Advances in eddy-current non-destructive evaluation

318 Ad ances in Elec ical and Elec onic Enginee ing ADVANCES IN EDDY-CURRENT NON-DESTRUCTIVE EVALUATION L. Janoušek, D. Gombá ska, K. ápo á Depa men o Elec omagne ic and Biomedical Enginee ing, Facul y o Elec ical Enginee ing, Uni e si y o Žilina, Uni e zi ná 1, 010 26 Žilina, Slo ak Republic, el.: +421 41 5135062, e-mail: janousek@ el.uniza.sk, gomba ska@ el.uniza.sk, capo a@ el.uniza.sk Summa y The pape p esen s a gene al o e iew o ecen de elopmen s in he eddy-cu en non-des uc i e e alua ion. Basic p inciple o he eddy-cu en non-des uc i e es ing is explained and possible applica ions o he me hod in non- in asi e e alua ion o conduc i e ma e ials a e summa ized. Ac ual issues o esea ch and de elopmen in his ield a e discussed. A no el me hod o dep h e alua ion o a de ec ed c ack using eddy-cu en es ing signals is p oposed. 1. INTRODUCTION Non-des uc i e e alua ion (NDE) o ma e ials is de ined as a g oup o ac i i ies aiming o he in es iga ion and he cha ac e iza ion o ma e ials as well as s uc u es wi hou hei mechanical damage. Di e en physical p inciples and phenomena a e u ilised o he NDE o ma e ials. One o he elec omagne ic me hods, o igina ing om he p inciple o elec omagne ic induc ion, is an eddy cu en non-des uc i e es ing (ECT) o conduc i e ma e ials. Inc easing ends o u iliza ion o he me hod in p ac ical inspec ions, especially in he las pe iod, impose g adually aising demands on esea ch and de elopmen (R&D) in his ield. The pape summa izes ecen ad ances in eddy- cu en non-des uc i e e alua ion. A e he p inciple o he me hod is explained, possibili ies o i s u iliza ion in di e en applica ions a e lis ed. Ac ual issues o R&D ac i i ies in his ield a e summa ized and discussed. A no el app oach in he eddy-cu en non-des uc i e e alua ion is p oposed and he p esen ed esul s p o e i s e ec i eness in dep h e alua ion o a de ec ed de ec . 2. PRINCIPLE OF ECT The p inciple o he ECT, shown in Fig. 1, unde lies in he in e ac ion o induced eddy cu en s wi h a s uc u e o an examined body [1], [2]. Fig. 1 P inciple o ECT A p ima y al e na ing exci ing elec omagne ic ield is gene a ed in he icini y o a coil d i en by an al e na ing cu en . Elec omo i e o ce is induced in a conduc i e objec which is in p oximi y o he coil and eddy-cu en s low he e acco ding o he elec omo i e o ce. A seconda y elec omagne ic ield gene a ed by he eddy-cu en s coun e wo ks o he p ima y exci ing elec omagne ic ield. The induc ion coupling he e o e exis s be ween he coil and he conduc i e objec . I can be simply conside ed as an in e ac ion be ween he p ima y and he seconda y elec omagne ic ield. The esul ing elec omagne ic ield o he coil and he conduc i e objec depends on geome ical pa ame e s o he sys em as well as on he elec omagne ic pa ame e s o he conduc i e objec . Fo he gi en exci a ion, i.e. con igu a ion, dimensions and o ien a ion o he coil(s) and i s eeding, he coupling is in luenced by he ollowing impo an pa ame e s: • posi ion o he coil wi h espec o he objec , • geome ical con igu a ion o he objec , • dimensions o he objec , mainly i s hickness, • he elec omagne ic pa ame e s o he objec (conduc i i y, pe meabili y), • na u e o he objec (homogenei y, linea i y, aniso opy). U iliza ion o he ECT in p ac ical applica ions depends on a possibili y o de ec luc ua ions in he esul ing elec omagne ic ield due o changes in he impo an pa ame e s. The ECT is he e o e applied in: • hickness measu emen s o conduc i e ma e ials, • hickness measu emen s o non-conduc i e coa ings on conduc i e ma e ials, • measu emen s o he elec omagne ic pa ame e s (conduc i i y, pe meabili y) o conduc i e ma e ials, • e i ica ion o conduc i e ma e ial ea men , • e i ica ion o selec ed pa ame e s o p oduc s (dimensions, e c.), • de ec ion and e alua ion o discon inui ies (de ec s) in conduc i e ma e ials, e c. Rising employmen o he ECT in di e en echnical applica ions imposes new challenging appeals on R&D ac i i ies. Ad ances in eddy-cu en non-des uc i e es ing … 319 3. ACTUAL ISSUES OF R&D ACTIVITIES The p inciple o he ECT has been known o se e al decades. Nowadays, he mos wide sp ead applica ion a ea o he ECT is he de ec ion and possible e alua ion o di e en discon inui ies in conduc i e ma e ials. The ha dwa e and he so wa e means o he ECT can be di ided in o ollowing g oups: • ECT ins umen s, • ECT p obes, • e alua ion o measu ed signals, • posi ioning sys ems and manipula o s. ECT ins umen s supply exci ing coils o he p obes and sense he ECT signals. F equen ly, se e al supe imposed ha monic exci ing signals a e used o d i e he eddy cu en s. I helps o gain in o ma ion abou a de ec unde se e al equencies a he same ime. Mo eo e , be e sepa a ion be ween he use ul signals and backg ound noises can be ob ained. Nowadays, he R&D ac i i ies a e inc easingly ocused on a pulse exci a ion [3]. I s ad an age compa ing o he con en ional ha monic exci a ion is in a wide equency spec um esponse ob ained wi h one exci ing impulse [4]. ECT p obes a e one o he mos impo an elemen s in he non-des uc i e es ing, because hey ans e in o ma ion be ween an ECT ins umen and a conduc i e objec h ough he induc ion coupling. Usually, induc ance coils a e u ilized o build ECT p obes. Howe e , magne ic senso s a e employed in sensing low in ensi y ields [5]. Fea u es o ECT p obes depend on numbe , shape, con igu a ion, o ien a ion, dimensions and connec ions o coils as well as on pa ame e s o a magne ic ci cui . Op imal ECT p obe should assu e [6]: • high sensi i i y o expec ed de ec s, • high p obabili y o de ec ion o expec ed de ec s, • possibili y o dis inguish pa ame e s (loca ion, dimensions, e c.) o expec ed de ec s. Many ypes o ECT p obes ha e been de eloped e lec ing special demands o pa icula applica ions [7]-[9]. ECT p obes wi h e omagne ic co es a e some imes u ilized o p ecise localiza ion o he elec omagne ic ield [10]. Howe e , he non- linea i y o he co es b ings ano he unknown in o he e alua ion. Sensed signal is he in eg al alue. I s phase depends mainly on a c ack dep h. Thus, a c ack dep h can be oughly es ima ed om he signal phase in o ma ion. Sol ing o in e se p oblems ep esen s mo e sophis ica ed app oach in c ack e alua ion. Selec ed pa ame e s o de ec s (dep h, leng h, posi ion, p o ile, e c.) a e econs uc ed based on he measu ed signals. Two app oaches a e u ilized o he pu pose [11]: • de e minis ic, • s ochas ic. Usually, wo a iables o a de ec a e es ima ed, i s dep h and leng h, while a p o ile, a wid h and elec omagne ic pa ame e s o a de ec a e adjus ed in ad ance. A di e ence be ween measu ed and simula ed signals is minimized in he de e minis ic me hods [12]. The p ocess is i e a i e and he e o e la ge numbe o o wa d simula ions is equi ed. Da abases o p e-compu ed signals [13] as well as pa allel compu ing on supe compu e s [14] can help o sho en he e alua ion ime. The s ochas ic me hods u ilize so called e olu ion algo i hms, o example neu al ne wo ks, gene ic algo i hms, o he in e sion [15], [16]. A p esen , i is possible o de ec e en e y small a i icial o a igue c ack and qui e p ecisely es ima e i s dep h in homogeneous o non- homogeneous s uc u es [17]. Cu en esea ch ac i i ies a e mainly ocused on he ollowing a eas: • design, de elopmen and op imisa ion o ECT p obes o sa is y se e e demands o non- des uc i e inspec ion o s uc u es wi h eal de ec s, • eliable de ec ion and localiza ion o unknown and in mos cases in isible de ec s wi h a iable o ien a ion, pa ame e s, p o ile and s uc u e, • p ecise es ima ion o main pa ame e s o a de ec , especially i s leng h and dep h, • new possibili ies o p ac ical applica ions o he ECT. The issues lis ed abo e a e closely ela ed o each o he and hey a e de e mined by ac ual p oblems and p ac ical expe iences. 4. NOVEL EVALUATION METHOD ECT signals a e in eg al alues and hey do no ca y explici in o ma ion abou he c ack dimensions. The e o e, e alua ing he dep h o a de ec om he ECT signals is qui e di icul [18]. In addi ion, he skin-e ec concen a es induced cu en s on he su ace o a es ed ma e ial which means ha ob aining in o ma ion abou he dep h is essen ially di icul . This sec ion p oposes a no el me hod o a c ack dep h e alua ion om he ECT signals. A new p obe [19] shown in Fig. 2 is used o he inspec ion. The p obe consis s o ou ec angula exci ing coils and a ci cula de ec ing coil placed in he cen e be ween he exci ing coils. The exci ing coils a e o ien ed angen ially ega ding he su ace o a es ed sample and he pick-up coil has he no mal o ien a ion. The inne exci ing coils and he ou e ones a e connec ed in se ies, espec i ely. A de ec ed c ack is inspec ed wice. A i s , only he inne exci ing coils o he p obe a e d i en and only he ou e exci ing coils a e d i en du ing he second inspec ion. Di e en dis ances o he inne and he ou e exci ing coils om he de ec ing coil assu e a di e en p o ile o he eddy cu en densi y along he ma e ial dep h unde he de ec ing coil. 320 Ad ances in Elec ical and Elec onic Enginee ing Fig. 2 Pic u e o a new p obe The p obe scan igh o e he c ack along i s leng h; he windings o he exci ing coils a e pe pendicula o he c ack leng h. The wo sensed signals o a same c ack gained du ing wo inspec ions a e linea ly supe imposed acco ding o he ollowing ela ions: 2211 2211 ImImIm ReReRe ⋅−⋅= ⋅−⋅= CC CC , (1) whe e Re 1 , Re 2 a e he eal pa s o wo complex signals; Im 1 , Im 2 a e he imagina y pa s o wo complex signals; and C 1 , C 2 a e eal posi i e numbe s de ining a a io o he supe posi ion: 21 CC= α , (2) whe e max (C 1 , C 2 ) = 1.0. The phase o he esul ing supe imposed complex c ack signal ϕ co esponding o he maximum absolu e alue is e alua ed wi h espec o he alue o he a io. The phase is gi en by he ollowing equa ion: m m Re Im a c an= ϕ , (3) whe e Re m and Im m a e he eal and he imagina y pa s o he supe imposed c ack signal co esponding o he maximum absolu e alue o he signal. A pla e specimen made o he s ainless s eel SUS316L is inspec ed o e alua e e ec i eness o he p oposed me hod. The elec omagne ic cha ac e is ics o he ma e ial include conduc i i y o σ = 1.4 MS/m and ela i e pe meabili y o µ = 1. The hickness o he specimen is d = 25 mm. Non- conduc i e su ace b aking no ches o he ec angula shape wi h a leng h o l c = 40 mm, a wid h o w c = 0.5 mm and dep hs o d c = 10, 12, 15, 20 mm model he c acks. The equency o 10 kHz is adop ed in he inspec ion. Figu e 3 illus a es dependences o he supe imposed c ack signal phase change on he a io o supe posi ion o ou c acks. I can be obse ed ha he phase o he supe imposed signal changes wi h inc easing he a io o supe posi ion, while his change depends on he c ack dep h. Thus, a unique ea u e alue o he a io o supe posi ion can be ex ac ed om he dependence gained o one c ack. I is a alue o he a io whe e he c ack signal o a es in a hal angle o i s o e all o a ion. 0 20 40 60 80 100 120 140 160 180 0 0.5 1 1.5 2 2.5 3 phase change [deg ee] a io [-] dc=10mm dc=12mm dc=15mm dc=20mm Fig. 3 Dependences o he supe imposed c ack signal phase change on he a io o supe posi ion o he c acks wi h dep hs o dc = 10, 12, 15 and 20 mm 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 5 10 15 20 25 a io [-] c ack dep h [mm] simula ion measu emen Fig. 4 Dependences o he ea u e alue o he a io o supe posi ion on he c ack dep h The dependences o he ea u e alue o he a io o supe posi ion on he c ack dep h gained om nume ical as well as expe imen al signals a e shown in Fig. 4. The nume ical esul s a e plo ed o a wide ange o he c ack dep h, om 0% o 100% o he ma e ial hickness d = 25 mm, o highligh ad an ages o he p oposed me hod. As i can be seen, he e is a unique dependence be ween he ea u e alue o he a io o supe posi ion and he inne exci ing coils de ec ing coil ou e exci ing coils Ad ances in eddy-cu en non-des uc i e es ing … 321 c ack dep h. I means ha he dep h o an inspec ed de ec can be de e mined om he dependence based on he ea u e alue o he a io ob ained om measu ed signals acco ding o he s eps desc ibed abo e. Mo eo e , he nume ically gained dependence is almos linea wi hin he in es iga ed ange and hus, he dep h o su ace b eaking de ec s ha a e much deepe han he s anda d dep h o pene a ion ( δ = 4.25 mm in his case) can be unambiguously e alua ed using he me hod. 4. CONCLUSION The pape deal wi h ecen ad ances in eddy- cu en non-des uc i e e alua ion. P inciple o he me hod was explained in he beginning in o de o demons a e a ie y o possible applica ions. Recen esea ch and de elopmen ac i i ies in his ield we e summa ized. No el me hod o non-des uc i e e alua ion o a c ack dep h using eddy-cu en es ing signals was p oposed. The nume ical and he expe imen al esul s p o ed he e ec i eness o he me hod e en o su ace b aking c acks ha a e much deepe han he s anda d dep h o pene a ion. Acknowledgemen This wo k has been suppo ed by he g an VEGA No. 1/0308/08 o he Slo ak Minis y o Educa ion. REFERENCES [1] Sun, Y.: An in oduc ion o elec omagne ic nondes uc i e es ing. P oceedings o he ISEM’97 con e ence, IOS P ess, 1998. pp. 145-152. [2] Janoušek, L., Ma ek, T., Gombá ska, D.: Eddy cu en non-des uc i e e alua ion o conduc i e ma e ials. Communica ions, No.1, 2006, pp. 29-33. [3] Sophian, A. e al.: Design o a pulsed eddy cu en senso o de ec ion o de ec s in ai c a lap-join s. Senso s and Ac ua o s, Vol.101/2002, pp. 92-98. [4] Tsuboi, H., Seshima, N: T ansien eddy cu en analysis o pulsed eddy cu en es ing by ini e elemen me hod. IEEE T ansac ions on Magne ics, Vol.40/2004, pp. 1330-1333. [5] Uesaka, M. e al.: Eddy-cu en es ing by lexible mic oloop magne ic senso a ay. IEEE T ansac ions on Magne ics, Vol.34/1998, pp. 2287-2297. [6] Chen, Z., Miya, K.: A new app oach o op imal design o eddy cu en es ing p obes. Jou nal o Nondes uc i e E alua ion, Vol. 17/1998, pp. 105-116. [7] Ku okawa, M., Miyauchi, R., Enami, K.: New eddy cu en p obe o NDE o s eam gene a o ubes. Elec omagne ic Non des uc i e E alua ion, Vol.3/1999, pp. 57-64. [8] Huang, H. e al.: Design o an eddy-cu en a ay p obe o c ack sizing in s eam gene a o ubes. NDT&E In e na ional, Vol.36/2003, pp. 515-522. [9] Janoušek, L., ápo á, K, Yusa, N., Miya, K.: Ad anced p obe wi h a ay o pick-up coils o imp o ed c ack e alua ion in eddy-cu en non-des uc i e es ing. P oceeding o abs ac s om The 12 h In e na ional Wo kshop on Elec omagne ic Nondes uc i e E alua ion ENDE 2007, Ca di – UK, June 19–21, 2007, pp. 33. [10] Mo ozo , M., No o ny, P.: E alua ion o eddy cu en p obes based on local ield exci a ion. NDT&E In e na ional, Vol.35/2002, pp. 147- 153. [11] Ca delli, E. e al.: New pe spec i es in non- des uc i e elec omagne ic diagnosis. In e na ional Compumag Socie y Newsle e , Vol.11/2004, pp. 7-21. [12] Chen, Z. e al.: In e sion echniques o eddy cu en NDE using op imiza ion s a egies and a apid 3D o wa d simula o . In e na ional Jou nal o Applied Elec omagne ics and Mechanics, Vol.20/2004, pp. 179-187. [13] Chen, Z. e al.: Recons uc ion o c ack shapes using a newly de eloped ECT p obe. Elec omagne ic Nondes uc i e E alua ion, Vol.2/1998, pp. 225-232. [14] Yusa, N. e al.: La ge-scale pa allel compu a ion o he econs uc ion o na u al s ess co osion c acks om eddy cu en es ing signals. NDT&E In e na ional, Vol.36/2003, pp. 449-459. [15] Rao, B.P.C: An a i icial neu al ne wo k o eddy cu en es ing o aus eni ic s ainless s eel welds. NDT&E In e na ional, Vol.35/2002, pp. 393-398. [16] Koyama, K. e al.: Fundamen al s udy o law es ima ion in eddy cu en es ing using gene ic algo i hm. On CD om he 16 h Wo ld Con e ence on NDT, 2004. [17] Yusa, N. e al.: Applica ion o eddy cu en in e sion echnique o he sizing o de ec s in Inconel welds wi h ough su aces. Nuclea Enginee ing and Design, Vol.235/2005, pp. 1469-1480. [18] Yusa, N. e al.: Cau ion when applying eddy cu en in e sion o s ess co osion c acking. Nuclea Enginee ing and Design, Vol.236/2006, pp. 211-221. [19] Janoušek, L. e al.: Exci a ion wi h phase- shi ed ields - enhancing e alua ion o deep c acks in eddy-cu en es ing. NDT&E In e na ional, Vol.38/2005, pp. 508-515.