9
Ultrasonic Measurement of Weld Penetration The use of pulse-echo techniques to determine weld pool dimensions is investigated BY D. E. HARDT A ND J. M. KATZ ABSTRACT. The automatic production of high quality welded joints requires a means of measuring weld quality in real time and a feedback control strategy for regulating that quality. Penetration is a good first order indicator of weld integri ty, and most efforts at weld quality  con trol have been concentrated on penetra tion.  In this work, a technique for using ultrasonic pulse-echo measurements to determine weld pool dimensions is exam ined. Although current research in nonde structive evaluation indicates that size and shape of discontinuities can be inter preted from the time history of an ultra sonic reflection, the thermal gradients caused by welding cause sufficient distor tion in the ultrasound reflection to pre clude use of these techniques for initial studies. In a more straightforward approach, a geometric optics framework is developed for the estimation of ultra sonic transit times between an ultrasound transducer and a stationary, hemispheri cal weld pool in a rod. Experiments were conducted to verify these predictions by performing mea surements of ultrasonic reflections from machined hemispheres and from weld pools in long rods. The results show good agreement between the measurements performed on the cylindrical rods and the geometric optics predictions for both machined surfaces and weld pools. Introduction The automatic production of high  qual ity welded joints requires the side-by-side development of techniques for position- D.  F.  HARDT and J. M. KATZ are with the Laboratory for Manufacturing and Productivi ty, Massa chusetts Institute of Technology, Cambridge, Massachusetts. ing the welding torch (seam tracking) and for regulating welding parameters in real time to obtain a desired level of weld quality. The work described in this paper concentrates on weld penetration as a measure of weld quality using ultrasonic pulse echo techniques to measure the penetration of a  weld  pool in real time. Several researchers have considered techniques for measuring weld penetra tion for feedback control.  In  1976, Vro- man  and Brandt (Ref. 1) used a line scan camera to measure the width of the top side of a  weld.  Since for a given geome try and fixed welding conditions the depth to w idth ratio of a weld remains somewhat constant, regulating the top side weld pool width regulates penetra tion.  The results reported were not  con clusive and indicated the need for further study of this approach. In a refinement of this technique, Richardson  et al.  (Ref. 2) recently reported the use of video mea surement methods to control the topside width of GTA welds. The unique aspect of this wor k is that the pool is viewed directly from above by putting the optical axis in  line  with the electrode. In  a more direct approach, Nomura ef al .  (Ref. 3) have used photodetectors placed along the back side of the  weld ment to measure the back bead width of th e  weld.  As the weld goes from partial to full penetration, the infrared radiation from the back side of the weld goes through a step transition. In this way, photodetectors can be used to detect a full penetration  weld.  Garlow (Ref. 4) and Reiff (Ref. 5) have used a simple photo- transistor to measure the back side weld bead width in GTA welds with reason able accuracy. This measurement was used in a closed loop controller that quite accurately regulated the back bead width.  However, back side weld bead sensing has an inherent drawback in that it is difficult or impossible to conveniently locate a sensor on the back side of the weldment for many weldment configura tions. In addition, back side sensing does not provide useful measurements when knowledge of partial penetration is desired. Hardt and Zacksenhouse (Ref. 6) dem onstrated that weld pool size could be determined by measuring the resonant frequency of a full penetration  pool.  The pool is modelled as a dynamic mass spring system, and they show that the natural frequency of this system is a function of the size of the weld  pool.  The existence of this frequency dependence on weld pool size has been verified experimentally, and the present work is attempting to show that arc voltage  fre quency measurements can be used to determine the weld  pool  natural frequen cy. Renwick and Richardson (Ref. 7) have also observed a pool resonance in the case of a partially penetrated  weld. All of the weld pool measurement techniques mentioned so far share the difficulty that they are attempting to mea sure variables that are not single valued indicators of penetration. In an attempt to provide a means of directly measuring the desired weld pool dimensions, the concept of using ultrasonic pulse echo techniques to directly measure weld pool dimensions was developed, based upon established technology for other applica tions. In this paper, the techniques avail able for ultrasonic measurement and the implications of application to in-process welding are discussed. This is follow ed by a critical experiment where the existence of reflections from a weld pool are  con firmed,  and some rudimentary depth measurements performed. Ultrasonic Methods for Defect Measurement Ultrasonic testing has been used effec tively as a nondestructive evaluation WELDING RESEARCH SUPPLEMENT  |  273-s

Ultrasonic Measurement of Weld Penetration

Embed Size (px)

Citation preview

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 1/9

Ultrasonic Measurement of Weld Penetration

The use of pulse-echo techniques to determine

weld pool dimensions is investigated

BY D. E. HARD T A ND J. M . KATZ

ABSTRACT. The automatic production of

high quality welded joints requires a

means of measuring weld quality in real

t ime and a feedback control strategy for

regulatin g that quality . Penetr ation is a

good first order indicator of weld integri

ty, and most efforts at weld quality

  c o n

t ro l have been concentra ted on penetra

t ion.

  In this work, a technique for using

ultrasonic pulse-echo measurements to

determine w eld poo l d imens ions is exam

ined.

Although current research in nonde

structive evaluation indicates that size

and shape of discontinuities can be inter

preted from the t ime history of an ultra

sonic reflection, the thermal gradients

caused by we lding cause suff ic ient distor

t ion in the ultrasound reflection to pre

clude use of these techniques for initial

studies. In a more straightforward

approach, a geometr ic opt ics f ram ew ork

is developed for the estimation of ultra

sonic transit t imes between an ultrasound

transducer and a stationary, hemispheri

cal weld pool in a rod.

Exper iments were conducted to ver i fy

these pred ic t ions by performing mea

surements of ultrasonic reflections from

machined hemispheres and f rom weld

pools in long rods. The results show go od

agreement between the measurements

performed on the cyl indrical rods and the

geometric optics predictions for both

machined surfaces and weld pools.

I n t roduct ion

The automat ic product io n o f h igh qual

i ty welded joints requires the side-by-side

deve lopment o f techn iques for pos i t ion-

D.

  F.

  HARDT and J. M. KATZ are with the

Laboratory for Manufacturing and Productivi

ty, Massa chusetts Institute of Technology,

Cambridge, Massachusetts.

ing the welding torch (seam tracking) and

for regulating welding parameters in real

t ime to ob tain a desired level of we ld

quality. The work described in this paper

concentra tes on w eld penetra t ion as a

measure of weld quality using ultrasonic

pulse echo techniques to measure the

penetration of a

  weld

  pool in real t ime.

Several researchers have considered

techn iques for measur ing weld penetra

t ion for feedback contro l .  In  1976, Vro-

man

  and Brandt (Ref. 1) used a line scan

camera to measure the width of the top

side of a  w e l d .  Since for a given geome

try and f ixed welding condit ions the

depth to w id th ra t io o f a we ld remains

somewhat constant, regulating the top

side weld pool width regulates penetra

t ion.

  The results reported were not

  c o n

clusive and indicated the need for further

study of this app roac h. In a refinem ent of

this technique, Richardson   et al.   (Ref. 2)

recently reported the use of v ideo mea

surement methods to contro l the tops ide

width of GTA welds. The unique aspect

of this w or k is that the poo l is v iew ed

direc tly f rom above by put t ing the opt ica l

axis in

  line

  wi th the e lec trode.

In

  a more d i rec t approach, Nomura e f

al .  (Ref. 3) have used photodetectors

placed along the back side of the   w e l d

ment to measure the back bead width of

th e  w e l d .  As the weld goes from partia l

to ful l penetration, the infrared radiation

from the back side of the weld goes

through a step transit ion. In this way,

photodetec tors can be used to detec t a

ful l penetration  w e l d . Garlow (Ref. 4) and

Reiff (Ref. 5) have used a simple photo-

transistor to measure the back side weld

bead wid th in GTA welds wi th reason

able accuracy. This measurement was

used in a c losed loop control ler that qu ite

accurately regulated the back bead

w id th .

  However, back s ide weld bead

sensing has an inherent drawback in that

it is difficult or imp ossible to c onv enie ntly

locate a sensor on the back side of the

weldment fo r many weldment conf igura

tions. In addition, back side sensing does

not prov ide usefu l measurements when

knowledge of partia l penetration is

desired.

Hardt and Zacksenhouse (Ref. 6) dem

onstrated that weld pool s ize could be

determined by measuring the resonant

f requency o f a fu ll penetra t ion  poo l .  The

pool is modelled as a dynamic mass

spring system, and they show that the

natural frequen cy of this system is a

function of the size of the weld

  poo l .

  The

existence of this frequency dependence

on weld pool s ize has been verif ied

experimentally, and the present work is

a t tempt ing to show that arc vo l tage

  f re

quency measurements can be used to

determine the weld

 pool

 natura l f requen

cy. Renwick and Richardson (Ref. 7) have

also observed a pool resonance in the

case of a partia l ly penetrated

  we ld .

Al l o f the weld poo l measurement

techniques mentioned so far share the

diff iculty that they are attempting to mea

sure variables that are not single valued

indicators of penetration. In an attempt to

provide a means of directly measuring

the desired weld pool dimensions, the

concept of using ultrasonic pulse echo

techniques to directly measure weld pool

d imens ions was deve loped, based upon

established technology for other applica

tions. In this paper, the techniques avail

able for ultrasonic measurement and the

implications of application to in-process

weld ing are discussed. This is fol low ed by

a crit ical experiment where the existence

of re f lec t ions f rom a weld poo l are  con

f i rmed ,

  and some rud imentary depth

measurements performed.

Ul t rasonic Methods for Defect

M e a s u r e m e n t

Ultrasonic testing has been used effec

t ively as a nondestructive evaluation

W EL DIN G RESEARCH SUPPLEMENT  | 273-s

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 2/9

c y l i n d r i c a l

d i s c o n t i n u i t y  -

ul t rasound

t r ans duc e

i  . s a -

o.om—

e.ae—

• * . » • —

'•—

©

I

n

-  i n i t i a l

p u l s e

¥

Mi .

I

nder

| j

1 1

bock

wall

\

i —

l i

Fig.  1

 —

 Ultrasonic reflections from a cylindrical

discontinuity (Ref.  71): A — ultrasound incident

upon cylindrical discontinuity;  B —ultrasonic

reflections from a  cylinder;  C - oscilloscope

trace of reflections from cylinder

t e c h n i q u e f o r l o c a t i n g d i s c o n t i n u i t i e s

below the surface of various metals. It

has become a standard technique for

locating cracks, lack of fusion, porosity,

and other discontinuit ies in fusion welds.

All

  ultrasonic testing methods are based

on reflection of stress waves caused by

changes in the material properties of the

medium in which it travels. This makes

single point range measuremen ts a s imple

matter of measuring the transit t ime for a

single ultrasound pulse to travel to and

from a re f lec tor . However, a d iscont inu

ity of dimension larger than the wave

length of the incident ultrasound pulse

wil l not only reflect ultrasound, but i t wil l

rearrange the phase relationship of the

incident ultrasound pulse. Thus, the

sh ap e of the reflecte d ultrasound pulse

is d i f fe rent f ro m the sha pe o f the

incident pulse —see  Fig. 1 and Freeman

(Ref.  8) .  The manner in which the pulse

sh ap e is altered is a func tion of the

size,

  shape, and material properties of

the reflector. Thus the reflected ultra

sound pulse contains information   c o n

cerning the size and shape of whatever

caused the reflection.

Many nondestructive testing problems

require knowledge of discontinuity s ize

and/or shape as well as the location of a

discontinuity, and several researchers

have been studying ultrasonic techniques

for directly determining the size and

shape of the defects in materials. These

techniques generally involve some

  deci

phering of the characteristics of ultrasonic

reflections from idealized defects l ike

spheres, cylinders and disks. These

studies include both t ime domain (Pao

and (Ref. 9) and frequency domain (Ref.

10)) approaches and include techniques

that determine size from a single reflec

tion (Ref. 11) or by observing the scat

tered intensity at several different loca

tions (Refs. 12, 13).

The results of these studies indicate

that the ultrasonic reflections from a giv

en tar get are a func t ion o f bot h size

and shape of the reflector. If the essential

features of ultrasonic reflections from

known surfaces can be characterized, i t

should be possible to develop a means of

measuring reflector dimensions from

ultrasound traces. Sachse (Ref.

  11)

  has

done this for the case of a fluid filled

cylindrical inclusion in an aluminum block

while Thompson and Thompson (Ref. 14)

have perf orm ed s tud ies us ing a f requen

cy domain scattering technique for identi

fy ing a more general target. In yet anoth

er approach, Rose (Ref. 15) has used

computer pat tern recogn i t ion techn iques

to identify certain postweld discontinui

ties in welded steel plates.

U l t ra s o n ic M e a s u r e m e n t o f W e l d

Poo l D imens ions

A weld pool constitutes a change in

phase and material properties relative to

the rest of the weldment and thus should

be a reflector of ultrasound. Tabulated

values of the density and speed of sound

of molten aluminum and copper (Ref. 16)

indicate that a weld pool in either of

these materials would indeed reflect

ultrasound. Since the wa velen gth of ultra

sound is smaller than the characteristic

dimension of any but the thinnest of

materials that one might wish to   w e l d ,  an

ultrasound pulse reflected by a weld pool

should contain information concerning

the size of that weld

  pool .

In this paper the condit ions for reflect

ing ultrasonic waves from a weld pool

are examined, and a means for interpret

ing such reflections is presented. A critical

exper iment is perfor me d to demonstra te

t r a n s d u c e r

m o u n t  X

the abil i ty of ultrasound to measure the

dep th

  of

  pe netra tion of a static we ld

poo l .  There are, however, several areas

of research that must precede the devel

opment o f an u l t rason ic weld poo l mea

surement system. The bulk of these are

the result of the large temperature gradi

ents that exist within a weldment.

The propagat ion speed o f u l t rasound

is a func tion of the tem pera ture of the

med ium in whic h it travels. In general, this

speed drops as the temp erature o f the

medium increases. For weld pool mea

surement, this will result in an ultrasound

velocity that varies with posit ion. This

velocity gra dient can cause shift ing of the

path fol lowed by an ultrasound pulse and

may make it d iff icult to se e reflections

f rom the we ld  pool .

High temperatures also cause an

increase in attenuation of the medium.

Thus larger amplitude ultrasound pulses

must be applied to the medium in order

to get a measurable reflection back. In

the heat-affected zone (HAZ) of the

w e l d ,

  the solid phase transition can result

in a large enough variation in properties

to cause reflections of ultrasound (Ref.

17). If the magnitude of this reflection is

large, it may not be possible to transmit

u l t rasound through the HAZ to the weld

poo l .

 This would l imit the ultrasonic

  tech

nique to one of measuring the dimen

sions of the HAZ. On the other hand, i f

f in ite but small reflections from the HAZ

exist, it might be possible to indepe nden t

ly measure the size of the HA Z and of the

w e l d  p o o l .  This would present the

opportun i ty fo r independent contro l o f

HAZ dimensions and the penetration.

Another prob lem is that an u l t rasound

transducer must be in acoustic contact

with the material to be inspected if any

energy transfer is to take place. This is

normally accomplished by using a l iquid

or gel to fill in the asperites between the

transducer and the object to be

inspected or by clamping the transducer

to the specimen with a large force (Ref.

18). A weld pool measurement system

such as that shown in Fig. 2 must be

capable of moving along the top surface

of a hot weld me nt wh ile maintaining

Fig. 2 — Proposed configuration of the transducer in a penetration measurem ent system

274-s I

 SEPTEMBER 19 84

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 3/9

ultrasound

transducer

G T A W E L D I N G T O R C H

heat sink  \

  ^

W

mild  steel sample

 weld pool

-^-locking  r ing

ult ra sound

t ransducer

to

  pulser

W

Fig.  3  — Configuration  for detecting ultrasonic

reflections from a weld pool

acoustic contact between the transducer

and the weldment.

Current ly , four approaches to the

problem have been considered. The f irst

technique consists of a standard contact

angle beam transducer that is lightly held

against the weldment by a spring force so

that it is still free to slide over the   w e l d

ment. The second techn ique would

em ploy a comm ercial ly available wheel

shaped transducer that is f i l led with

water. The th i rd techn ique would

employ a l inear actuator to l i f t the trans

ducer off of the weldment, discretely

advance the weld torch, and then com

press the transducer back down on the

we ldm ent (Ref. 19). The four th and by far

the most e legant techn ique would be to

em ploy an electrom agnetic acoustic

transducer (EMAT), see Thompson and

Thompson (Ref. 14), as the ultrasound

source. EMAT's prov ide a non-contac t

method o f performing u l t rason ic inspec

tion that is clearly desirable for the case

of we ld poo l d imens ion measurement.

Unfortunately, state-of-the-art EMAT's

have low transduction eff ic iencies.

Fig.  4  — Transmission of ultrasound through a flat ended rod

Weld Pool Reflections in a

Cylindrical Rod

In order to measure the dimensions of

a weld pool using ultrasound, i t is neces

sary to determine how the presence of a

weld  pool reflects an ultrasound pulse

and then how the dimensions of the

reflecting weld pool can be isolated from

a time trace of the ultrasound pulse. To

th is end, a s impli f ied we ldme nt geom etry

(a cyl indrical rod) combined with a ray

optics wave analysis was employed. (The

results of Schmitz (Ref. 20) are encourag

ing with regard to the uti l i ty of this

approach.)

The case of ultrasonic weld pool

dimensional measurement lends itself to

the geometric optics approach, s ince the

locat ion o f the weld poo l is known. By

physically connecting the transducer

mount ing s truc ture to the weld ing torch,

it can be guaranteed that the transducer

wil l track the weld pool at al l t imes. Using

the ray tracing techniques of geometric

optics, it is possible to reconstruct a

surface pro f i le o f the weld poo l f rom

ultrasonic t ime traces. Important

  con

cepts for such a reconstruction are dis

cussed below.

Cons ider the conf igura t ion shown in

Fig.  3 where an ultrasound transducer is

clamped to the end of a cyl indrical rod. A

GTA welding torch is placed above the

rod end oppos i te to the t ransducer, and

a weld pool is established in the rod end.

In this configu ration , the ultrasound trans

ducer wil l be directly v iewing the weld

poo l as i t fo rms. Ho wev er, be fore

  con

sidering the reflections of ultrasound

f r o m a w e l d

  p o o l ,

  it is well to look at the

ultrasonic reflections from the end of a

plain cylindrical rod of length L, and

diameter d.

In Fig. 4, an ultrasound transducer wit h

maxim um b eam angle is sho wn placed o n

the end of the rod (posit ioned such that

its center is aligned with the rod axis and

that the geometry is axial ly symmetric).

This wil l a l low th e use of a tw o dim en

sional model of the ultrasonic ray paths.

In order to simplify the discussion, it is

assumed that Lftanf/y))  is less than half the

rod diameter so that the ultrasound pulse

wi l l reach the oppos i te rod end before

the beam spreads out enough to reach

the sidewalls of the rod.

The ultrasound beam can be described

by a series of rays traveling down the

length of the rod. The rays wil l make an

angle 0 with the axis of the rod such that

9 is less than th e absolut e value o f th e

transducer beam angle. For an arbitrary

ray with angle 9, the t ime required for

ultrasound to travel from the transducer

to the opp osite ro d end is given by:

t i  =  L / ̂ cos(9) where c-\  is the lo ngitudi

na l wave speed.

The rod end is a bounda ry betw een a

solid and air so refraction from the rod

end will not take place. From Snell's law it

can be shown that the longitudinal pulse

is reflected at the same angle as the

incident pulse while the reflected shear

pulse has an angle given by: 9

r t

  = sin

(c

t

/Ci

  s in 9) where  c

t

  is the shear wave

speed. The radial posit ion at the rod end

is given by:

r = L tan 9

(1)

It is possible for ultrasound to return to

the transducer along several different

paths,

  and Fig. 5 depicts four possible

Fig.  5 -   Return paths o f reflections from  a flat ended rod: A  — longitudinal wave return path without  sidewall reflection; B - longitudinal return path wit

sidewall reflection; C —return path with one she ar wave reflection; D —return  path with multiple shear wave reflections

WELDING RESEARCH SUPPLEMENT | 275-s

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 4/9

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 5/9

I f we assume that an ultrasound ray

with angle 9 to the rod axis is incident

upon the we ld  p o o l ,  an arbitrary point on

the ray path can be described by the

equat ion:

r = x tan 9.

(8)

The po in t o f in tersec t ion between the

ray path and the weld poo l  (xi,

 rj)

  can be

found by solv ing equations (7a) and (8)

simultaneously with the result:

x; = L cos

2

9-L  cos 9  (r

2

/L  - s i n

2

  0)

and,

Vi

n

  = 0.5L sin 29

  -FL

  sin 9

  ( r

2

/L

2

  - s i n

2

  9 ) '

/2

9)

The angle of incidence of the incoming

ultrasound ray is a function of 9, and the

tangent angle of the

  weld

  pool surface

(Fig.

  7) and is given by:

9  = 90 deg + 9 -  a

In turn,  the tangent angle

 a

  is related to

the slope of the weld pool surface by:

=  t a n

- 1

a\ = tan

  1

dr,/dxi

L - x

2

  +   2Lx - (x

2

  + L

2

(10)

The ang le o f re f lec t ion f rom the weld

pool can be found using Snell's law again

w i th 9

r

  = 9j for longitudinal reflections

and eq uation (9) for shear reflections. For

convenience, the reflection angle can be

expressed as an angle relative to the rod

axis, 9 (Fig. 7) using:

9

f

  = 9; - 9 + 9

r

(11)

There are 4 basic return paths back to

the transducer from the weld

 p o o l .

  These

return paths and their respective transit

t ime equations are l isted below; the

return paths are also depicted in Fig. 8.

1.  Longitudinal reflection wi th direct

return to transducer:

t =

*i

Ci cos  9  CT COS  9

12a)

2.   Longitudinal reflection which then

reflects from a sidewall, goes through

n-shear wave reflections, and then

returns as a longitudinal pulse:

« - r — — 1

I  cos 9 cos  9f I

\  Q C , J ( c

2

• +

l

(12b)

cfcos

2

9)

A.

where : n <

[Xi

  (tan 9

f

  + tan 9) -

  d /2 ]

(c

2

  -  c

2

  co s

2

  9 )'

/2

do cos 9f

(Note that the equations for reflections

from a f lat ended rod are a special case

of this equation.)

3. Shear refiection wh ich returns

directly to transducer:

t =

[

1

• + •

1

CT  COS 9  c

t

  co s

  9,

]

12c)

4.

  Shear wa ve reflectio n w hich is

mode converted back to long i tud ina l

wave after n reflections:

cos 9

  Cl

  cos 9f

_ d cc

r c i ^ l t a n e l

c-,c

t

  J  sin  9  J

12d)

+

Fig.  9 — Comparison  between the path length

to weld pool bottom and rod end

[

A,

  -  c

2

  I  (2n + 1)d

d c ,  I  2 sin  9f

where: n < [ tan 9 f + tan 9 ] -

(0

The relationships between the differ

ent transit t imes are very much a function

of the radius of the weld pool and the

length of the rod which serve to deter

mine the coordinates of the reflection

po in t  (Xj,  n), and the angle of reflection.

The amplitudes of the reflected pulses

are also functions of the angles of  inci

dence a nd the pa th lengths. The shape of

the f inal ultrasound time traces thus varies

with the radius of the weld

 p o o l .

  It is clear

from the equations that there wil l be

some typ e of mult ip le peak structure as in

the case of th e flat en ded ro d. This is a

result of the difference in speed of sound

between longitudinal and shear waves

and the difference in path length

between the various possible paths.

If u ltrasonic reflections from a weld

poo l are used to determine the d imen

sions of the weld

  p o o l ,

  it will be neces

sary to isolate characteristics of the ultra

sonic t ime traces to al low one to deter

mine those dimensions. Equation (12) is

somewhat cumbersome for th is purpose.

However, a s implif ication provides an

estimate of the radius of a hemisphere

shaped weld

  pool

  in the end of a rod.

Cons ider a sm al l hemisphere on a

rod end such that the incident ultrasound

beam wil l be reflected from part of the

flat porti on of th e ro d en d as in Fig. 9.

This wil l result in a reflection f ro m the rod

end whic h shou ld be a dominant charac

terist ic of the ultrasound time trace. The

®

Fig.

  8  — Ultrasonic reflections from a  hemisphere shaped weld pool in a cylindrical rod: A -  longitudinal return path; B — longitudinal  reflection mod

converted to

 shear

 and then back to

  longitudinal; C — shear

  return path;

 shear

  reflection mode converted to longitudinal

WELDING RESEARCH SUPPLEMENT  | 277-s

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 6/9

sample

rod

o c k i n g

ring

shortest transit t ime to the f lat portion of

the rod end is given by:

t r a n s d u c e r

Fig. tO —A cylindrical ro d  sample  showing the

transducer mount and  clamping  ring

t

P

  =

2 (L

2

 +

  r

2

  *

However, the f irst ultrasound reflec

t ion that returns to the transducer in this

case fo l lows a path down the center o f

the rod a t 0 = 0 to the b ot t om of the

weld pool and back. This path wil l have a

transit time given by:

t

P

  =

2(L

 -

  r

p

)

I f the ultrasound peaks corresponding

to these tw o transit times could be iso

lated and their difference, At, measured,

the radius of the weld pool could be

determined as fo l lows:

A t

  =

  t

e

  -

  t

r

=

  1 /

C l

  [L

2

 + r

2

  - 2 ( L -

  r

p

)]

which can be solved for r

p

  yielding:

r

n

  =

LciAt  +

  A\

  t

2

/4

2L +

  CiAt

13)

The validity of this expression has been

evaluated experimen tally an d is discussed

be low.

Cylindrical Rod Experiments

The cyl indrical geometry analyzed

above was used in a series of experi

ments with the objectives of:

1.  Establishing whether ultrasound is

indeed re f lec ted f rom weld poo ls .

2.   Establishing the validity of equation

(13).

To separate the effect of temperature

gradients on this result, tw o differe nt

exper iments were per form ed. In the f i rs t ,

hemispheres of various radii were

machined into the rod and then the

radius was measured from the ultrasound

reflections and equation (13). In the sec

ond set of tests, a GTA torch was used to

create various size weld pools on the f lat

end o f a rod. Ul t rasound measurements

were again made and the depth of the

weld pool estimated by application of

equation (13).

The test pieces were turned from 1020

hot rol led steel to a diameter of 15.8 mm

(

in.) and a length of 152.4 mm (6

  in.).

The weld poo l exper iments were per

formed us ing a DC weld ing power sup-

1 1  r

30.-4  S-4.5  72.7

T  I mas Cm I c r o - « e c )

a .  Hot  ended rod

c  5/16  in dia hemisphere in rod

9

3

Q.

£

<

3 . 3 —

2 . 2 —

1.1

 —

0 . 0 —

- 1 . 1 —

- 2 . 2 —

'

1

.

1 1 1 1

  I

l

  @

l

1  '  1

18 .2 3B.-+ 5H .5 72 .7 S3  . 9

T  I

 m«

  m I  e i - o — « « e }

i

  r

30.-4  5-4.5  7 2 . 7

T I  me Cm  1 c r o - s e c )

d-  3 /8   in.  die. nemispriere  in rod

b.  3/16  in.  dia.  1/4  in. deep  hemispnere  in rod

Fig. 11 — Recorded reflections  from  hemispheres o f various radius rods: A  — flat ended;  B — fie

 in .

 diameter,

  ]

A in. deep hemisphere; C — e

 in.

  dia

hemisphere; D— Ya  in. diameter hemisphere

273-s I

 SEPTEMBER 19 84

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 7/9

ply connecte d to an a i r -coo led G TA W

torch wi th a 3 .2 mm   (V s   in.) 1 % thoria

tungsten electrode. Argon shielding gas

was used for al l of the experiments.

A Krautkramer USM-2 ultrasonic pulser

receiver unit was used as the ultrasound

source. The transducer was a Paname

trics A611S, 10 MHz, 12.7 mm  ( in.)

diameter delay l ine transducer. The

A611S is thread ed so that a delay l ine can

be connected to the transducer using a

lock ing r ing to ho ld i t down, as dep ic ted

in Fig. 10. The transit time of ultrasound

through the rod is expected to be o f the

order of 50  micro-sec.  To ex trac t we ld

poo l d imens iona l in format ion f rom the

signals, t ime re solution of the o rder of 0.1

micro-seconds(s) is necessary.

To record and display the ultrasound

reflections, a system compris ing a Tek

tronix 7854 digital osci l loscope coupled

to a DEC

  MINC

  lab computer was

em ployed . By a l lowing com municat ion

be tw ee n these devices o n an IEEE Gener

al Purpose

  Interface

  Bus (GPIB), the high

speed digitizing and signal averaging

capabil i t ies of the oscil loscope could be

exp lo i ted.

  At the same time, the MINC

was used to control the functions of the

oscil loscope during the experiment,

record the data on disk,

  and

  plot the

resulting wa vef orm s. Since the 7854 has a

bandw id th >100 MHz , reco rd ing u l t ra

sonic reflections was well within i ts capa

bilities.

The f irst set of measurements were

perfo rme d on 8 rods: 1 f la t ended rod ,

and 7 wi th hemispheres o f d iameter

be tween 4 .76 mm

  ( e

  in.) and 12.7 mm

(V2 in.) ball end mil led into the rod end.

Typical results are shown in Fig.   11.

Figure

  11A

  depicts the ultrasonic

re f lect ions f rom a f la t ended rod . Here

the transit time from the initial pulse to

the first reflected pulse is 51.56 micro-s.

From equation (1), this is equivalent to a

rod length of 152.1 mm (5.98 in.).

The spacing between the trai l ing peaks

is 4.17 micro-s. Equation (6) was devel

oped to pred ic t the spac ing between

these peaks. Rearranging this expression

to es t imate rod d iameter f rom the peak

spacing, we find that d = 16.01 mm (0.63

in.). The estimates of rod length and

diameter fro m e quations (1) and (6) are in

good agreement with the actual values of

length (152.4 mm or 6 in.) and diameter

(15.86 mm or 0.62 in.).

Typical t ime traces of ultrasonic reflec

t ions from hemispherical ball end mil led

surfaces are contained in Figs. 11B and D.

These samples were manufactured using

ball end mil l ing cutters ranging in diame

ter f rom 4.76 mm

  ffie

  in.) to 12.7 mm ('/2

in.) in incre me nts of 1.59 m m  (fte  in.).

Examining the data sequentially, the first

thing to notic e is that a shoulder fo rms on

the f irst reflection and breaks off into a

second peak as the hemisphere becomes

larger. This first small peak is postulated

to be a re f lec t ion f rom the bot tom of the

hemispherical surface as depicted in

Fig.  9.

The next fea ture to notice is the sm ear

ing out of th e refle ctio n as the size of the

reflector increases. Figure 11A shows a

set of very c learly defined peaks. As one

goes through to Fig. 11D, there are more

and more re f lec t ions between the la rge

peaks. These are the result of reflections

from d i f fe rent port ions o f the curved

surface which result in an almost continu

ous variation in transit t ime. Intermediate

peaks exist between the major peaks

found in the f lat ended rod. These are

probably results of the hemisphere

re f lec t ions be ing mode converted to

shear waves in the rod.

In an a t tempt to measure the d imen

sions of these hemispheres, the t ime

difference between the small in it ia l peak

or shoulder and the f irst major peak of

the ultrasound time traces was measured.

The rods were then sect ioned down the

middle and the depths of each of the

hemispheres was measured using a Nikon

toolm aker's micro scope . Figure 12 is a

plot of the measured t ime difference vs.

the depth o f the hemisphere. A lso sh own

in Fig. 12 is the theoretical relationship

be tween the t ime d i f fe rence be tween

the f irst tw o peaks of an ultrasonic t im e

record and the hemisphere radius

according to equation (13) and using the

simplest approximation:

A t =  2

Cl

r

p

14)

There is reasonable agreem ent b e

tween theory and the measurements . I t

appears that either of these approxima

tions could be used to give a f irst order

estimate of the relationship between

hemisphere size and the spacing of the

first two reflected peaks.

Af ter comple t ing the s tud ies wi th the

milled hemispherical surfaces, a set of

exper iments were performed produc ing

weld

  pools on the end of some of the

sample rods and obtaining ultrasonic

reflections using the  MINC-7854  data

acquisition system. Figures 13A and D are

ultrasonic t ime traces obtained while

welding the ends of four of these sample

rods.

The first noticeable feature in these

traces is the large amo unt of noise

present between the init ia l pulse and the

first returned reflection as noted in Fig.

13A. This noise does not appear in the

traces between pulses and appears to

attenuate with t ime. This seems to  indi

cate that the noise is in fact caused by

reflections from something present in the

r o d .  This could be the result of material

variations or inclusions of oxide or carb on

present in the rod. At low temperatures

.10 .20

Radius ( in inches)

Fig.  12 — Shoulder-to-peak time difference vs.

hemisphe re and pool radius. The lower line

represents the prediction of equation (14)

while the upper is from equation (13)

the material properties of such inclusions

may not vary enough f rom the propert ies

of steel to cause reflections. However,

the material prope rties of these inclusions

at elevated temperatures can be different

from those of steel result ing in reflections

from the inclusions.

The amplitudes of the rod end reflec

tions in Figs. 13A and D are much smaller

than those in Figs. 11A and D. This is a

result of the increased attenuation as the

rod heats up. The transit times to the first

rod end reflection are on the order of 53

micro-s, roughly 2 micro-s s lower than in

the case of the cold rods. This corre

sponds to an average speed of sound in

the rod o f 5750 m/s as oppo sed to the

speed o f sound in a rod a t room temper

ature which is 5900 m/s. Finally, notice

the presence of mult ip le pulses in the rod

end reflections. These resemble the

reflections from rods with mil led hemi

spheres, although attenuated in ampli

tude.

The time difference between the small,

initial peak and the first major peak was

again measured, and equation (13) was

appl ied to es t imate the poo l depth . (A l

though identify ing this in it ia l peak

appears tenuous from Fig.  13 ,  it was

easily and consistently distinguished from

noise by examination of the digitally

stored ultrasound signal.) After welding,

the rods were sec t ioned and e tched wi th

a Ni ta l so lu t ion and the w eld p oo l d im en

sions measured using the toolmaker's

microscope. These measurements have

been plotted against the corresponding

t ime d i f fe rence between the f i rs t two

reflection peaks in Fig. 12. These results

clearly fal l a long the theoretical curve to

first order. While there are not many data

points, it seems clear that the ultrasonic

reflections from the region of a  weld

WELDING RESEARCH SUPPLEMENT |  279-s

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 8/9

2 . 7 7 —

1  . 8 6 —

•p

  -

3

  8 . 8 2  —

g-  e.ae —

- e . 9 2 —

'

1

\^J^/^^^

]  

®

1

0 . 0

a.   -13  in. deep

1 8 . 2

  3 0 . 4

  S-4.S  7 2 . 7

T m «  Cm i  c r o - s e c )

18.2

c.

  .19 in.deep

1 1 1 1 1

  r

3 8 4  5 4 . 5  7 2 . 7

T I

 me

  Cm I

  c r o - s e c )

- 1 1 1 1 1 1 1 1 1 1

. 0

  1 8 . 2 3 8 . 4

  5 4 . 5 7 2 . 7

  9 0 . 9

T i

 m«

  Cm i

  c r o - s « c )

b.

  .17  in deep

3

-

a

E

2 . 7 7—

1 .

  8 S —

0 . 9 Z —

0  2 0 -

- 0

  .

  0 2 —

- 1 .  8 S —

r

©

I

 

I

 

0 . 0  1 8 . 2 3 S . - 4  6 * . S  7 2 . 7 9 0 . 9

TIm*s

  Cm i

  c r o - » « c )

d.

  .27 in deep

Fig.  13 —Reflections  from weld pools of various depths: A  — 0.13  in. deep;   B-0.17  in. deep;

  C—0.19

  in. deep;   D-0.27  in. deep

p o o l a r e b e h a v i n g a s e x p e c t e d .

C o n c l u s i o n

T h e c o n c e p t o f u s i n g r e f l e c t i o n u l t r a

s o u n d m e t h o d s t o m e a s u r e t h e si ze o f a

w e l d p o o l in - p r o c e s s h a s b e e n a d v a n c e d .

W h i l e t e c h n i q u e s f o r d i r e c t s h a p e m e a

s u r e m e n t b y d e c o d i n g t h e e n t i r e u l t r a

s o u n d r e f l e c ti o n ar e u n d e r d e v e l o p m e n t ,

t h e a d d i t i o n a l p r o b l e m s i n t r o d u c e d b y

h i g h t e m p e r a t u r e s a n d l a r g e t e m p e r a t u r e

g r a d i e n t i n w e l d i n g r e q u i r e s t h a t s i m p l e r

a p p r o a c h e s b e f i r s t c o n s i d e r e d .

I n t h i s p a p e r , t h e r e f l e c t i o n p a t t e r n

f r o m h e m i s p h e r e s in t h e e n d o f  long

c y l i n d e rs h a v e b e e n p r e d i c t e d a n d v e r i

f i e d e x p e r i m e n t a l l y . M o r e i m p o r t a n t l y ,

w h e n t h e h e m i s p h e r i c a l s h a p e r e p r e s e n t s

th e w e l d p o o l - s o l i d m e ta l i n t e r f a c e , i t h a s

b e e n s h o w n t h a t a c c u r a t e r a d i u s m e a

s u r e m e n t s c a n b e m a d e . S i nc e t h e c o m

p le t i on o f th i s w o rk , resu l ts o f a s im i l a r

c o n c u r r e n t s t u d y b y L o t t ( R e f . 2 1 ) h a v e

b e e n r e p o r t e d , w i t h si m il a r c o n c l u s i o n s

r e g a r d i n g w e l d

  poo l

  s i ze m e a s u r e m e n t .

Th u s t h e i ni t ia l s t e p t o w a r d r e a l i z a t i o n o f

r e a l t i m e w e l d p o o l c r o s s s e c t i o n m e a

s u r e m e n t f o r in - p r o c e s s c o n t r o l p u r p o s e s

h a s b e e n s u c c e s s fu l .

cknowledgments

Th e w o r k d i s c u s s e d in t h i s p a p e r w a s

s u p p o r t e d b y t h e U .S . O f f i c e o f N a v a l

R e s e a r c h u n d e r c o n t r a c t n o . N 0 0 0 1 4 -

8 0 - C - 0 3 8 4 .

References

1.  V r o m a n , A .  R„  and Brandt, H. 1976.

Feedback contro l o f  GTA  weld ing using

  p u d

d le width measurements.  Welding lournal

55(9): 742-749.

2.

  Richardson, R. W .,  C u t o w ,  D. A., and

Rao, S. H. 1982 (Nov.). A vision based system

for weld pool s ize contro l .

  Measurement and

control for batch manufacturing,

  ASME Special

Publication,

  pp. 65-75.

3 . Nomura ,

  et al.

  1980 (Sept.). Arc l ight

in tensi ty contro ls current in SA weld ing sys

tem.  Welding and Metal Fabrication:

  457 -

463.

4.

  Ca r l ow ,

  David. 1982 (June).  Closed loop

control of full penetration welds using op tical

sensing of back bead width.  S.M. Thesis, Dept.

o f  M.E.,  MIT.

5.

  Reiff, I.

  R. 1983 (Feb.).

  Closed-loop con

trol of backside pudd le width in the gas

tungsten arc weld process.

  M.S. thesis, Depart

ment of Mechanica l Engineer ing, M.I.T.

6. Zacksenh ouse, M., and Hardt, D. E. 1983

(Oct.) . Weld  pool impedance ident i f icat ion for

s i ze measuremen t and con t ro l .  Trans. ASME,

lournal of Dynamic Systems, Measurement

and Control  104 (3).

7. Richards on, R. W ., an d Ren wick , R. ).,

Exper imenta l invest igat ion of GTA weld pool

osci l lations.  Welding Journal   62(2):29-s to

35-s.

8. Freeman, A. 1962. A mechanism of

acoust ic echo formation.

  Acoustica

  12: 10-

21.

9. Pao, Y. H. and Wol fgang Sachse. 1974

(Nov.) . In terpretat ion of t ime records and

power spectra of scattered u l trasonic pulses in

solids.  J.A.S.A.   56(5): 1478-1486.

10. Ger icke, D. R. 1963. Determinat ion of

the geometry of h idden defects by u l trasonic

pulse analysis testing.  J.A.S.A.   35: 364.

11 .

  Sachse, W ol fg ang . 1975 (Apr i l) . Deter

minat ion of the s ize and mechanica l propert ies

of a cyl indrical fluid inclusion in an elastic

  solid.

Materials Evaluation

  33: 81-88.

12 .  Tittman, B. R. ef

  al.

  1978. Scattering of

longi tud inal waves incident o n a spherica l cav

ity in a

 sol id.

  J.A.S.A.   63(1): 68.

13. W hi te , R. M . 1958 . Elastic wa ve scatter

ing at a cyl indrical discontinuity in a sol id.

J.A.S.A.  30(8) : 771-785.

14.

  Th o m p s o n , D.

 O.

  and Thompson, R. B.

1979.

  Quant i ta t ive u l trasonics.

  Phil.  Trans. R.

Soc. London.

  292: 233-250.

15 .

  Rose, |. L. ef

  al.

  1980 (Aug.). Flaw

classi f icat ion in welded p lates using a micro

processor contro l led f law detector .

  ND T   Inter

ims  I SEPTEMBER

  1984

8/10/2019 Ultrasonic Measurement of Weld Penetration

http://slidepdf.com/reader/full/ultrasonic-measurement-of-weld-penetration 9/9

national 13(4): 159-164.

16.

  We bber, G. M. B., and Stephens, R. W.

B. 1968. Transmission of sound in molten

metals.

 Physical acoustics: principles

 and

  meth

ods,

  ed. by Warren P. Mason. New York:

Academic Press.

17.  Papadakis, E. P. ef

  al .

 1972. Ultrasonic

attenuation and velocity in hot specimens by

the momentary contact method with pressure

coupling and some results in steel to 1200 C.

J.A.S.A. 52(3): 850-857.

18.  Cam evale, E. H. ef  al.  1964. Ultrasonic

evaluation of elastic moduli at elevated  tem

peratures using momentary contact.

  J.A.S.A.

36(9): 1678-1684.

19.

  Rumbold, |. G. and Krupski, S. 1981.

Ultrasonic thickness variation measurement of

hot forged cannon tubes.

 Materials Evaluation

39 9): 939-942.

20.

  Schmitz,

 V. U„  and Becker, F. L. 1982.

Scattering of shear wave pulses by surface

breaking cracks —time and frequency domain

analysis. Materials Evaluation  40(2): 191-197.

21.

  Lott, L. A. 1984 (March). Ultrasonic

detection of the molten/solid interfaces of

weld pools.

 Materials Evaluation

  42.

WRC Bullet in 291

January, 1984

Fracture Control  of  Pressure Vessels  Up To

 2

l

A

  Inches Thick

by

 P. 0. Metz

T h i s r e p o r t w a s p r e p a r e d i n r e s p o n s e t o a r e q u e s t f r o m t h e S u b g r o u p o n T o u g h n e s s o f t h e P r e s s u r e

Vesse l and Bo i le r Code Commi t tee o f the Amer ican Soc ie ty o f Mechan ica l Eng inee rs . I t s pu rpose i s to

p r o v i d e a g e n e r a l f o r m a t f o r a s s e s s i n g e l a s t i c - p la s t i c f r a c t u r e i n a f r a c t u r e - c o n t r o l p l a n f o r s t r u c t u r a l

s t e e l v e s s e l s o r m e m b e r s  2Vz  in . o r less in th ick nes s . A gene ra l ove rv iew o f the s ub je c t o f e las t i c -p las t i c

f r a c t u r e i s p r e s e n t e d .

P u b l i c a t i o n o f t h i s r e p o r t w a s s p o n s o r e d b y t h e S u b c o m m i t t e e o n F a i l u r e M o d e s i n P r e s s u r e V e s s e l

M a t e r i a l s o f t h e P r e s s u r e V e s s e l R e s e a r c h C o m m i t t e e o f t h e W e l d i n g R e s e a r c h C o u n c i l .

The p r i ce o f WRC Bu l le t in 291 i s $12 pe r copy , p lus $5 fo r pos tage and hand l ing . Orde rs shou ld be sen t

w i t h p a y m e n t t o t h e W e l d i n g R e s e a r c h C o u n c i l , R o o m 1 3 0 1 , 3 4 5 E . 4 7 t h S t . , N ew Y o r k , N Y 1 0 0 1 7 .

WRC Bullet in 290

December , 1983

Factors Affecting Porosity

  in

 Aluminum

  Welds—A

  Review

by J. H. Devletian and W. E. Wood

T h e e m p h a s i s o f t h i s r e p o r t i s o n t h e v a r i o u s f a c t o r s a f f e c t i n g t h e p o r o s i t y i n a l u m i n u m w e l d s

d e p o s i t e d b y t h e G T A W a n d G M A W p r o c e s s e s . A t o t a l o f 9 4 t e c h n i c a l p a p e r s w e r e r e v i e w e d a n d

e v a l u a t e d i n t h e p r e p a r a t i o n o f t h i s r e p o r t t h a t s u m m a r i z e s t h e s t a t e o f t h e a r t o n t h e c a u s e s a n d

f o r m a t i o n o f p o r o s i t y in a l u m i n u m w e l d m e n t s .

P u b l i c a t i o n o f t h i s r e p o r t w a s s p o n s o r e d b y t h e A l u m i n u m A l l o y s C o m m i t t e e o f t h e W e l d i n g R e s e a r c h

C o u n c i l .

The p r i ce o f WRC Bu l le t in 290 i s $1 2 .0 0 pe r cop y , p lus $5 .00 fo r pos tage and hand l ing . Ord e rs s hou ld

b e s e n t w i t h p a y m e n t t o t h e W e l d i n g R e s e a r c h C o u n c i l , R o o m 1 3 0 1 , 3 4 5 E . 4 7 t h S t r e e t , N e w Y o r k , N Y

1 0 0 1 7 .

WRC Bullet in 287

September ,

  1983

Welding  of  Copper and Copper-Base Alloys

by R. J. C. Dawson

T h i s I n t e r p r e t a t i v e R e p o r t d i s c u ss e s t h e c u r r e n t s t a t u s of f u s i o n w e l d i n g t e c h n o l o g y f o r c o p p e r - b a s e

m a t e r i a l s o f m a j o r i n d u s t r i a l i m p o r t a n c e . C u r r e n t w e l d i n g p r a c t i c e s f o r e a c h g r o u p o f c o p p e r - b a s e

m a t e r i a l s a r e d i s c u s s e d . L i t e r a t u r e r e f e r e n c e s a n d s u g g e s t e d f u r t h e r r e a d i n g is a l so p r e s e n t e d .

P u b l i c a t i o n o f t h i s r e p o r t w a s s p o n s o r e d b y t h e

  Interpretative Repo rts Com mittee

  o f th e

  Welding

Research Council.

  The p r i ce o f WRC Bu l le t in 28 7 is $1 2 .0 0 pe r cop y , p lus $5 . 00 fo r pos tage and hand l ing .

O r d e r s s h o u l d b e s e n t w i t h p a y m e n t t o t h e W e l d i n g R e s e a r c h C o u n c i l , R o o m   1 3 0 1 ,  3 4 5 E a s t 4 7 t h S t r e e t ,

N e w Y o r k , N Y 1 0 0 1 7 .