NONDESTRUCTIVE EVALUATION OF DOUBLE BEVELT-JOINT BY TANDEM ARRAY ULTRASONICTRANSDUCER

H. Shirahata1, C. Miki2 and R.Yamaguchi2

1 Department of Civil Engineering, Musashi Institute of Technology1-28-1 Tamazutsumi Setagayaku, Tokyo 158-8557 Japan2 Department of Civil Engineering, Tokyo Institute of Technology2-12-1 O-okayama, Meguro-ku, Tokyo 152-8552 Japan

ABSTRACT. The double bevel T-joint is one of the most fundamental joints of steel bridges.Double bevel T-joint can be seen at beam-column connection of bridge pier, In the Japanesespecifications, the welding should be full penetration. However, weld defect of incomplete pene-tration could be left in the joint due to the lack of quality control in welding. Fatigue cracks canbe propagated from the weld defects. The authors developed a tandem array transducer. Thetandem array transducer consists of 10 elements aligned in the same direction. Tandem scanningcan be simulated by the transducer. Image reconstruction of incomplete penetration by syntheticaperture focusing technique was carried out. The test results showed sufficient detect ability ofincomplete penetration by the tandem array transducer. Height of incomplete of penetrationcould be estimated.

INTRODUCTION

T-joint is one of the most fundamental joints of steel bridge structure. T-joint isapplied to column-beam connections and to main girder and lateral girder connections.There are two types of T-joints. One is full penetration weld and the other is buttweld. Pull penetration is applied to joints where fatigue should be considered. Buttweld is applied to joints where fatigue doesn't have to be considered.

Partial penetration isn't allowed by the Japanese specifications. In partial pen-etration, root face remains in the welded joint. It is difficult to control the qualityof root face such as length, gap and roughness of the surface. For a T-joint of fullpenetration, it happens that root face remains due to incomplete penetration. In suchcase, fatigue cracks can be initiated and propagated from the incomplete penetration.Incomplete penetration must be eliminated. Reliable inspection technique is needed.

The objective of this study is to detect and estimate the height of an incompletepenetration in a T-joint. Ultrasonic tests were carried out. Considering a column-beamconnection of box section, radiographic test isn't applicable. Incomplete penetration isa planar defect. Tandem technique is effective to detect planar defects [1-3]. A tandemarray transducer was developed. For the display of weld defect, synthetic aperturefocusing technique (SAFT) [4] was applied to reconstruct the image of incompletepenetration.

CP657, Review of Quantitative Nondestructive Evaluation Vol. 22, ed. by D. O. Thompson and D. E. Chimenti© 2003 American Institute of Physics 0-7354-0117-9/03/S20.00

1164

254mm4————————————————^chl 2 3 4 5 6 7 8 9 chip" i LI I i_— ..

70deg10 elements alignedeach channel incident angle 70 degrees

frequency 5MHzFIGURE 1. Tandem array transducer.

(mm)A

600

>

•< 50° ^*ti S

\ \

f-

500* r"

^ ———— ,^ t

""X^V

60deg25

partial pen<

parameters:^,/!

(mm)

full penetration

t:thickness of plate/irheight of root face

FIGURE 2. Specimen.

EXPERIMENT

Tandem Array Transducer

A tandem array transducer was developed. Figure 1 shows the transducer. Thetransducer consists of ten aligned elements of angle beam. Frequency of each elementis 5MHz and incident angle of each element is 70 degrees. Incomplete penetration isa planar defect. Incident waves are reflected specularly at the defect. It is sometimesdifficult to receive spcularly reflected waves by one probe technique. Tandem techniquehas shown effectiveness for detection of planar defects. There are ten patterns fortransmitting and receiving. Totally, 100 patterns of wave forms can be obtained bythe transducer without moving back and forth.

Specimen

Figure 2 shows a specimen. Two plates were welded at the center of the spec-imen. The parameters of the specimens were thickness of the plate t and the heightof incomplete penetration h. Two types of specimens were made. Parameters of thespecimens are shown in Figure 2. Both types of the specimens have almost the samethickness as the plates used in column-beam connections.

1165

obtain 10x10 patternsof wave forms in 2 secswitcher/multiplexer

specimenFIGURE 3. Experimental setup.

joint

x10 1

>9 8 | 7 6 h 4 3 2 i t

x closest distance between the front oftransducer and column

obtained 100 wave formsFIGURE 4. Experiment.

25 (mm)

Setup

Figure 3 shows an experimental setup. The system consists of a personal com-puter, ultrasonic flaw detector, switcher/multiplexer, and transducer. The flaw detec-tor generates pulse signals. The switching device switches the signal to each of thechannels for transmitting and receiving. The computer calibrates the flaw detectorand acquires wave data by digital I/O interface. 100 wave forms can be obtainedwithin two seconds.

In the experiment, the transducer was put at the closest position from the weldtoe, as shown in Figure 4. Transducer position x is defined as the distance betweenthe front of the transducer and the column plate. 100 wave forms were obtained atthe closest position from the weld toe.

DETECTABILITY OF INCOMPLETE PENETRATION BY THE TAN-DEM ARRAY TRANSDUCER

Figure 5 shows some of the wave forms. The horizontal axis is distance fromthe transmitter. The vertical axis is related to echo height. Ten wave forms of onetransmitter and different receiver channels are shown in each of Figures 5(a) and (b).The thickness of the plate t was 25mm. Figures 5 (a) and (b) are wave forms whenthe transmitter was channel 1 and 2, respectively. When the transmitter was channel1, a higher echo can be found when the receiver was channel 4.

1166

(a)

2±

T^rn? 45(b)

|525

4 |3 | 2 | 1 £" *

x= 1 5nm transnitter ch2

receiver

±±Lrfe-»/U. i|. i IkJL3L

10100 200 300 400 500 TOO 200 300 400 500

distance of propagation (mm) distance of propagation (mm)FIGURE 5. Wave forms, the thickness of the plate t:25mm, height of incomplete penetration

n, (a) transmitter channel 1, (b) transmitter channel 2.

(a)

o•gj-

transducer position;x =15mm :transmitter chl "receiver ch4 :

(b)

100 200 300distance of propagation (mm)

FIGURE 6. Defect echo and the path, thickness of plate t:25mm, height of incomplete penetration/i:8.5mm.

(a 24mreT8l7l6l5l4l?TZ]T40|

|| Iran• rec>HHMAMhMhJMM*4MkMMMUte

transnitter chlreceiver

10

x=24irmtransnitter ch2receiver

——— lo

5100 200 300 400 500

distance of propagation (mm)

__ _________10100 200 300 400 500distance of propagation (mm)

FIGURE 7. Wave forms, the thickness of the plate t:40mm, height of incomplete penetrationfollmm, (a) transmitter channel 1, (b) transmitter channel 2.

1167

*L£13)*<343.§o

CD

: transducer position:

j-

x=24mm :transmitter ch2receiver ch7 .

(b) x=24mrc9 f 8 17 I 6 I5I4I312T11

200 300 400distance of propagation (mm)

FIGURE 8. Defect echo and the path, thickness of the plate t:40mm, height of incomplete pene-tration to: 11 mm.

receiver transmitter receiver transmittern n

receiverr\transmitterr\

*. reflectorreflector

transmitter receivern n

transmitter receiver' r\

transmitter receiverr\___r\

^reflector3* reflector

FIGURE 9. Six paths to be considered for image reconstruction.

The wave form of combination of channel 1 and 4 is shown in Figure 6 (a). Anecho can be seen around 180mm from the transmitter. The wave path is shown inFigure 6(b). The incident wave was reflected at the defect specularly and reflected atthe bottom and received by channel 4.

Figure 7 shows wave forms obtained from a specimen of which thickness t was40mm. Figures 7 (a) and (b) are wave forms when the transmitter was channel 1 and2, respectively. In case of combination of transmitter channel 2 and receiver channel7, an echo could be found. The wave path is shown in Figure 8.

Figures 5 and 7 show that defect echoes could be seen clearly. Specularly reflectedwaves at an incomplete penetration could be received at least one of the channels ofthe tandem array transducer. Those defect echoes could be obtained in two secondswithout moving the transducer.

IMAGE RECONSTRUCTION

For the display of weld defects, synthetic aperture focusing technique was ap-plied. Comparing with one probe method, the wave path of defect echo becomes morecomplicated. Six paths of wave propagation were considered as shown in Figure 9.

Figure 10 shows the reconstructed image. The thickness of the plate of thespecimen t was 25mm and the height of incomplete penetration ft was 8.5mm Animage could be seen around the position of incomplete penetration.

It is possible to pick up some specific wave forms and reconstruct an image by thereconstruction program in the computer. Figure 11 (a) shows an image reconstructedby three wave forms. Figure ll(b) is an illustration of the image. Figure ll(c) shows

1168

25|SJ4|3

"&&$!• JSH $$~M?M

ch2 eh Icolumn

25

reconstruction area

FIGURE 10. Reconstructed image, thickness of the plate £:25mm, height of incomplete penetration/i:8.5mm.

(a)

transmitter chl:•receiver

_C2W)

'53&£9

:' I transmitter chl:f 1 receiver ch4i: - Xx - - --- i4 "

•a lower tip echo

distance of propagation (mm)TOO 200 300

distance of propagation (mm)

transmitter ch2receiver ch2l L .TOO 200 300

distance of propagation (mm)

FIGURE 11. (a) Reconstructed image by three wave forms, thickness of the plate i:25mm, heightof incomplete penetration /i:8.5mm (b)illustration of the image (c) geometry of the joints (d) waveforms.

1169

; i mm

d*2 chj

1J

^

k

40

f

^

j

i

>

^ (itint)145

k11

f

145k :

FIGURE 12. Reconstructed image, thickness of the plate £:40mm, height of incomplete penetration

actual size of the incomplete of penetration. Figure ll(d) are wave forms picked up.Those wave forms were combination of "transmitter channel 1 and receiver channel1", "transmitter channel 1 and receiver channel 4", and "transmitter channel 2 andreceiver channel 2". Echo of "combination 1 and 1" could be an upper tip echo fromthe incomplete penetration. Echo of "combination 1 and 4" could be a specularlyreflected echo at the incomplete penetration. Echo of "combination 2 and 2" could bea lower tip echo from the incomplete penetration. The height of incomplete penetrationestimated from the image was 9 rnm. The actual height is 8.5 mm. Prom those threewave forms, the incomplete penetration was detected and the height of the incompletepenetration was estimated accurately.

Figure 12 shows another example. The thickness of the plate of the specimen twas 40 mm and the height of the incomplete penetration h was 11 mm. As well asFigure 10, Figure 12 is an image reconstructed by all the 100 wave forms. Figure 13(a)shows an image reconstructed by three wave forms. Each of the wave forms is shownin Figure 13(d). Echo of "combination 1 and 1" could be an upper tip echo from theincomplete penetration. Echo of "combination 2 and 7" could be a specularly reflectedecho at the incomplete penetration. Echo of "combination 2 and 2" could be a lowertip echo from the incomplete penetration. The height of incomplete penetration wasestimated accurately.

It takes less than one minute to reconstruct an image. The system makes quickand accurate inspection of T-joint possible. As the procedure of inspection, the authorsrecommend to (l)obtain 100 wave forms, (2)reconstruct an image, (3)check wave formon by one, and (4)pick up some of the specific wave forms and estimate the height.

1170

(a)

200distance of propagation (mm)

400distance of propagation (mm) TOO 200 300

distance of propagation (mm)FIGURE 13. (a) Reconstructed image by three wave forms, thickness of the plate £:40mm, heightof incomplete penetration /i:llmm (b)illustration of the image (c) geometry of the joints (d) waveforms.

SUMMARY

The authors developed a new tandem array instrumentation and image recon-struction program by synthetic aperture focusing technique. The results show thatthe instrumentation can be successfully applied for inspection of T-joint. Weld defectscould be detected. Height of incomplete penetration was estimated accurately.

REFERENCES

1. Bowker, K.J.,Coffey, J.M.,Hanstock, D.J,, Owen ,RC. and Wrigley, J.M.: CEGBinspection of plates 1 and 2 in UKAEA defect detection trials, British Journal ofNOT, Vol.25,No.5,pp.249-255,1983.

2. Barbian, O.A.,Engl, G.,Grohs, B.,Rathgeb, W. and Wiistenberg, H.: A secondview of the German results obtained in the defect detection trials, UKAEA, BritishJournal of NOT, Vol.26,No.2,pp.92-96,1984.

3. Barbian, OA.,Grohs, B. and Kappes, W.: Evaluation of time-of-flight data withthe ALOK-method, British Journal of NDT, VoL26,No.4,pp.214-217,1984.

4. Miiller, W.,Schmitz, V. and Schafer, G.: Reconstruction by the synthetic aper-ture focusing technique(SAFT), Nuclear Engineering and Design, Vol.94,pp.393-404,1986.

1171