GB2396427A - Eddy-current sensor arrays - Google Patents

Abstract

Inductive sensors measure the near surface properties of conducting magnetic material. A sensor has primary windings with parallel extended winding segments which impose a spatially periodic magnetic field in a test material. The extended portions may be formed by adjacent portions of individual drive coils. Sensing elements provided every other half period may be connected together in series while the sensing elements in adjacent half periods are spatially offset. Certain sensors include circular segments which create a circularly symmetric magnetic field that is periodic in the radial direction. Such sensors are particularly adapted to surround fasteners to detect cracks and can be mounted beneath a fastener head. In another sensor, sensing windings are offset along the length of parallel winding segments to provide measurements over different locations when the circuit is scanned over the test material. An image of the material properties can be provided as the sensor is scanned across the material.

Description

GB 2396427 A continuation (72) cont Andrew P Washabaugh Vladimir

Zilberstein Vladimir Tsukernik (74) Agent and/or Address for Service: Page White & Farrer 54 Doughty Street, LONDON, WC1N 2LS, United Kingdom

F.DDy-('rNT SENSOR ALWAYS B tCKCRO-u3D OF Tr-18 DWELT [ION The technical field o this invention is that of nondesmctire materials

czrac.e iza.ion, par,icularly quarti,aL;'e' model-bas_d ca-acer bayou. of surface, near-s-face, and bulIc -,narerial condition for fiat and curved parts or components using; eday-cll-rent sensors Characrelizaiion olbuLk rnaeriai condition includes (l) meas=-ement of changes in material state caused by farigue damage, creep Carnage, thermal exposure, or plastic deformation; Art) ass_ssmerr of residual stresses ar1d applied loads; and (3) assessrnen: of processing-reIated conditions' Lo example from 10 shot peeruncr, roll bumisning, the.nnal-spray coating, or heat treatment It also includes measurements characten.zing material, such as alloy type, and material states, such as porosity and temperature Characterization or surface and near-e,-ace condi+.ions includes rneasurell,e-,ts of slfaco roughn.ess d.splacem,e-t or changes in reiar;\re position coating +. hickness and coaxed condition Each of 1: these also includes detection or electromagnetic proper changes associated hi h Male or muitipie clacks Spatially periodic field eday-cment sensors have bee;:

used LO measure Ioil 'niclmess, characterize coatings, and measure porosity; as well as to mevsllre properly pronIes as a Notion ol depth i ato a part. as iiscIosed in _ S. Pat. Pros 0:,91 and 4o 6v29 270 Co renhoraI edr-cull ant sensing involves the excitation o a conducing vrindin=, We p=.ima with an electric cur.-ent source of prescribed Frequency This produces E Diane-,=r7"a r -.aelc.'ield --t -the He n _, Fluency w Rich Wry is cGvrecred cairn a sensing v-Lnd7nG, We secondary Tlae spatial dish rigor- of Lhe macne-ic field and We laid meas, Ed by -. I.e seconds or is ir^iv7e:lovd by- flip

2: pro- Alibi =c pia'siAval '-ope1,es (eiectricE1 co;lcuctivirSr 2nd a=eT.ic peleaoilit) o lineal he rnaeaIs. Vi7he3 -a e senso. is i3ten+ ior1all,r DIaced -a close p;-c ir.'i-.r to a rest a:;e-al, he physical prope.T.ies O,A -Lne or" i2I c2 1 be ledll__-

^ o: easurer.l-,Ls of r:r!e i^-A?en^-rlcv be.;ev:l +.:-le ?im2 and seconAa.-y v innings

-2 Trad ionall-, scn g olAeddy-cul-re;lt sensors across the 1a,eial allrace is r";acn used LO detect flaws; such as cracks For the inspection of structu al 'embers in an aircral. power plant, etc., it is desirable to detec, and monitor material damage, crack ir it ation and crack g-cvh due lo fatigue, creep stress corrosion craclin etc. in the earIies, s-.ages pcssiole in order t0 'reriT)7 the irlTegnty of 'one stcnJre This is par, lcuiarly -syndical for aging a reran, where reedit - and commercial aircraft are being Rio;. well beyond their ori_-irlal design ii es. This requires increased inspection maintenance, and repa r of aircraft components, which also leads to escaiaimg costs For example, the usem1 1Q line of the ch-rerlt inventory of aircraft in the U S. Air Force (e g, T38, -i6 C-130E/r, \-10, AC/RC,rC-13:, IJ-9, E-3, B-1B, B-2El) is being extended an additional I): years a. least [4ir Force Association, 1997, Committee, 13975.

Similar inspection capability requirements also apply to the Iietne extension Olr engine cornonents tGoldfine, 1998] t: Safely supporting 1iIe ex'Lension for sties requires botl1 rapid and cost effective inspection capabilities The necessary, inspection capabi Fries include rapid mapping of fatigue damage and hidden cor1 osion over Inside areas, reduced req--=vlmems for cali'vrauon and neic. starldaras, monitoring ol^cifcuit-to-access locations wi+ouL disasserlD7; corAL,rluol!s on-I5ne monitori,,g A^or crack iritiatio rho and g:o-v Lh, detection of cracks beneath mull ple layers of material (e g z second layer c-acl; detect or.) ; and earlier detect on of clacks benea'n fastener heads rife fewer false alarms general, each -,nspecr7:on caa'nilirv- rem; es a dir-^-rD:lr sensor coi-.n-V,Tlrano7,. Ike use of edd--cllrrent sensors:or nspection O r C regal _ocatio=s 'S an 2::I!tegr^1 ccr.nponer' OT r,he damage olean.ce a d r; "truce t for cause n,-le-hoes "sea 1o: CC1. nerciai and military a7-cral. Lyle acceptance and s7arcessl np,e7merca.ion obese me Clods ov=-!he last fierce decades Sac riabied i^e en.e vision and sale, op-vatiO 1or 1n71mv-O7S a -clay The co respo,-idirlg aces 12tior OF ^a,ioue damage in c-i-1 scrl-al me^'nbe.s c -dies e agirl- a -c-a,. homer e is pi i-.c. _asglr C, Amid =-d cvI]_-^v In omit probl m Mcr^-y.-or^0v:l-^ts a.--_rv oigima By ties gnu_ Lo ash -he des g life o ^ -he aircr_* -1'i'-hvL- e'lvi-)g vl-a-Li.g /. sails 1 1^- So-.po,e-l-s) A- nor- '^;L:l' ri^- -A se-in-).J-'L''! 'vaL'se __._Cl^'

rerr,ain in ser,;ice beyond or.grAnAal dsign fire and; for mLi; a=; aircral, ivhecause excandedArrAission reqtliremerAts expose s7rucn^A-es to unAanh:cipated loa1g scenaios.

NeN tee e>.ension pro-lams and reconnended repaTr and repIacement activities are o.en excessi;ely burdensome because o f lirn^.i.a.ioAnAs in technology avai7ahle toda, for fargue deection and assessmen'. Ma=nagers o-a,Ae A ircral, S±ctural Integng' Program" (A31Pj are 0 ^ _n ^aced -V'Thi dirI-7C'L7lt dec-'sions AO einAer. eplace coAnponen,s on a fleet-wide basis or introduce cos iy inspection programAs.

1hlermAore, there is owing evidenc- tha. (1-) mulriple site-daTnAa-,e or multiple eIemel^r darlage ma, compromise fail safer;v in oIder aircraf+, and () IO Sini'tCaAAL+ IafigU daArAnAage' vryi+Lh subsequent forr^!laion oicracks, may oCcu^- at locations no' considered cr;r.ical in ori_'rAal 1atigue evaluarions. IAA application of d-=^age tolerance, irAspecr.ion schedules are often overIy conserva-:ive because of IimTtaTions in ia. igue detection capability for early stage dal^:^ae. reAn so, Iill ed inspection reIiabiIity has Ied to nuTnerolAs corAn^A=A_rciaT and mAili+ ary compo^=^ent 1: fai-.-res.

A be,,er uAnAderstaArAdiAnAg ol^ c^-ack miriation rAd short crack grow h behavior also affects both,he folula,ion oldrr.acre toIeranAce methodoIogies and design rmodilica+iois on new rcra* cLnd aiAAg aircra*. ror sale-ii e col=^ponents, designed tO Iast the Iife of the aircra, ^-7o inspection requirernen7;s are rpivaly pla;=^ed for rhe 2Q - -s desi Iie. Life e';teTlsior. proor77^S nave in=.-o'7uced req-v:.^relL7^eAnts to inspec ÀhAse safe-iile', corA-Apon2nls An se.vice sinAce ey aAre noW oper272Ln^C7, beyor rhe original design iil^e Xore7er,.heA"e are also numerous e^xaT^npIes of coAn^^ponel=^Ls orimnally desigArAed on a sa-e-Tie basis at ha; e raiIed pr70 to or ne' their or _AAA CiAy SD ec1 LA ed desi e r1 D otn rrAiiitaL "y Ad C orA7.A eTci al aircraf t.

5 ror saAfe-life cor pone-rs at -m.-L7SL AAOW be AmAanaged bNt da=Aa5e rOIeAaAnCe -Ae-os' periool'c inspections 7re generally far nore vos-.iy 7naA=^:or co^-npo en s oA=v-7rlally desi ed WiLr1 DlanArAed AnS,eC7270nAS. O1Len Lne h5C-1eSt COS, iS 2SSOC-, ated i-r7 ^sas s A:.b "y a^^d s -1 ace, 27, r 7 T 7 - T. D_ _ VA aioA A. idci.i 0-Aal rvaL.es 7 0- [D e i:I ee; AS r 7 2 r-y litvd by AA7-e ou of se 77 2 2 - - 7 7 v r An ice; 1a - _n __ -- - 55iDA- --- LV?_S as GA_ vf L __ -

-<i 1 iT7suvction re^uireme!ts becoe rnor- b7lrn-^so - e. 7-Lhvlore rhe 7arer an L_ A5C-ion vnco - S L_ iV Je da 7 v - r- 7 7 - -,, _. 1 a- v o r 2 = _.__ _ V C AA A.

S. _ __L_ _ - ' - _

or lAeiN; replacern-v7r is requred. Tk' 5, inS?5ction 0 - 2-'i'Se ?ocaio:. s -'.O-J

- - disassemDiy and surface preparaLior is of significant advantage also, -Lhe capability to detect fatigue damage at early stages car provide alternatives 1or corn?oner repair (such as Anal material removal and shotpeening) Flat will peanut lithe extension at a lower cost than current practice.

5 In general, fatigue damage in metals progresses through distinct stages.

These stages car be charact;zed as tailors [S. Suresh. 1998: (1) subsruc. =LIral and microsacrLral changes which cause rucleacion-of permaner. damage, (2) CreaTion of microscopic cracks, (3) growth and coalescence of -these microscopic flaws TO form dorninan:' cracks, (4) stable propagation OI the dominant macrocrack, and fS) 10 structural instabiliv or complete fracture.

Although there are differences of opinion, thin the fati sue analysis community, Suresh dennes the the-d stage as the demarcation between crack initiation and propagation. Thus, the first Two of the above stages and at least Me initial phase of Stage are generally rnolght of, from a practical enneeLing 15 oerspecrTve, as the crack initiation Phase.

In Stage 1, micro?lasic sums develop a. the surface even at nornnal stresses in the elastic range. Plastic deformation is associated win movement of linear defects known as dislocations in a given load cycle, a microscopic step earl form a; Me sT-face as a result of localized slip owning a C slip line" These slip lines 20 appear as parallel lines or bands cononly called 'pe-sis:^vn slip bands" (PSE s).

S'ip'v2=Tld irlL!slons become stress co,c-r.rarion sires where,misroc.acirs car' develop. FiisoncvallY, X-ray diffraction and eiectr c1 resiscivlt:, are among me few nondestructive methods -h.a' haste bee? explored,^o' de action of fatigue damage in 2: he i='iation stages. <--ay am ac; on e'ovds for deT=cion or 1^atigv Darlene prior to mic ocac^g have been iresti rated since the 1930 s rRecrle', 19 7; Me rle.

I99J. these --esT.s, fir g e damage 'rumbas L-"nC -50 be rei2ted LO n 2C ion lme oroade ring. More recently Taira,1966, drainer [19743 and Weiss aria Osida 19.3 have no the, d5$relv d.ne X.==y i:-apace Rod. T "' 1> opossum a 3Q sel--.ere en. fig sys,eml^cr crarac,srizz lot o^larrage, larnely-,;e,-aiio of dislov^-tior1 densit es as neas -ad 1 'rat -ricrol=.. srs 'eels hale s,-,ace c runt r,-as--ed is - JO icror-ere- D5lsr ^5 solad=. -lv5.^--a 02-ev'.3 c'aL" s=cr5s

-s - À'hat in high str_lTlgth aluminum alloys tne probabiliw; of fan:gue failure is zero fo disiocar!on density raiios of 0.6 or bolov.T. no,ever, i. is generaIly impr_cricaI to make sach r..easulrement in he leld.

Electrical resisrvity also provides a potential ir aicatio; of c=mulative 5 ra crTe dar;,age. Tllis is supported by theory, since an Trrcrease in dislocation densirv rslT-s ir ar incroase in elecrical resis.ivi:. E, !irrates sugGeb -hat, in he case o aln inum, dep enaing or1 ibe increase in -:he density o f aislo cations in tne aigue-damage zone, the resisfivlty in the ratigae-affecfed recrion may increase by up to 1% prior o formation orrmicrocracks. Tllese esfna.es are based on 10 disloca:on dens.;;es in:.e fgue-darllacre zone lip T.0 ber,ee 2(lOl! crn tG 10'7 crl2 and a resisTivity factor of 3.(1 o-l9 ((cm3,riedel, 1954] .

SIJARY OF TTIE T.NTIOIxT Aspects of ihe inrentions descnbed herein involve novel inauctive sensors or ^e rneaslL e.TnDnt 0 ' e near s.ace propel,ie5 vf cond-lcTT rlg and magne ic 1: maTeTials. These s-nso s ease novel i;lding geo neT=es that promote accurate À- rnodeling of the response, elimirlate nany of-e undesired behavior in -,ne response of the sensing ele.nents in existing sensors, provide increased depth of sensit-; vity by eliminating the coupling of spaLia1 magnetic field modes fha+ do not penPrrfe deep

inTo,'ne Tnaterial,.,r7der tes; crr), and ororide "-lanced sensir!Yi, or c.ack 70 detection, localization crack oren.anvn, and lengh character zatio;. The focus is specificany on lateriai character zation and also the de ecion and rnonitor.;lng of precraDk fatigue damae, as well as deection a nd Tnor!iToring of cracks, arid o-,her :a_iaI deada-ion om testma or seice expos7re.

\'f__ods are descr,bcd for oT=ung eddy cen" sensos nai p- L!ary A' -inclLlGs 1or imposinC a spa-,iaL7periouic ma=gn=-=ic:iei in70 atestrnare=ai. Ln one e='coaiment, the prirnar-y wlnnimc inco,rpo-ates parallel exter7ded windiT7g se=ents oed y acjacPnt ex+.ended DonioLis OI i:l7idua' ci>}- coi7s. The Lr!ve co Is -e cD'= ed so har f.he v,LL^1 err passin_ [LUO'lh adja0-nr "v';lended NA mdin= see "s s in a comon &rsction and a sDatiallyperiodic mameric.leic c lwsed in rhe 3 -. n noher erlbou:nen, a svle;-ne=^n=nc conl_e o- hz\-i= ex "n ed portio-s n1 one plze is conneced ir se-ies.o =ohe = nd5!ng C0:ld-C''O'- r a

-5 second plane. The conductir g rneande s ae spatially olriset om one another so hat the current passing through adjacer,t extended winding secnents is again ir a common direction.

:E7or sensing the response of Eke Chug to the periodic magnetic f old, sensing elements are located within the prima= - v winding. one er. Ibodiment, the sensing eierr,ents hay_ extended Domons pal-aIle to -the extended portions of rile Scrimp wincin g and link incremP-la1 areas of magnetic flux Irvin each- Hal mender. The sensing elements in even, o'er half-waN,elPng h are corrected together in series while the sensing elements in adjacent haIfwavelengths are spatially offset, parallel I O to the extended portions blithe prmarv. The sensor can be scanned across the surface of the 7\7T to detect flaws or the sensor can be mounted on a pal, for detecting and determining '-he location of a flare. preferably, the longest dimension of the flaw will be suba=hatiaiiy perpendicular to 'the extended portions of the p.mn<r winding.

1: 7'.Ierhods -ire also described T-Or owning circular eddy c-rrent sensors having pa windings JO- imposing a spa7;iaIly periodic magnetic field into a Lest

material. The spatial pattel-n can be created 1rorn a plurality of concentric circular serpents, where cur-rem flow -rough these segments creates a substanaily circularly syr^etric nonmagnetic Weld Mat is periodic An the radial Infection.

^0 response of the LuTI to the magnetic held is de acted nl'-nA ore or rno-e sensing elerileAn^s vIaced be.eer1 each coerced Tic loon.

The emended po.,ions of each serAsing element aT conc3n -. c with The concentr c cT,-cular segments of uhAe pT-,Tnc wading. The sensing elements relay also be in a diTf-renl vl?ne than "e pin-3 w rid fig. These dings may also 2: ^o a su'vs.ar,iia v- closer loop outlet Man as a circle to follow a co;i:o + mAatP-r-al 1 rider rest.

1l e s ens g el Mets ca t v _i sub v ted in 0 _ho e n - Am a y v ^= _ meters. one PA bo=-nr a s nAgle seals rig elernen. Is placed i- iirA each hat vvas;eleilct or trlepla v -Heidi r -Ad, Beg -ale G'vc'L Cal tOvriv iS cal1 vet-age TO LOG each sons lag Plrne l, -,o area e 2 3=nso.-. he senS rig el=mPr1ls curl = Recked og=e Lo.-QN:iie of =l.n sisal-s.,Tq orh=-r elllbOv! I.. -.ke s=:.s-._ -ices Cal= lick =as: rule ED l Cal a--- ado- -. he ir,-^ in ^= he

-7- primary winding senen,s The sensing elements can have the same angular arTnensions and all every offer halt wavelength Carl be connected -together in series to provide a common output. These are examples of circular spanally periodic field

eddy-c1LTent sensors. Chinese circular s-nso s can be eased in either a surface mounted 5 or scanning mode Another embodurlent of an imaging sensor includes a preach vrind;nc of parallel extended winding sene,nts that i npose a spatially periodic magnetic field,

with at least trio periods, in a est s' bstrate when driven by electric cTn-rent The array o1 sensing windiness for sensing the response of the MAT includes at least two 10 of me sens:rng windings in difieren. half-a-.;eieng.s ol^ Ale pumas y winding. i-nese sensing windings lit incremental areas olive rriagnetic BUT and are offset along the length of the parallel winding segment to provide material response measurements over different locations when the circui, is scoped offer the test material in a direction perpendiGular to the extended ndmg seal, Dents To minimize unmodeled 1: ear acts on We response, extra conductors can be placed at the ends Grime sensing elements and 3,iuDin the endmost pram winding meanders, and the sensing elements can be spaced at ices, a nalfwaeleng from the ends ofthe pnma7 name addition the distance from the sons rig elements to he ends of the pa iindir.g can be kept constant as me o corset spacin_ berwGen sensing 90 elements within a single meander is varied An image of t..e mater properties c he obtained when scar he sensor in a direction perpendicular to the extended portions owe Putnam - idly-

Ihe sensing elements can provide absolute or c-, rerentiaI responses, which can provide a ailIererice ir LOUT properties parallel to' perpGnicuiar to o- at =n ?, mterrneate auntie to the ex^,=ded portions of the-;pl-m winding The spa.i2iiy periodic sensors car be.-abr.cated onto O=^ible conformable Slash-atesforheLesoGcLlo Olc-mwepaFS.tirerriati;-ely =es_rsorscan or =.o,=-ed on nerd flat or cleaved sv,oslates so- non-collLac scnrg. ProteC,lve or sacri-lcial coats can also be used.o cover;. e sensor.

3G The sensors Cain be moon Ed a_airlst auricle surfaces for tile Geteco Ol T, --as TO Q =1c:raii one'.. Do.' all -or =. 'a Cal: =-are + _ see soor --o if ao di,ion? i - rllorla.io-! o- Me O'Dell Easy rer- en or e>; 2?1e the sir sor

\ -8 l.-off the -T temperature, and the l\irL}T permeabiIi7v can be varied.

Remeasurement grids or databases can 7oe used to derennine the electricaI and georr.eic properties of interest at the location measured by each sensing element;.

Tile electrical or geometric properties can also be correlated to of her properties of 5 interest for the MUT, such as crack size or depth. Multiple Sequence- measurements can also be performed to determine propel=; vaia.ions with depth flow We surface owe MUT.

one embodiment, Film age near fasteners can be nolmtored hip spatially periodic field eddy-current sensors. The sensor should be rnoun7,ed near the fastener

10 so that nonage in the MA can be detected through charges in tile eIecrlical properties measured with the sensor. The sensor can be mounted beneath the f&srener head, between structural layers attached by Me fastener. or at both ends of the fastener. The damage may be in the form of a crack. Circular spa7-aIly periodic sensors halting hollow center regions can surround fasteners to detect and locate I damage Mat may emanate radially. I\Io7nred on. o, wi-n an a cyiinaricaI suppo t material in the form of a washer faciilrates mounting under a fastener head. 1the suppose. material nay also support compressive loads. The damage from nearby fast-n-rs call be snow tory] sirni+eosly i-UhA multiple seTlsoTs. Each sensor can haste a sir.gIe, absore output, or pairs old sensor responses can be used tG provide GIG aiiI-vrenrTal responses. SimilarIy, for mAUl._?le SeT4SOTS, the d-riv-- conductors may be connected wi h a colon drl;ve Sinai o, the sense conductors may be corer d together for a con=on output cormec loll.

Methods are also!1escriDed for creating databases of mease.mAelA responses :or m-LAl-tTple Iay-er sensors arid using these databases ror covertmo sensor 1 espouses 9, T!ro coronaries oaths >my. ibe responses can be dee.invd -row r.alyrical, nm.e LA er rAc- O.- __A_t_ __ _Ar!le MAO ITS.

Ca?zbilTtTes formo''i-oATng raL;gue damage as tar occurs or test 2=icIes also o-ril- no-r-1 methods '^or faDr carting 1a=;gue s.-dars..aClng a^-7^ electroma'enc sensor ^.''at p-cid-s Ire ab,olu, e rneas-re^-=rnr c. He lectcal !0 prop--iA=S C--'Ag -A7AeC7rAarlA_a1 ioaUAng fir _-i7Je rAeStiAng a Ic-.,s he Inca Anal consist on o be rAorA,ored as Me Vantage occ.. s. Moi-ol= o a:- c.aanGes 72A..'ie -lec.;cal joules ±A7A17 adobe ' tar The 10aG' Lo oA, =7-oTr- >-L = SO Ie's o, ='7'70. Eye

-9 - damage coin take the form OT a fatigue crack or pre-crack damage Once 'he cram has formed, the sensor CaOt be used to monitor the change in crack length Nt'h the n,nber of faticrufe cycles Multiple frequency measurements can provide a measure of crack depth. These changes in material properties can be monitored with multiple sensors to coNrer several inspection areas and create spatial images of the damage In one embocLner t the sensor is a spatially periodic "leld eddy current sensor and the MAT is a metal Al+,err.ativel>T, 'he sensor could- be a aielectror^f-eter and the 1vIITT a dielec.c material or cornpos te In another er.bodimPvot either eddy current sensors or divlectrometP-rs can be monkey under patches or bonded repairs.

10 For he fabrication of fati=-fe standards, He geometry of the fatigue a'.ticles can be altered to share the stress distribution so that the fatigue damage initiates Underneath the sensor This can be accomplished by Among 'e center section of typical dogbone specimens, by providing reinforve;nent ribs on ibe edges of the specimen to prevent edge cracks From forming, and by providing radius cutouts on 15 -me sides of The nh;. =ned center section E3REF DESC21PT1ONT OF THE DRANING-S

Me foregoing and owner oDjec-.s, reatures and adN;anrages olrche ve0G: will be apparent from ale foIlowng more particular deson:pion of preferred emboss TI^?nts ofthe in enTion, as illustrated irA rJ]e accorrl, =Asing drawings in which 9 0 lye reference characters A efer '.o We same pans r=o-^gh3Ylc The different Views The dr_r.gs Are not nec=-ssril,N to scale, en Stasis rlsread being placed upon ilius:.rar=^g The pr,'=^cipies of,he invert ions Fly 1 is a r>Y,a-1 -fen; or a 54--r,rT-r-^I^g \/YTloing \4aome+=r senator FIG is al ill,s-aion or.5e hl.IiM measured corlduc:ry dean deuce o 2: he percent o- total intrigue life 1^or T 4 s-,airless stem and alu^YulAA lIov 9'v.!9''. -, r I- -,.r -lo A- - v- " ' Y Y ^' I& sho s A v A I v as e beet. s c Baas al on al!inaA A Silo j 0 :^^o' -- glass Spe-All^=-AnS -_fv^-e BANS at-. --a ice.esr Lo 2rv^Y2s pe^vA=l--S ^- ct^l .. at go: iiI_.

3 FIG is As. iI!ns ^a=3rl c'^ *r^-d:ne:lsio-l-l al =as -ed asotr.Te cr' wS5-i- scales a'cr!_ -n' s- -rap Pa my a'- -v9 'denim also e

- -10 coupon widi extended portions othe windings (a) perpendicul,Air to macrocrack or entation (i.e., perpendicular to the bending moment axis) and fib' parallel to macrocrack orientation.

FIG is al illustration of two-dimensional \,IN\M measured absolute conductivity scans along the surface of a military aircraft component with windings oriented (a) perpenaic-lar and O parallel to -,ne bending mortem axis.

rich- G shows scans of i-du-ecronal magnetic pe.-me-abilitV along rwo austere stainless steel specimen s. One specimen was not la.igue tested and the other specimen was fatigue tested.

10 FIG 7 is an illus ration of moidple frequency m=-easlrem.ents on a fleeing 73 ' fuselage as the Whim is scanned (a) hor zontally above flee lap join b beneath the passenger windows and (b) vertically from a window to ibe lap joint.

FIG is (a) a plan view of a sensing element and OWE- Array with one meandering primary minding and an Way of secondary sensing elements with con,nec.hions to each indiid,al element and fib) An expanded view of Ll3e sensor windings. TIC; 9 shows art illustration of six A- rra,vs mounted inside a'^ld on the surface ol^ a fondue lest coupon.

TIC- 10 SI"\AOWS an M4]\4A-AL alar mGU=Aited inside a fat;.crue rest coupon.

O:FIG 11 shows &) e>;arnDle of the \NTM Tneasured conductiN Styli -yranar, ion yin -QuOue keel.

FIG 12 shovers Al- ex2-nple office l'v'YM,l.eas-ec' i l.-o'-lvaia ion 'flits fatigue level.

IGI3snoNxrs--,.e3?Ieofthe 1yr\/yfilL3ea^-edCO-UCNr'r>N'C-;i-,oi ?: -w-ieasily stage Unique damage.

FIG 1 i; shoNx. s the M\..4 TneaS=rea Coniu-TtiT- --axioms -,-^, -ATE ai<r He cycles ror specine;s a) 'A (b),','3A, gild (c),' At? FiG is shovers the >'r\4 measured onductiNrir ran.a-ior -i-2rl serLs;ng elegize pcsi;io.-' for sped,l As (2),,; -), 'A, rind (C). '_.

A,,3 BIG- i6 snows ran i:la=_-!ion of An aigo-.-.,.^ for aecec.ic- 0 ' he onset o ^ fatigue damage us;=g a surface noted edd-clrent sensor.

-11 FIG 17 iliustraes +he reiarionstip befween he M57r/f measured conduc'. 'vity changes and crack length estirna.ed -om SEM.

FIG 18 shows an englne^=ng 2,r.\rin.cr for a faiia-e spec;nnen has ing a reduced iciness center section and r eirorcement ribs on the sides.

5:FlG 19 shows an engineering drawing for a raugue specimen hating a reduced thickness center section' and sow ethical radius cu+oucs on bow sides of the reduced thickness area.

FIG 20 shows an engine-.ing L-awng for a fatigue specimen having a reduced thickness center section, reiDiorcement ribs on the sides, and symetfica7 10 r_A;us cutouts on bo+,h sides o. the..nled arm.

FIG 91 shows (a) a ra.igue test configuration win tine IVI-MTay mounted at a steel fastener installed on the Al 2024 test speci nen and (b) a side view of t'ne fatigue test configuration.

FIG 2 is an illus+a+.ion of he lose of an MOM sensor for measuring c'_cr i-. ienh 'ear a fastener.

FIG 93 is (a) a plan Field or a linear _,A Tray for crack depletion End

detraining crack location and go) an e>;panded View of a sP-ns:,ng element in the knew hI-Vi;-ray.

FIG 24 is (a) a pliers Niew of all 17M-Roseffe for crack detection and 20 delliliiL>r crack- circ-mlervnti21 (7im7lhal) local on And (b) an eXDan.dd -I ion-' Oir some of me -winding conmecions in an -Rosette.

rIG 25 STONES "ady-cl,ent Arrays mol="ed between layers Of a -mcur_.

:F'' G 96 shows an edd-clrem array mounted underneath a fastener.

FIG 2 s pa) a o1= view cl an -cost e '-or crack detection and v- ack 7: lent mvasurem-vr! rid fib) an exp;unacd vim; o'some o.- -he winding cennern:ons ran a:1 14 r iV'-ROS - LLe.

FIG "8 is ala, ill,.,sration of apa- of2lv-Rose us placed aroma lAasLene . A,aA' near a o leer.-^L..nr.

fA, MA T, 5 r,, 7. _ r. v c A - - - _ _5 is a sc_s!a lc ?1 n 1 c a Brim la: Lr Wang err = l tons - o^sevond2 elements. On one side,-ins seconders eIem=-l.s ^--5 ccr,Aevted Liz ' Llal x, the 1-^'-n.s cr. Try-- ??osiT.= sibs 0. - le=-l(iel - ? _! V __, ad Jr con ec era i I Hi I Alla if

-12 FI(3 30 shows aplar, view of a Manured Mi7i7f-.Mray.

FIG 31 shows an expanded view of an absolute sensing element.

FIG 32 shows an expanded view of a Miff rential sens;,ng element.

FIG 33 shows an expanded view of a differential sensing element.

5 FIG 34 shows an alte-na2ve Method for correcting 'LO an absolu7,e sensing element. FIG 3: i Sates al? ateA=at.-e design or a mended lo, p-a^r = nc;mg.

FIG 36 shouts a meas.^reArnenr and for z layered winding resign.

FIG 37 illustrates a design for cross-conAnecing the meanders of We primp 10 winding -'rich greatly reduces the necessary number of bond pad connections.

FIG 3 is (a) a plan view of a multi-layer electrode Comedy and (hi) an expanded view o We,ind7'r g segments.

FIG 39 is a plan view of a sensor similar to Cat shown in FIG 3 8, except the grouping of sensing elements coffer dilerent sections of the meandering primary 15 footprint. FIG 40 is a schematic plan 1^OA a layered primary winding design.

FIG A, 1 iS an illustration of the temperature dependence of We 17 mAeas^vd electrical conducti-riW.

FIG 42 is an illustration of the absolute conducTivirv dada NOT,, repeated 20 7 scans in slots (a) 29 and (b) 23 of a Stage 2 fan disk.

FIG 3 is In iiiusTa-!ion of rue absolute co lcactiNTi data Mom hum scans ail no slots in a Stage ? Tan disl;. Grows indicate slots -mat had cracks detected by brie,,rv arid OTT. ^A^ci-cled slot! - 1''' hers denote C,=^v' detected by -rho Own -but not Tip 9: FIG 44 is ar il:-,lsaiio Qua one ^ormaTi2;=ff conductive ty data GONE -s oamg to urge data Gf rlG 4.

rIG I. (at is an -,ll;is'rar'.on olive, e,nc2ion Oh rAoImalized cond- cirl - T de?endvnce or c am length Or the slots listed in Table 1, No:ninal -hesholas fo c-ar dtectO:1 is Audi '2 id. (it) -is ^-n e.?=dd N-'es- al Ike:es?ose f-=e 3v mall r cracks.

r G A 6 is a pled i' a. 0 or ELi _' a ive-, Dodiin?vrt. o a I no= s-nso- i a.

C' A7 I P -1} It- 5 ^ 21 P.2-\re 3, 9Ct'el or 2 guinea: s=::s., Blew;.

-13 FIG A, shows measuremen, scarfs across a "clean" w-eld and acrcss contaminated titanium welds FIG 49 illustrates the effect of shielding gas contam nation on rhe normalized conductivity of titanium welds 5 FIG 50 illustrates several measurement scans across three engine disic slots, along With nominal detec-,ion thresholds. FIG:1 illustrates Me variation in Me normalized conauc,i,i due to -e

foliation of cracks in endue disk slots.

FIG:2 illustrates the effective relative penneabiliTy ariaion zenith position 10 along lee a, is OF Ale barrel.

FIG 53 illustrates the GRIM measured efIective relative permeability in two regions and possible behavior between the two regions along the axis of a 75 rum diameter partially overheated gun barer.

FIC? 54 illustrates hidden crack detection and sizing ir a nicke',-based alloy 15 sa!,lple5 usmg a two-=reqtlency re;hod 3ErAILE3 D=SC RlPTIOL- C) F TEE,TlIOT A description of preferred embodiments office Invention follows

To safely support life extension LO. awns st=c.Lres and to reduce -.ei_ht and an ainLenancefinspection costs for new so cures requires both rapid and cost u efTecrisre inspection canabiiities in anica; co7un7ao77S monitor rig o, c _ck 1rntiaiion and gl-or requires the penna7ent moon -ng of sensors to 'e comoner.t being monitored and severely lirn,ts the usefulness o calibration or eferer'ce strides, especially --hen placed in ciffic-ul -.o-access locations Cal a- r g or nsv stews. 2; P-.a-er.t and s '-face.loui of coren-;ona1 ed^"-cL Lrvll sel sons is nO-L per orated One reason for this is 'e calibration requ-1 emits for -=e easure;;lenTS and a over is T 7e Arias lily hew een probes. Co leer -tonal. edc -c, Brent Lecrr-iq1=s r5v7 he 7,ra,i:l0 the 7oimirv of-ne sersoT (or ii,-olJ tc me TOSS ILa.e2a' reference pal by -oc.g -ale se..so baa:< and 1o,- o: sing across a s- an- = to art corgu-e tine _,,-_e7 se. logs and disola,; Tic --AT Ale To- Fact ee7,lo 1 if_ --0^ V'77-_z7;is.S e-er.--iit disP; avv'4 as;'O-! __'' due -'-'--'2'7g-i 7 _ --.: À0

-14 left, so that cracks or other material properry variations appear on ine vertical axis.

hoaxing or mounting the sensors attains, a test surface precludes this calibration routine. The probe-to-vrobe variability of cor-entiorAai edaycurent sensors prevents calibrating -filth one sensor and then recornec-. ing the instrumentation to a 5 second (e.g., mowed) sensor rror the test material measurements. Measmred signal responses from no^n,laly identical urges handing inductance variations less than 2% Hattie signal;ariarions Greater ah- _:5% [wild, 1999. These shortcomings are Overcome with spatially periodic bleed eday-current sensors, as deson:bed herein, the; provide absolute property measm-ements and are reproduced reliably using I O micro-fabrication techniques. (::aliDrarions can also be Deformed with duplicate spatially periodic field sensors using We response in air or OVA reference parts prior to

marlin=' the correction the surrace mounted sensor.

The capability to cha^-ac,er,ze fatigue damage in stAructrAraI ma'enals, along wi-h the continuous monitorin, or oracle initiation and growls has been i 5 demonstrated. A no Gel eddy-cent sensor suitable for These measul-em^^nts, the Meandering Winding Magnetometer Foray (57M'M--ray), is described in U.S. Patios. 01:,91, -543,689 &nd5,793,205. The L57Misa'planc,'' CO1,,rvl,lable edd,-c,ent sense- a: was designed to supports qllanuanve and autonomous data intepre,arior! m--ehods These m-3rhods5 called Ed m-easwrer-'ent rho methods. permit cracl; detection on curved surfaces without one use of crack standm-ds, and provide clean itari\e images of absolute eIecmcal groceries (corductivim and pemean liry) and coat-lug thickness Thou reouir,ng filed reference standards vi. e., ca librarian is pello mod in Dairy' away from conduct no -faces). The use of Me Mow- A -ray for ^aiGue mapping and on-Ilne fatigue rJ r-'Gr. itonn==,:.as also been described LGoldrme, 9gv N-' US 63. This inSp=CnQn ca:osb '-try is s-LllLavie F^or ol-, ne, a-igle rests or CO'lDO: rS 2-rlG compiox comooncrl=s, as Neil as fcT rnoITi,D=,,,c of cite,cui,- o-access locations on bo- Nile and co:r-rv 21 a-o,ar..

C 1 to IC _ i1 so, a, e, e s Ode d gvo am; fib HI s sor and i S AN. A A' A VIA A an - - - - T, F1 _ v CS iv 0, c -Lv:1,v,aig_e c._cge East_ Us & l used gee-meows- 0 -r PI server 16; detaiIec desolations off Mice 215 teen III.S T _ A _ _ A _ A A e A, _ 1 V t. Us v 9. _.6v9, _ - 793r05. Tk S_ASV- AIRS a

- - - -13 primary winding I O having extended portions for creating the rnag;etic field and

meanderirAg secondary windiness 12 within the primal winding for sensing Me response The pnrna^Tv heir ding is fabricated in a squa!-e wave pattern with the dimension of the spatial periodicity terrr.ed the spat;. al wav-elenc ( 4 current i, is applied to the pruner-, winding and a voltage v2 is measured at -he tenninals of-the 3-^onda windings The secondary elements are pulled back ^romA:he co-7necr.ing po Lions OI the primarywinding to minimize end efiec' coupling of the magnetic fie'd and a second se; of secondary windings can meander on We opposite side of the primary or dry elements 14 can be placed between the meanders of the primary 10 lo maintain the sew elm of do_ nra_lletic Held, as describer' '- pending applica,lo, Q9/' 29,693. The magnetic vector potential p, educed by the current Tn he prepay can be accurately modeled as a Fomrier series summation of spatial sinusoids, with the dominant mode having the spatial wavelength ( For an A1-AF, ay, The responses from individual or combinations OI the secondary Grindings can be used to 1:, proscribe a P111r I1TJ OT sense si,,als for a single p-,rnlal-'r windi-.g co so act as described in U S. Patent:,733,206.

The IBM structure can be produced using rnicro-l^abricaion Techniques TL7jCa1-Y e1=P1OYed in in.eg.rated Circlet &ICI flexible CilCUit manuactme This results in nighly reliable and highly repeatable like, esserliaily identical) sensors, 20 -airsick has irreverent advantages over the coils used TT7 C^AnNreifiOla1 edCl.)r-CTT-re' sensors Us indicated by Auid arid Boulder, for corven.ional eddy-cvnrent sensors rlorninal',!- ide.'lfical probes have been round o Give signals -ma- hi fer by as Couch as 3%, eNren 'though the probe inductances we,-e identical o better than 7%' Would, 2L399j Ives l_cl: of reproducibihrN; vir.h conNren,ionaI coils. romances seN'e-e 9: requireme;7Ls 1^o. ca:ibrarion of tone sensors (e.g. matched set so-/caLibraLio=' block sv,s). in Contras, duplicate sensor tips have nay identical magnetic field

ismIbutions ZTOL-ld the No T1gS as s.&nQard ric-o-rabica.io Eclats) ecDmiuDes haN7= Moth H' r, sp&21 e,, oducibiliy arid rso7-Lior. A. 5 the SO lSO- N'2S also Gesinea O pi ounce a seat ally He ionic ma=, Chic Bilk in hi- He._ elfin'' aeA Lest 3u <oval). tile senscAes:'c.lse can be accurael.:;:oii_ied --'ci al-ama--vally -educes ca i^-aio-l.eq- evens. Foexile n so.le sir.Iai" ons an "a:, ca7io-E.;ol'' can be -use 7 to measly= abso',u'.e e' apical conGuc..Ni,,;;-i:ol catib--,icrl s,_!aa-as

-16 ich main-es e M sensor gome,ell-suied to surface mo=,ted or embedded applications where calibration requirements will be necessarily relaxed An efficient method for con Y-erting the response of the MWM sensor into material or geometric properties is to use gad measurement rrethods Obese methods map he magn.ude and phase of the sensor impedance into the properties to be dereimined and provide for a real-t,^=e mleasurernen. capability The measureinen. grids are No-dImensional databases- ala. can Toe visualized as "grisly ala. relate two measured pa'neters to wro unknowns, such as tne conducfNriv arid lift-on fNsh-re lift-off is deemed as the proximity of the to the plane of the 10 MVit\f unwindings). For the characterization of coatings or surface layer properties, threedimensional versions of 'the measurerner. grids can be used lternatir,ely, -the surface layer parameters Can be de.elined from ner,cal algo' Iowans that minimize 'e leasf-squares error Derween he meas=--mens and the predicted responses rrom the sensor 1- An adTaDtacre of the mvasuremer End method is *.a i' allows for -eal-!me measurements of the absolute electrical properties of the rnateriaI. The database of The sensor espouses can be generated prior to -e data acquisition on Me part itself, so mat only table iooLup ope, arion, vrilic:r1 is reiaively fast, needs to b peiormed.

FlLrreore, grids can be cr,-nerar.ed for the -ndividual el-rnew S in In array so that 2Q ea h:ndiidual element can be lift-off: corn.raensaed To provide absolute proper meas--ments, such as -e elec.llcal coT'du_-.i ity. - s again -educes the need for e^:rsns;re calibration s andarAs Lo coat as., conve.^tiona1 eddy--clTrP-rt rnetholr, that 1lse empirical _Grelation tables boat Diane -the ampIit;lde arid phase of a h,L-ofL" cold Mensa ed signal to pa-a riveters or rop-;ies of merest, such as crack size or 2: hardness, rea,:lre ex-vasi;e caIibra-ons and =.snm-,nt pr=ra-;.

hIG and rIG 3 iLuslrate -!in-e capabin,r of Rio L<7yl\{ sensor.o Drowse a rnsvasl-e 0,^iCle damage prior to' Toss fo!.aio o,- cracks 3-tecTaole by Decisional =odes_ucnive ins?ertion methods Eio=glass ^ og-bon=-" shaped sue linens -A e-- 5,.?ose LO -A mung-actions G,- Nuclei. rZ ice life al z i=.oNA al'-=g stress - ''e.- We,,7 co lucr-,-:y measured -i--- _3n iTy.; cat,,- Van, his -^o' s.nles,- steel a:, alp-i-l- alloys CQ- ela-es -I ale, -Cle fire action as SCONCE- 7? --2'C- OTT it' Elects c-_la.iN-e --a.iol Gp" parve A 1 I-, -h- x<vr.\

i/ measurements detect fatigue damage at less Clean:0 percent of-the specimen's 1satiDrue '-e. For Type 334 staunless steel specimens, the decrease in effective condTlcvit,v stalls much earlier (which can be a:huted to a change in magnetic perrneabili= due to a gradual Connation of 1narcensite of deformation) and continues to decrease, almost linearly, with increasing fatigue life fraction, as defined by the cycle ratio /F' i.e., (curluiaive cycles)/(cycles to ailurQv). The nor,lineariTy of the damage with cumulative fatigue life for Al- 2024in a iypicaioenng fatigue coupon is well depicted by Ivy measurements illustrated in both rIG A; and FIG 3.

FIG 3 shows the ability o. LvrM sensor Lo detect the spatial aistnbution 0 51A argue dew, age as +,he servo,- alas scan Fled along.+he length of colons exposed to fully reversed bending. These rneasurerr,enrs redcap a pattern of fan Rue damage focused near rye dogbone specimen transition region for both he 70 and Me 90 percent cumulative tile specimens. The minnun1 conductivity at 'he 3 cm point on the specimen the: reached 90 percent o, its atigue life corresponds precisely with 1, -Lhe ioca.ion of z vis;ole crack-. these measllr=-nts were et-en -a-,- a sensor haN'ing a footprint of 1 inch by I inch. The presence OI a damaged region in 'he VlC.-its: of -'e cracl; is indicated Rev the depressed condlctivit,N; near the crack, even when the crack is not under the fooip=. of the sensor. Thus bending faii.g.le produces an area damaged by microcracs prior LO the fo,acion of a dominant macrocracl: and 20 -brat damaged area is detectable IS a siiDcant reduction in -he I'M measured conaucris-ilNr. Photomicrog.-a?i,s have shown a- clusters oT^microvracks, Q.307 To 0.003 Aches deep begin to '-or.1 at it s stage. AlollO,h Alec able with Me 'veal, these ric.ocrack clusters, teemed wide-s,read Latigllv damage (FAD) , were slot Ge.es+LaDie Nv-i-h liquid nenerrant.es i:g, except at We N'VA j edge or Me 90 oeler.4 if e 25 spec men. TO s same behavior As been ovser;'ed Tor Levi measurements on military and commercial aircraft, s actual members.

rIGS ma aTlG I'D provide rw0-2irner,siona'' triages of 'e pleased Co=criN lo; over the 90 pe-c_ l. hi:- rapine specimen Alit in Eke 'vl-\M Go diff=r=-' O-l'._iO=S.!-= '-''S 'Xqv: the v d ^OO=l_-' -was. rlcclesby' t -'at ic2nes. -id Olive 'shoe CX5v'-lded DGt-io.-iS Go -=e -tv.-- r--;= settlers are ore-, _] peenci^- la, o '.ne _-ac s, -e.x,f-\\'' r.as ma..;r- = S5rlsi V, \lT 820,h.m-crocrack - Tic-oc acid -v. - stars G $ - i. \,- _1,e bled All Rio,s Of We. i. Ye_:5

-18 oriented parallel ro the crack7 the MW1d has rninimu l1 sensitivity o e macrocrack and microcrack clusters (FIG fib). The aurecr onal dependence of the sensor response in the fatigue damaged area adjacent to the macro cracl: indicates Thai. the rnicrocracks that form at early stages of fatigue damage are highly Lrecuonal Arnold, in 5 -this case, ar" aligned with the bending moment axis. Similar measurements on conpiox aircraft s7;rv!crural merits have shown sirdiar behavior at e.-l-y stages of lea ig7,le damage, Before -e.ec.abIernacrvclacs amp Remend. -Nor-e that ID microcrack density and size increases are maicated by a larger reduction in tile - 'M absolute conductivity rneasuT'ements. Thus, as e}; pec7,ed, the rrcroc;=k size 10 and density increase near the coupon edges and are 1o7vv-er at the ceT ten Similar nto-dunensional images of Me measured conductiv have been obtained on actual Millard- components. FIGS 5a and 5b show the space scan mapping of fatigue damage on a military aircraft bullhead for MWM windings segments oriented both perpendicular and parallel to Me bending moment ads. One 15 portion of the bullhead was found to eons a localized conductivity excursion characteristic or early stage fatigue rnicrocracking. A conventional eddy-c'==re insDecdon of this area found only discrete -macrocracls. IIG,ever, the width olrt:he area of Me h]vi measured reduced conductivity beyond Me rnacl-ocrack area indicates that there is a region of rnicrocrackhrlg;,n aliLion '0 *le miscue O mac, ocracl;-s.

FaTigu damage Carl aisG c.ez,e variations iT1 Me ravc oemAeabili.v, as indica ed in rIG 6 rror two auseniic stainless steel specimens. One specrnen was 'augue tested vr.hile the other as not. S=rraC" scars:i-.h e 4 ids ordered 2': e. -Ldic''lar find phrailel so 5e 1erig.tl o the c?ec:me:ls show a bl-uirec!Lonal malefic penneabiiit; in The -alighted specimen. =. He -l_eti^ sl sc-pti'il; is urges tile ion cli-x-i5 as the fa.;gue 21-.tS -e Eric Dslc--e ol^+.le sr_ bless s+e l, c;-eat:=g a;nalle+ .ic phase such as 1nalensi,e loo.= the ini;ailv nonmagnetic ma._,-;al. 3C -IONS. a and, _ sac- Lne rats Us o -en Mona ions o- S'=iC5 e=ose sac ores c a Sheik_.37 Fuselage. W =-as,.e._en-.s Beds snag. on the lz:o Join pea. he ass-ge.- -rdo-/s urn c.n -,hrr a;;il pan- s --tic'_ -ha- Alp- -v^. -.u-- pus. iT1' -''n\.'f

detected several areas with subsrantial conductivity variations the: could be identified as areas of wide-spread fatigue damage, i e; extensive fatigue r.nlc,-oc,-ac7,ng. FIG 7a shows a horizontal so n several inches above file top rest ner row- of the lap joint "fine.! measured conductivity has rnini^Tna that correspond consistently with the vertical edge locations OT the windows. Thus, substantial bending fatigue damage was detected by the M-. several inches above 'the la? joint fastener rows The benefit; fatigue coupon dare suggest 'hat this region is beyond 60 percent of its fatigue tile, although it probably does not con.aun nacrocracks which would be detectable with conrvotional diffelent. ial eddy-c=-rent IO rnekods or tTrri'.h liquid pvrlGvtra-.. tes n,g, FILET 7b sho-^rs a Helical scalp down, 'he panel The damage begins near - the botrorn of the windows and Increases steadily, with the maximu-n damage occutr.ng at the fasteners. A key observation from these measurements is that this damage is detectable more than six inches away from the fasteners was rarer srer.fied hat cracks new fasteners ware correlated with 1 1e- - tOrlS O=ledUCed C3'-ldUt-,lil= r71d Scythe 'Intel--A'r 321 ^C'=3vS a= rar Torn Any fasteners Five out of 6N,e iocaons in which ^-c. oc.acs had been doc7'ne77d at fastens s had been in areas similar to those id-vn.fied by the AM detection of List. ibuted damage away from Me fas.enes Ibis ability lo map the spatial extent or the while area fatigue proN.-ides Rio;nTo,a. on hat can be T,sed TO '- rnrOVe he selection OT pecan. loca+ior, and size, thereby?oten.ia7, 1NrimprovirLgherel7abilt;07^-=e epriTs arid ed7lc;.7lg--oilow-o 1ma ntenance cos, s Mine -/-fll r,leasTed condo, ry n-onrahon may also be used to iden7;i-,, specific regions that rec. ire fastener ir.s?e^.ions, as well as to s7=v?or inspection, rma-i.enace scnedultlg and,-design el-; or.s -!-ilis is irnpor.

25 bec 7 7se Me locaicns o'- T' -se areas al - nOT a Nrars -,n;T:;re, since die STrucurai -esponsc is allec.-d by desicl-i "eateries such 25 w-;-do-v add stipple s, lap joins, and doodlers, and by malera:Lce flea, res SJC2^ as patches and repairs in sorrl--irnes --o-esve.. ways.

FIGS Sa -d 8b sov,.-nde --sic7s,-,ar eir--eler=.el7 a an By Con l cLioils are baa-lo each of be eL! ViL, --i s- undo i_=e+s 4 t or use -^'i'-1 Sir CalTvRa-iCIl dills! I' el'-r;7O l 5 27)'t' e Panacea or One o-rsiae 1=a. C=-s 0 ^ lee p.-rr;=y ^o s d-sc^;b -, 7 music A?,-1-,ai-,n 'J-','1'-2 6,ale se 0 blames ---e s

-20 back from the pr mary winding coinmectors 250 and the gap berween fhe leads to the secondary elements are It ir';mized This flexible array can be inserted iron a hole within the gage se' non of a fatigue specimen to monitor crack initiation and initial cracl: propagation or placed flush against a surface -to monitor cracl: propagation.

5 FIG 9 shows an example application of six 7M-ra>s f om FIGS 8a and fib 'v\7ith.v-^.ro,llo==ed i 1side a hoi" Sold tour mounted on 'e adjacent da. side - srAArraves of a fatigue test co-cAon. fine M--fl- BAA-Aays to tried i hi TO 1,' e hale c^^An be used to detect shallow part--through wall cracks (e.G., nuLnelinG cracks that have initiated inside the hole but have not propagated to the outside surface). The 10 MWM-A\rrays can also be placed around the circus-reference of a cylindrical or hyperbolical gage section. Multi-frequency \1 measurements can provide diagnostic armor nation to monitor cracl: propagation;m bosh and depth dil-econs. The /l-A,.ays on the sides are used once a "corner" or d;rough- wall crack (i.e., one that has reached either or bow outer surfaces) forms. Idle crack 15 ten Oh can be in f en red worn the MIMI measu-enA ethic isle conduct v since the Af measured coductiT;ty chance correlates with clacl: length, as shown for example in FIG 17? era en for relarise-: short surface cracks and ror cracks deeper than lilac lF'vl pene-.la;ioI1 depth. The correlation, Libel Zinc h is expected to be even more obust IO. Lhrouh-waIl cracl;s so that a single sensing element \/f DIG may be eased 1or regions outside -e hole as Neil. This,-,vpe Of app'mcatio:a is su*aDle orAnorlitorirg clack propagation with "arable.x,rcles (daidl<r) d=7=cr complex component testily,. For vxampie rnoi.ong of Bride areas (e. g. between sksl an a',-craft component may not be possible o,icallv or vain poteniiai drop methods.

Finis t\7M capab-lirv can provide a new tool to dernonsra-= damage tolerance o 9: S7T7C-n Eves ^^nd es'a-!ish less r,,-densome inspeGTon alad ret -ardent for,imC policies.

Surface mounted Sri'YlYr--ra-, s Cave also ielllor s, a ed an on-:,ne carnal Il -; to rlor,or cl7ulaLile Teatime namace _=in load C,Nicl? rim 10 shows -;'!e plac==ent of ^-- --rat., r-, FIGS Pa and So into a Q.- - Cal cia;=ee hole 3 located is ah- center o a i -inch NX Ike rN; 0 ? Ants thick,(?:,4 l57n SX i]'_ 3; 6.,:-= --in:) SYNC en, rs'e c, an --' -'I-; L-_1<.1 ^-T3,1' cl C. he ha.

s?ecure^ with is genri-ils- nIe-e'5.-2''es 31 5e.--e7 one test secrecy rr'1 the ends Xe,-e fessed,:?de _OnS{^A' cyclic sess A- rIilAe -a tension; l at no. fine

- - l -21 cetal hole represents an elastic stress concentration factor of 2.4. The i-Array had eight sensing elements (1 tram by 9.5 mm n area) located at 1-rnm ncrernents along the array len=+,h A, the periodic dile_tiorl.,'; of the eight elements were mounted in contact with the internal cylindrical surface OI Lhe hole Mile the two outermost elements were intentional:; outside the hole. The fixtu.-e 36 holds the MWI-.rray inside the hole and the probe electronics 39,&r aping and mul.iple>;mg the measured signals to allow continuous monitoring th,-oughout the test. Several specimens were run to failuie lo determine the response + hroucrhout the fatigue life i.e., from cracl::litiarion to failure, while fatigue tests of other 10 specimens were stopped al various stages of c, aC I'iriat,on ar.= propagation as il,ust-ated or example in FIGS 11 '-o-h 1.

FIGS 11a, 1 lb, Ma, and 12b show Me MAIM measurements during a fatigue test. The Gird element channel failed in this furst test so the dare for the -bird element is not provided. FIGS 11a and 11 D show -the absolute electrical cc,nucTiy measurerne-ts for each ele-lea of the:7ivr.iyI-A-ay. FIG- l la shoVYs -=e conductivity as a 1!luc,ion of Ike r,lmber of fa+. iue cycles 1or each elemerit r,x.rhile -RIG 1 lb shows +,he concucniLy as a function of the elerrlerlt position across the iclmess of the doled hole for several fatigue levels. I he p,-onoanced decrease in conaucLivi at around 2,000 cycles indicates crack nitiaLior. The crack appears to 90 irate near:Eleert 2, as rants was -,he Best cl-rnent -,o exhibit a decrease in He conduct ity. The crack then apuicli; propagates Lo -one edge al Elernelnt 1 and -hen gradually propagates to the oaths, edge Enf i is ae,ected be; Element. 5. 1-nis pant liter test was stopped when Element o began to detect the crack. Upon an examination -1- E-1 optical rnicroscoe Et maica!ioO oft 100 limes, no crack w as apparent on 2: the s7e -face near Element 6.

rIrS i2a and 12D show.ne,--o l Easements:or each element of the 1-- -12: using a -La-7T-Or,-:L Progeny rnG eel. DIG lea s rows he Ii5-o:^ as a -cio- of hen=lbervTfa-.i eccles ioreaciele.ert---ile DIG 12v s, ows-he limo crib as a r'=lcion or the elernel),- 70,itio n across -e 'earth of-me clirl ca' bole 30 or sever ra fa ice levels. The,riiai decrease a-lG i \Tenv- ol^ Ale il -o Strata dl roar he ilrlal esi;lg, ess -churl l,CCO cvcivs) ill- sea s me se. Lo-' o'^ He 1vP as One s nsx acjls.s to -e sly ace. he c-ease o-::r!e e:-ev.l,v li --G:,^-,= iA^=-r

it j -2 stage testing shows ibe effect of the o?ening of the cracI:. Al trough this lift-off data shows that the uniform property mode! can represent the crack, improved models of crack interactions with spatially periodic field sensors should enhance crack

detection sensitivity and also provide depth measurements. Also, morutoring of "effective ft-off" signals using the EM- Gray for deep cracks (over 0.1 inches) Provides information about We "compliance'' of large cracks and may be useful 1-or crack depth estimates.

rune ability to con,nousIy monitor fatigue specimens Chile being loaded provides a capability to create sampIes with veldt early stage fatigue damage. FIG-S 10 13a sand 13b show the response OI an. MM-ray inside a Al 2024 fatigue test specimen and provide an image of the crack initiation sand.=, ouch as a function of fatigue cycles and position. In this case the specimen was removed *om the test after the decrease in EM measured conductivity indicated the formation of a sizable crack at one location within the hole (Element 2) and the possibiIit: of 1- mirvrocracking at mul.lipie locations along the zips of the hole (elements 1 and a).

MetalIo,graphy performed on -is specimen after scaring electron microscopy (SEM) identified a crack near Element about o.ox inches deep and substantially smaller cracks fu7.er away from Ei7er.ent 2. The SEM examination of the area monitored with me EWE- Army revealed mul-i-site darnaoe with predon TnantIy 90 axial cracks ranch G Troy 0.004 inches to over 1/16 inch in leafy. Adjacent to -the sizable crawl- detected by the 1\I, the SEM e xa7 niT:at7 c? red ealed a ser. es Ol" :n+sions parallel to.h_ crack and normal To -e machining marks thorn renting.

These intrusions raid it be associated With persstn+ slip bands (P8B). The 7n;,orm -educ+ion i 7 absolute conductivity across the sex sensing elements as Me fate 25 comport warms lip (why increase load cycles) is ais.mguishable fro m the local re--cons conductivity by maivicnai Ie news En aIIows for compensation or the ted permit_,a-ia+ions am-7i?g Ale measl7reme. 1 he^ocoUple3^.heiso s or o*er ten?erare monitor mo methods con be used for tips emperacre CO1-leCtion.

rT-S i Ma, 1 4'U, 1 AC, lea, IS3, and I:c s''o-:e roaiized electTica 0 Con-L}CaVi-LieS ^07 severallmore aieTes7; specimens. Specimen.-5xasa707 alma - =n allo--hiI specimens.37 and.'.'3. --_ 709A aIlos. LO or=.= to hem defy _ e _ e A_ vshold 0 _ v Stir 1 al 1a i Sue d_ age,,.1es v t s s y en e s. ? d z

-) -23 different levels or conduch-r reductions. the case of Spervhmen =32, the fatigue test was stopped when the VM conducvi drop (reiaLive to Me "background"

level at neighboring channels) at Channels 92 and 3 There considered Tndicat ve of either microcrack formation or advanced stases of fade damage accumulation 5 prior to formation of microcracks These samples were examined thoroughly with an S EM by scanning the surface of file hole a. magnifications up to 1,000- across the entirarea-nrorliro-red Or the fatigue fests wi-u:-MWM:A-As. A number of areas were examined at higher mancations up -o 1 O,OOOX. The S=M examinations are extremely time consuming, since one anus cover substan,Tal 10 surface area looking for cracks on the order of O.OOo inches arid smaller Since She cracks for each of these specimens did not reach the outside surface of the component, it appears that the rnoDitoring capabiIify with the MWM-Array allows tests to be stopped with various crack sizes Milan 'he hole a-d pa- ticuIarly at Bayous early stages of "pre-crack" accumTaIated "atigue damage, during Me "short 15 cr_ct' grov,,h stage US well as -;ng "long cracks' growth stage.

SEM examinations cordoned the presence and locations of cracks in the specimens SEM examinanons of Specimen 34,e;eaIed a few microcracs, ranging from O. jO0A to 0 0036 inches (10 To 90 (in) on the surface o ^the hole monitored by Amp. -he 0 0336 inch long inteltten+ crack w as in the area 20 rnoitored byEi_rnens 3 End 4 of'he 1\I A clack in This location is consistent wi'Lhellresponseo FIC-Si4bandl5b. exarinanonofSpecirner1 '34 by an N3 Level 3 nsoec+,or, using a vex- sensi+ .i-re conrenlGnal eday-c--lent probe, did Do+. reveal any crack-iil e ir; dicz.ions in the area monitored by Me _\'j34--ray dating the 1zT:gue rest Floev-v., the edy-rL!e examination 9: leiec.ed small crac;-Ie inaicaLiorls on Me opposite side of the hole net was 'not moo co-ed by the 'I;, hi S.7 nd ng Act ovides an addit;.cIlaI coi=natio r.a. mic- ocacks 1OT de.ectabl,e byz traditional _dav-c,Tent method DUE de.ectale and detected by M]vI sensor shoed nave existed on.:h" side alto i-o ed by -he -lay. _A 7=,er creD:';7y v-os-sQvc' odd= tile sp_cirren to -the osinon o -the 30 O.QQ35 inch crac'<, e-,ttia-,oi-s owe crack r^-rea 7r'2i PI! o=Lica1 r-iciroScope aL se--elal ^ac=caLion Isle s lie i led Ale presence c-' Me crack evailoaphy --tea tha Lee clack a=-Lh --=s aarcl72Nr:,n,. c,es;: (-a). l=.ila S=i/{

J -24 examinations performed on Specimen,.'5 indicated two cracks, vhich is consisren+.

with 'he MWM data of FIG 1:a SEM examinations of Specimen 32 revealed a few cracks ranging in length from O.Q005 to 0.006 inches (12 to 10 (m), with two distinct cracks that were less than O.G02 inches long. The longest detected crack was 5 intermittent, i.e., consisted of a few adjacent continuous cracks. Assuming a s-:nicrrcuiar geometry for tile cracks, the estimated depth of individual cont nuous cracks rangring7lengfh-frorn 0.000: tO 0.0074 i-nc-hes t-12 to 60 (-ret) would-be between! O.OOQ2: and O.OG125 inches (6 and 30 (m).

FIG 17 slrma sizes the results on 'e tested specimens in terms of crack 1Q length compared to the hum rneasTired data The dare for specimens.'. '32 and =34 are difficult to analyze because there are rnuliple crack indications and -he longer cracks (e.g., the 0.005 inch long crack in spvcirnen r'39) appear to be Intermittent (i.e., Torrid from several shorter cracks). Frtherlnore' the depth of perpetration or file EM maedc fields at 1 MHz is on',he order of 0.003 inches so Hat cracks

15 skallow---r than 0.003 Riches will produce a A1 concTlcnmy dependence 'cased on depth as well as length. For these crac'-s; a higher frequency measurement (e.g. 6 o,- I O Mud) is expected TO provide a more reifiable meas are of crack led Ah as well as a better signal to noise for improved sesi-,ivi to nicroc,acI: detecuorl. lVIaI-ple frequency rnleasmements should then silo IOr esuma4.g crack propagation in both 0 length and depth directions. The reliable defection of Me obese. of -ridge damage and He number of

cycles to crack initiation, 5 C^n be per owned aomaficaIly using trend detection algorithms. BAA Ample de action algorl-Ll,,ll is LO use a simple hypothesis test to build a first set o,staTistics (e.g srardard deviations) Ior He no damage 9: conduct sty data at fire beg o -the test and also a second set OI S:arisucs for a Cooing window of most -ecen. data. Penis combing of data is ilInsT,ated in FIG 16 for An e>;am.ple corldilotIN;Tty vafia-Lio,, with nllrnber of fatigue cycles. The data must trot be collected for tneImaI At, ei-hLe by- using eocollries or by I kerir g the (nea-Iy ITTlear) tmperafT're 'rend = -e damage related conductivity changes vs. 0 nTlTb=r 0 f _a.ig,ie cycles data. 4 s Me hy No hesis test 1ligh'requTre -Lila. L3le vFir.34 corriactiw; change D' al leas =:ice the s m o'^the san&d de-iaior of he Rio Da,a=e rV76r Data -- ld He ost Recen.vA DaTa. 4= au Named As

(: -2: would determine the confidence leYeI of the statement that "the most recent da a

shows a conductivif drop not related to mend re;nperahre changes, compared to the eerier no damage data 7' The confidence level AT depend on the sf,±sticaI separation of the,.;o sets of data Sibyl&- techniques are commonly used to detect downward trends in noisy data, such as the stock market An a-ornated test is an improY?rnenf over the huma interpretation or visual data as human operators t,vpicaDy have an expectation of results based on prior knowledge of the coupon material or expected number OI cycles to initiation, that can influence the results Another aspect of the invention described here relates to 7mique georne+. ries 10 for fatigue specimens the';!ter,ionally shape f.';,e stress disfn:bu+ior so That the damage iritia+.ion sites Fill lie within the area Strider inspection by a surface rr.ounted eddy-curr_nt sensor With a traditional dogbone design, fatigue damage starts in the Diddle of The specimen b7lt is not localized along Me length of the samples Thus, - there is no 2, u7-antee +ha! Thfafi=Te dPrnage will ini=-e beneath he surface mounted sensor The new specimen geometries described here, and illustrated in FIGS' 18, 19, and 20, localize fatigue damage both Ienwise to ensure it occurs in The reduced center section of The specimen 30 and in the middI- of the reduced Thickness center section in order to avoid cracks at the edges of Me gage section Irene lens localization GO is accomplished by 'Lt;;rig across the centerporhon ofthe ecen 30I Reduction of The formation o f cracks at the edges is accomplished win reirorcement ribs along the edges 302 andior with sy,neical radius clltoms 303 on bodh sides of the specimen, above and below the gage section FIG 18 shows a dOgDOne S,DeCirneD 300 is tl:t thong at Me center section of'he spxirrln 301 and 7: reinforcemenr ribs 302. We rninning a. -Lhe center section can also be accomDsned -IBM carton, secuo;ls or each side in order to avoid bending mmen.s FIG 19 shows a do bone svecrlerl 300 w-lih hinnincr at - the cenre' OI the s?ecnnen 3ul -\r;-h.-adiu3 c repours 30_ C 1 Doth sides of me ir.;rrned seer7on. T'TG 20 shows a dogbone specimen 00 with.-i n at one cents section 3G1 and both rein ^o-c-lme:lt ribs 302 and 0 ^-aaIu3 curious 303 Each of these designs sii,lca-.iyreduces the stresses at the - ares and rh ere-D >, Exerts 7iciaion Of fatigue I Image fit tile edges in -hi- early s.a_es OI Rena CUe.

/ ''N -26 FIGS 21 through 41 illustrate new embociments for the 1-Array sensor structure and applications ofthese st=ctrres. These embodunents provide greater sensitivity to the flaws being investigated and can be applied to both surface mounting on and scanning across rest materials.

FIGS la arid 2 lb show a sample configuration for the detection of cracks near ras,eners wi^LIl MIMI se floors mounted on the surface. A steel fastener 42 is attached to the fa-.- Me rest veiny on 4O 0 ^Ai ?-924 at- a semiircu-lar notch. The-

- mounting bracket 44 holds the WM sensor against the surface of the rest coupon throucrbout the duration orthe tension-tension fatigue rest. The electronics package 1 O 6 provides signal amplincation of Me sensing elements in the 1'M sensor' as necessary. EM sensors can be permanently mounted at fasteners in aiincult-to-access locations and elsewhere.

FIG 22 iIIusra.es the positioning of an MWM sensor 16 near the hole 6 used for a steel fastener 57. A crack 61 formed beneath the fastener as a result OI -Lhe 1: tension Ia:igue load cyst on the test colon of rIC-S 21a and Blb. The cIack 61 originally -mated at the notch of the hole beneath the head of the fastener and was detected when it extruded approximately 0.070 inches (1.75 Starr,) beyond the edge of-le fasene- head 6. How ever, this crack propagated only O.G20 inches -Leader the iAoo+prlt of the sensor array defined bye region covered by the active sensing 20 element, as iliustra.ed in -DIG 22. Tale signal rneasurvd by the Maid\<, and hence the effective conducive+ and mA, o A, -neared by the sensor, viI1 change as he crack propagates across the sensing elernerl.s 7 8. Orienting the sensor so that the extended potions of the At naggings are perp_ndicil 'to the crack provides ma,imn,At sensitiNri- to The presence office crack, as illustrated, FIG 4a. The earliest 95 deteCf.lon orAthe crack OC^-LrS as the crack! p arp-oacnes the position 0 A hi_ nd-AmAoAt sensing element. TS suggests mar it is desirable o locate -one tats. sensing eleme7., (as opposed to a dummy Lien, denoted by 14 in FIG 1) as close as possible to to_ ed e OlAl.he,,ilA.a;Anfn.Al CAM r T,, À ' Al n a, an g A ry Ag mea AdeA s. I ho Ago eI_AiAAati Ag he d y ele ent or The edge He -'nilueAAce -he Amelia- to pa-own an -ir calib-a;o= mAeas1 rer-eAr, it o0 can p-o--.de al-! earl =: innciol1 o ^ Me presence 0^2 clack benea k Lye fastener.

F; rDnem?ore. although ells eve 7 sensor does Cot locate +,ne position of 'he crack along a meander he len--Lh off he crack can b- es+,.aTe o-rAlee auction n -=e

f - - / -

elec,ive conduced- as the crack propagates across each mdi unideal secondar,v-

elenent. FIG 23 iIlustraes an alternative embodiment for an MWM- Array. This linear sensing M'M-A:ray has a pr-na winding 52 for creating 2 spadalIy 5 penocic magnetic field for inie1 - 1ogating the MUG and a plurality of seconder>'

elements 54 along the length of each meander. The primary winding j2 is sphit into Go palms' who lead co':nec-no-ns 66 and 68 on either side OI Me sensor. This configaration for the plmarv winding -uses Tao conducting loops to impose a spatially periodic magnetic Klein, sirlIar to the single loop meandering winding 10 I O of FIG I. This conuTurat on an Ionizes 5 e Elects or spay magnetic fields Ohm thee

lead connections to the primary Indian, winch can create an extraneous large inductive loop that influences Me measurements, maintains the meandering winding pattern Ior the primary, and eectiveIy doubles fee current through the extended portions of the meanders, as wiI1 be discussed with reference to FIGS So, 3 7, and 40.

I Secondaw elements that couple to the saLle direction of ibe magnetic, mid generated by me privacy winding, such as elemerl.s 54 and:6, are connected with connections 7O, perpendicular to the plenary winding meander direction, so 'at the sum of the secondary- eiement-esDonses appears at the winding leads 64.

À To provide complete cove. age.r-rrhen the sensor is scar,,led across a pay or 20 when a crack pro,aates across the sensor, perDendicu1a' -,o the extended pomons of Me pilot -WIIlOLng secondary eIements SO in adjacent meanders o f Lhe pi EMIT are o= set alG^.g file ienGrh of "he meander. The dI7T=myr eIemeI:ts 60 ale used to manta n he periodic S-Nnetr of the magnetic fieId and the extension elements of are used to 1-n77=e ferOllOC3 Hal the co,apng of the Anemic,,eld to Me r=o.:s So sensing eIe'Tneln.s, as described ir1 Patent AppIica+ion 03/1 8.6r3. Additional primates Nriindig meander loops, N,rhi only contain dandy elements, cart also be ?7 aced a. Me edges of the se:lso1 to help 1main+ain:-he piodc'.; v OI Me magnetic À ield ^o1 me sense: cr elevens:lees. The sensor ad Yes. The secondaryelernerl7s =-e set back." Furors -=e los-r or,, e_-,ion par ions 5 OF the plenary- art LlOiDg To Pandas to 3 Resize end e.-:ec,rs of he Teas=-vme Is.

he naectio-1 1-ads 6 to -ah" secondary eiemen-.s ale pelLledicnIm- to the -ey- w-n-m, mele-s, -,-ni., i Creates a!' shape and n_cess care,,h_ -us_ 02

- r,Q - no -

muIti-layer skuctu.-e m fa,Dricang the sensor. The sensor of FIG 23 has Me layer contairr ng die prim&y winding 52 separated Worn a isyer containing the secondary wind;mgs by a layer of nsulatT on. GeneraIIy, layers Insulation are also apphied to the top and boredom surfaces of the sensor to electrically insulate the primary Aid 5 secondary windings from the MUI. An of the leads to the secondary elements can also be reached from one side of the sensor. In contrast he basic sensor geometry of FIG 1 has a single layer structure and connections to sevonday elements, Men placed on opposite sides of the primary wmding meanders, requr-e access to hot:h sides of the sensor.

O, - advantage 0,r I've sense,- 0 I 'Iffy 23a End 23b ove the se.n.sor geome±y Or FIG 1 is rhatit can de+.ect cracks and determine the crack location iibn the footprint o, the sensor. When a crack propagates perpendicular to the primary Owning meander direction, only the secondary elements directly over the crack v;;11 sense a significant charge in signal or reduction in e.ffecu To conductivity. As Me 15 c.r_rk continues to propagate, the signal om addir anal secondary eiemens will be affected. principle, file cracl: length can be date milked from the reduction in effe rive conduct vim. En contrast, the secondary elements 17 of FIG 1 span the length of rhepT=Tnary winding and cannot distinguish Me crack position along the Pariah of Llernemlde.

20 rI&S 94a and Fib shover a circularly syretric ernDodient OT an -Aciay. Ibis Mvv14-Rosette orpeiodic field eddy-c,Trent-rosette

<PC-osette) maintains the spatial periodicals of the magnetic field in line radial

direction with primeval' winding 82. The characeris,ic dimension for this radial spatial pelioGicl- is file scalar wavelength. Th p:'rafi of secondary elements 84, 2;' 86' and >v8 provide complete cove' age around Ike circuTnference of the sensor and can He used To detect cracks aria cede the crack location. The 82p Og 'between -Lhe priorly r,ndiI7g coIlalic-ors 5 and 87 is minirrlized to reduce any sway Tna=edc fields CON. a TecTrg me. easLlremets..T-'IGS,7a =d 27' snorer a

c-,, cola- y sly Crier c var anion ol a s,Tdard Ivrw-.ar. As --A.'i-Ln P G-S 2,ra arid 30 2 Abe.lle p Ma -; -bird rig 90 na:-airls he spac.ai pe ionic 0,e rnagneic. eld me radial dire_-. on. Tile secondary elements g2, 9$ 96, Cold 93 provide complete co.relage aroltad rile circm=!fererce olive se:lc.cr a d can be used to de,ec. c.acL's

- J -is and determine the crack length. The first active sensing, (secondary) e.le-:nerlt is located as close as possible to the inside of me sensor to enable early detect on of cracks. The primary wind, ng 90 is Fabricated onto one side of a substrate Al while the secondary elements 92, 94, 36 and 98 are fabricated onto the opposite side of 5 'e substrate. Ir dividual connections 93 are made to each of the secondary elements for independent rneasureme=.ts of the response of each eIernent. Ite. rnatively, the net signal from all of the elements can be obtained by cornectir g the loops together.

The rosette confi,c,ura'.ion is most useful for crack de.ec'ion arid location around circularly symmetric regions, such as around fasteners. The rosette iO coriuraion can also be used in areas where the stress distr Caution and Ale crack initiation point and growth direction may not be known because of complex component geometry or service related repairs The \ A-ay comr-igranons of FIGS 23a, 24a, and 27a car be surface recounted on a paw, as has been demo?sared for 'tale standard M31i'M arid :};-ArrayofFIGS 1, Sa, and 8b. This mounting can take the fonnof a clamp or pressure Ming against the surface or Me sensors can be mounted with an adhesive arid covered Filth a sealant. Since Me f sensors do no, require all intimate mecharical bond, compliant adhesives can be used to improve durability.

The li7v1 sensors embodied FIC-S 1, Sa, 23a, 2/!a 27a5 USA, 39a5 46 and 20 =7 can also be packaged on a roil of a Inesive tape. Indi7iduai lengths of me tape may be clot to meet the length requirements of park cellar application For exa-nple, a single sip of tape containing numerous LWM-Rosettes may be placed along a row of. ast--ners relanve?v rapidly. Electrical connections can toe -' ade to bor d cads fo the indi-vid-, al sensors or Promos of sensors. 'When mounts?- against a s-, rface, the 7: adhesive can be provided along one shoelace of the siphon rig membrar e to bond the selected Ien1 of the sensor a--ay to a Part to be rested. When momrl,ed bungee layers, the awl- esive should be =ro-id--d aio g born -'e untie' and lowe exposed S?,tTac=s. Ice sensois c - a so - -Tin- do ad e..s= last-,s ON a s ruc.?ar, such as 30 ?oe./een ayersofz!'p,oito nd--rrepai-sus-r=Gro posiesormealdoublers, r SS2-IT Or ' ' ah__ 'I_ _ (Jo, -. __ _, ___ e __ MICA 7 -\' T UStlaTe _A_.:je crcss-sec-ioal view c.- RIG '): -for MINI- Arrays 266 eroded

( - -30 the sealant 252 bvhteen structural panels 26G and around a fastener 764. I: also follows that the rosette configurations can He formed into 'smart' washers that can be placed a;.rec+ly beneath file heads o i - fasteners. This is illustrated in the cross-sectional view OI FIG 25 for an MWM-Rosette 272 placed between the head or- a fastener 270 and a structural panel 260 The sealant 262 may be placed between the structural peels between the M7M-Rosette and the fastener Leads or over he entire Fastener head Since processing of the measured responses through the measurement gods provides the capability for each sensing element to be individually I-fit-off 1 O cormpensA^ted and access.o each elerne.t is not required for calibra, on, She sensor can be covered win a top coat of sealant to provide protection Coma any hazardous environments Furnnore, the sensor can intentionally be set of Fa surface, or fabricated with a porous (or liberally perforated) substrate martial' -to avoid or minimize interference with Me environment causing 'he corrosion process to occur 1on the surface and to Drowse conk ^-.!o mo- toeing and nspec.lon JO s Ross corrosion cracking or corrosion r^atizue.

FI(? 2v iIlust,-ates an example cons Oration m Chick two closely spaced MWM-Rosetes 97 are placed around two fasteners 9 The fasteners are also nea- a comer flog 1 Oi. INS iS meant lo illustrate that Me rosettes ca 1 operate when next 20 tooreano+=er,nd.hey can bedriseneiibers ul:aneo sl' o sequer..-ally T he wi lding patterns for the pri-m-aries help cancel the magnetic fields outside the

IOOtpr't OI' each sensor so treat -the cross-coup) ng 0 fields berweAn rosettes is

minimal. A dinted architecture Carl be used for the electrical connections to each O Or Me rosettes T. He eIec ronics 103 call be distributed so -that each Rosetta has 95 independent amp:ificciiorl 2nA connection cables. Alerna;rely, multiplexing o' parallel process-- O Ir each OI VIA i,ldividuai se:lsir eiemercs, as appropriate, ca reduce the nurnoer of independerlr amplifiers and cables. sine eleciro lies can be iocarec rear the sensing elements or at file opposite end orrhe cormec flog cables, liar o n The sensing "lern--nts, as Lne^essar. Id addict, Me eleccics car also be 0 made A at gild rleibie for el.lbeain ir! Me sincere so a, relative y ^ew signal zinc,.ovVe- i-ne conLle=-iolns a e required for each ros Le.

f - - / I These configurations, particularly when applied in a surface recount application provide new capabilities for fatigue damage moritor ng. For example there is a stated requirement in both military and commercial sectors to more accurately determine the number of cries to crack T'itia-,ion, N;, in fatigue test 5 coupons and component tests. For coupons, this is necessary' to determine the fatigue behavior of new alloys and to Mali roducrlon rags for materials used in aircraft structures. For fatigue tests of complex structures, determination of both the number of cycles to crack initiation-and-moritorinc of crack propagation and crack-

propagation rates, da/dN (depth vs. cycles) and dI/dN (length vs. cycles), is required 10 and would provide essential reformation for both aging aircraft rnanaernPvnt and newer aircraft design and monincation. when cracks initiate in diincuI,-to-access locations, however, crack propagation hates car no. be determined during fatigue testing. Thus, e ther costly disassembly is required during r^atigue tests, or very conservative damage olernce-based inspection scheduling for n=serrlce aratc 1: will re-.llt. Surface molontg or We sensors substani-aiIy reduces the disassembly requirement and allows for more periodic inspections.

-:FIC} 29 shod s an alternative enooaament for a sensor 212 having a prunary winding 21$ and a pluralizer of sensing elements 216 mounted onto a common substrate 213. The sensing elements 218 o the sensing eiemeTlts 216 on one side, zO -Loose in the ch;A-nneis opening to the bottom of FIG 29, are smaller sensed elements.

The sending elements 218 are offset, startling at the top on the left of rIG 79. The offset is perpendicular to me scan direction to support image outlying ol^the "crack" response. The staggering of the secondary; positions provides -or complete coreracre when the sensor is scanned over the tJT a direct on DeroendicuIar to flue primmer 75 meanders. divianal cormections LO each o,;e staggered secondary eIernen+s 716 also support Lhe construction of images of the measured properties. Flongated extrusions 226 to the secondary elenels (294) caI1 Help to:rmn Lze variations in he parasitic So Splint between tnG Pn7na)r old -file secondary elenGnrs]:?y el or Gv 1,_5 __2 ci a.s v;, v ad do _ TO - _ - _ _A rr IA v_ 0 ?-..,ean s 2S.Z_;_1n _ O ale t c t n n A,; A _ P _ A _ Ppli lo.!lv_ 69. he v.vrnvTl.s 9 on Theo pus Le sideo he a _ _ r A _ P 1 A 1 \ r A = P -= Dzcro -=d pro?er.lGvs o,^rhe mate-lal whic'1 CA- COnA-lIe-r1e.- the Pi rI:Le Peso' at on

J _ property unage obtained from the smaller sensing elements. FIG.6 and FIG 47 show two additional embodiments for linear sensor arrays where a single prwary winding creates the Lposed magnetic field and individual connections are made to

each secondary element m the array.

5 FIG 30 shows a schematic for a multilayer sensor array that provides high unagring resolution and high sensitii y ro hidden macrocracks and distributed microcracks. This deep penetration array desi is suitable for the detection of hidden fatigue damage al depths mole 'an 0.1 inches. The sensor array contains a smile primary winding 104 and an aray- of secondary or sensing elements designed 10 for absoI:;te 106 or dime Dial iu8 measuLererl.s as described below critic respect to FIGS 31 and 32. this tapered PHI- Array current flout Slough -one primary wielding creates a spatially periodic ma raeuc held that can be accura.eIy modeled The voltage induced in We secondary elements by the magnetic field is related to He

physical properties and probity to the MUT. Except for 5 Most sens(r 15 elements, two se-:-'g elements are located w-'diin each meander of the primary Ending. He absolute ale Bents are onset in Oh- x d reckon from o=cner absolute elements to provide an overlap and complete coverage of one MIJT when the a-ray is scanned in Me y direction. SirniIarly the difTerv,iaI ele:Tlents am offset from one another to also provide complete coverage.

29 This sensor also uses a single p -cam 'iYi riding -Lana. extends beyond the sensing elc;ner-s in f he x and y directions. This has the specinc advantages of emanating One problem of cross-co-apiing between individually driven sousing elements and, educing parasitic enects at the edges or the sensor. These parasit c e -ec,s Are All Her reduced by Me Tntroductior1 OF passive, dummy elements that 9: maintain the pe-oicity The sensor geometry. These elernP-rLts me illustrated n rIG- JIG in the end meanders i 10 aTld wi En ':ne meanders cona=g the sensing elernenTs 119.

Fieore, He A;stnve berg een he sensing elements aloud tulle p==nary, (dike) winding is large enough to,,- We copping 0 ^shon s?C:ii wax clench 30 magnetic 1leld modes. As result, Lrle sousing Sterner,-esorse 725 pi_ nar Iy =-,;ve to the doi-lent Periodic m o'-5 l,nis Droti ICES llOT depth of sensi-ivitytotilep-ope.iesoan /-lJ.

rat -3 lhe design of +Lhe sensor in IG 30 aIso mi umizes anferences in coupling or tide magnetic field to the sensing elements. Idle taper of the primary winding in the

y Affection maintains file distance between the sensing elements and the edge segTne.nts of Me puma winding I 1 A and 115. This also effectively balances The 5 Mingling field coupling to the electrical leads 118 for comnectil-c' to the sensing

elements. These lea;ls are kept close togeler to raze Deluging field coupIing.

The leads for the primary winding l 9.0 are kept close together to miTrnize the creation of flinging- shields. The bond pads 12 ar d 124 provide the capability for connecting the sensor to a mounting:;ture. The trace widths for the prelacy i O baaing can also be increased to nnTnTmi7e ohmic heaTirlg, piculariy for large pere"ar-.on depths that require low frequency arid hi rh current ampIifude excitations. In order to maintain the syrrnefr for the sensing eIernenfs, mtipIe layers are required for the wmding pattelms. LO FIG 3 Q the Or many winding Is fabricated 1: on one side or^aln electrical insuIa,or iO: wee the secondaries are deposited onto the opposite side Gl-the insulator. Mine rbree-iayer strucurare is then sandwiched between two additional layers of i;:sula,ion, with adhesives bondir g -e layers together. This deposition can be performed using standard microlabricaorl echiq-ues. Tale.:nsuIation used for the layers flay depend upon the application.

90 For corlfola7-Ie sensors, the insulating laye,-s can be a exibIe marer, al such as Kapton -, a polyimide available horn, E. I. DuPont de lemons Company, while for high remperarure applications the insulatin, layers can be a c-.a=c such as alumina.

Although the use of lmulvIayer ser.so.-s and senso. always is widespread in the lirera.-e. one unique approach here is the oiise. combna.rion of absolute and 9: d;.,e, eri ial eIemen-s vY,irnin a meandering windung stare that provides a spatially periodic irn?osed 'nawe.ic field and has been desimed to minimize!TT7rnoeISd

prasi'.ic effects. Spec:c ac;-anrages of As design ale -at (1) it zIlo,s col-nple-e coverage -bairn o,n ypes OI S_SI:1C' _1=._U ' w-:nen Me array is sca.rm_d over an Am, (2N ''-nl -espouse of Me in-iia.7a leEUe;IS can De ac_--ra sly modeled, An al10wing Juan i-aT lie easlre-tents vie -muproDe.-ties arid p, ox=ity, arid (3) i ro\2des -iC-=SS=l ^'-^?-t AT -I -','- =4 pa lo, Anise.. Cam..-,7 up -caches of me use o'n'ero -is seris.rg eiemiems vixen each Jeanne- or z primp,

- -4 winding, the design of FIG 30 illustrares how the Iayout of the piunary and secondary;, windings can provide improved measurement sensitivity.

FIG 31 shows an expanded View of one of the absolute sensing elements 106. Electrical connections to the sensing loop are made through the leads 130 and 5 the bond pads 122. The durnray elements 132 maintain the periodicity of the Ending structures and reduce element to elernen, variability. The distance between e-p'imary winding senerl,s 13 A- and the secondary win-ding s-e=ne-nJs 13-6 can Adjusted to mprove measurerrenf se Activity, as described in Patent Application O9il 82,693. It is particularly advantageous to have this distance as large as possible 10 wh n attempting to detect deep defects; far from the SU1-face. huh each awesome sensing element independent of the response of the other elements, the measured signal can be processed with measurement grids, as described in 'uS Patent 5,343 689' to independently measure the local mater al propelled, and proxirnty to the MET. The measured properties from each absolute sensing element can then be 1: co nbined rogethe to provide a rvso-Tnensional napoing of the marelilal properties.

FIG 39 shows an expanded view of into dinrenriaI sensing elements 140 placed adjacent to one another, between two pray winnings 142. Each dfi^erericial element includes Two sensing coils 1 ^ A, Rich assocatea correction leads 1 4. 6. The meander ng p^-ttem oriole leads provides essen+.ially Me same coupling 20 areas and nerds ac oss the sensing region between he sensing coils. Dnnrn.,r elemeTlis 148 are placed on Me sides and between the pairs of diferenal coils closest to the carte- of the sensor the x direction to Her Atomize any Of erP-nces between He coils. By maintaining the symmetry erween the coils and the sensing leads, He coil differences can be aken al the bond pads 124 or with 2 electro tics =xel-nai to tale so-so- i,seif. S;rnilar to =e absolute coils, 'e gap spacing between the primary w--nd;=gs and -:he second; cod can be adjusted and op,izea ror a par - icllar neasur=.-n. application. v;he scarred in he 3, ci;rection, the onset o ^rh_se Iements in He x di--ion provides use capabiiir Tor creating a two-:ne-sioaI mapping oldie d- le. Qltial response, --''ch indicates local 30 -vanl!ics -A the math proposes sold p cut-.

FIG 33 shores an al.er:c-. -rr- or en a ion rid he dir--eeLial set el me, s 14OSe.r_Pnthe?rr2r--^ri-lcrs 2, 42. T7, h;5 case, -.h ci--idlal--naIce 1'4., 0

f - (- -35 the sensinG el-rnerts are piaced symmetricalIy on opposi e sides of the cenrerIine between t7ne plenary Windings and perpendicular to tibe extended portions of the pr. many Grindings. 1;, thus orientation the d.-ereniiaI response is parallel to the scan direction for the sensing arrant.

finis combination of both differential and absolute sensing elements vYitkin the same footpnut of a meandering primary winding is novel and provides new imaging capabilities. The di^erenciaI elements are sensitive to sIighf -variations in the material properties or p.oximirv Willie the absolute elements provide the base properties and are less sensi7.e to small property variations. one -.mbodunent, IO Hera dl.fiereniial sensor, easements can he c nrnb7'led wit One, solve or HI of the raw absolute meas,ernents to provide another method for c ea7.inG a - two-dimensional mapping OI We absolute material properties (including layer thicknesses, dimensions of an object being imaged, and/or other properties) and proxi7, v. In another e nboduneni the property and proximity information obtained 1- from He absolve Tn-asT remends can be used IS "lp"S So. nodeIs that relate -rule differential r espouse to absolute proper- variations.

DIG 34 snows an expanded Alien of an altenave method for connecting to a 1 absolute skein element 3 04. EIectncal contractions to;'he sensing loop are made through he leads hi O7 which are offset Mom lie centerline 14 between 20 adjacent conductors for heplarywding3Q9. 4 accord set oficadsz16 a-e o reset the same distance mom the centerline on he owes sid- of the centeIine and connected together to Solon a '5- '7nki;c loop drib conductor 318. The colneo. or leads 310 to the sensible ele.nent are 'en co7ne_ted to the second set of Imps 315 in a d=l^erennal fond to so Mat -e flux linked By the second set of leads esser7zlIy 9: subraces Irom tee flux linked by We leads IO We sensing elerne:. Tins is particuIa-ly usel=,il when the sensing elerlen.s are retake relaLireIy small -to p,ovioe a Ah spar'iai resolution and the,= (or area) linseed by the loop created by the connection leads becomes com.paraole -to the f!-nx tor =52) of the sensing element'.

The GlSrcS 319 oe;e^ me clos3-coec.ion Silo on the second set oilcans 3 She sensing eIee should be milky izei to ensure -la. the,. ax ir;7iced by -Lhe coT=ecDon leads is eiy- cot ple.el c^^TlcOied. D-h='y elerner,s Cain also De used.

as illus area in -YG J13 to he p Dam ain he pelocici,, OL^ the COT1=UC"CS. - ^'L _ >_ One of the issues with planar eddy-current sensors is the

placemerd of the current redden for the primary wading. Often fine ends of the planar wading are spatially distant from one another, which.^, reates an extraneous Rind large inductive loop 'hat can influence the measurements. One el-nbodiment for a layout for a primary winding that reduces the effect of this inductive loop is shown in F 33.

The primary Grinding is segmented with Me width of each segment 10 determiniArAg the spatial wa:elen,g,di /. Ihe se:,nents ol^the primary wndngare connected to bond pads 14 through leads isle, where me lead^ are brought close togenher to rnirumize the creation o.^stray magnetic heirs. after wrapping the leads and bond 10 pads benched the Face of the poll lacy w a!. icing, me -it -dual sees fir he!, connected together iATA series. The arrows Helen indicate the:mstantaneous c-alrent direction. The space behind We sensor pray can be filled with rigid iAnsulators' focal, ferrites, or some combination of the above. This +hree-dimensional layout for Bee sensor efr^ectively creates a meandering wading pattern for the pn-nAAary win 1: e "o:,ely rv.rice fee cleverer in +l.e extended or.h.ons 0 Teach see ' =d Alcoves,he large inductive loop for the primly winding co;:necrions far Mom the sensing region. The sensing elements 156 and dumn?.y elements 18 are then placed in anouler layer over the primary wending. This design can also be applied to the tapered Twill away forinat of rIG 30, where Me prunary windings become 70 trapezoidal IQOPS.

C)r d measurernentme,:od3 can also be applied to rnuIti layer sensor constructs. For Sample FOG 3 6 snows a m.easuren-.t god IGr the two layer:A sensor of FIGS 3 8z send 38b. T: ifs measurement grid provides a database of the sensor response (one - =a-siTnec&nce between fine second v winding voltage and fine 25 pr nlary Sir ding currently to variation s n,r9;ro pararnete,-s to be Deere on ned Tn rim 35, L'lese t arameters are he Ii.-Gl,-d ale test ma, _ila1 conducive. The sensor espouse Values ale wpicly created with a rnodel rucn iterates each parameter Virtue over Me ange o ^inte-est to cz7,_7aIa:-.27-le sense- response 7: cicstances where cater sive -e rence mars me - vaiIaole wry ch Dp! the oropercs vat Clarions OI of,,-i e5', espy scat responses can be used o c 2:e -=e grid;. Fred meas7-rang the sensor response on z Sir rnaTer.21. Me parameter; ta:ues me dee=n-med i-t=TpoIa ing be-^Teen the lines on.:ae ne^-s-e,-lent go, A.

(I A - / On alternative method o making connections to the v^-rious components of the primary winding elements is shown in FT& 37. In this case, the cross-connections 1 SO between Me various segments of the primary winding reduces We number of bond pad connections I 54 for the primary windings. This greatly 5 simplifies he electrical connections to the sensor as only four bond pads are required, indepnderlt of the nLrnber OI mear,de,s in 'e footprint oT the sensor. The same concept can be adapted for the secondary; elements as the connections 182 indicate. This is usefiil -whenever a- combination of secondary elements is desired or independent connections lo each OT the secondary elements is not required. FIGS i 0 3 8'a and 3 8b illustrate another earn Die of tos "split" prima y winding design.

nmTny elements i32 near -he ends of he sensing elements are also included in -this case. Fuhelunore, the dreamy elements 18 are extended along almost the entire length of the primate; hnd;ng loops in o der t o main file design sv.nPtr.

An embodiment of an l-A-ray rarity r_ulfple sensing elements is shown! 15 in FIG- 39. The primly wincing =sanders 230 nave connections sirrniar to the prunary shown in FIGS 3 Sa and Gab. Seconded- element connections 232 are made to groups of secor,dar, elements 236 -blat span different regions of the plenary winding sincere so that scanning OI the &-ray over an MUT in a direction parallel to tile meanders o, the primary provide measurements OI span ally distinct yeast 20 Dummy elements 234 and 238 help niz" parasitic coupling between the primary and secondary eIemen.s o improve an- calibrations.

Mother embodiment for a layout of -die planar pr Tr ary wading reduces he e:Tect of the pr mans winding inductive loop as illustrated in FICT A. Thee sepsis Grindings 1 i2 with downy elements 170 are sandwiched between a meand rmg 2: -Andy 162 in the rust layer and a second meander ing sNrindbTg 168 in rue third layer Wilt electrical ns71ation between each Aye-. Vias 164 b-7,een the -first and -d layers provide an electrical cs!n_c ion De-.reen Me meanae s. Ine co rlec+.iors to Me ? -7Tna:) ale - ade at the bond bat such as 150. When stacked tore-,!ne ^ -it in me.7=ia7 Audi g is en "c:.ve7y ---he tong c. Rae- o Ha s -gic have: 33 pl-=ar, 7. =A;nc T 7,::37-r,..7- Or, a' _r" -a _T=7 - r -h-__7- 7.r c - I.. z _ Arch- 'G An-1ay-s by uiliz,,-.?n- ac.7'a 1' modeled and rGp7 onci51e al- TV Peon -. y

- -38 and measm-ement grids so thar extensive sets of refernc- p rrs are no. reqwred. A initial "air" caIibrafion is performed prior to mounting on 'the surface. This involves taking a measurement in air, for each array element, and then Boor ng the caIibratior.

information (e.g., in a computer) for later reference after mounting 'he sensors.

Af,er 'e sensor has beer mounted to a surface, the ins;r Grunt and probe electronics car be calibrated by correcting to a duplicate sensor so that an air castration can be peer-formed. After connecting die surface mounted sensor -to the ins:rarnentation, fine sensor operation and calibration can be brined D. rneas=-Tg the ii5-off at each element. The sensor is not operating property i f T.he lift-o readings are too high, 10 which may result from the sensor De=.g de+ache''.,om the s Mace, or Or e rneasureme points no longer fell on a measurement mid, which generally corresponds to a lack of continuity for one or the windings. A nnal vercatior involves cornpanag baseline measurements to other Measurement Iocaons that are not expected to have fatigue damage or cracks. This reference compisor can 1 ver TV sensor operation anal ma v assist in compens_ti=.g lo- noise vocables SUCH as emperatT-e drift. Uris.nay involve using elern.ents of the array Mat are a:,s+2:t -crow the areas of high stress concen-aon.

The elec+ical conductivity of many test 1-natenals is also temDerae dependent. This telnpe.-a-e dependence is usually a noise factor that requires a 0 correction o-,he date. tore>:ample7FIG4lsborsa representative set of conductivity measuremelnLs from the eIemGIls of The Mum- i-ay of rIG 8 ise+ed Aside a hole in a famine test coupon as the coupon,-nl;?e.a.are is seed and doctored with a -thermocouple. The A WAS designed -LO be insrsitie to variations its O-!l '_mpGra;'' '-e, as desc} e,r.S. aL=^nL loos. 5,'.3,6PS add 2:,793.706 alla LT.3. Pa-in-L ^4Dp iica+ ion loo. 09/1 87,Q93. I he lemp-r_=Llre OI The clone; call be caged.l1 a TrEie+y GI we. S: With +Lhe Client. co1ldi,ions in the room, win fine rnecha-,cai loacins as she condone,: is latigLlen' by Asps_ cr it ^ri-Ln a h=-,d, Lord by movie g a hot or cold an jet across i.. FIL. T 1 sir ows ha he cod,JciN;iCr h as an esseli inky:near,. e=7=eratrurv de-ncer1ce, o Ore -is ance 0 flu' 'is. ieraDlres SO r-Lat Cot -iCUC+LiVi; me=L2red By v_ck eieen. can DC CG.-Ziegler For +e?=rar! Ire C- AL.

- -39 Thermally induced changes in the electrical conductiviry also provide a mechanism for testing the integrity of the sensor. Heating the test material locally, in the vicinity of Me MWlV-Array should only iced to a change in conductivity, riot nft-off, when the array is compressed agaunst the part. Monitoring the conductivity changes with temperature, without significant lift-off changes then verifies the calib-arion of the sensor and also -hat the sensor eiemen2s themselves are intact.

Another component of the lue extension program for aircraft is the rapid and cost-e, 1ective inspection of engine components such as 'one slots of gas turbine Aisles and spools. Cracks often fonn in regions of fretting damage. The fretting damage i O often i_ads to false positive crack detections with conventional eddy-curren sensors, which set erely limits 'e usefulness of conventional eddy-c--rent sensors in this inspection. For a number of disks/spools, uitrasorc (UT) inspection is We current standard inspection method. The current TUT threshold for "reliable" detection of cracks;m fretting damage regions is Fought to be beF.-een 0. 1SO afield Q.:O inches

1: A- there is art ongoing need to reliably detect signaler cracks, possibly as Somali as O.Oc3 to 0.080 inches in ten=. The INTER C} ndStation( System we; the conformable I'M eddy-cu:rent sensor and grid measurement methods of Iers the capability to detect these small cracks in fre,.ing regions, while elimirla+.ing the need for crack caIibl-ation standards other than to. ergs performance. Calibration can be 20 performed with the sensor in Me Eddie OI any slot on Me engine disk. A scan of tress slot is then performed first 'to verily- that no crack e> isted at Me calibration location. Then all slots on a disk are inspected withou,.-ecaIicraion.

For the ins?ecuon of norane-uc disks, such as ti.cruTm disks absolute electrical conduc+.ivi and proximity- (lifter mecsurernens can be performed with 2: l-lA sensors. Yi'en a crack i'Lhin a slot is encourlte edit it Tnan;fes's itslfty a distinct cud repealable drop in condactiiy. FIGS 42a and 42b shows an example o^-veaced irSpeCriGns OTI 'the sane slots for a Slave 2 an disk. No calibration standards were used to ne:for; n -these ir specTioIls. t 'e stat: o he i:lspec';on3 a selected =-e2 \'rith n S '= V le slot {fine' he Tradie) fleas used. for. V-^er=-nce allbaio.

* 30 and eras -one only calibration -eolired r^o1 ':he Inspection of a 1 of he slo s. Mali_ i-,^-e^,.-on C07?Siqed.^T=linU each Slot vita the \.-A,4pobe ache ells e ion lath -o wlth:n aToproxi:nately 0.08 -filches '^ u;r the edge. To Mete SC=iS Cal- or

- - -4o- perfoluned in an uncrernental mode, where the sensor positioned is moved increments of 1 to 2 mm, or in a continuous mode, where a posiDon er coder automatically records the sensor position as the sensor is moved along the slot FM 43 shows the results of the slot inspection in all 46 slots, with some slots 5 showing the characteristic decrease in conductivity associated wi h a crack Both FIGS 42a, fob, and AS present the absolute electrical conductivity without any norrnahza.i flue data from FIG AS after nollua-lzation-t-o acco=. for ed,,e effects are Even in FIG 44 The slots Mat contained a distinct conductivity decrease and indicate the presence OI a crack are marked in the Ievend far each plot The arrows 10 mark the slots where the UT inspection reported reject indications; the sloes where the Minim detected cracks while the Tub inaicatiors were below the reject threshold of 30% ale encircled In addition to conductivity As slot Iocatior infor;narion, the grid measurement methods provide iii-off vs slot iocaion inforrnaion The linof: data appear to indicate the extent and relative seventy of fretting 1 Table 1 Toupees the findings of the MN7.7M Inspections with the UT :mspecuor The 1 report identi Bed rejected indications (>30 /0) In nme of he 46 slots (slots " 9, 10, 11, 13, 29, 34, 3:, 36, and 4) The disk slots had regions of fretrina drainage and, accorama to The 17T inspection report, some of the slots contained crack s in the fretting damage regions l n Contras the MWM win Grid 20 methods reliably detected cracks within f citing damage regions I Of slots, including ail Rhine slots wife rejected US indicat ons and five additional slots (slot 1, 8, 14, 93, and 41) Fcr verifications, thee -A,eIl-lmo. proce'-==e for t,qn acer,q.e replicas' that provide a "fingerprint' image of -he surface, w as adapted for we character nation office surfaces condition 'ithin'e slots These replicas coned 9: 'e LiPi' Wrings and showed images Of cracks in freeing damage regicus

- -41 TABLE 1

(:omparlson oicrack detection by MWM wii:h reported UT indications for an F110 Stae 2.f3:n A; k S10T IJT UT __ Crack Length as Distance from 5 rr Accept Response Detectiou Verified slot edge to the ance % by RepIicas nearest crack tip _ _ _

I 4ccepr '3 - Yes rr)- - O.l -n. 0.=:r.

:.....,, _

2 4ccept 20 ?(4JAT/^T) 0.0: in. 0.16i7, | 3 | Accept zO | brO (-) rVo cracs lvo creeks 4 Accep. 20 7/o () 0. 015 in. 0.2 6 in iO | j Accept | Z3 I No(Aj | 0.045 i. | f02 t t | 6 | ccept 20 ?fA/RT) 0.080 | >0.l2in.

7 Accepi 2Z No,f/) Nocrec! ocrac ! I_ CC=p! 2i _s ('E) | 0 16 n. | 0.32 in.

_ 1

| 9 | ReJecT | a Yes(3 0.2Gin. 0.26in. l 1: 10 Regec 116 Yes fEj 0.21. 0. in. I _1 1.. 1

| i 1 Rejec. 52 I Yes r o.z: in. | o. 28 in. l j. _; --. 1 | 12 | r | 9 i l\TO J | POSS D1Y <0 0 i 0-4 n.

I-- ' - T --- - - E i I ! i | ejec | 7 i P () | O Z8 O G.n.

14 |.-cceP'! 13 I Yes r) | u. Tn. j 0.2 i.

- -4- 4ccept | 10 | No() I Possibly2 022 in.

adjacem cracZ,s (combined lengZh 60.03 i7lJ _. 16 Accept IO ? (/ART/ERT) 0.00: to 0.015 i.m 0.13 i7.

long inte7mtterzt cracks over Q.1' _ I! 7n _ 17 Accept 1 7,To (AJ ATo cracZ o cracks - ! Accept I,7\,Jo f 4) xYo C'--/7cis ho cracks _!.! | 19 ccept 9 7 () Possibly o7ze O.03 0.29 in.

i in. crack:? _,....... _

0 dccep' 0 io (.4) hho mrecZ:s /Vo cracZ3 ! ,1 | Iccep! 1o | trO () I tTo cracks o crack I I ^' Reject E 63 I Yes 0.44 in. 0.18 in.

| 3 | 4cce7, 1' T'es, 0.19;.' j 0. r6 in.

I _. E _

10 I 99 | ccept | 7 77\rO,{/J | .00: o 0.: r]L 1! 0.29 in.

| 'a7lgi7zlermz'7-ent | | | | crac-s oPer 0.165 f ! 30.cce | 7 | rlRTj r3vo adiacent | 0.26 zn.

I I E | cracks (cor7z f le7zg 40. 04,'72.) 1 | pilS t'a a.: | iI cracks .. ., --, r 1 | 33 accept! 17 | ?,',/AKJ7 Possi31), 9 c.'acrs/ | 0.02 in. Back, I aoli O.l n., l l ll | ?rt E l r----'--- t - i ---- ti! j;' iteJ'ec. J j,D5 i ( O. / ii, j,3 _J,' I I i _ _. _..

- / -43 35 1 R.iect 68 1 Yes 1 (0 440.:n 1 0.76.n 36 Reject 54 Area- I Not reflected Not replicated ,.. _ _

41 Accept 12 Yes 0.15 in. 0.36 in.

--- 1. 1

z Reject 41 Yes j 0.15 in. 0.21 in.

i l _ I 5 Note: A - accept;: - eval7ua.e Subject. to an e-,aluaior for re?atleure dins); ART - accep. on retest; MAT_ evaluate on retest These decisions depend on We Tnr^Shold settin as in ice apF7icatlor module. Additional measurements were also performed to ilustra.e the" use of an encoder for determirng the position in a slot and sequential -:nreshoids ror 10 determining the acceptability of a disk slot A tw7caI set of measurement scion results is illustrated in FIG 9. The normalized eleefn:caI conduct laity measured with the EM. is plotted agains, -the sensor position measured With -the linear - - encoder. For each scan, the initial position OI me sensor in the slot is set visually, usually by aiming a "corner'' of the shuttle with the top s7mIace of file slo+. The 1 - conductivity is then measured as the shuttle is passed Trot gh me slot a. a r easonably constant rate. The presence of a crack in he slot causes a reduc ion in the electrical conductivity as the server approaches He sIo: edge; as the sensor leaves Me slot and goes oaths edge, the effective eleGticai conducf-v-ity dips aid Decor: es very large (eve ruaIiy going off or-7rhe measurement god). The measured eiectucal 20 conductivity is noI;nal7zed by the amperage conductivity, near the center o1^7;ne slot, p-7O1 do reac in g -Lhe region of interest near We riot edge. TicaIl We a-v erasing was performed over Me 0.8 to 1.3 mch r Gion while the edge of the sick was in the 1.7 to 1.9 inch region; based on 2 IiTnited rl7mnoer o'scarls, a-veagma i^rom 0.5 Lo 1.3 inches does not zpp ear to affect -Lhe measurement es alts. AIrhough LO e c aces 7 95 some once slo s e';ind Fore fine edge into ibe average rector, the siaI ob aided 1 he cracks still iral1 -to the "oval- ate region Or he espouse, AS desks-iced ielo:;. lc I; 17,al le Feast ed To- ADD 0_.al2e'7 _ eerily co-- : VT fir IS used to do er7ne th- presence oft crack.

t L ==rr = -l To' I of ant= Ar air+ Ts -

33 0-12 sea En Al decision proses-. i TO esnods -as e e i. sed i-1 -- s pce,s.d a-e

- ( -44 deno7ed by the 1-bels AI and A2 in FM:0. Ln the decision process, each slot scaD is compared to the No thresholds. A1 is Me Retest/Evaluate threshold while A2 is 7 he Accepetest threshold. If We normalized conductl'r7ty is above An, then the decision is ACCEPT (e.g., both A1 and A2 pass). IF the normalized conductivity is below A1 on Me initial scan, Me slot is thought to contain a flaw and EVALUATE is the final decision (e.g.; both Al and 42 do not pass). If the mimrr7n normalized conductivity falls between Al- and 42 (e.g., b--7 pass. 4 does notpass), the slot must be retested se-v-eral times. Then He average of He inspection scans is used to reach a decision on He slot. Now, IF the average is below- A2, the final decision is 10 EV^A!17 A TE upon -3 fist. Otherwise; the outcome Will be ACrT-PT upon retest. In He case a slot is accepted upon retests. a slope.nrisor concurrence and signature are requite ed. Thus, rear the case of; A CCEPT," no Fourier action is required other = malting a record. For 7ne case of'RETEST," the slot has LO be re-iaspected several rimes. The Retested slot will Hen be labeled as either Accent or Evalua7 e.

1: 'REVALUATE", Pearls that the slot is likely- to have a Sit -ZClf if.' Ma needs lo be evaluated by other methods.

These thresholds are based on saf'sics for he disks being measured and the Raining set population. this case, the ^mreshoId level AI was set 70 provide an Evaluate decision fir a 0.16 inch long crack while the threshold level A? was set to 90 be near the rain; in normalized conducivi- for a 0.080 Inch ion, crack. As the nor Or disks and s1O7s inspected increases, He r.reshoid levels can be deermmed vv-:,h s aTistica1 methods based on He probabiIitr of de'Rctior1 for a given cracl: siz=.

Representative -threshold levels are Ai=0.992 and A9=O.99: Mile lmin7mAll^m in the normalized onduc-! vim loo all of-ie slots or 2 d'sl are -3 iiTrateu Ida FIG:1. The column bars denote the average values wale 'he or or ba s shove the siancard deviation of the m asu^-=A=^eris. The ellec. al 2Ic_r'--lg He ---s'noId 'eveIs can be seen. 1ne Al (ic.-vrer) = eshold is t-picaIly se. so la. larger cracks (V--eater -Han 0.1 inches long) a-e evaluated after he nest scan. The ^4 Hoer) resilold is set to cil vrer,iate -the smaller cTac^ks, om I've noose ire,!r'rlaN$red -0 slots Ag>273A, He Valor bars denote He v2^=lao ii ^'^rA m-- ^7easure^=en s so choosing an _ nresho'd -splat passes through (or n ar) the PA, o bars -A in '-a- he ala in, _reia.e e.' teem err zero and one) p-obabiIity or v_.ectio. Once pore cracks ha e been

( -45 charactenzed (e.g., replicared), better statistics can be applied 'o determAining the thresholds that should be v,Ased ror detection of a given crack size.

FIGS Ma and 45b illustrate the crack length dependence of the miuAmA in the normalized conducrTity for the slots of Table [which had been replicated. In 5 this case, three to I 1 measurements were performed ore each slot. Three difIerr inspectors inspected each slot. The av--lage and standard deviation or',he measarem-v-r^ts onea-cEsfor are iIlusratedin [lGS 43a and fob. The vertical error bars -epreser'-t the standard deviations in the measurements between the operators and illustrates the operator va-iabiy in the measurement results. The horizontal 10 error bars denote the effective crack length due ro --ul-..ple cracks or char s of cracks greater -men 0,0(55 inches long. The slot number Is given on the ri,ht side of each data point. The thresholds indicate the evaluate (AI) and retest (A2) levels for the rninimumA in the normalized conductivity. CicaAri, adjusting the r eldest level (A2) slightly will affect the probability of detection of the smaller cracks, such as the 15 - C.080" sad 0.05G" Iong cracks (slots 5 and 0, respectively). The miri^:r^ detectable crack size depends uonLe seIectio of cue detection tnl-eskoIds and the - variabiIir ofthe instrument, opera ors, and other noise factors. The detection thresholds set -e mirum allowable reduction in the normalized conuctiviiy for an acceptable scan. Choosing,l-esholds beyond the measur=-mer "noise" level that 20 n=nizes the n'=Tn7-er offaIse inoica!ions also sets the rum delectable c ack size. foe use of sensors and Grid m=-asurernent methods Cain also provide a resole meamn=G, al assessment of weld quality than convennonai inspection -needs.

Ike -huh cost arid cor'1pie't.tr of,.tarT wed me caused by social cIea.-g 75 End shielding procedures to precinde coTamin2-.on. Quarry control o t:T&= welds includes, among other thistles, iss-Dtion for cori:anirlaon. (:n,le,,t'y, War. an welds are accepted or rejected based on surface color inspection results, even though the sir, ace color has not Dee a reliable rdic?.',or of -weld con arnina:on level.

30 The capab'1 in; of -the 7V rAl\! -LO C acel7ze cG=ia,'natio of +ij9 welds was tie eons rated o,-, several lest s Cecil! as. 1;togenoJs Gay -ils were avrlca:[ed in s Ma lo Crab- 'ares i-= Retain erases that included -i_h n-=i Ron,

it'\ -46 tee levels of alr contanunarion, and rwo levels of (70 cotanination. The measurements were performed in a point-by-point ' scanting" mode across each weld so that each scan included the titanium, Grade 2 base metal, heat-anected zones on each side of a weld, and weld metal. The footprint OF the VM sensor was 1/z in. by 1/2 in.

FIG48 shows anMWM measured electrical conductivity profile across Ike welds obtained at a reenc,v 400 Adz. All measured conductivity values were normaIized be., the mamrrn conductivirv in the base metal. The dip in conduchvity in each cT-e corresponds to tlie -'S.,eldriietal, whereas the left and Shalt "shoulders" 10 correspond to -the base metal. Tn the specimen containincr the weld fabricated win pure argon as the shielding gas, Me conductivity of Lne weld me.ai is only slirttIy lower than conductivity of the base metal. There is a general trend of conductivity decrease with contamination level. This Trend is illustrated in:FIG 49, for excitation frequencies of I. 00 tlZ and 1.5S Whiz, as air contamination in the shielding gas i 5 reduces the conduct o'fhe 'titanium weld metal. this plot he con deceit ty of weld metal is normalized by the mnim7r. measured conduc.ivi7:r of weld fabricated in pure argon.

Periodic field ecdy-c7 rrent sensors czar also be used to detect overheat

damage in An barrels or other steel co nponents Mat nay be coated milch another 90 material or uncoated.

As an example, meaS-LLrernen7s were pe-Io Lned on ONTO seini-cy.i.ncal samples from a long77d:mally sectioned 25- gun barrel. lithe section of his prriC7'lar,Ln barrel, loca. A be77i eon 2xiaI pos7+.7or.s m. and 24 in. avray, From Ale spa. oaths rir'iincr, had xp_ie7nced ov--v-heating Sample ha (in FIGS 5 end 5jJ 9: was removed from. the overheated section and Tro he O&T of pie gun brreI between Me 7-in and 16-7. Exit pos7fions. Staple (an FIGS -0 and 53) is a section of the gun bagel not a Tected bit orerhearir= zrd Arm om th part OT Ale Elm barrelo mreenthe41-i. znd:1-. zzlposi^ions. neguntz.eis-elernadeoa lovv--allo-y stem v.,hica was heatrreated origiaIIy to oD+zir tempered.-ar.t_sire 30 tri-.os+r'c re. -.he ore heated section. -were -eras z dis inset Ma -zf,ec.ed zone cld thee bore where the resuming le-rlatic-D-rtic-, ic-os c:u,= suggests -one repn-s So- ye Seen at leas 'Jo -LO J. He us He Face or An

A\ - -47 barrel was pIa+ed wirk electrodepositeo chromium where the chiclmess ranged rom O. 10 rum to 0.20 mm.

FIGS 52 and 53 show a representative set of Maim measurements on gun barrel samples. These measureme,-irs were performed Pith a JENTE GridStation 5 using magnetic permeabilit,v-lif+.-off^neasul-ernen. grids at a frequency of 100 loTIz.

Axial scans along the ieng.h of Me samples were performed svith the MiN7M sensor wirrdins oriented b-o,hparalel (Onen+atfon pi) andperperi&c-ular (5rient-anon JO to the gun barrel axis. FIG::2 shows the results of +,he L1VI axial scans 1 t rms off effective relative magnetic permeabiIir vs axial position (sithir each sample) along 10 Me barrel axis. Note that the I1i/ is mos+, sensitive to permeabihi; in the direction perpendicular to its forger Grinding segments. The data reveal that the longitudinal effective permeability measured with Orientation.'. 2 in Sample 5 (not affected by overheating) is nigher than the transverse permeability measured w In Orientation rcl, indicating some anisotropy. The \.'M data for Sample Pa show! 1: that ove:hea,ing na.icallyr inr:easod oh_ loncrir.diArAal ef^^ecive pe^=^eab T. Rev rnAeaslLred win+ O. i octagon.2.'2 in s^A^pIe 2a cougar d to the transverse elrlec+,isre pe^T^nAeability, rAeasured with Orien.aion.','1. FIG 53 shows Me elfecTiv-e permeability is plowed its aistce Mom -the snarl of ri-;ng along The Darrei axis. The lvIW!f measured results are shown in solid L nes while the doped lines indicate a 20 possible trend relet ve magnetic '3rmeaoii+,y in Me region between Sample 2a and amp le 5.

base ^easuremeT Is indicar alar Me MVi'.f probe.-esporse Was characteristic OI a ferromagnetic material. Note that bile low-alloy steel is a fxro^^zgr et c maenaI To:vreas the eIertrod posited c^.-orni pIat=g is rt; T,o:,rnagnet c unless Me plai--ncr had be:l exposed lo no rh temperatures or suf Ficierly long Tome to effect Gi--r-sion OI iron into the deposited piarin=. A a frequency OA- 100 Akhz ulnae esirnlateo c p h O er S^?lSit;Vity in pure Cnro, ninths= is _S7rnAated LO be p,- oximaely 0. == w.'ich., greeter man The 0iici=ess Ol-]le leclroaeposited c-om' r,la-=,r.g. s -espy 'L.;!e ' '7 fi ';sees ' oyonl the plated 0 ladeol-rl-!, oi:r.!u.'n and -he mease.vl-l,s reflect fine I ecT.i ire pereab,iir kiln nc-c5-cwal eoTl-tios of-: lo--r-a"oy stee. 1--ls, the -mucus bid.reco. al p--eb',-;,eaS=-e=n cat-, se--heA,i?\4p-cNil sensilNi N- o Me

-48 properbr, char!ges caused by ov-ernea.irlg For r apid inspections of g ba, els, Gylirldical pro yes having A - M sensors in both parallel and peendicuIar orientations car be used so that 2 single measurement scans provides both measurements of the erfectire permeability i Periodic held eddy-cuent sensors can also be used to detect and quantify the depth o^subsurface clacks As an example, consider Me measurement iIlastrated in FIG 54 In Ads case, two-equency condTlcivitr--irt-or^^measurements were perfiorrned on flea back sTrface ova nickel alloy sam,pIe having notches that sim.uIate c ack-I-,ke flaws on the wont surface FIG- SAT showsa schemlaric of die f aw pattern 10 in the sample and the EM measured conductivity scan at two frequencies A spike ratio of the two-equency absolute cndact:l,i measurements (after passing the rarer data tough Me two-row rneasurernent grid) provides a robust correIatT:on with distance fi. em the Jaw tip to the back surface Tliis method can be used to detec, and deTelllline depth or distance to hidden cracks for both 'fatigue 1: cracks and, for some components, crack:mg associated with ccosion fatigue While this inv_n-Lion has been Reticularly shown and described with references to preferred embodiments Hereof, it wiT1 be understood by those stilled Me am -that various changes in form and details may be made therein w-ithou departing from the scope of -he Benin encompassed by Me appended cI=ms 90 elerences incorporated by reference in their enr!rery: Air Force Association (1997), "-rose lnac",!Ia, lg97 _A -aid, A mid Moulder, J C (1999), 'reviews old kdv.ces in Quantitative ddy-CurreTH N7ordesucTire 7-val-,la-,ion," Journal of NoIlles ructive E-al7larior, vol. is, lio 1 9: Committee O Aging of l;S 4 r Forc- Aircrew (Ig97)' "Aging o US jr Force rcrar", ISBX O-309-0-93-6, ' 997 Breeder, J. (195), Dislocations, =rgcqrnon ress -o'fEne, -x, A t asiava-iv-h, '\adl3-_' ? Zorlbo, and Tic 'miller cl99Cj, CCo:;=oable -did -C-.-.r Sensors mid Ale hods o,- C-as bill- msaeLio2 and

i -49 Health Monitoring", ASM InternationaI, Gas Tmbine Technology Com^erence, Materials Solutions '98, Rosemon', IL.

Goidfine, 1Y., D. Schlicker, and A. Vashataugh (1998 MESA), "SurfaceMounted Eddy-rent Sensors for On-Line Moritoring of Fatigue Tests arid forAircraft 5 Health Monitoring," 2n NASA'F4AJ:DoD Conference on A Erg ALrcrat.

Kramer, I.R. (I974), ffvtallurgicai Transactions, V.57 p. 173 c Regler, F. ('3-7), Ze-hrft-fu"+ F-lektroenne, v. At, p.:5 Regler, F. (1939), Verformung und Ern^.; dung Metallischer WerkstofT=.

Suresh, S. (199S), Fatigue of Materials, Second Edition, Cambridge University 10 Press. Tara, S., and Ea:ashi, K. (1966), Ploc. 96q Japanese Congress of Testing Materials.

Weiss, V. and Oshida, Y. (1984.), "Fatigue Damage Chalaccerizationusii^.g X-Ray Diffraction Line Amalysis-7 in rerigue 84, p 1151, 3utterwo.

RELATED DOCTuS 1: This present m;-enton is related to: 1. Nat y Phase I Proposal, titled ' A-pplicaion of '.ne Meandering W re Magnetometer to Detection ma Quanr.ificanon of Cumulative ra:tiG-ue Damage Aircraft S+ ructT=ral Co.nonvrlis ', Topic =1-9-Q33, dared 1/12/95 9.:Nay Phase I Final Report, tilled " Application OF Me Meandering Afire 20 Mame-.omete1 to D-t-c+.ion and Quan+ncanon of Cumulative raTigu_ Damage In Aircraft S, cr-Tl-al (components", dated Air I 30, 1995, Contact rlLoDol 9-9-C-0290 3. Nary, Phase 2:I Proposal, titled ';AuL;ca-:o of r.!le Meandering wire Magn-rome+e, to 3e+ection and Quail Canon o'^Cl="nula+.ire Fa:time Damage n 25 Aircraft, Struc=-al Components ', 1 opic. 9-033, dared 5,1 //36 4. Navy Phase Final Report, Titled AppIicarlon of Me Meanders WY== vIaguletometer to e,ecuion and Qua=-inca.ion of C=l12ci e a-. Me DzIr2=Ge in -ca,= S=lc.*ura Components", tea e-'C.r 2' c'' 399, Compact 5C'-21 97-C-1 23

-so- 5. Air Force Phase I Proposai titled "Porabie AccumuIated Fatigue Damage Inspection System Losing Permanently Mounted and Wide-\rea unaging ^7M-Arrays'', Topic.-,AP99-286, dated 1!11/99 6. Air Force Phase II Proposal, titled "Portable Accu;nulated Fatigue Damage Lnspection System Using Permarlen,1y Mounted and Wide-Area Irnamug -Arrays', Topic.-.MF99286, dated 12/3/99 7. Fo ce Phase I Final Report, titled "P ort-ab l-e ccT:Iated Fatigue Damage Inspection System Using PerrnanentIy- Mounted and Viride-Area Lmaging --rays", dated March 10, 2000, Cortractr',2F090099-l\I-i328 I O S. Technical Paper titled "Su-rrace-Mounted r:ddy-OuIrent Sensors for On-line Monitoring of rename Tests And for Aircraft Health Monitoring", presented at the Second Joint N-ASA/FAA/DoD Conference on Aging Aircraft, August 1998 9. TED Sensors Trip Report to Tinker GLIB, dated July 5, 1999 1 O. Technical Abstract titled 'mew 1 Always Keith High Resolution and! 1: Lncreased Depth of SensitlN in for lani.ati-ve Irma2ng of ' Hidden ' Par gee and Corrosion offer Wide A reas, sorted to the Third Joint K- AhDoO Conference on Aging Aircraft, September i999 i I. Tecllrical Parer titled C'ecrli Explications of Me? d,rrg JO ending Manetomet=s to Materials Characterization", presented at The 38 Annaz1 British 20 Conference on N7:DT, September 13-15, 1999.

12. Technical Paper titled "I- isotopic Conductivity Measurererlts for Quality Control of C-1301P-3 Propeller Blades Using 77M(-Sewso.-s Ah Mid Methods", presented at the olden Joint DoD/r A. SA Conference on AG1I1g A, -Cra1L, May 1 6, 2000.

o: 13. Presort ation Slides ti led "4nisoopic Conductivity Mea='ere=ts.or (aIity Control orC-130,-3..opeIle: Clacks lis no 747.(-SeTso.-s -with <3. d h,erhods", Presented at the Fourth Join' Do3.rA 'WAS A, Corler-nce or Aging -irc-alrt, May 6, 2000.

1 A, A- A Ye- Iwo Acne Report. titled "--velopment of Conform.abie 0 L^-C,L-nr Sensors L-0r EIig''-3 Ccronet Tnsection,' dared August LOGO w.o,-,.. acL '.'DT \0398-D30008.

- 1- 15. Technical Paper titled "application oft-Array;?N--Cr eat Sensors to orrcsion Mapping", p eserAtd at Who Ala I0LelatiO3a1 A irvraft CooSlo: NJi'OTt-S6Op, August 22 200Q, which Are incorporated herein by rererenc=.

Claims (1)

1. CL4T\lS 1 A method for monitoring damage at a fastener composing: mounting

   a spatially periodic field eddy-current sensor

   to a test substrate near a. fastener, and 5 sensing response of the test substrate to a magnetic field imposed by the eddy-current sensor

   2. A Method as claimed in Claim i where the sensor Is mounted

   under the head of the fastener.

   A method as claimed in Claim I where the sensor is mounted between the layers of the structure attached by the fastener.

   4. A method as claimed in Claire I where a sensor Is mounted at 5 both ends ofthe fastener.

   5. A method as claimed In Claim I where the sensor Is a circular spatially periodic field eddy-current sensor surrounding a

   fastener. 6. A method as claimed in Claim I where the damage is m the form of a crack.

   7. A method as claimed m Claim 1 further comprising at least 2 5 two circular spatially periodic field eddy-current sensors each

   mounted around a fastener and a single cable connects the drive and sense conductors to the data acquisition system.

   A method as clammed In Claim 7 where the each sensor provides a separate output.

   9. A method as claimed in Claim where the output is an absolute property measurement.

   5 10. A method IS claimed In Claim 7 where the sense conductors from pairs of. sensing elements are connected together to provide a differential measurement 1. 4, method as clailmled in Clairol 7 where separate drive 10 connections are made to' each sensor.

   i2 A n:lethod as claimed in Clarion I 1 where the sense conductors are connected together to provide a common output connection 13 A method as claimed In Claim 7 where the drive conductors are connected together to provide a common drive signal.

   1. A method as claimed in Claim.3 wllere trio sense conductors 2 0 are connected together to provide a common output connect on 1 S A method as clamped In Claim i where the sensor Is mounted in a cylindrical support material shaped In the fiords of a 25 washer for mounting under a fastener head 16. A method as craned In Clarrn 5 wliere the support. material wrt.hstands compressive loads.

   3 0 17. A method for estimating material properties from an Inductive sensor composing creating rrleasureillent grids that provide primary and sense wrradings In rnuihple layers; storing the measurement grids as databases of sensor 3 5 responses to a predetcrmmed range of at least two unknown properties; aIId using the databases to convert sensor responses into estimates of at least two unknown propeitcs

   5 l S. A method as clamcd In Claim r7 where finite elcn'ent modeling is used to gencraT.e databases of responses.

   19. A method as claimed In Claim 17 where anal,vtica] morsels are used to generate databases of responses.

   20. A method as claimed in ('la.im l 7 where finite df.ference modeling Is used to generate databases of responses 21. A method of fabncatmg a damage standard comprising: 15 attaching an electromagnetic sensor to a. critical surface of test material, mechanically loading the test material and measuring a change in the electrical properties under tllC surface of the sensor; and removing the m1echarlical load vvLen the charge in electrical 2 0 properties indicates a prescribed level of damage.

   22. A method as claimed in Claim 2 I where the damage Is a fatigue crack.

   25 23. A method as claimed In Claim 21 where the sensor Is a spatially periodic f eld eddy-current sensor and the test r1atcnal is a metal.

   24. A method as cIaimed in Claim 23 where the sensor Is flexible 3 0 for confonning to the shape of the surface of the test material 25. A method as clanr1cd In Claim 2A where the sensor is an array smith rr ultlple sensing elements for producing a spatial Image of the damage.

   26 method as claimed In (Claris 21..hcre the scissor Is a Electrometer and the test material Is a delcctuc natcnal

   s; 5 27. method as claimed in C land 21 further compusmg the use of a temperature measurement sensor to con-ect for electncai properly variations with temperature.

   28. A method as claimed in Claim 21 further comprising the use 0 of multiple sensors to monitor multiple regions 29. A method as clain:1ed in Clan 28 v.1icre the sensors Include arrays with multiple sensing elements for producing spatial images of the damage.

   30 A method as clawed In Claim 22 further comprising the use of a sensor to monitor the change In crack length with the number of fatigue cycles.

   2 0 31. A method as claimed m Claim 22 further composing the use of multnle f renuencv measurem. ents to or alter crack de.pf.h 32 A method as claimed m Claim 22 where the damage Is pre-crack darnage.

   33. A method as claimed in Claim ? further comprising mounting the sensor between lancers of the test material.

   34. A,.ncthod as cleaned In Claim 23 further comprising the use 3 0 of a sealant to provide mechanical support.

   35. A.rlethod as claimed in Clam =1 further comlpnsing shaping the test material to create a stress dstr button so that fatigue damage minatcs under the sensor.

   36. A method as claimed In Clanm 35 where the test n,^latenal Is formed mto a dogbone Sloane and the center section is thinned to localize fatigue damage.

   sC 5.,7. method as claimed In Claim 3<j whore the test material further comprises reinforcement ribs on the edges.

   38. A method as claimed in Claim 37 where the test material further c07lprises radius cutouts on both sides of the thinned section. 39 A method as claimed m Claim:5 where the test material further comprises radius cutouts on both sides of the thinned sect7 on.

   40 A test circuit comprising: a primarN winding having parallel extended portions for imposing a spabaily periodic magnetic f eld of at least two spatial wavelengths in a test substrate when driven by an electric cu7 lent; 2 o an array of sensing elements for sensing the response of the test substrate to the imn:>sed magnetic field' at least one sensing element

   positioned between the extended port707is of a half wavelength. of the pnma7y winding located every other half wavelength of the primary winding, extended portions of the 7ndividua! sensing elements being 2 5 parallel to the extended portions of the primal winding; and a series connect70n between the sensing elements in every other half wavelength perpendicular to the extended portions of the pn7nary wndmg to group the Individual sensing elements, the series connection being in a d7f.erent plane than the primary windings; and 3 o separate output connections to each of plural groups of sensmg elements located along the length of the extended portions of the pumas v, winding.

   41 A test circuit as claimed In Cla7m 40 wherein the individual 3 5 sensing elements are located In at least two adjacent half wavelengths of the primer vNindi7l2 42 A lest circuit as clairllcd ill Cla7711 41 wherc7n the individual sens7ng elements In adJacenthalfwavelerlgths are spatia]l.;

   s] s oirDsetpara]lcl to the extended portions oftli_pi,nary winding. 43. test circuit as claimed in Claim 42 whercn the spatial offset is one half the 'engtli of an individual sensing eleinent parallel 0 to the extended portions oil the primal winding.

   44. A test circuit as claimed us Claim 40 wherem all sensing elements are positioned at icast one halt wavelength away from the ends of the extended portions of the primal winding. 45 A test circuit as clammed in Claim 41 further comprising additional conductors near the ends of the spatially offset sensing elements, aioiLg tl.e length of the extenled portions of 2 o the primal winding, to mamtam the spatial penodcty of the conductors. 46. A test circuit as c]amcd in Claim 4i further comprising extensions of the midmost mdvilual spatially offset sensmg 2 5 elements between the sensing elements in the adjacent half wavelength. 47. A test circuit as claimed In Claim 40 where one individual sensing element is located every other half wavelength to 3 o provide a single response 48. A test circuit as claimed in Claim 40 where the sensmg elements are in a different plane than the primal windings.

   35 49. A test circuit as claimed U1 Clann 41 where the array of sensing elements has one mdvidual sensing e lenient located every tither half wavelength but an array of sensing elements located in adjacent half wavelengths

   so 5 50. A test circuit as claracd in Claim 40 NN bcre every other half wavelength of the primary wmding is coimccted together in Been e _ _. _.

   51. A test circuit as claimed in Claim 50 wherein a single pair of 10 connector leads Is connected to each set of primal windmg halfwavelengths. 52 A test circuit as ciamed nil Claim 40 where the primary winding is distributed:n two planes with extended portions of 15 the primary- windings located over one another and the connectors between the extended portions of the primary windings offset by one half-wavelength.

   5,. A test circuit as claimed In Claiir. 40 where each sensing 2 0 element provides an absolute response.

   54. A test circuit as claimed in Claim 40 where at least one of the sensmg elements provides a differential response.

   2 5 55. A test circuit as claimed In Claim 40 that Is conformable to inspect curs ed parts.

   56. A test circuit as claimed In Claim 40 placed on a curved and compliant substrate to inspect a curved part.

   57. A test circuit as claimed in Claim 40 that is scanned across the surface of a part for the detection of flaws.

   58. A test circuit as claimed in Claris 57 where the longest dimension of the Saw is substantially perpendicular to the extended portions of primary winding.

   sq R 59 A test circa as claimed in Claim 40 that is mounted against the surface Of a part for detcchng and determining the location o f f tows 0. A test circuit as claimed in Claiir 59 where the longest dimension of the flaw is substantially perpendicular to the extended portions of pnrnarN,/ winding.

   6i. A test circuit as claimed In Claim 4G where the iift-offis varied during calibration.

   rS2. test circuit as claimed in Claim 40 where the lift-of1 is varied during measurements.

   63., test circuit as claimed ill Clairol TO where the par+. under test 2 o temperature Is varied to v ary the part conductivity for calibration. 64 A test circuit as claimed in Claim 40 where the part under test temperature Is varied to vary the part conductivity for 2 5 measurements 65. A test circuit as clair:Ied In Claim 40 where the part under test permeability is varied using a bias held during calibration 3 o 66. A test circuit as claimed in Claim 40 where the part under test permeability Is varied using a bias field during measurements.

   (;7. A test circuit as clerked in Claire 40 where measurements are made at multiple operating conditions, including multiple 3 5 lift-offs, tempcraturDs, or bias fields.

   68 A test circus as claimed m Claun 40 where measurements grids with one or more dimensions are generated In advance and used as databases to look 'Jp and ntenoiate the electrical

   to 5 and gcometac prol:,er',es of Interns' at the location measured by each individual sensing element.

   69 A test circuit as claimed In Claim 68 where the electrical and geometric properties at cacti sensing element location are O correlated with dependent properties of Interest.

   70... test circuit as claimed In Claim 68 where the array Is scanned to amid Images of eiect.ncai propcrtAes across the surface of a part.

   71. A test circuit as claimed An Clairol 69 where multiple frequencies are used to measure property variations with depth at each scns;ng element.

   0 72. A test circuit as claimed m Claim 70 where multiple frequencies are used to create three-dimensional Images of properties. 73 An apparatus comprising 25 apnAnary wmding of parallel winding segments that impose a spatially periodic magnetic field, verity at least two periods in a single

   plane, in a test substrate when driven by electric current, one or more sensing windings that link flux over regions of Incremental area along the length of a dr ve winding segment, with 3 o the sensing elements located An a second plane, with a senes connection between sensing elements An every other half wa.vclength; and leads to the sensing elements miring the sensor footprint In a direction perpendicular to the direction of the drive winding 3 5 segments.

   74. A test circuit comprising: a neandeiilg primary wndmg having concentric substantially closed winding segments For hobos rig a spitball: periodic.nagilctc field in

   6l 5 the radial direction of at Icasi two spatial -wavelengths In a test substrate; and at least one sensing elemen! for sensing the response of the test substrate to the imposed magnetic field.

   10 75. A test circuit as claimed in Claim 74, where the closed winding segments are circular.

   76 A test circuit as claimed In C'lam 74, where the closed wmdu1g segments follow a shape In the material under test.

   77 A test circuit as claimed in Claim 74, with at least one sensing element positioned between the con.centnc circular senents of a half wavelength of the primary winding and located every other half waveleiigth of the pnnlaiy winding, and with 2 o extended portions of the individual sensing elements concentric will the conc.entnc circular segments of the prmarv winding 78 A test circuit as claimed in Claim 77 where a single sensing 2 5 element Is placed within In each half wavelength of the primary winding.

   79 A test circuit as clanged In Claim 78 where separate output connections are made to the sensing element In each half 3 o wavelength.

   SO..^ test circuit as claiined in Claim 79 where at least two of the sensing elements are connected together to provide a common output. 8]. A test circuit as claimed In Claim 80 where all of the sensing elements are connected together to provide single output

   Gt 5 82. A test circuit as claimed in Claim 81 where the sensing elements are m a different plane than the pnTrary windings.

   83. A test circuit as claimed in C1aiTn 77 where the circumference of at least two half w a v elengths of the prilnary winding is 10 slammed by more than one sensing clement and the sensing elements spaTming the same angular dimensions in every other half-wavelength of the primary wading arc connected together; and separate output connections are made to each group of sensing elemeTlts spanning the circumference of the 5 primary winding.

   84. A test circuit as claimed in Clang 83 where the sensing elements are connected together with a series connection.

   2 0 85. A test circuit as claimed in Claim 84 where the series connections are In a different plane than the nrimarv wTndmg.

   86. A test circuit as clammed In Claim 83 wherein the individual sensing elements are located in at least two adjacent half 2 5 wavelengths of the primary winding 87. test circuit as claimed iI1 Claim 86 wherein the individual sensing elements in adjacent half wavelengths are rotationally offset from one another 88. A test circuit as claimed in Claim 87 wherein the rotational offset is one half the angle spanned by ZT', individual sensing element. 3 5 89. A test circuit as claimed In Claim 88 further compusmg extensions of the nner-most rotationally offset sensing elements between the sensing elements In the Inner adjacent Dali wavelength.

   G. 90. A tes, circuit as claimed in Claim 77 where the sensing elements are In a different plane than tile primary windings 91. A test circuit as claimed in Claim 74 that is conformable to + -,A in, L WL11 Van gal I. 92. A test circuit as claimed in Claim 74 placed on a. curved and compliant substrate to inspect a curried part.

   93. A test circuit as ckm.ed in Claim 74 that is mounted against a 15 surface of a part for the detection of flaws 94. A test circuit as claimed in Claim 77 where the part under test temperature Is varied to vary the part conductivity for calibration. 95 A test circuit as claimed in (Claire 77 where the pat under test temperature Is varied to vary the part conductivity for measurements. 2 5 9G. A test circuit as claimed In Claim 77 where measurements grids with one or more dimensions are generated in advance and used as databases to look up and interpolate the electrical and geometric properties of interest at the location measured by each individual sensing clement.

   97. A test circuit as clamored In (claim 96 where the electrical and geometric properties at each sensing element location are correlated with dependent properties of mtcrcst.

   3 5 98 A test circuit as claimed in Claim 96 where the array Is scarified to bulks images of electrical properties across th surface of a part.

   Gas 5 99. A test CfCUt as claimed in plains 97 where multiple Trequences are used to measure property variaions with depth at each sensing element.

   100. A test circuit as clamored in Claire 9O -where multiple 10 freque1lces are used to create threc-dmensona] images of properties. 101. A test circuit as claimed n Claim 77 wherein the sensing windings Lila flux over regions of Incremental area along the length of a dove winding segment, the sensing windings are located In a second plane with each sensing winding link rig magnetic flux every other half period, and the leads to the seas ng elements exit the sensor footprint radially, pCrperldicular LO the direction of the drive winding se,_lnerlts.

   109 A test circuit as cla1rned Or Claim 77 further conmpricin-g a hollow center reg on for placement around a fastener shaft.

   1(13 N test apparatus cornprisng: 2 5 concentric circular winding segments that Impose a radial spata]]y periodic magnetic field, with at least tWCi periods m a s ng]e plane, in

   a test substrate when driven by electric current; and one or more sensing windings that link flux over each region bordered by the drive winding segments and concentric with the drive w riding 3 0 segments.

   104. A test circuit cornpusing: a pnma-rv winding of parallel extended winding segTr ents that impose a spatially periodic magnetic field, with at least two periods, in a test

   3 5 substrate when driven by electric current; an array of sensmg windmys for sensing the response of the test substrate to the imposed magnetic field, at least two of the sensing

   windings In different half-N'avelengths of the pnrmary winding linking mcre,nenta! areas of the rnagnet1c flux and being offset along the

   GO 5 lengths of the parallel -indmg segments to pro-vice material response measurements over different locations when the circuit is scanned over the test material in a directc?n perpendicular to the extended winding segments and the sense vvindiilgs confined to a single plane and the drive windings 10 confined to a single plane.

   lO5. A test circuit as claimed in Claim ()4 wherein the pa.railel extended winding segments of the primary winding are fondled by parallel extended portions of adjacent drive coils 106. A test circuit as claimed In Claimed 105 wherein every other drive coil is connected In series on one side of the primary And the remammg drive coils are collected In series on the <.A. for = I: wyyvSlL Slow V1 1 111111).

   107 A test circuit as claimed In Claim 05 wherein a. single pair of com. ector leads is connected to each set of primary winding half-wavelengths. 5 1 08. A test circuit as claimed in Claim 104 where the primary windmg is distributed in two planes with extended portions of the primary writings located over one another and the connectors between the extended portions of the primary windings offset by one half-wavelength 109. A test circuit as claimed In Claim 104 where the sensing windings linking incrernlental areas are located in eve.: other half-wavelength while additional sensing windings that extend along a substantial length of the primary wmdngs are 3 5 located in the intermediate halt:wavelengths of the primary winding.

   _ 11 O. A tcsi circuit as Darned in ClanTn 109 where output comlections of at least two of tile sensing windings of substartial length are corrected In series.

   1 11. A test circuit as c]zimled in Claiir 109 where extra conductors 10 are placed m the endmost hall' wavelengths, without secondary elements, to maintain the spatial perodcity of the conductors m the test circuit.

   1]2 test circuit as claimed in Clang 109 where extra conductors 15 are placed at the ends of the sensing elements to maintain the spatial peiodcity of conductors In the test circuit.

   1 13. A test circuit as claimed In Claim 109 where the sensing virldings linking, increTnenta1 areas have extension Mops on 2 o the winding side opposite the collector leads to provide essenrially equval ent coupling of magnetic flux between of'i-'set sensing windings.

   1 14. A test circuit as claimed In Claim 104 where the sensing 2 5 windings are offset a distance one-half the length of extended portions of the sensing windings so that sensed responses cover overlapping areas.

   115. A test circuit as claimed in Claim 105 where the array of 3 0 sensing windings has at least one sensing coil every half wavelength, distance between sensing coils within an adjacent pair of extended portions being iiicreased by the length of the sensing coils at least each half wavelength, each individual sensing coil having separate output connectors.

   116,\ test circuit as claimed in Cla' m 1 15 where the distance from the ends oi'the sensing coils to the connectors between the extended portions of the primary Is kept essentially

   GO 5 constant as the distance between the sensing coils is increased. 117. A test circuit as claimed in (:']aim] 16 where the sensing coils provide an absolute response to the material properties.

   l 18 A test circuit as claimed in Claim l6 where the sensing coils provide a diffcrental response to the material properties para]iel to the extended windings 15 119. A test circuit as claimed in Claim 1 16 where the sensing coils in e-v cry other half wavelength provide an absolute response to the material properties while the sensing coils m the Intermediate half meanders provide a differential response to the Lateral properties para''el to the extended endings.

   l 20 A test circuit as claimed In Claim l l 9 where extra conductors are placed m the endmost half-wavelengths, without secondary windings, to maintain the spatial periodicty of the conductors m the test circuit.

   A test circuit as clanged m Claim l 19 where extra conductors are placed at the ends of the sensing elements to maintain the spatial period city of conductors in the test circuit.

   3 0 192. A test circuit as claimed In Claim 1 19 where e>;:tra conductors are placed between the Inner coils for pairs of differential elerrcnts to ensure that each Individual cod] Is surrounded by the same conductor pattern.

   3 5 123 A test circuit as claimed in CIann 22 where the individual coils for the diffcrenha] elements are separated by more thaTl a coil length

   G' 124. A test circuit as claimed in Claim 119 NvLerc at least onc of the difTereIltal elements is rotated so that the coils are offset equal distances C\7.' either side of the centerline of n half wavelength of a primary meander.

   125. A test circuit as claimed In Claim 124 where the rotated differential clement is perpendicular to the extended windings of the primary 126 A test circuit as claimed Z1 Claim 1 16 where rmeasuremeIlts 15 grids with one or inore dimensions are generated in advance and used as databases to look up and interpolate the electrical and geometric properties of interest at the location measured by each individual sensing element.

   70 127. A test circuit as claimed m Claim 126 where the electrical and geolne.tnc properties al; each sensing element location are correlated with dependent properties of interest.

   128. A test circuit as claimed in Claim 126 where the an-ay Is 5 scanned to build images of electrical properties across the surface of a part.

   129. A test circuit as claimed In Claim 127 where multiple frequencies are used to measure property variations Nvth O depth at each sensing elerr.ent 130. A test circuit as claimed in Claire, 128 where multiple frequencies are used to create three-dimensional Images of properties. 131 A test circuit composing: a meandering primary winding havmg extended portions for Imposing a spatially periodic magnetic field of at least two spatial wavelengths

   in a test substrate; and

   5 a sensing winding array of sensing elements, with a sensing element positioned every hair wavelcngti1 of at least two spatial wavelengths between respective adjacent extended portions of the prir ary winding for sensing the response of the test substrate to the imposed magnetic field, the sensing eleir,ents including sensing coils for fir' wing flux

   0 over incremental areas between adjacent extended portions, at least one absolute sensing coil every half wavelength and at least two diffcrenta.l sensing coils every other half wavelength, distance between sensing coils within an adjacent pair of extended portions being increased each wavelength, each individual sensing cod] having 15 separate output connectors, distances from the ends of the sensing coils to connectors between the extended portions o. the primary being substantially constant as the distance between the sensing coils IS 1nCreaSO6.

   2 0] 32. A test circuit as claimed m Claim 126 wherein the extended portions of the primary wading are termed by individual drive coils having parallel extended portions with every other drive coil connected in series on one side of the primary and the remaining dove coils connected In series on the opposite 2 5 side of the primary.

   133. A test circuit as claimed In Claim 16 that Is confonnable to inspect curved parts.

   3 0 134..N test circuit as claimed in Claim 116 placed on a curved and compliant substrate to Inspect a curved part.

   135 A test circuit as claimed in Claim i 28 Mere one part being scanned is an engine disk slot.

   1.36 test circuit as claimed In Claim 35 where two or more scnsng coils are located at the same position along half-wavelength but In different half-wavelengths of the primacy to provide property measurement redundancy at a

   5 specific test material location with a snglc measurement scan.

   137. A test C"'C'J't as claimed m C'lain1 136 where five sensing coils are located at the sarr e position along a half wavelength but in different half wavelengths of the prirnlary 138 A test circuit as claimed in (claim 117 further composing connection leads to the sensing e].cment that are offset from the centerline o1 the half meander, a second loop having extended portions substantially parallel to the connection leads and symArnetucal]y located on the opposite side of the center]ne, and with a difterental connection between the sensing element and the second loop 139. A test circus, as claimed in C.lailn 119 further comprising connection leads to the sensmg element that are offset frorr the eenterLne of the half Meander, a second loop having extended portions substantially parallel to the connection leads and synmletrical]y located on the opposite side of the centerline, and with a differential connection between he 25 sensing element and the second loop 140 An eddy- current sensor fair creating a spahal]y periodic magnetic Held of at least two periods composing: a primary winding having parallel extended winding segments and 3 0 driven by an electric current: and each half wavelength of the primary w1ndmgs being forrAled by an individual drive coil having parallel extended portions with ever; other dime coil collected in series on one side of the primary and the remaining drive coils connected in senes on the opposite side of the 3 5 primary 141. method for creating a spatially periodic magnetic field, with

   at least two penods, for eddy content sensors comprising

   Al 5 providing a primary winding of- parallel extended winding segments, each parallel extended winding segments being Donned by p.aral]el extended portions of adjacent Individual drive coils; driving current through individual drive coils in alternating clockwise and counterclockwisedirections such that current through 10 immediately adjacent extended portions of the drive coils are in common directions and a spatially periodic magnetic field, with at

   least two periods, is imposed in a test substrate.

   142. A method as claimed in Claim 140 wherein a single pair of 5 connector leads is connected to each set of primary winding halfwavelengths. 143. A method as claimed in Claim] 41 further composing providing parallel extended wir,ding sediments distributed 2 0 over two planes, the extended windings of the primaries located over one another and the correctors between the extended portions of the primary windings offset by one half-wavel ength.

   2 5 144. A test circuit comprising a plurality of adjacent individual dove coils having parallel extended portions with current flow In any adjacent extended portions being in a common direction and all the drive coils together imposing a stationary periodic magnetic field, with at least two periods, in a test

   3 substrate.