A NOVEL ULTRASONIC PHASED ARRAY FOR THE NONDESTRUCTIVE

Transcription

1 A NOVEL ULTRASONIC PHASED ARRAY FOR THE NONDESTRUCTIVE EVALUATION OF CONCRETE STRUCTURES Lawrence Azar and Shi-Chang Wh Department f Civil and Envirnmental Engineering Massachusetts Institute f Technlgy Cambridge, MA INTRODUCTION The prus, inhmgeneus, and attenuative nature f cncrete requires the use f lw frequency ultrasnic nndestructive methds fr the purpse f cnditin assessment due t the disturbed wave characteristics in the medium. Current ultrasund techniques are limited by a small aperture width, resulting in minimal reslutin and static directivity. In rder t vercme such deficiencies, a nvel lw-frequency ultrasnic phased array has been develped and its feasibility was tested fr the assessment f cncrete and ther cementitius materials. A phased array allws fr dynamic steering and fcusing capabilities, which cannt be realized with current ultrasnic methds. The enlarged near field zne resulting frm the immense size f the array, which is a cnsequence f the use f lw frequency, authrs in a "dead zne" where steering is nt effective. It was shwn theretically [1] that fcusing is advantageus ver steering in the near field and the characteristics f fcused beams in the field appraches that f simple steering. A systematic apprach using an autmated test assembly was used t experimentally demnstrate the perfrmance f steering and fcusing f the array in cementitius media. The experimental results agree very well with the numerical simulatin presented in the cmpanin paper [1]. It was shwn that the phased array culd be used as a primary imaging and scanning device fr large-scale cncrete structures. EXPERIMENTAL SETUP A nvel apprach fr experimental study f steering and fcusing behavir f the array in a cementitius medium was undertaken. This sectin discusses the system used fr the experimental characterizatin including the array transducer, the system cmpnents that drive the array, and the autmated cntrl and acquisitin assembly. Review f Prgress in Quantitative Nndestructive Evaluatin. Vl. 18 Edited by Thmpsn and Chimenti. Kluwer AcademiclPlenum Publishers,

2 System Develpment A device was designed and assembled fr calibrating and characterizing the phased arrays used fr assessing cementitius materials [2]. The directivity patterns fr bth steering and fcusing were btained using the assembly illustrated in Figure 1. A disk-shaped cncrete (r mrtar) wheel, with an array psitined within the grve, rtates abut its center, and the transmitted energy is received by a statinary receiver lcated at the base f the assembly. Detectin f energy alng the periphery is recrded as the wheel is rtated frm the full -90 t 90, maintaining a set delay fr the measurements. As the receiver is nt in cntact with the wheel, partially immersing the material in water is required fr cntinued transmissin f the energy. The cncrete and mrtar wheels fabricated measure cm in radius and 6.35 cm in thickness. Selectin f the disk's radius was based n the apprximate near field frmula given in the cmpanin paper. This yields a ZT R f 11.5 cm fr 8 elements, and a ZT R f 50.0 cm fr 16 elements in cncrete. As we were interested in measuring directivities in bth the near and far fields t examine the benefit f fcusing, a directivity length (R) f cm seemed t be a gd medium distance. The thickness f the disk prvided enugh clearance s that the first received signal was nt interfered with any sidewall reflectins. The bttm half f the disk was placed inside a Plexiglas water tank, with the water level set slightly abve the rim. A custm-fabricated 150 khz fixed-fcused immersin transducer, situated directly beneath the center f the disk, was aimed at the rim t receive the transmitted energy. Phase steering and fcusing effect was prduced by exciting each element at predetermined time delays. The delayed pulses used t excite the transducer were created using a 16-channel independently prgrammable multiplexing circuit. The initial TTL signal was furnished using a functin generatr (HP 33120A) and dispersed t delay banks which are cmpsed f tapped digital delay lines and multiplexers. The delays fr each Figure I. Cncrete wheel in the water tank with a fcused immersed transducer placed at the bttm fr receiving the ultrasund, and the phased array transducer placed within the center f the wheel. 2154

3 channel were adjustable in 5 ns steps up t a maximum delay f 20 JLs. A 192-bit prgrammable digital I/O bard (Cyber Research ) was used t interface the cntrlling cmputer t the inputs f the multiplexers. Each delayed TIL signal subsequently triggered a high vltage negative spike pulse which was issued t a specific array element thrugh an individually shielded cable. As will be shwn, directivity plts will be attained at a distance (R) f cm fr steering and fcusing, the latter als requiring a fcal length (F) f cm. The maximum angle attainable, fr an array with 16 elements, is limited t apprximately 15. T increase the angular capabilities, 8 elements were used, while maintaining the inter-element spacing when using 16 elements. This enabled steering and fcusing capabilities up t apprximately 40. A piezelectric ceramic, lead zircn ate titanate (PZT), was utilized in the array fabricatin. With a thickness f 0.75 mm, the resnant frequency f the ceramic was apprximately 270 khz. It shuld be nted that with the cmbinatin f backing, ceramic, and lad, the frequency was reduced t 140 khz and 170 khz in cncrete and in mrtar, respectively. The elements f the array were elastically islated by fully cutting thrugh the ceramic, int the backing, which was a mixture f tungsten pwder and epxy resin. Fr mst f the experimental results (i.e., directivities), n matching layer was needed fr the type f data required, althugh a prtective layer f Kaptn tape was applied. The array used in the tests has a required lateral dimensin (L) f 1.6 cm t fit in the testing assembly. Althugh this dimensin is t small fr the wavefrnts t behave as a line surce in the lateral dimensin, it will be shwn that this did nt deleteriusly affect the results. There were 32 elements cut, resulting in an inter-element spacing (d) f cm, and an element width (a) f 0.52 cm. The experiments were made using every ther element, which resulted in a d f cm. With a frequency f 140 khz measured fr cncrete, the spacing resulted in a d f 0.582A. With a frequency f 170 khz measured fr mrtar, the spacing resulted in a d f 0.7 A. (Remember that the ptimum spacing t avid grating lbes was apprximately 0.5A). The verall dimensin D was 23.3 cm. Test Prcedure There are essentially three cntrl paths in the verall cnnectivity f the varius cmpnents and sub-assemblies used in this experiment which include (1) ultrasnic beam cntrl, (2) mtin cntrl and (3) data acquisitin. A central cntrlling PC enabled and mnitred these rutines thrugh a 192-bit parallel I/O bard, RS-232 serial prt and GPIB, respectively. The I/O bard was interfaced t the delay circuit mdule and selected the inputs f an array f multiplexers as lgically high r lw. These bit patterns were used t create varius time delays s that the ultrasnic beam was steered r fcused. The phased array was psitined in the grve f the circular calibratin fixture whse rtatin was cntrlled by the step mtr. The transmitted sund field was detected by a fcused transducer and the signal was amplified with an external receiver unit (Panametrics 5072PR) and displayed using a digital scillscpe (Tektrnix TDS-21O, 1 GHz sampling rate) which was synchrnized with the rtatin f the wheel at every step. Each time the wheel is rtated at a set increment f 0.5, 2155

4 a wavefnn appearing within the gate is acquired frm the scillscpe, and stred. Once the assembly has rtated a full 180, all the acquired wavefnns are prcessed t attain their respective maximum amplitude (psitive r negative). These maxima are stred in a data file, and directivities are pltted. RESULTS AND DISCUSSION This sectin prvides the results f the experimental and simulated parameter study and discuss the directivity curves btained fr varius steering and fcusing angles. Experimental results were initially btained fr mrtar, fr an array with 8 elements and 16 elements, and then fr a cncrete medium. The experimental directivities shw excellent agreement quantitatively with the predicted steering and fcusing characteristics. Figures 2 and 3 shw the measured and predicted directivities. Fr all figures btained, experimental directivity curves fr steering were cmpared t their respective predictins. With a directivity btained at a fixed distance R = cm, the behavir f beam directivities within the near field and far field f the array was bserved by changing the verall lateral dimensin D f the sensr, nting that the same d was utilized thrughut. Fr N = 8, experimental results in mrtar fr steering were attained fr 0, 15, and 25. With cncrete, an angle f 40 was als btained. Fr N = 16 an N = 8, experimental results in cncrete fr steering and fcusing were acquired fr 0. The majr bservatin drawn frm Figs. 2 and 3 is that phase steering and fcusing is nt nly pssible in a cementitius medium, but it des s with excellent agreement t predicted thery. As mentined earlier, directivity is the key parameter used t measure the accuracy and steerability f a phased array transducer. Gd accuracy implies a clse match f the main lbe width f the experimental results t the simulatin directivity. A ntatin fr the main lbe sharpness factr q-6db will be utilized t evaluate the accuracy f all the experimental results. q-6db is an angular measure f the main lbe width at -6 db (50%) nnnalized by 1T (180 ). Steerability is a criterin used t evaluate the crrelatin f the measured Os t the expected Os. Values fr q-6db. 180 and Os were cllected frm all experimental and simulated directivities btained (nt just thse shwn in Figures 2 t 3), and summarized in Table 1. These values are given fr steering, fllwed by fcusing, fr each representative parameter. Several bservatins can be inferred frm the infnnatin summarized, and frm the experimental directivity curves. Cmparing the expected Os t thse cllected frm the experimental results, the array is prven t exhibit great steerability, with an average variatin f nly Withut prper steerability, angular errrs will exist when imaging the received wavefnns. This array has great steerability, as gd as any cmmercially available high frequency arrays used in current medical and NDE applicatins. Als bserved, is the negligible steerability differences between steering and fcusing. With the case f steering in the near field, the resulting pr accuracy negated the pssibility f attaining steerability values. These were marked as "Nt Applicable." The reasn fr attaining q-6db values was t bserve the accuracy f the experimental directivities as cmpared t the simulatin. Excluding steering in the near 2156

5 .E 0.7 -g 0.5 0:1 E _. : ,.....ajwi'i..u_ 0.0 '----'--'-'-----' '---'-' Angle, (J, (deg.) 1.0 r ,r , :E Q) 0:1 E " " " " - - " ' > ' Angle, (J, (deg.) (a) 1.0 r r-----, :E Q) 0:1 E = : = :.... : :.... : :. - ~... :.... -::=' 1.0 r , , :E 0.7 -g 0.5 0:1 E L-_-'-... _O_--'- ----"" (b) 1.0 r ,.E Q) ~ ~ r ,, , a 0.7 j ~ : : ~ ~ L-:::--=---':--L..=---:----:::---:----:-:-""="'" (c) Figure 2. Experimental and simulated results demnstrating steering behavir in cncrete, fr a (). f 0 & 40, (R=30.48cm, N=8, c=3650m/s, f=170khz, and d=0.7'x) and mrtar, fr a (). f 25 (R=30.48cm, N=8, c=3500m/s, f=140khz, and d=0.582'x). 2157

6 1.0, ~ _, CIl O.B. ~ 0.7 a: 0.5 t 1.0, r , CIl O.B ~ CIl t 0.0 I '! '. : ~ = ~ ~ ~ -, - - ~.. ' ". ".":-._ ~ - > ~ < ~ -~- - L L L _ - ' - - ' - - " BO BO Angle, IJ, (deg.) Angle, IJ, (deg.) (a) 1.0 r r , CIl 0.8 ~ CIl t 1.0 r------r-r , 0.0 " ' - ~! J. = _ : _,., _ - -., - -, - :. 0.0 L="--.:..,- _ : : _ : : - - " ' ~ CIl O.B ~ CIl t,., --,------:., --::--=----::>=J (b) 1.0 r - - ~ y _ , CIl 0.8. ~ 0.7 a: 0.5 t ) ~ ~ ~! : : : :!,! l! < :. A. ~ ~ ~..... : :! 0.0 ~ 1...!! =!- -! ~! < 0.:. :.. ~ L ~ Angle, IJ, (deg.) 1.0 r y , CIl O.B. ~ CIl t... + _ ' _ - ' - -. :.... L.... ~. L.. I. - " - ' ' ' ' - - ' Angle, IJ, (deg.) (e) Figure 3. Experimental and simulated results demnstrating steering and fcusing behavir in cncrete: (a) fcusing (N=8), (b) steering (N=16), and (c) fcusing (N=16) (R=30.48cm, F=3048cm, c=3500mls, f=140khz, 0=0 0 and d=o.582'x). 2158

7 Table 1. Summary f simulated and experimental results, demnstrating the accuracy and steerability f the directivities. Fr each case, results are given fr steering fllwed by fcusing. Parameters Simulated Results Experimental Results mrtar/cncrete N O. O. q-6db 180 O. q-6db 180 mrtar (steer) mrtar (fcus) mrtar mrtar mrtar 16 0 NA 39.6 NA mrtar NA 39 NA cncrete cncrete cncrete cncrete cncrete 16 0 NA 36 NA cncrete NA 38.5 NA field, which exhibits inherently pr reslutin, the average deviatin frm the simulatin in q-6db 180 was nly 1.4. The directivities btained shwed excellent accuracy. Fcusing ultrasund is a key functin f phased arrays, especially within cementitius materials. There is a significant increase in the near field length when the number f elements is increased frm 8 (ZTR=I1.5 cm) t 16 (ZTR=50 cm). Using 8 elements, Figure 2(a) demnstrates the negligible imprvement fcusing will have ver steering within the far field. When the number f elements is increased frm 8 t 16, R appears within the near field f the sensr. The directivity acquired when steering the ultrasund, shwn in Fig. 3(b), demnstrates a prly shaped main lbe. When the ultrasund is fcused at the rim f the wheel, as shwn in Fig. 3(c), dramatic imprvements fr the directivity is bserved. The fcused directivity shws excellent steerability, and excellent accuracy as cmpared t the simulated result. 2159

8 SUMMARY AND CONCLUSIONS The cnclusin drawn frm the experimental results is that phase steering and fcusing ultrasund is pssible in a cementitius medium. Theses results shwed excellent steerability and accuracy, and prvide excellent agreement with the numerical simulatin. It was shwn that the array exhibits great steerability, with an average variatin f nly 0.75 thrughut the experiments. It was als shwn that the directivities btained shw excellent accuracy, with an average deviatin frm the simulatin in q-6db. 180 f nly 1.4. Als bserved is the fact that increasing either d r N imprves the characteristics f beam directivities fr bth steering and fcusing. Hwever, grating lbes were bserved if the inter-element spacing reaches abve apprximately half the wavelength. This paper experimentally prves the arguments made in the cmpanin paper: A benefit f fcusing ver steering within the near field f the array and the negligible effect f fcusing in the far field. In summary, by fcusing in the near field and steering in the far field, a new cmpsite full field imaging scheme can be intrduced, which transfnns the "dead" zne f the transducer int a "detectable" zne, enlarging the useful area f inspectin. ACKNOWLEDGEMENTS This study was supprted by the Krea Highway Crpratin. We are grateful t Dr. Chang-Guen Lee, Prgram Manager and Mr. Ken-Chang Ch, Directr f Highway Research Institute f the KHC, fr their encuragement and supprt. REFERENCES 1. S. C. Wh and L. Azar, "Phase Steering and Fcusing Behavir f Ultrasund in Cementitius Materials," Review f Prgress in Quantitative NDE, eds. D. O. Thmpsn and D. E. Chimenti, (Plenum Press, New Yrk, 1998), in this vlume. 2. L. Azar, Ultrasnic Phased Arrays fr the Cnditin Assessment f Cncrete Structures, MS Thesis, Massachusetts Institute f Technlgy, (1998). 2160