ], m=,..., M. (8), m Eq. (0) may be nerpreed as he correcon of he error a each collocaon pon m, wh he error resdual beng opmally weghed by λ. By usng a se of orhogonal polynomals Φ = { φ0, φ, φ,...} T as bass funcons, he ral funcons u and u can be consruced as N u = a φ, and u, = a,, φ, (9), e, en, n n= 0 N e en n n= 0 Where u,e and u,e are elemens of u and u respecvely. a, en, and a, en, are coeffcens o be deermned. From Eq. (9), we have and where T [ u ( ), u ( ),..., u ( )] = B [ a, a,..., a ], p =,, (30) T pe, pe, pe, M pe,,0 pe,, pen,, T [ u ( ), u ( ),..., u ( )] = LB [ a, a,..., a ], p =,, (3) T pe, pe, pe, M pe,,0 pe,, pen,, 8

9 B φ0( ) φ( ) φn ( ) φ ( ) φ ( ) φ ( ) φ ( ) φ ( ) φ ( ) + 0 N = 0 M M N M ( N ) M, LB φ 0 ( ) φ ( ) φ N ( ) φ ( ) φ ( ) φ ( ) 0 N = φ ( ) φ ( ) φ ( ) + 0 M M N M ( N ) M. (3) Normally, he number of colocaon pons M s seleced as he same as he number of bass funcons N +. I can be found ha he values of u pe, a collocaon pons are relaed wh hose of u pe, by T [ u ( ), u ( ),..., u ( )] = ( LB) B [ u ( ), u ( ),..., u ( )]. (33) T pe, pe, pe, M pe, pe, pe, M u In an analogous way, he values of dτ a collocaon pons pe, m, m =,,..., M can also be obaned from hose of u pe, hrough he followng ransformaon. where M T T upe, dτ upe, dτ upe, dτ upe, upe, upe, M = L BB, (34) [,,..., ] ( ) [ ( ), ( ),..., ( )] L B φ dτ φdτ φndτ φ dτ φdτ φ dτ 0 0 N = φ dτ φ dτ φndτ M M M 0 ( N+ ) M. (35) A schemac presenaon of he large me collocaon s gven n Fg., wheren he me nerval ( + ) can be several hundreds of mes larger han he sep sze of HHT-α algorhm. In each me nerval o +, M collocaon pons are seleced o nerpolae he approxmaed soluon. In hs paper, he wo ends of he me nerval are seleced as he frs and he las collocaon pons,.e. M =, = +. I should be noed ha even he me sep beween wo neghborng collocaon pons can be larger han he me sep used n he HHT algorhm. 9

10 + + Knowns Unknowns Unknowns m x( ) x( ) x( + ) ( ) m x x( ) x( + ) M M + May be he used n he HHT- α Algorhm Fg. Schemac presenaon of large me collocaon n he nerval +.4 Wang-Alur (WA) Algorhms.4.. WA Algorhm- Alhough we have already obaned he erave formula (8) hrough collocaon, he marx of generalzed Lagrange mulplers λ n sll reman a puzzle. The possbly of drecly solvng he consran equaons (5) for λ s unlkely for mos nonlnear cases. Forunaely, here s an approach o bypass hs dlemma. Dfferenang Eq. (9) leads o where R ( τ ) s he resdual funcon: R( τ ) d u() d u() λ () () ( τ) dτ d () = + + d () λ R R, (36) u u c u ( τ ) u ( τ ). (37) = + u ( τ ) M Cu( τ) + M N( τ) M F( τ) Accordng o Eq. (5) and usng he heory of Magnus seres, can be proved [7] ha λ() = I, and λ = J() λ. Subsung hem no Eq. (36), we have d u() u u u () dτ d () = J + c + λ. (38) u M Cu M N M F u M Cu + M N M F 0

11 u u u u Nong ha λ dτ + = u M Cu + M N M F() u c u error-feedback eraon formula (9), Eq. (38) s rewren as accordng o he opmal d u() u u u = () d () J. (39) u c M Cu + M N M F u c u Afer rearrangemen, s rewren as d u() u() u u() () () d () + J = + () J (). (40) u + c u c M Cu M N M F u Eq. (40) can be regarded as equvalen o Eq. (9). By collocang n he local me nerval [, ] algebrac erave formula s obaned as. m m m m J( m) ( ) J m ( ) ( ) ( ) ( ) u m c m + m m m m M Cu M N M F u c + u ( ) u ( ) u ( ) u ( ) + = + ( ) ( ), (4) u =, [, ] where m,,... M m +, and s a fne large me nerval. + Afer rearrangng he sequenal of he collocaon equaons and usng he relaonshp n Eq. (33), Eq. (4) s rewren as where U U U U ( E + J ) = + J, (4) U c M CU + M N M F U = u ( ) u ( )... u ( ) u ( ) u ( )... u ( )... T p p, p, p, M p, p, p, M, an, p =,. (43) The confguraon of he marces E, J, M, C, N, F are provded n Appendx C. I should be noed ha he nal condons are no ncorporaed n Eq. (4). For ha, whou loss of generaly, we usually selec he frs collocaon pon a he nal boundary, of whch he values u ( ) pe, are gven. By dong ha, Eq. (4) becomes overdeermned, hus s necessary o drop excess collocaon equaons a me. Fnally, Eq. (4) s modfed as r r r r U = ( E + J ) E + c + U U U, (44) U U U M CU M N M F The symbol [ ] r denoes he remaned marx afer he ( lm + ) h rows and columns n [ ] are dropped, l = 0,,,.... If [ ] r s a vecor, we jus need o remove he ( lm + ) h rows n [ ]. r

12 A flow dagram s presened n Fg. o llusrae he WA algorhm-. Consan Marces Consruc C, M, E, F Le U p Inalzaon be he sarng guesses: 0 U p Sae Updae U = ( U ) p p c Evaluaon of Varyng Marces The resorng force vecor: N n Eq. (4) he Jacoban marx: J n Eq. (4) Incorporang Inal Condon Drop excess collocaon equaons a n Eq. (44) Correcon Ge he correced soluon ( ) usng Eq. (44) U p c Incremen = + Correcon Norm ε = ( U ) U U p c p p Yes Soppng Creron No < max Sop No Soppng Creron ε δ, where δ s he olerance Yes Fnshed Fg. Flow char overvew of he WA algorhm-

13 .4.. WA Algorhm- Accordng o Eq. (5), he Lagrange mulplers can be approxmaed by Taylor seres as λ T T T, (45) ( τ) = 0 + ( τ ) + ( τ ) +... where T = I, 0 T = J (), formula s obaned as T J, and so on. Subsung no Eq. (8), he correconal = () u( m) u( m) m J() u u [ ( )( τ ) ( τ ) ] dτ ( ) = ( ) m I J u m c u u M Cu + M N M F (46) In mplemenaons, λ ( τ ) s commonly approxmaed by runcaed Taylor seres. The smples could be he zeroh order approxmaon or he frs order approxmaon λ( τ ) = I, (47) λ( τ) = I J ( )( τ ). (48) Hgher order approxmaons are possble, bu hey are rarely used n pracce consderng he compuaonal complexy. Wh he zeroh order approxmaon of λ ( τ ), Eq. (45) becomes u( m) u( m) m u u dτ ( ) = + ( ) m. (49) u c u m u M Cu + M N M F Wh he frs order approxmaon of λ ( τ ), Eq. (45) becomes u( m) u( m) m u u [ ( m)( τ m)] dτ ( ) = + + ( ) m I J. (50) u c u m u M Cu + M N M F =, [, ] where m,,... M nvolvng m and τ, Eq. (50) leads o m c m +, and + s a fne large me nerval. By separang he erms u( m) u( m) m m m ( ) d ( m) m ( ) d ( m) ( ) d ( ) = τ τ τ τ τ τ τ ( ) R + J R J R, (5) u u m Afer some rearrangemens, Eq. (5) can be rewren as U U = HR + JTHR JHTR, (5) U U c 3

14 where H s he ransformaon marx correspondng o negral, and T s he marx relaed wh me. The confguraons of marces H, T, J and R are provded n Appendx C. Fg. 3 shows he flow char of WA algorhm- Consan Marces Consruc H, M, C, E, T, F Le U p Inalzaon be he sarng guesses: 0 U p Sae Updae U = ( U ) p p c Evaluaon of Varyng Marces The Jacoban marx: J n Eq. (5) and he resdual marx: R n Eq. (5) Correcon Ge he correced soluon ( ) usng Eq. (5) U p c Incremen = + ε = Correcon Norm ( U ) U U p c p p Yes Soppng Creron No < max Sop No Soppng Creron ε δ, where δ s he olerance Yes Fnshed Fg. 3 Flow char overvew of he WA algorhm- 4

15 .4.3 WA Algorhm-3 Consderng Eq. (38), f he Lagrange mulpler λ s approxmaed by Taylor seres, we have d u() u u u ( ) [ 0 ( τ )...] dτ d () = J c + T T u M Cu M N F u M Cu + M N M F (53) If λ s smply approxmaed by T 0, Eq. (53) becomes d u() u u u () 0 dτ d () = J + c + T. u M Cu M N M F u M Cu + M N M F (54) Snce T = 0 I, Eq. (54) s furher rewren as d u() u u u () dτ d () = + J + c + (55) u M Cu M N M F u M Cu + M N M F By usng u ( p =, ) as ral funcons and makng collocaons, he WA algorhm-3 s obaned from Eq. (55) as p u ( ) ( ) m u m m u u = ( ) m dτ + J ( ) ( ) ( ) + ( ) m + m m + u m M Cu M N M F u M Cu M N M F c (56) where u ( ) and he negrals can be obaned as s saed n Eqs. (35, 34). Afer rearrangemens, Eq. (56) p m can be wren as U U E = + U M CU c + M N M F JHR. (57) The confguraons of marces E, M, C, N, J, H, and R are provded n Appendx C. Noe ha he nal condons are no guaraneed by Eq. (57), jus lke Eq. (4) n WA algorhm-, hus should be furher modfed as r r d r U U d U E = + JHR E U c M CU + M N M F U c, (58) The symbol [ ] r denoes he remaned marx afer he ( lm + ) h rows and columns n [ ] are dropped, l = 0,,,.... If [ ] r s a vecor, we jus need o remove he ( lm + ) h rows n [ ]. The symbol [ ] d 5

16 denoes he par of [ ] ha were dropped o oban [ ] r. The flow char of hs algorhm s provded n Fg. 4 Consan Marces Consruc H, M, C, E, T, F Le U p Inalzaon be he sarng guesses: 0 U p Sae Updae U = ( U ) p p c Evaluaon of Varyng Marces The resorng force vecor: n Eq. (57) he Jacoban marx: J N n Eq. (57) and he resdual marx: R n Eq. (57) Incorporang Inal Condon Drop excess collocaon equaons a n Eq. (58) Correcon Ge he correced soluon ( ) usng Eq. (58) U p c Incremen = + Correcon Norm ε = ( U ) U U p c p p Yes Soppng Creron No < max Sop No Soppng Creron ε δ, where δ s he olerance Yes Fnshed Fg. 4 Flow char overvew of he WA algorhm-3 6

17 3. Nonlnear Vbraons of a Buckled Beam The governng equaon of a buckled one-dmensonal Bernoull beam s wren n non-dmensonal form as, (59) v w + w + Pw + cw w w dx = F( x)cos Ω 0 BC s: w= w = 0 a x = 0 and x =, where w s he ransverse dsplacemen of he beam, and P s he axal load on he beam. F( x ) s he ransverse dsrbued load on he beam and Ω s he frequency of he appled load F( x ). The overdo denoes he dervave wh respec o me, whle he prme denoes he dervave wh respec o he spaal coordnae x. To solve he precedng paral dfferenal equaon, we frs assume he spaal modes: wx (, ) = w( x) + vx (, ) = w( x) + φ ( xq ) ( ), (60) s s n n n= where w s s he sac posbucklng dsplacemen, vx (,) s he superposed dynamc response around he buckled confguraon. φ ( x ) are he mode shapes of vbraon around he buckled confguraon, and q () n are he ampludes of φ ( x ). n N n The buckled confguraon ws ( x ) can be obaned by frs solvng he sac bucklng problem where he me dervaves and he dynamc load are removed n Eq. (59). + = 0. (6) v w Pw w w dx 0 Mahemacally, here could be varous buckled mode shapes, dependng on he correspondng load. However, n srucural mechancs, he frs buckled mode shape s mosly of mporance, from whch ws ( x ) s obaned as ws ( x) = b( cos π x). (6) The nondmensonal ransverse dsplacemen b a he mdspan of he beam s relaed o he load P va b = 4( P P) π, (63) c where P c s he crcal load correspondng o he frs Euler buckled mode, namely Pc = 4π. Subsung he assumed soluon of wx (,) no he governng equaon and droppng all he nonlnear, dampng, and forcng erms, we have he followng lnear paral dfferenal equaon ha can be ackled usng he lnear vbraon mode heory. 7

18 v 3 v + v + 4π v b π cos πx v sn πxdx = 0 0. (64) By assumng vx (,) = φ() xe ω and subsung no Eq. (64), he mode shape s obaned as φ( x) = φ + φ = ( c sn s x+ c cos s x+ c snh s x+ c cosh s x) + c cos πx, (65) h p where s, ( π 4 π ω ) / = ± + +, and 5 c should sasfy he followng equaon: h ( b π ω ) c = b π φ sn πxdx. (66) Usng he boundary condons and Eq. (66), he mode shapes φ ( x ) can be obaned. The resulng lnear vbraon mode shapes φ ( x ) are used o consruc he soluon vx (,). Usng he n n mul-mode Galerkn dscrezaon, where he weghng funcons are he same as he ral funcons φ ( x ), he paral dfferenal equaon (59) s hen reduced o a sysem of coupled Duffng equaons. N N n ωn n n nj j njk j k n, j, j, k q + q = cq + b A qq + B qq q + f cos Ω, n=,,..., N. (67) In hs paper, four modes are reaned n he reduced model. The bucklng level s seleced as b = 4. Correspondngly he naural frequences of he four lnear vbraon modes are obaned as ω = , ω = , ω 3 = 08.33, and ω 4 = The parameers A nj, and B njk are provded n Appendx A and B. 4. Numercal Resuls and Dscusson n Consderng he dscrezed nonlnear sysem (67), several ypes of nonlnear resonances of he buckled beam may occur due o he exernal harmonc excaon. The prmary resonance s mosly observed when he excaon frequency Ω s close o one of he mode frequences ω, whch ofen leads o a perodc moon of large amplude n ha mode. For he exsence of quadrac nonlneares and cubc nonlneares, he subharmonc resonances and superharmonc resonances may also occur for Ω ha has an neger relaonshp o ω. n In he followng, he resonance of he buckled beam under harmonc excaons s nvesgaed. The frequency Ω s seleced close o he naural frequency of he frs vbraon mode ω. The exernal force s supposed o be unform over he lengh of he beam, hus F( x ) n Eq. (59) s consan. Through Galerkn dscrezaon, he forces mposed on he four lnear vbraon modes are obaned as f = F, f =, f3 = F, and f 4 = 0. 0 In he numercal smulaons, he force-sweep as well as he frequency-sweep processes are used o oban an overvew of he nonlnear dynamcal behavors of he buckled beam when subjeced o a prmary n 8

19 resonance excaon of s frs vbraon mode. Consderng ha he magnudes of he coeffcens ba nj, and B njk are roughly beween force F s se o vary beween and ω n, 4 0 n he resorng forces, he amplude of he exernal F = and F = 600 n he force-sweep process. In he frequency-sweep process, F s fxed a F = 400, whle he frequency Ω s vared beween Ω= 30 and Ω= 8. There are abundan ypes of bass funcons for choce n leraure. In hs paper, he bass funcons used n WA mehod s seleced as Chebyshev polynomals of he frs knd and he collocaon pons are seleced as Chebyshev-Gauss-Lobao nodes [5]. 4. Nonlnear Dynamcal Behavor ncludng Bfurcaon and Chaos Wh he dscrezed equaons derved above, he nonlnear vbraons of a buckled beam are frs nvesgaed under unform harmonc excaons, n whch he frequency of he exernal load F( x)cos Ω s Ω= 30. A force sweepng approach s adoped heren o capure he bfurcaon phenomenon. For smplcy, a perodc moon s referred o as perod- n moon f s perod s nt, where T = π Ω. I s shown n Fg. 5 ha a perod-one moon s obaned for F = 40. Fg. 5 (a) s he Poncare map ha ndcaes he evoluon of he dsplacemen q () and he velocy q () from he nal values o he fnal sae, whch s referred o as he snk. I can be found ha he ransen process of perod-one moon s raher smple. Snce he snk s he only aracor n hs area, all he pons around are araced o asympocally. As he excaon force F ncreases, he frs perod-doublng bfurcaon occurs a abou F = 40. A ypcal perod-wo moon s presened n Fg. 6 for F = 440. In Fg. 6 (a), he Poncare maps of he ransen process are obaned usng WA algorhms of varous negrang accuracy. I s found ha he wo snks sandng for he perod-wo moon exs n a srange aracor. Ths means ha he ransen response of hs moon s chaoc-lke. For he exsence of snks, he moon wll evenually sele down o a seady lm cycle oscllaon. However, could be dffcul o deermne he exac me when he moon seles down f he ransen sage s prolonged. The accurae negraon of he sysem wll also be a challenge, because he rajecory n he chaoc regme can easly dverge even wh a very small negraon error. By furher ncreasng he force amplude, he second perod-doublng occurs a abou F = 454. A perodfour moon s presened n Fg. 7 for F = 460. I can be observed n Fg. 7 (a) ha he four snks also coexs wh he chaoc aracor, hus he ransen moon s also chaoc. By comparng he chaoc regmes n Fg. 7 (a) and Fg. 7 (a), here s no much dfference before and afer he bfurcaon occurs. Therefore, seems ha he chaoc aracor s barely affeced by he bfurcaon of perodc moons. The nex bfurcaon occurs a F = 46, leadng o he perod-egh moon. The correspondng Poncare map, phase porra and response curve are ploed n Fg. 8. The resuls n Fgs. 4-7 can be obaned by boh he HHT-α and WA algorhms. However, he me sep sze has o be seleced very small o accuraely oban he seady perodc moons by usng he HHT-α algorhm, of whch he compuaonal me wll be much more prolonged. On he conrary, he sep sze of WA algorhms can be seleced relavely very large, bu sll easly acheve hgh accuracy n predcng hese lm cycle oscllaons. I wll be shown n he nex subsecon ha he WA algorhm has a far beer performance han he HHT-α mehod n erms of accuracy, compuaonal me, and speed of convergence, n predcng he nonlnear dynamcal responses of he buckled beam nvolvng bfurcaon and chaos. 9

20 Fg. 5 (a) Fg. 5 (b) Fg. 5 (a) The Poncare map for F = 40, (b) The perod- lm cycle oscllaon. Same resuls for HHTα and WA algorhms, bu wh dfferen compuaonal performances, as s dscussed n subsecon 4.. 0

21 Fg. 6 (a) Fg. 6 (b) Fg. 6 (a) The Poncare map for F = 440, (b) The perod- lm cycle oscllaon. Same resuls for HHTα and WA algorhms, bu wh dfferen compuaonal performances, as s dscussed n subsecon 4..

22 Fg. 7 (a) Fg. 7 (b) Fg. 7 (a) The Poncare map for F = 460, (b) The perod-4 lm cycle oscllaon. Same resuls for HHTα and WA algorhms, bu wh dfferen compuaonal performances, as s dscussed n subsecon 4..

23 Fg. 8 (a) Fg. 8 (b) Fg. 8 (a) The Poncare map for F = 46, (b) The perod-8 lm cycle oscllaon. Same resuls for HHTα and WA algorhms, bu wh dfferen compuaonal performances, as s dscussed n subsecon 4.. 3

24 Fg. 9 (a) Fg. 9 (b) 4

25 Fg. 9 (c) Fg. 9 (d) Fg. 9 The same chaoc moon revealed by HHT-α (wh very small sep sze) and WA algorhms. (a) Phase porra, (b) me responses, (c) Poncare map, (d) FFT curves. 5

26 Fg. 0 (a) Fg. 0 (b) Fg. 0 Bfurcaon dagrams obaned by (a) force-sweepng, (b) frequency-sweepng process. Same resuls obaned for HHT-α (wh very small sep sze) and WA algorhms. 6

27 As he perod-doublng bfurcaon proceeds, more snks appear n he chaoc regme, and evenually he moon becomes compleely chaoc. In Fg. 9, he chaoc moon s presened hrough phase porra, response curve, Poncare map, and FFT curve. Smlar perod-doublng roue o chaos s also observed for vared excaon frequency wh he amplude fxed a F = 400. The bfurcaon dagrams obaned hrough force-sweepng process and frequency-sweepng process are ploed n Fg. 0. In predcng chaoc moon, he rajecores obaned usng he consdered wo mehods dverge afer a shor me. However, he Poncare maps and he FFT curves concdes very well, whch shows ha he same chaoc moon s revealed by hese wo mehods. 4. Comparson Beween he HHT-α and WA Algorhms The dscrezed model s solved usng boh he HHT-α and WA mehods. I s found ha boh mehods can predc he lm cycle oscllaons and chaos. However, he compuaonal performances of hese wo mehods are very dfferen. In he analyss below, varous sep szes are used o es he sably of he algorhms. I s found hrough smulaons ha he WA mehod converges for boh perodc and chaoc moons wh sep szes as large as T 5, where T s he perod of he frs vbraon mode, T = πω. The larges sep sze of HHT-α mehod depends on he nonlnear algebrac equaons (NAEs) solver. By usng Newon-Raphson mehod, he sep sze should be no greaer han T 0, oherwse he soluon of nonlnear equlbrum equaons of HHT-α may no be found. We red o reveal he seady perodc moons of he buckled beam under dfferen exernal excaons, usng HHT-α and WA mehod separaely. Afer he moon seles down, he exremes of he dsplacemen q () obaned by hese wo mehods are recorded n Table. The exac values are assumed o be provded by ode45, and hey are fully conssen wh he resuls obaned by WA mehod, wh he me sep sze beng T 5. The number of collocaon pons used n WA mehod s N = 7. Increasng he collocaon pons can furher mprove he accuracy of hs mehod. The HHT- α mehod fals o work for = T 5 and T 0, because he Newon-Raphson eraon scheme canno converge for such large seps. Wh he sep sze beng T 50, he HHT-α mehod provdes seady perodc moons, alhough hey are que erroneous compared o he exac ones. Parcularly, n he parameer regon where he perod-8 moon domnaes, he HHT-α mehod only provdes perod-4 moon, whch s due o he negrang naccuracy. By shorenng he sep sze, he accuracy s sgnfcanly mproved. However, here are sll some observable dscrepances beween he resuls of HHT-α and he exac ones, even wh he sep sze n HHT beng as small as = T 000. In he smulaons, dfferen values of α are used. In he case of α = 0., a moderae numercal dampng s ncluded, whch could fler he hgh-frequency componen of moon and avod dvergence. For α = 0, no numercal dampng s nroduced and he HHT-α mehod becomes he average acceleraon mehod. By comparng he values of exremes wh he exac ones, able ndcaes ha he perodc moons are beer predced wh α = 0 han wh α = 0.. Ths resul s reasonable snce he numercal dampng nroduced n HHT-α mehod makes he sysem behaves lke ha exra dampng exss n he smulaon. The lm cycles of perod-4 and perod-8 moons are obaned usng WA mehod wh = T 5. They are compared wh he resuls obaned from HHT- α mehod, where α = 0. and = T 50. I s shown n Fg. ha he HHT-α mehod canno predc he rue dynamcal behavor wh such a large 7

28 sep sze. In Fg. (a), an obvous dscrepancy exss beween he resuls of WA and HHT-α mehod, whle n Fg. (b), he HHT-α mehod gves a perod-4 moon a where he perod-8 moon should be revealed. In all, boh he WA and he HHT-α mehod can sably negrae he nonlnear sysem of buckled beam. However, he sep sze of HHT-α mehod should be seleced very small o oban vald soluons. Table Exremes of he seady perodc moons obaned by HHT-α and WA mehods Mehods Perod- Perod- Perod-4 Perod-8 WA ( T 5) HHT-α α = 0. α = 0 α = 0. α = 0 α = 0. α = 0 α = 0. α = 0 T 5 \ \ \ \ \ \ \ \ T 0 \ \ \ \ \ \ \ \ T \ \ T \.988 T T Fg. (a) 8

29 Fg. (b) Fg. Comparsons beween he lm cycle oscllaons obaned by he WA ( = T 5 ) and HHT-α ( = T 50, α = 0.) algorhms. Generally, he evaluaon of compuaonal accuracy s nonrval for numercal mehods, especally n nonlnear problems. Heren, he numercal soluon of MATLAB bul n ode45 funcon s used as he benchmark o esmae he compuaonal error of HHT-α and WA mehods. The absolue and he relave accuraces of ode45 are boh se as E 5. The ransen responses of perod-one, perod-wo, perod-four and perod-egh moons are obaned usng HHT-α and WA mehods respecvely. Then by comparng wh hose obaned by ode45, he dscrepances of he numercal resuls are shown n Fg.. To evaluae he hghes accuracy hese wo mehods may acheve, he seps sze of HHT mehod s seleced as = e 4, whle ha of WA mehod s = T 6 = wh 3 collocaon pons n each me sep. As s llusraed n subsecon 4., he ransen response of perod-one moon s relavely smple and nonchaoc. I s refleced by he conssen numercal dscrepances over he smulaon me n Fg. (a). As s shown, he WA mehod acheves very hgh accuracy wh respec o ode45. The dscrepances beween hem s fve magnudes lower han ha beween HHT-α and ode45. For he mulple-perod moons, he numercal dscrepances accumulae exponenally for boh HHT-α and WA algorhms before = 0, as shown n Fg. (b-d). Ths can be explaned by he fac ha chaoc regme exss around he perodc moon. I s shown ha he accuracy of WA mehod s sll much hgher han he HHT-α mehod n he negraon of ransen chaoc moons. Alhough he resuls of WA mehod dverge from ha of ode45 afer = 0, does no mean ha he WA mehod fals. Acually, he chaoc regme s so sensve o he nal sae ha even an error occurred on machne precson could blow up and cause sgnfcan dscrepancy n he fnal sae. 9

30 Fg. (a) Fg. (b) 30

31 Fg. (c) Fg. (d) Fg. The compuaonal error of he HHT-α and WA algorhms 3

32 Table lss he compuaonal me, eraon seps and sep szes of HHT-α and WA algorhms. I s found ha he WA mehod consumes only 0% of he compuaonal me requred by HHT-α mehod. Ths cos savngs are expeced o be much more dramac for large order dynamcal sysems. The sep sze used n HHT-α s 0.000, whle ha n WA s Accordng o he performances demonsraed n Fg. and Table, he WA mehod s much more accurae and effcen han HHT-α mehod, n addon o requrng only much larger sep szes. For furher comparson, he performance ndexes of ode45 are also presened n Table. I s shown ha he compuaonal effcency of he WA algorhm s far superor o ha of ode45. Nong ha he ode45 s a bul-n funcon n MATLAB ha has already been opmzed, whle he WA algorhms are mplemened n MATLAB wh a roughly desgned program. Moreover, he WA algorhms can be easly realzed usng parallel programng, whch wll furher mprove he compuaonal effcency. Table. Compuaonal cos of HHT-α and WA mehod Cases Compuaonal Tme (sec) Ieraon Seps Sep Sze HHT-α WA ode45 HHT-α WA ode45 HHT-α WA ode45 Perod e-05 Perod e-5 Perod e-5 Perod e-5 To nvesgae he energy conservaon properes of he HHT-α and he WA mehod, a smple undamped duffng equaon s used for demonsraon. The Hamlonan energy of hs sysem s 3 + =. x x x 0 H 4 x x x = +. 4 Sarng from he nal sae x (0) =.5, x (0) = 0, he sysem s negraed usng he HHT-α, he WA and he ode45 mehods. The sep sze of HHT-α mehod s = 0.0 and he smulaon s carred n he me nerval [0,000]. For he WA mehod, he sep sze s seleced as =, wh 3 collocaon pons n each sep. The absolue and relave accuraces of ode45 are boh se as E 5. The compuaonal errors of he Hamlonan are recorded and ploed n Fg. 3. I can be seen ha boh he WA mehod and ode45 are much superor o he HHT-α mehod on energy conservaon. Noably, he WA mehod behaves even beer han ode45, wh a neglgble error of Hamlonan of E 3. Among all hese mehods, he HHT-α mehod wh α = 0. s he wors on energy conservaon, as can be seen n Fg. 3 (a). Afer droppng ou he numercal dampng by seng α = 0, he performance of HHT-α mehod s much mproved, wh he error of Hamlonan beng E 5. 3

33 Fg. 3 (a) Fg. 3 (b) Fg. 3 Compuaonal error of Hamlonan usng he HHT and WA algorhms, and ode45 33

34 4.3 Dscussons of he relave performances of he 3 WA algorhms Alhough he WA algorhms-,, and 3 are derved from he same opmal error-feedback eraon concep, hey have some dfferences n mplemenaon and compuaonal performances ha he users of hem should be aware of. In Fgs. -3, he flow char overvew of hese hree algorhms are presened. In he mplemenaons of WA algorhms- and 3, he nal condons have o be ncorporaed n he eraon formula and he excess collocaon equaons need o be removed. In WA algorhm-, he nal condons are naurally sasfed, hus he mplemenaon of s more sraghforward han WA algorhms- and 3. The convergence speed of he WA algorhm- s supposed o be he fases, snce s equvalen o Eq. (), where he marx of weghng funcons λ s opmally derved. However, he WA algorhm- nvolves nverson of he Jacoban marx ( E + J ), whch s varyng durng he eraon process. For sysems wh hgh dmensons, compung he nverson of he Jacoban marx could be very me consumng. In addon, f he Jacoban marx become ll-condoned durng he eraon, he WA algorhm- may easly dverge. On he conrary, he WA algorhms- and 3 are free from nverng marces durng he eraon process. Consderng he buckled beam problem n hs paper, he performances of he WA algorhms are compared wh each oher. The comparng resuls lsed n able 3 are obaned durng he smulaon me of 00T, where T s he perod of he frs mode. Table. 3 Comparson of WA algorhms Mehod Larges Sep Sze Number of Ierave Seps Compuaonal Tme WA algorhm- T s WA algorhm- T s WA algorhm-3 T s As s shown n able 3, he larges sep szes of he WA algorhms are he same, whle he number of erave seps and compuaonal me of WA algorhm- are he leas among he hree proposed algorhms. The compuaonal errors are no provded heren because he WA algorhms behave smlar n erms of accuracy, whch s already presened n Fg.. Overall, alhough all he proposed WA algorhms can be effcenly appled o solvng he nonlnear dynamcal equaons of he buckled beam, he WA algorhm- s he mos recommended, no only because ncorporaes he nal boundary condons nherenly and s free from nverng he Jacoban marx, bu also for ha has he bes compuaonal performance among he proposed mehod. 5. Concluson The Wang-Alur Algorhms are used o solve a nonlnear dynamcal model of he clamped-clamped buckled beam. The numercal resuls show ha he proposed mehod s very accurae and effcen n predcng nonlnear dynamcal behavors ncludng bfurcaon and chaos. Through force-sweep and frequency-sweep processes, he perodc moons and he perod-doublng roues o chaos are successfully capured by he Wang-Alur algorhms. I s also found n he smulaons ha he seady mulple-perodc 34

35 moons are acually surrounded by chaoc moons n he neghborhood. I s he reason for he chaoc ransen moons o be hardly predced precsely. The robusness of Wang-Alur mehod s ndcaed by he hghly accurae negraon usng a large sep sze = T 5. Compared wh he Hlber-Hughes-Taylor-α mehod, he proposed mehod s superor on many aspecs nvolvng accuracy, effcency, sably and energy conservaon. I s shown n hs paper ha he Hlber-Hughes-Taylor-α mehod canno accuraely predc he perodc vbraons of he buckled beam unless exremely small sep szes are used. An obvous dscrepancy can sll be observed beween he exac phase porra and ha provded by Hlber-Hughes-Taylor mehod-α even when a relavely small sep sze = T 00 s used. The example of Duffng equaon also ndcaes ha he energy of he sysem s barely conserved usng Hlber-Hughes-Taylor-α mehod afer he smulaon me = 000, whle he error of energy n he resuls of Wang-Alur algorhm s neglgble. Appendx A A,, = 4.905, A,, = 0, A,,3 = 6.554, A,,4 = 0, A,, = 7.098, A,,3 = 0, A,,4 = -4.55, A,3,3 = , A,3,4 = 0, A,4,4 = ; A,, = 0, A,, = 34.95, A,,3 = 0, A,,4 = -4.55, A,, = 0, A,,3 = , A,,4 = 0, A,3,3 = 0, A,3,4 = , A,4,4 = 0 ; A 3,, = 8.768, A 3,, = 0, A 3,,3 = 84.45, A 3,,4 = 0, A 3,, = , A 3,,3 = 0, A 3,,4 = , A 3,3,3 = 64.7, A 3,3,4 = 0, A 3,4,4 = ; A 4,, = 0, A 4,, = -4.55, A 4,,3 = 0, A 4,,4 = 50.38, A 4,, = 0, A 4,,3 = , A 4,,4 = 0, A 4,3,3 = 0, A 4,3,4 = 430.6, A 4,4,4 = 0 ; Appendx B B,,, = , B,,,3 = , B,,, = , B,,,4 =., B,,3,3 = , B,,4,4 = , B,,,3 = , B,,3,4 = , B,3,3,3 = , B,3,4,4 = ; B,,, = , B,,, = -04., B,,,3 = -7.8, B,,3,3 = , B,,,4 =., B,,,4 = 369.8, B,,3,4 = , B,3,3,4 = , B,,4,4 = , B,4,4,4 = ; B 3,,, = , B 3,,, = , B 3,,,3 = , B 3,,,3 = , B 3,,3,3 = , B 3,3,3,3 = , B 3,,,4 = , B 3,,3,4 = 967.3, B 3,,4,4 = , B 3,3,4,4 = ; B 4,,, =., B 4,,, = , B 4,,,3 = , B 4,,3,3 = , B 4,,,4 = , B 4,,,4 = , B 4,,3,4 = , B 4,3,3,4 = , B 4,,4,4 = , B 4,4,4,4 =

36 Appendx C Table 4 The Consan and Varyng Marces n he WA algorhms Consan Marces E = IL L ( LB) B, M = M IM M, C = C I, M M = [,,..., ] T, M Varyng Marces J = J () ˆ, N = N (), U U R = E + U M CU + M N F ˆ = dag(), = IL L, H = IL L ( L BB ). F = F () M s he number of collocaon pons n each me nerval; L s he lengh of varable vecor x n Eq. () Acknowledgemen The auhors hank Dr. Lawrence Schovanec, he Presden of Texas Tech Unversy, for hs suppor of hs research, hrough he Presdenal Char and Unversy Dsngushed Professorshp. References [] Kerschen, G.; Worden, K.; Vakaks, A. F.; Golnval, J. C.: Pas, presen and fuure of nonlnear sysem denfcaon n srucural dynamcs. Mechancal Sysems and Sgnal Processng, vol. 006, no. 0, pp (006) [] Srogaz, S. H.: Nonlnear dynamcs and chaos: wh applcaons o physcs, bology, chemsry, and engneerng. Addson-Wesley (994) [3] Fehlberg, E.: Low-order classcal Runge-Kua formulas wh sepsze conrol and her applcaon o some hea ransfer problems. Techncal repor, NASA (969) [4] Newmark, N. M.: A mehod of compuaon for srucural dynamcs. Journal of he Engneerng Mechancs Dvson, vol. 85, no. 3, pp (959) [5] Hlber, H. M.; Hughes, T. J.; Taylor, R. L.: Improved numercal dsspaon for me negraon algorhms n srucural dynamcs. Earhquake Engneerng & Srucural Dynamcs, vol. 5, no. 3, pp (977) [6] Xe, Y. M.: An assessmen of me negraon schemes for nonlnear dynamc equaons. Journal of Sound and Vbraon, vol. 9, no., pp (996) 36

37 [7] Wang, X.; Alur, S. N.: A Novel Class of Hghly Effcen & Accurae Tme-Inegraors n Nonlnear Compuaonal Mechancs, Compuaonal Mechancs, vol. 59, no. 5, pp , (07) [8] Wang, X.; Alur, S. N.: A New Feedback-Acceleraed Pcard Ieraon Mehod for Orbal Propagaon & Lamber s Problem, Jounral of Gudance, Conrol, and Navgaon, onlne. [9] Dong, L.; Aloab, A.; Mohuddne, S. A.; Alur, S. N.: Compuaonal Mehods n Engneerng: A Varey of Prmal & Mxed Mehods, wh Global & Local Inerpolaons, for Well-Posed or Ill-Posed BCs. Compuer Modelng n Engneerng & Scences, vol. 99, no., pp. -85 (04) [0] Alur, S. N.: Mehods of Compuer Modelng n Engneerng & he Scences, Volume I. Tech Scence Press, Forsyh (005) [] Kreder, W.; Nayfeh, A. H.: Expermenal Invesgaon of Sngle-Mode Responses n a Fxed-Fxed Buckled Beam. Nonlnear Dynamcs, vol. 5, no., pp (998) [] Tseng, W. Y.; Dugunj, J.: Nonlnear vbraons of a beam under harmonc excaon. Journal of Appled Mechancs, vol. 37, no., pp (970) [3] Emam, S. A.; Nayfeh, A. H.: On he Nonlnear Dynamcs of a Buckled Beam Subjeced o a Prmary- Resonance Excaon. Nonlnear Dynamcs, vol. 35, no., pp. -7 (004) [4] Ba, X.; Junkns, J. L.: Modfed Chebyshev-Pcard Ieraon Mehods for Soluon of Boundary Value Problems. The Journal of Asronaucal Scences, vol. 58, no. 4, pp (0) [5] Fukushma, T.: Pcard eraon mehod, Chebyshev polynomal approxmaon, and global numercal negraon of dynamcal moons. The Asronomcal Journal, vol. 3, no. 5, pp (997) 37

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

3. OVERVIEW OF NUMERICAL METHODS

3. OVERVIEW OF NUMERICAL METHODS 3 OVERVIEW OF NUMERICAL METHODS 3 Inroducory remarks Ths chaper summarzes hose numercal echnques whose knowledge s ndspensable for he undersandng of he dfferen dscree elemen mehods: he Newon-Raphson-mehod,

More information

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems Swss Federal Insue of Page 1 The Fne Elemen Mehod for he Analyss of Non-Lnear and Dynamc Sysems Prof. Dr. Mchael Havbro Faber Dr. Nebojsa Mojslovc Swss Federal Insue of ETH Zurch, Swzerland Mehod of Fne

More information

Chapter Lagrangian Interpolation

Chapter Lagrangian Interpolation Chaper 5.4 agrangan Inerpolaon Afer readng hs chaper you should be able o:. dere agrangan mehod of nerpolaon. sole problems usng agrangan mehod of nerpolaon and. use agrangan nerpolans o fnd deraes and

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

Handout # 6 (MEEN 617) Numerical Integration to Find Time Response of SDOF mechanical system Y X (2) and write EOM (1) as two first-order Eqs.

Handout # 6 (MEEN 617) Numerical Integration to Find Time Response of SDOF mechanical system Y X (2) and write EOM (1) as two first-order Eqs. Handou # 6 (MEEN 67) Numercal Inegraon o Fnd Tme Response of SDOF mechancal sysem Sae Space Mehod The EOM for a lnear sysem s M X DX K X F() () X X X X V wh nal condons, a 0 0 ; 0 Defne he followng varables,

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

Let s treat the problem of the response of a system to an applied external force. Again,

Let s treat the problem of the response of a system to an applied external force. Again, Page 33 QUANTUM LNEAR RESPONSE FUNCTON Le s rea he problem of he response of a sysem o an appled exernal force. Agan, H() H f () A H + V () Exernal agen acng on nernal varable Hamlonan for equlbrum sysem

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Chapter 2 Linear dynamic analysis of a structural system

Chapter 2 Linear dynamic analysis of a structural system Chaper Lnear dynamc analyss of a srucural sysem. Dynamc equlbrum he dynamc equlbrum analyss of a srucure s he mos general case ha can be suded as akes no accoun all he forces acng on. When he exernal loads

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Time-interval analysis of β decay. V. Horvat and J. C. Hardy

Time-interval analysis of β decay. V. Horvat and J. C. Hardy Tme-nerval analyss of β decay V. Horva and J. C. Hardy Work on he even analyss of β decay [1] connued and resuled n he developmen of a novel mehod of bea-decay me-nerval analyss ha produces hghly accurae

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations. Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

Implementation of Quantized State Systems in MATLAB/Simulink

Implementation of Quantized State Systems in MATLAB/Simulink SNE T ECHNICAL N OTE Implemenaon of Quanzed Sae Sysems n MATLAB/Smulnk Parck Grabher, Mahas Rößler 2*, Bernhard Henzl 3 Ins. of Analyss and Scenfc Compung, Venna Unversy of Technology, Wedner Haupsraße

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy

More information

Supplementary Material to: IMU Preintegration on Manifold for E cient Visual-Inertial Maximum-a-Posteriori Estimation

Supplementary Material to: IMU Preintegration on Manifold for E cient Visual-Inertial Maximum-a-Posteriori Estimation Supplemenary Maeral o: IMU Prenegraon on Manfold for E cen Vsual-Ineral Maxmum-a-Poseror Esmaon echncal Repor G-IRIM-CP&R-05-00 Chrsan Forser, Luca Carlone, Fran Dellaer, and Davde Scaramuzza May 0, 05

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

2.1 Constitutive Theory

2.1 Constitutive Theory Secon.. Consuve Theory.. Consuve Equaons Governng Equaons The equaons governng he behavour of maerals are (n he spaal form) dρ v & ρ + ρdv v = + ρ = Conservaon of Mass (..a) d x σ j dv dvσ + b = ρ v& +

More information

CONSISTENT EARTHQUAKE ACCELERATION AND DISPLACEMENT RECORDS

CONSISTENT EARTHQUAKE ACCELERATION AND DISPLACEMENT RECORDS APPENDX J CONSSTENT EARTHQUAKE ACCEERATON AND DSPACEMENT RECORDS Earhqake Acceleraons can be Measred. However, Srcres are Sbjeced o Earhqake Dsplacemens J. NTRODUCTON { XE "Acceleraon Records" }A he presen

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015)

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015) 5h Inernaonal onference on Advanced Desgn and Manufacurng Engneerng (IADME 5 The Falure Rae Expermenal Sudy of Specal N Machne Tool hunshan He, a, *, La Pan,b and Bng Hu 3,c,,3 ollege of Mechancal and

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

VEHICLE DYNAMIC MODELING & SIMULATION: COMPARING A FINITE- ELEMENT SOLUTION TO A MULTI-BODY DYNAMIC SOLUTION

VEHICLE DYNAMIC MODELING & SIMULATION: COMPARING A FINITE- ELEMENT SOLUTION TO A MULTI-BODY DYNAMIC SOLUTION 21 NDIA GROUND VEHICLE SYSTEMS ENGINEERING AND TECHNOLOGY SYMPOSIUM MODELING & SIMULATION, TESTING AND VALIDATION (MSTV) MINI-SYMPOSIUM AUGUST 17-19 DEARBORN, MICHIGAN VEHICLE DYNAMIC MODELING & SIMULATION:

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon If we say wh one bass se, properes vary only because of changes n he coeffcens weghng each bass se funcon x = h< Ix > - hs s how we calculae

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary

More information

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

Handout # 13 (MEEN 617) Numerical Integration to Find Time Response of MDOF mechanical system. The EOMS for a linear mechanical system are

Handout # 13 (MEEN 617) Numerical Integration to Find Time Response of MDOF mechanical system. The EOMS for a linear mechanical system are Handou # 3 (MEEN 67) Numercal Inegraon o Fnd Tme Response of MDOF mechancal sysem The EOMS for a lnear mechancal sysem are MU+DU+KU =F () () () where U,U, and U are he vecors of generalzed dsplacemen,

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

Li An-Ping. Beijing , P.R.China

Li An-Ping. Beijing , P.R.China A New Type of Cpher: DICING_csb L An-Png Bejng 100085, P.R.Chna apl0001@sna.com Absrac: In hs paper, we wll propose a new ype of cpher named DICING_csb, whch s derved from our prevous sream cpher DICING.

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

First-order piecewise-linear dynamic circuits

First-order piecewise-linear dynamic circuits Frs-order pecewse-lnear dynamc crcus. Fndng he soluon We wll sudy rs-order dynamc crcus composed o a nonlnear resse one-por, ermnaed eher by a lnear capacor or a lnear nducor (see Fg.. Nonlnear resse one-por

More information

Single-loop System Reliability-Based Design & Topology Optimization (SRBDO/SRBTO): A Matrix-based System Reliability (MSR) Method

Single-loop System Reliability-Based Design & Topology Optimization (SRBDO/SRBTO): A Matrix-based System Reliability (MSR) Method 10 h US Naonal Congress on Compuaonal Mechancs Columbus, Oho 16-19, 2009 Sngle-loop Sysem Relably-Based Desgn & Topology Opmzaon (SRBDO/SRBTO): A Marx-based Sysem Relably (MSR) Mehod Tam Nguyen, Junho

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

Lecture VI Regression

Lecture VI Regression Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

Physical Simulation Using FEM, Modal Analysis and the Dynamic Equilibrium Equation

Physical Simulation Using FEM, Modal Analysis and the Dynamic Equilibrium Equation Physcal Smulaon Usng FEM, Modal Analyss and he Dynamc Equlbrum Equaon Paríca C. T. Gonçalves, Raquel R. Pnho, João Manuel R. S. Tavares Opcs and Expermenal Mechancs Laboraory - LOME, Mechancal Engneerng

More information

Anisotropic Behaviors and Its Application on Sheet Metal Stamping Processes

Anisotropic Behaviors and Its Application on Sheet Metal Stamping Processes Ansoropc Behavors and Is Applcaon on Shee Meal Sampng Processes Welong Hu ETA-Engneerng Technology Assocaes, Inc. 33 E. Maple oad, Sue 00 Troy, MI 48083 USA 48-79-300 whu@ea.com Jeanne He ETA-Engneerng

More information

APPROXIMATE ANALYTIC SOLUTIONS OF A NONLINEAR ELASTIC WAVE EQUATIONS WITH THE ANHARMONIC CORRECTION

APPROXIMATE ANALYTIC SOLUTIONS OF A NONLINEAR ELASTIC WAVE EQUATIONS WITH THE ANHARMONIC CORRECTION THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY Seres A OF THE ROMANIAN ACADEMY Volume 6 Number /5 pp 8 86 APPROXIMATE ANALYTIC SOLUTIONS OF A NONLINEAR ELASTIC WAVE EQUATIONS WITH THE ANHARMONIC

More information

Gear System Time-varying Reliability Analysis Based on Elastomer Dynamics

Gear System Time-varying Reliability Analysis Based on Elastomer Dynamics A publcaon of CHEMICAL ENGINEERING TRANSACTIONS VOL. 33, 013 Gues Edors: Enrco Zo, Pero Barald Copyrgh 013, AIDIC Servz S.r.l., ISBN 978-88-95608-4-; ISSN 1974-9791 The Ialan Assocaon of Chemcal Engneerng

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

A Simulation Based Optimal Control System For Water Resources

A Simulation Based Optimal Control System For Water Resources Cy Unversy of New York (CUNY) CUNY Academc Works Inernaonal Conference on Hydronformacs 8--4 A Smulaon Based Opmal Conrol Sysem For Waer Resources Aser acasa Maro Morales-Hernández Plar Brufau Plar García-Navarro

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

Lecture 9: Dynamic Properties

Lecture 9: Dynamic Properties Shor Course on Molecular Dynamcs Smulaon Lecure 9: Dynamc Properes Professor A. Marn Purdue Unversy Hgh Level Course Oulne 1. MD Bascs. Poenal Energy Funcons 3. Inegraon Algorhms 4. Temperaure Conrol 5.

More information

Response of MDOF systems

Response of MDOF systems Response of MDOF syses Degree of freedo DOF: he nu nuber of ndependen coordnaes requred o deerne copleely he posons of all pars of a syse a any nsan of e. wo DOF syses hree DOF syses he noral ode analyss

More information

Dynamic Model of the Axially Moving Viscoelastic Belt System with Tensioner Pulley Yanqi Liu1, a, Hongyu Wang2, b, Dongxing Cao3, c, Xiaoling Gai1, d

Dynamic Model of the Axially Moving Viscoelastic Belt System with Tensioner Pulley Yanqi Liu1, a, Hongyu Wang2, b, Dongxing Cao3, c, Xiaoling Gai1, d Inernaonal Indsral Informacs and Comper Engneerng Conference (IIICEC 5) Dynamc Model of he Aally Movng Vscoelasc Bel Sysem wh Tensoner Plley Yanq L, a, Hongy Wang, b, Dongng Cao, c, Xaolng Ga, d Bejng

More information

Research Article Numerical Approximation of Higher-Order Solutions of the Quadratic Nonlinear Stochastic Oscillatory Equation Using WHEP Technique

Research Article Numerical Approximation of Higher-Order Solutions of the Quadratic Nonlinear Stochastic Oscillatory Equation Using WHEP Technique Hndaw Publshng Corporaon Journal of Appled Mahemacs Volume 3, Arcle ID 68537, pages hp://dx.do.org/.55/3/68537 Research Arcle Numercal Approxmaon of Hgher-Order Soluons of he Quadrac Nonlnear Sochasc Oscllaory

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

NATIONAL UNIVERSITY OF SINGAPORE PC5202 ADVANCED STATISTICAL MECHANICS. (Semester II: AY ) Time Allowed: 2 Hours

NATIONAL UNIVERSITY OF SINGAPORE PC5202 ADVANCED STATISTICAL MECHANICS. (Semester II: AY ) Time Allowed: 2 Hours NATONAL UNVERSTY OF SNGAPORE PC5 ADVANCED STATSTCAL MECHANCS (Semeser : AY 1-13) Tme Allowed: Hours NSTRUCTONS TO CANDDATES 1. Ths examnaon paper conans 5 quesons and comprses 4 prned pages.. Answer all

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

2/20/2013. EE 101 Midterm 2 Review

2/20/2013. EE 101 Midterm 2 Review //3 EE Mderm eew //3 Volage-mplfer Model The npu ressance s he equalen ressance see when lookng no he npu ermnals of he amplfer. o s he oupu ressance. I causes he oupu olage o decrease as he load ressance

More information

Motion in Two Dimensions

Motion in Two Dimensions Phys 1 Chaper 4 Moon n Two Dmensons adzyubenko@csub.edu hp://www.csub.edu/~adzyubenko 005, 014 A. Dzyubenko 004 Brooks/Cole 1 Dsplacemen as a Vecor The poson of an objec s descrbed by s poson ecor, r The

More information

Pendulum Dynamics. = Ft tangential direction (2) radial direction (1)

Pendulum Dynamics. = Ft tangential direction (2) radial direction (1) Pendulum Dynams Consder a smple pendulum wh a massless arm of lengh L and a pon mass, m, a he end of he arm. Assumng ha he fron n he sysem s proporonal o he negave of he angenal veloy, Newon s seond law

More information

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current :

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current : . A. IUITS Synopss : GOWTH OF UNT IN IUIT : d. When swch S s closed a =; = d. A me, curren = e 3. The consan / has dmensons of me and s called he nducve me consan ( τ ) of he crcu. 4. = τ; =.63, n one

More information

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys Dual Approxmae Dynamc Programmng for Large Scale Hydro Valleys Perre Carpener and Jean-Phlppe Chanceler 1 ENSTA ParsTech and ENPC ParsTech CMM Workshop, January 2016 1 Jon work wh J.-C. Alas, suppored

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

THE method of moment (MOM) is widely used to extract. Fast Green Function Evaluation for Method of Moment. arxiv: v1 [cs.

THE method of moment (MOM) is widely used to extract. Fast Green Function Evaluation for Method of Moment. arxiv: v1 [cs. 1 Fas Green Funcon Evaluaon for Mehod of Momen Shunchuan Yang, Member, IEEE, Dongln Su, Member, IEEE arxv:1901.04162v1 [cs.ce] 14 Jan 2019 Absrac In hs leer, an approach o accelerae he marx fllng n mehod

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β SARAJEVO JOURNAL OF MATHEMATICS Vol.3 (15) (2007), 137 143 SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β M. A. K. BAIG AND RAYEES AHMAD DAR Absrac. In hs paper, we propose

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

Response-Spectrum-Based Analysis for Generally Damped Linear Structures

Response-Spectrum-Based Analysis for Generally Damped Linear Structures The h World Conference on Earhquake Engneerng Ocober -7, 8, Beng, Chna Response-Specrum-Based Analyss for Generally Damped Lnear Srucures J. Song, Y. Chu, Z. Lang and G.C. Lee Senor Research Scens, h.d.

More information

WiH Wei He

WiH Wei He Sysem Idenfcaon of onlnear Sae-Space Space Baery odels WH We He wehe@calce.umd.edu Advsor: Dr. Chaochao Chen Deparmen of echancal Engneerng Unversy of aryland, College Par 1 Unversy of aryland Bacground

More information

Tight results for Next Fit and Worst Fit with resource augmentation

Tight results for Next Fit and Worst Fit with resource augmentation Tgh resuls for Nex F and Wors F wh resource augmenaon Joan Boyar Leah Epsen Asaf Levn Asrac I s well known ha he wo smple algorhms for he classc n packng prolem, NF and WF oh have an approxmaon rao of

More information

Introduction to. Computer Animation

Introduction to. Computer Animation Inroducon o 1 Movaon Anmaon from anma (la.) = soul, spr, breah of lfe Brng mages o lfe! Examples Characer anmaon (humans, anmals) Secondary moon (har, cloh) Physcal world (rgd bodes, waer, fre) 2 2 Anmaon

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information