Study on Relationship between Amplitudes of Reflection Signals and. Positions of Magnetostrictive Transducer

Transcription

1 17th World Conerence on Nondestructve Testng, Oct 28, Shangha, Chna Study on Relatonshp between Ampltudes o Relecton Sgnals and Postons o Magnetostrctve Transducer Lhua SHEN, Yuemn WANG *, Fengru SUN, Fan ZENG, Chuanjun SHEN College o Shp and Power; Naval Unversty o Engneerng; NO. 717, Jeang Road, Wuhan 4333 Chna Tel: , Fax: E-mal: yuemn_wang@163.com Abstract: As one o key technques or guded waves NDT based on magnetostrctve eect, the transducer s a brdge whch plays an mportant role n practcal nspecton. In the experments, we nd that the guded wave echo-sgnals detected n ppes are aected by the poston o the magnetostrctve transducer. Frst, guded waves generaton mechansm and ts propagatng characterstcs n ppes are theoretcally analyzed. Second, several experments are carred out to very ths theory. At last, the relatonshp between ampltudes o the reecton sgnals and postons o the transducer are dscussed. For a gven detected steel ppe, when the transducer s nstalled on derent poston o ppe, the varetes o guded wave sgnals are detected and analyzed usng computer. The varous curves between ampltudes o the reecton sgnals and postons o the transducer are obtaned,and the wave crest and wave trough appeare perodcally. In the practcal nspecton, the postons where the wave crest appeares n the curves should be selected as locaton o the transducer. To generate and receve the most perect sgnals whose wave energy s stronger and the sgnal-to-nose rato ncreased greatly, the postons o guded wave transducer should be optmally selected accordng to derent type ppes. The expermental results are very well agreed wth the theoretcal analyss. Keywords: Guded wave, magnetostrctve transducer, NDT, ppe 1. Introducton As one new dgtal non-destructve testng technology, the guded wave NDT, whch has the advantage o nspectng the whole wall thckness and spreadng long dstance o ppe rom a sngle probe poston, has been appled wdely n long ppe nspecton [1-3]. At present, two man guded wave testng methods, whch are based on pezoelectrc eect [4] and magnetostrctve eect [5-7] respectvely, are studed mostly by the researchers. These derences are the methods o exctng and recevng guded waves, whle the propagatng characterstcs o guded wave are same. The generaton o guded wave and nspecton o echoes are realzed by the guded wave transducers, and the modes o guded wave generated can also be controlled by the alteraton o transducer s conguraton. In the experments, the guded wave echo-sgnals detected n ppes are changed as the poston o magnetostrctve transducer moves along the axal drecton o the ppe and the ampltudes o echo-sgnals and SNR are also aected by the transducer s poston located at the ppe. The detected sgnal waveor are receved when the transducer s nstalled on derent poston o a gven steel ppe. The varous curves

2 between ampltudes o the reecton sgnals and postons o the transducer are obtaned by analyzng these receved sgnals. 2. Fundamental o magnetostrctve guded wave transducer A tme-varyng magnetc eld s appled to the erromagnetc ppe, whch has been magnetzed by statc bas magnet, and mechancal guded wave s generated along the ppe va magnetostrctve or Joule eect. When the elastc (or mechancal) wave passes by the recevng transducer, the magnetc nducton o the ppe changes because o the nverse-magnetostrctve (Vllar) eect. The changes n magnetc nducton brng about an electrcal voltage sgnals n the recevng transducer va Faraday eect. The nduced electrcal voltage sgnals nvolve the normaton that whether cracks corroson or deects exsted n the steel ppe. The exctng and recevng o guded waves, and the magnetzaton o nspected steel ppe, are all realzed by the magentostrctve transducer. Exctng transducer, recevng col and statc bas magnetc eld, whch make up o the sensor, realze the correspondng uncton respectvely. A bas magnet apples a DC magnetzaton to the steel ppe and mantans erromagnetc materals n a magnetzed state. The DC magnetzaton s necessary to enhance the ecency o the sensor, whch converts electrcal energy nto mechancal energy, and mantan a lnear transducton between the electrcal sgnals and the elastc waves [7,8]. When the gude wave propagates along the steel ppe, there are three modes: longtudnal mode L(,m), torsonal mode T(,m), lexural mode F(n, m). When the bas magnetc eld s parallel to the col magnet eld, longtudnal guded wave can be generated along the erromagnetsm ppe. The general elastc wave equaton wth a magnetostrctve orcng term s expressed as Eq.(2.1) [9]. ρ& = σ (2.1) u k em + x + k where ρ s the densty o the metal; u& s the -th elastc wave dsplacement s the magnetostrcton orce; em s the electromagnetc orce; σ k represents the k-th element o the elastc stress tensor; x k s the Cartesan coordnate. In our experment, the low requences o the guded wave (up to a ew hundred khz) are excellent or nspecton [8]. At low requences, s gnored n the equaton (2.1) generally. >> em. So the em The magnetostrctve orce z can be calculated by Eq. (2.2): z mz = 1 (3 ˆ λ + 2 ˆ)(1 µ 2υ ) λ (2.2) 2 M z Where λˆ µˆ are Lame coecent, υ s Posson s raton, λ s lnear magnetostrctve constant, M s the statc magnetzaton, m z s the alternatng magnetzaton. λu l r jk ( ξ x+ ct ) u( k, x, t) = H ( ξ ) e dξ 2E (2.3) Where E s Young s modulus, c = E ρ, the velocty o the longtudnal sound wave n the medum.

3 H ( x) s the magnetc eld dstrbuton uncton, l s the length o the recevng col, µ r s the reversble permeablty, The voltage detected by the recevng col can be wrtten as ollows: d l u V k t r ns + 2 (, ) = 4πµ λ e( x) dx (2.4) l x t Where e(x) s the ecency rato o the recevng col; n s the crcles o recevng col per mle, s s the cross secton area o col. The above equatons are ntegrated together to the undaments o the magnetostrctve longtudnal transducer. From above unctons, we can see that the transducer poston d s one o actors whch aect the recevng voltage o guded wave. In order to nsure the best nspectng poston, the next two parts o ths artcle would be devoted to dscussng the eect to recevng o guded wave because o the transducer poston. 3. Expermental arrangements A steel ppe o nsde dameter 33mm, outsde dameter 38mm and length 6.5m was used as the nspectng object. There were three artcal deects o derent dameter(φ5mm, φ1mm, φ12mm) n the expermental ppe as shown n Fg.1. φ 5mm φ 12mm φ 1mm Steel ppe A B Magnetostrctve transducer PC 2. 2m 1. 1m 1. 1m 2. 1m MsS guded wave NDT system Fg. 1 The sketch o steel ppe used n experment Fg. 2.The schematc dagram o the expermental set-up The above ppe was nspected expermentally by the magnetostrctve guded wave NDT system developed by us. The undamental rame had been detal dscussed n [6, 7] and a schematc dagram o the expermental set-up s shown n Fg.2. The 2kHz requency guded wave was generated n the ppe by makng use o the guded wave NDT system. The magnetostrctve transducer poston changed along the ppe by 1cm per step relatve to the end A as shown n Fg. 1. The correspondng guded wave sgnals, when the transducer located at derent poston o ppe, were receved by the mangentostrctve sensor. 4. Results and dscussons The representatve relected guded wave sgnals, when the transducer located at the poston d=cm d=9cm d=13cm d=17cm d=34cm d=16cm rom the end A, were gven as shown n Fg.3. By analyzng the normaton gven by the waveor, we can see that the receved sgnal o relected guded wave vared markedly when the transducer was nstalled on the derent detectng poston o ppe. The dssmlartes were manly the ampltude o relected sgnal and SNR mostly. As the mangentostrctve sensor located at the optmal detected poston, the ampltude o sgnals was bgger and vvd to dstngush the law sgnals o derent dameter as shown n Fg. 3-a Fg. 3-c. However, as shown n Fg. 3-b and

4 Fg.3-e the sgnals ampltude and the SNR decreased obvously to dstngush the law sgnal dcultly when the sensor located at the poston where t was unavorable or detectng. The relected sgnals o two ppe ends shown n Fg.3-, whch receved by the transducer, covered the deect sgnals completely as the transducer moved to the mddle o ppe gradually. The exstence o detects can t be dstngushed n the sgnal waveor. (a) d=mm (b) d=9mm (c) d=13mm (d) d=17mm Fg. 3 (e) d=34mm () d=16cm Detected guded wave sgnal waveor when transducer nstalled on the derent poston o ppe Fg. 4 The varous curve between the ampltudes o ppe end Fg. 5 The varous curve between the ampltudes o detect sgnal

5 relected sgnal and the transducer s poston and the transducer s poston The relaton curve between ampltude o ppe end relected sgnals and the poston o the transducer was shown n Fg.4. It ndcated that sgnals ampltude presented the collapsng trend as a whole wth the changng dstance between transducer and ppe end A. The wave crest and wave trough appeared perodcally. As shown n Fg. 5, the relaton curve between the ampltude o Φ12mm deect echo-sgnals and the changng dstance d was smlar to trend expressed n Fg. 4. Durng the changng process, deect echo-sgnals, the wave crest and wave trough appeared perodcally n the same way. The deal waveor should be smlar to Fg. 3-a and Fg. 3-c n the practcal engneerng, n whch the ampltude o sgnals s hgh, and the sgnal-nose rato s also hgh. Whether the deects exst n the ppe can be convenently and clearly dered rom the waveor by analyzng experment results wth the computer. It s concluded that the poston o the magnetostrctve transducer had the great nluence on the response o guded waves NDT. For ths reason, the characterstc o the magnetostrctve transducer should be consdered adequately n order to mprove the testng eectveness and the testng response. I the poston conguraton o the transducer s mproper, t may cause the undetected error and dspensable loss. The ppe end sgnals cover the deect sgnals as denoted by the waveorm o Fg.3-, and t s dcult to dstngush deect sgnals rom other sgnals. We should optmze constructon o transducer to control the propagaton path o guded wave, whch only receve the echo sgnals rom one drecton and constran the sgnals rom the other drecton, so that t s convenent or analyzng the sgnals and recognzng all knds o sgnals. Accordng to the regulaton o sgnal ampltude appearng perodcally that denoted by the Fg. 4 and Fg. 5, the proper transducer s poston, where the wave crest appeared, should be selected as the detectng regon to gan the best waveor whch had the hghest ampltude and best SNR. 5. Conclusons The transducer poston s one o actors aect the recevng voltage o detected guded wave sgnals by analyzng the undaments o exctng and recevng guded wave. Plenty o experments are conducted to prove the consequences n ths artcle. The experment results agree very well wth the theoretcal concluson. For a gven steel ppe, the varetes o guded wave sgnals are detected and analyzed usng computer when the transducer was nstalled on derent poston o ppe. The varous curves between ampltudes o the reecton sgnals and postons o the transducer are obtaned by analyzng expermental data. In the curve the wave crest and wave trough appeared perodcally. At the same tme, the ppe end relected sgnals covered the deects normaton when the sensor moved to the mddle o gven detected ppe. In ths case t was hard to dstngush the laws. In the practcal nspecton, the postons where the wave crest appeared n the curves should be selected as locaton o the transducer. To generate and receve the most perect sgnals whose wave energy s stronger and the sgnal-to-nose rato ncreased greatly, the postons o guded wave transducer should be optmally selected. Reerences [1] M.J.S. Lowe, D.N.Alleyne, P.Cawley. Deect detecton n ppes usng guded waves [J]. Ultrasoncs, 36(1998): [2] Y.M. Cheong, D.H. Lee, H.K Jung. Ultrasonc guded wave parameters or detecton o axal cracks n eeder ppes o PHWR nuclear power plants[j]. Ultrasoncs, 42(24): [3] Kwun H and Holt C M. Long range guded wave nspecton o ppe usng the magnetostrctve sensor techbology-easblty o deect characterzaton[c]. SPIE. 3398(1998): [4] J. Barshnger, J. L. Rose, Mchael J. Avol, Jr. Guded wave resonance tunng or ppe nspecton[j]. Journal o Pressure Vessel Technology, 124(22): [5] Wang Yuemn, Kang Yhua, Wu Xnjun. Theoretcal and expermental study on generaton o longtudnal gudeded

6 waves n crcular ppes based on magnetostrctve eect[j]. Chnese Journal o Mechancal Engneerng 25, 41(1) : (In Chnese). [6] Wang Yuemn, Kang Yhua, Wu Xnjun. Applcaton o magnetostrctve generaton o guded wave to non-destructve testng o ppes[c]. Proceedngs o the 7 th Internatonal Conerence on Electronc Measurement & Instruments. 25, Vol.5: [7] Shen Lhua, Wang Yuemn, Sun Fengru. Numercal Smulaton and Expermental Research o Guded Waves Based on Mangentostrctve Eect [J]. Chna Mechancal Engneerng. 18(27): (In Chnese) [8] Wang Yuemn, Kang Yhua, Wu Xnjun. A model or magnetostrctve sensor generaton o longtudnal guded waves n a hollow cylnder. Conerence Proceedngs o the 7 th Internatonal Conerence on Electronc Measurement & Instruments. 25, Vol.5: [9] Boltachev V D, Pradn L S, et al. Electromagnetc-acoustc exctaton n erromagnetc ppes wth a crcular secton[j]. Russan. J. NDT. 25(199):