Topic 2 Project 1) event detection 2) fault classifieaHon @ Test system © Method using @ Toot 3) Harding Detection 4) Ant oy ’ ') Basie Intwoducton +0 3/9 2(0T) > tot sete er ROIS] = undesined 0 = _ ; (ue gel sta nave only) V J This woltld no: rat Hold ned be a tetangulan Iq hy ne prefem sin wave 2 Also siqs can be defined a5 founien of the sine Paw only x(mT) = Xs(HT) Xt) = 0 . nation Disencte s/q ¢ pee VS livcction sreconsinuction of s/g¢ weg OO $4 mth) Sarmpiy See, Oly. &‘ er = S_ (Antes (Oat P.)# Bn cost Oo t%)) fi , ; seen eueeesl OS cingle pulse sa} * . aire bem tars Then Wwe can + nnile tne eq, ; | 4 Fm BU ' tm fs d M(t) = Aeos (oEt@) = eos (24. a: 2 nt 4%) = Aeos(2x04+9) cos0, = Atos (2xn—9) ie pty fe => sampling theonem (contraction ) te back te that s/g again 7=4Ch)7 up- sampling donn- 1 “Say se PIES GS Aan = R20) also constraction zm) =4(%2) “AB -2 1 25.09.23 # Ovenvien of poner systems slqs in powew sytem, 4 oppiteation of signal processing os4 A intake 2 ad Transdues 5 ¢T 3 pT + if any dala 4s missing we con ge) 1b from slg processing Hawrente—> 1 appawt fnorn needed 2/g Whicheven component ts thene th the, founien sentes > hanmonies (Tower sys 9/4 processing fn sword qnids)| eo sounces of hanmonies in PS FET eR eqben atlacks He Few ‘lucie oton MATLAB. and Hime period 0 the Uluste atom of sampling Frequency tute 220 V- 7. Define a /F of freq. 5° He ef magni y. sinuscidal $/4 ap, sample tH ot SH. ue the 8/9 measurement freq, bi ofe Let us to el MHZ. teat Siqnel - meas F e406 5 tm = 503 tet 2 2" pi [stqnal- meas 5/5 5 mega sample in one eyele meas. t qe 220* sin (2 pi * fm} meas.£) 5 plot (meas to 4) Osnaee24 Founien series SourtenT, soy Ew » Transform o aoa ser oP ' tne * Zen o Leelune upcoming 5 F.tee> SCF) = J sCtye-J278 Hee 1p statonany 3/9. 8 nel stavonony Stqnal_meas_F2 4e6 5 Sin = 595 Fin > 10° > meas. b> 0:2" pi / Signal-meas ts2e pts ys rz0% gin (2* pi* fm? meas_t) 3 yt = 2204 sin (2# pi * frm% meas_t+ 2% pt/6) > yates” ae CSente ’ oS yor av aets yar ytytts fequre () subplot (47101)plod (meas-4, y> Subplot (a, 4,2) plot Cmeas 4,44) subplot (4,1,3) plot (meas-4, y, ) fon stationamy 5/9 > Freq Fe cons} C4114) -> Mb) ret , } : the chanacleniste of the s/q component on hanmenie / frequency changes. tL, non slatfonamy vy Ss. <——— SOW?» 250 Ha. Sour, Y= #840495 Fas Fountem Transfonm 8 502. 8 toong. cmeate tio slay, a Sepawstely take FE¢ then add the tye $195. (4=0 to 19 8) WS tp 2.5 100 Ha, 0 tes ks pane Bong.1-09,23 1) Baste 8/9°+ 2) Powe sys 5/q5+ 3) Fournier tgenete Hme s/g 2 * 5 type of 478 ‘9 ste * DetenminisHe — Non-determinish randomness tn chanaclents-He ative reat tical ined Tn y A Peniodic— BamApeniodie x(n) = 2(MtN) 9 VN I9 jodie) Cif follons then peniodic */9 ° themutse A penie Todie op not. eos (5 n periodic tnd the #/qs ane periodic 4 RR) = 1) - ii) zLnj = ess (Sn) + 300 pent Wy e(ml= sin (= a) Apentodie 5 = 248 w, = arf = 2A fen “eit fe the peniodietty 1 sd cogs = $7 75 3 mm) Aeos(ver + 9) Wee eas fom any expenental funk pentod ts 2xF. _ ko i¥ symmetote & Ant-symmetnic Vatm=atny = xtnj= -xl-a] D Even € Odd pamt of the 6/9 ) xem = 30!" My) ROMY) = (2,-2,6,-2} * a hes 1) xen) = 30!" yen x(a) = - get? Ti) em) =42,-2,6,-2 An rae) Xn) = 40,0,0,2,-2 6 -4 Even pant ¥ Xtm1 = 2el0] $2.(n] xetny = L (aCe 4xpng [239 + xe] xn}eh (e2eeir Bey 2s 2d strate a cote, Berets) = (743, t ‘ one 3/9 Fs & Enengy A Ronen M9 Y nan zene) ac aij __—_ [P= bm 12 jecal'| \" | Now ZN+1 80 \ 2 a fintie, VO on poner sq. s Find either H's emevgg y=) um) 3 xo} yee nz0 e- 2, [( ty uc) bk se fererp BP! _ Zo Joe 25 |" ne = Le = boca = 0.6625 : CGP soluti sy ore Chnite value] * qausal— nencausal-ant-eausal : xtny ¥ ny ) xin] Yo ngo a(n] ¥ eo (n can be any ting) Disenete time sys,+ Disemete Rme sys. the sys, bleh howks in disende Hme a device om algenitm that openates on teas Hime 3/ge- called h , 4 ° 6 aceonding 2 some Hell defined mule AAT ae ancinee ele ae 4 r 4 yinds-Zay(n-m) + Zbrcn-™ ) mel m=0 12,0923 n=0 —————_ ycor= 4 Cece) > 3 ket WOpd aA Lyte, 200 e aay Time-Tnvanient sys. 923 = £049 (42, yoy xt] 21497 Very easy to Tinplemend on design. xin}; yim] 5 Tle] 2 fp} —_fel y, Tr] 2s} oe 4 NCD = (%] > Time invanient g wimg <23,8] } check the s/q is Hme-invawient gx nol, Tenge ed +2 > Tin] =wy, inde utd, trad = x(nJt+2 sition theonem 3s maintained for those near 9y5. —LT1 super po 1 n wn S amyen-™) = S bma(n-m) Homogeneous solution ino mre panfieulam ozeno state yesponse wHeuler solution ve iq PS A k AB” kom ans Anokent hana. - ke) Ane Kort thom. + ee poser Leesan + hasinur Aan ve Ky Cosma sin on, Method > ne yond = Ynihd + Fe(H) 9 Total mesponse ; mine hemeg enecus 1s + (in Delewmine qeneens sel yaeny Wilh constant c,,¢2 .. Cx ao” the 7™ yen= - O-SyCa-D FC) Hhen t/p wes 7s und step WN =o yeD = 49 von = sycr) (-0.6)"% Vy a. [Fre 4H] o0r3 Fietle impulse response CIR) wet ycny= >_ hom (a-m) ‘mo hO=0 5 fom nco £ adn : y NkRS yes - Zeer) bm (n- mm) ne € yim s SZ hom) x m-m) m0 Be transom Reeunsive & Non-neeunsive t 3 f ofp doesn't depend on past ofp as well as present only depends on present tp.zo tnansfoom ese ee 2,! ven) 22 fae 3 leet 1 gnu) —? Da @t-az! ? je >fol DTFT 2 © gimp te & x(n) o % noe -jon <— DTFT XeIM_ FRc = Z xm) art 8 oue E-£. x(n) =0.5™u(n) x@- be 10.527! =! y= aaa 1-0-5287 ums4 1-05e70e" X (8%) 2 F (XM)) =3 19.09. Z XCM) = 0-9 exp (FAA™ 5 ¥ OE NLIO. peel 22 28 Find the DTFT using MATLAB. KH) 2S xeme® nee & zEnI= Co.9)"N OC HELIO Z open ties of DTFT > Propenties of DT FO fourier Thanfonm 1 Lineanity F Lym Ry preety] 7 Fe Bett bem) 2 BELR OTF ii tip Thme shifting xtny ——» x(ei¥) —» eLn-KJ Fs #{m-h] —Fexele UND Freq, shifing o F TxcnpedO-") 9 F +> Xe Ie-o Yo eonjugaken FLeuyts xX -Se AO Folding 5 FIxemy = xcets) Wt) convelutien FOLXrtn) -@ x,tnd = Pp Lxitn) FLX) 2 xp eIF ie) (Wi MulHplicaton 9 FOU) KUT PIELMOI® Flaw] Wii Energy - a a juny” \ yResponse of LTL system zon +—~= Y ONE X00 @ him eh tn] 6 elem uw exponen Hal signal? yoed= bond 6 elu osu cr) ytrd= gZ. Ak) ret -& pray e PP]: xEInJ = ee fo. e FER = FCatn ese?ws! _janke xtny > Zee Janke xen =o tes[G ott? DFTA RCHIS x0) PW She —> xCny nae LETC aC} 2 Frequency speetaum using DET ? < wy : ss Xk) © Xwlk) + ¥tth) Magnitude speelnum ¢ Ixug] = [ew Kru ert) ) (x00) 2 [os wea Phase speclmum > Z¥CK) = ton! [23 . Propenties of DFT? Linearity t x(n) +t eK 7 aX CH) + X20) Pentod! city * x(n tn) = x On) Ss KCKAND = x(k) cin eulan time ohift t nim, jon HE-M) J KHIe Time nevensal * x(N-n) > Xx (NH Conjugator : zACny > Xt(N-K) eineulaw freq. shits jax x(mye T > y(lem)) N NIMultplfeation - ( mon) etn) —9 5 THU @ x) J einculan convoluHon- yO) @ x) FU KU) Cinculan co-nelaiten- 1 Bytm = 3 LOY COO XW YA (k) m0 Pan seval ¢ nelatfon = . 2 reernrey LB a) wey fi: neo Analysis of LT1 Systems using DET x) — LIke] 5 qe. ROG bw) G5 DET supports only etneulan convolu to the fllowing progedune~ 1 Ofp can be obtstne ist 7 s Take (NMS) point DFT of ip stg xen) given impulse mesponse hen) Hox OS XC) dof thy * The (ineaw convolution in mre domain x(n) Dhow) can be bat ned ' obtained by Yoking the preduel y(n) HOw) tr foeg, domed » domain, * The output in Hime domain can ADFT of XCk) H(k), Seabee “elCompute 4- point DFT 2& & potat DEI of causal jhnee sample seq. given LD : elseuhene \ 20 (pled) 2 Bl xe N=4 are s xem) n=e yoyebt haere =4 x)= FET? method of computing DFT with wedueed no: of caleulat on padix- » FFT? Ne r™ use by broker down dpe wedix. padix-2 Fe wadin-2 5 vous 2 caleulat ons? N-=potet prt Ne of colevlaten fn modm-2 FET > N fog eom plex 53 e orn plex amaull p lien ton NON-U) complex addition numben pd Niog.n cemiex addition. ro ulfipliectionsPSS... Q Ee” .& + FT e Devinatton in Time (DIT) Raddtx-2 F 2,100] we fl,2,2 94s 4 2) & 3st fy fa ae [3 3,1 o4 2 6 15 3 F CES ge aed Tria Mey Mak * Ain) 25 vet Oy = 2 vacrrer M09 25 > OFT (Va) = 4 va 4¥ute) 4 ior “ iy Wy Fos), we, heal > x(k’) ' vf ' Ruy FiU)= yi 40+ WH), Vian) FO = Vin 4 Wy, Va k= 0,1,2,3 “Ot WOO Soa Vuln) oP fon K= 0.1 oe 2A 9 by = em Me Number or lumben orden BH mevensed onden x0) 000) x0) xtese) ' *O) C001) % (4) XC100) OY x (oe) xc % (ore) xO % (100 Ge % (e01) x (3) x (ol) > ale) x UWE y+ Thaee- stage computation of & point DFT? Combine 2 ied te potet DFT Combine Y point g [— * (D peted Dry + Flow qnoph few & point using Radix - 2 OIT DFT. I Vy (0) Vu Co) Butterfly Slo graph Va) Tr ws 2 = =I As aor Bea —bWe VinC0) Ae s Same fon oy we xO) Vacs) * vach) x 24) aa uw «sy Vevey 20 Vax C4) cS 4 Vay (0) ey Vin zu G we x» (6) ea Vea Cl)Ft) V9) RCD Vy) we FD Vi) + 3 vat) toa i eomputation of tnvense DET The following Procedume can be. employed fom 1DFT caleulation s Take N- potnd freq domain Sequence X(K) as % 1p sequenee. cempate FET by Using conju gate of phase febons 7 Divide the ofp sequence gbtatned fm FET computation bY Ns P get the sequence xen) S Ang point sequence ts qiven by ROYSh Ut 2s1,4 Compute 8~ point DFT 22,haa, of x(n) by wading DIT FET alqgonithm, RM) ={25,21, 02, V2pone pom? sequences i in. adi xe DIT- FFT algon'th: Requived — 9 point DFT using nadix-2 gorithm. 4 jew “x (0) 2"Z aim eine nae ~"Z xm n=o+ Ftntde Impulse Response Filtems 5 camera shee he Petpet Mites 04.10.23 *stottonany & Hen oekt dsm s/q¢ how te use “a using MATLAB *Genenating Analyteal signals ¢ Genenatfon of stattonany signal F_gampling= 4e3 3 — wee : T_ sampling = 1/F- sampling 5 | 2505 c/- siqnal fnequency ! core) To step = 55 Y Me es 4-0: 7. 20th Tees is 4 Rene) Ven = 2293 . i . ae pi/G > i ‘ yee Vent sin(2® pi > fo mt aphids Siquee (+) pr cee 42 Teale CishHonery 50 Ht signal’) Ganenaton of a non-statfonany gral a a, THe fixed ' b. Time vonying Foy, | ating vet s/q pleat (cee PGT aoe anh Teatiang ve ne Fred. 19, preesssing cee Gs sreamd guid), (Re ule Fewnande Bi be los Auguste | pesgn a s/é, Foeg.vantes HitK Heme. iy 4 nade of i rang lineanly preportenal with ime Tineaw variation to tig. /o-ls foequeney of s/q > SoHE: th amplitude = 220V , phase = 3o duet of Stinate ossCT. , RAIK DFT. FFT. 2 nh DSP . 7 healy He sly “Tose Ipao kA Tet Ta PT W sone. 0-02 see f.= 1000 He. > no, of samples = 508 1-3 1000 a oer > gt xo es one eyele > 50 samples - 5° N= 50 O12 ay,-t 0,4 ,0,-4] Uy = 10 Ka qoua _ 20088 sana = 100° Ss 5 x0) =6 : 5 ° Seon ane Tao N= 200 Hz. 20) (sampling rate) x)= 0 xc1) = | x (5) = 4 2Q)= x(A=-1 G9 O00 SE: -j4 ° : Foo:3s 24. 200/¢ . 20 25 58 Hy 100 ° tar ise ty j ple hcs, me 6 sous Uo aKAM A NAGIOR why window ? 220 oe lon ay gf Rechogu = *) talonquiore ° $e 4K Hied*) = 1 pet Hatele Jat 5 —OeLW LWe ba (my “° Cindons contains tntetow exoeh faeg Pesponse Walvent postutate)220K 3-9 balanced Imsin Cot + ®) aD = v(t) © Vmsin wt Let'us take, as eunrer t, & vod as vol. 3/9 previ by PT (T). alg provided by ded no component > 50H. should be trene The fog Uld be qnaaten than The wen + pone!) —> nyquiet aneonemy lottmes of the men. fmeg. cxmpenert > valvent portulote somple & toad process e lead, one Une den a sdvuple 2 Hus Pa" com pristng of 3-— nesishy tn the figure 7 is primarily te det nuous AQ provided by Tavelves using sample oA whieh the sampling ghould be dome is determine) by Somme! Led, us const QR 2-tfs CT & er of PS SP aod, The cont The objectve eernine the pamameters like, § + ReGks Vows er & PT convert 4 ch5 Yisenete sig. This & toad ele. The freq. 40 Theonems — precess 2 Nyquiet sarepling thesper™ 1) Nyquist sare pling aaqully yeted from the sampling slg. Sampling & gon thie theonere s/g TP be netonr gf eontuined In tote e/g ghould be < geiele of THE ™ma% $, > 2hm 31> Velvan postulate gafe the s/q , tre sampling Freq. shoul ontains jin this avea, ve 0” J be ta orden te presse? G the man. Sng. component & atleasd 10 +imes © ava) N=200 0.02 leg samples/eyele Be Bq sovenl ap ing freq ) fp to I Me poted pry Po %& 340 Cewvelep will ess the data) (sam Cra ba tan reas) ive DET. (P-120) unsive DFT! 'L) sample 500 He Samples /see, - oJo2 x56 “ 3if 20 samples, L= 10, how much windows ? n— window no. Mule) = EL aterm] WYCNm-v