Modification of the grain structure of austenitic welds for improved ultrasonic inspectability

S. Wagner, S. Dugan, S. Stubenrauch, O. Jacobs

MPA Universität Stuttgart

38th MPA-Seminar October 1 and 2, 2012 in Stuttgart

476

Abstract

Austenitic stainless steel welds, which are widely used for example in nuclear power plants

and chemical installations, present major challenges for ultrasonic inspection due to the grain

structure of the weld. Large grains in combination with the elastic anisotropy of the material

lead to increased scattering and affect sound wave propagation in the weld. This results in a

reduced signal-to-noise ratio, and complicates the interpretation of signals and the

localization of defects. The aim of this project is to influence grain growth in the weld during

the welding process to produce smaller grains, in order to improve sound propagation

through the weld, thus improving inspectability. Metallographic sections of the first test welds

have shown that a modification of the grain structure can be achieved by influencing the

grain growth with magnetic fields. For further optimization, test blocks for ultrasonic testing

were manufactured to study sound propagation through the weld and detectability of test

flaws.

1 Introduct ion

1.1 Motivation

For pressurized components operating at high stresses and at high temperatures, the

integrity of the welds is critical for a safe and reliable operation. In the nuclear industry,

(petro-) chemical industry, and also in conventional power plants at higher temperatures,

austenitic welds are widely used. Non-destructive testing is an integral part of the quality

assurance for these welds during the manufacturing process and in-service inspections. For

volumetric inspection and inspection of non-accessible inner surfaces, ultrasonic testing or

radiography can be used. In many cases, ultrasonic testing is the preferred method due to

large wall thickness, limited accessibility, or cost considerations.

The challenges facing ultrasonic testing of austenitic welds result from elastic anisotropy of

the material in combination with the grain structure of the weld. As long as the grains are

randomly oriented and sufficiently small compared to the acoustic wavelength, the anisotropy

of the material does not affect the acoustic properties on a macroscopic scale. Austenitic

welds, however, consist of large grains that are oriented depending on the cooling conditions

during the welding process – so-called dendrites, or dendritic structure, Figure 1. An example

for an austenitic stainless steel weld with distinctly dendritic crystals is shown in Figure 2.

This inhomogeneous, anisotropic structure of the weld strongly affects ultrasound

propagation. The following effects can be observed: local variations in the sound velocity,

deviation from propagation direction expected for isotropic material, scattering at grain

boundaries and at the fusion line, mode conversion. In addition to a decrease in signal-to-

noise ratio, this leads to difficulties for the interpretation of ultrasonic signals with respect to

localization and sizing of defects, and the distinction between reflections from real defects

and false indications.

477

The diff

number

develop

The aim

welding

be achi

dendritic

1.2 C

To avo

approac

approac

optimizi

narrow-

cases,

Figure

ficulties with

r of researc

pment or the

m of the pro

g process in

ieved by cr

c crystallite

Current sta

oid the abo

ches involv

ches uses

ing the weld

-gap welds

however, t

1: Schemati

h ultrasonic

h projects,

e improvem

oject presen

n order to im

reating a m

s.

ate of rese

ove-mention

ving modif

geometrica

d design to

has lead

this type of

c view of gro

Figure 2: a

c inspection

for example

ment of ultras

nted here is

mprove the

more fine-gr

earch and

ned difficul

ications du

al modificat

improve ac

to improve

f weld geo

owth of dend

austenitic we

n of austen

e: [1], [2], [3

sonic inspe

s to influenc

accoustic p

rained micr

technolog

lties for ul

uring the

ions, such

ccessibility

ements con

ometry can

rites. The nu

eld with distin

itic welds h

3] and othe

ection and s

ce the grain

properties o

rostructure,

gy

ltrasonic in

manufactur

as reducin

of the weld

ncerning u

result in v

ucleation star

nctly dendritic

have been t

ers. Most of

ignal analys

growth in t

of the auste

which mea

nspection, t

ring proces

ng the volu

for the insp

ltrasonic in

very long g

rt at the fusio

c structure

the target o

f them deal

sis techniqu

the weld du

nitic weld. T

ans less di

there are

ss. One o

ume of the

pection. Th

nspection. I

grains (thro

on face of the

of a high

with the

ues.

uring the

This can

istinctive

different

of these

weld or

e use of

In some

ugh-wall

e weld

478

extension) in the center of the weld which cause a reflection of the ultrasonic pulse

comparable to an interface reflection [4], [5].

The second type of approach is based on efforts to influence the microstructure of the weld

during the welding process, i.e. the solidification of the weld material. The possibility of

influencing the microstructure in order to create finer grains by the aid of electromagnetic

fields or ultrasound has been demonstrated in different research projects.

1.2.1 Impact of electromagnetic fields

Effects of alternating electromagnetic fields on the liquid state of an aluminum, austenitic

material, on a model liquid (Ga-In-Si-alloy), and on a Ni-based alloy are described in the

literature [6], [7], [8], [9]. Each of these studies was conducted in a laboratory set-up using

melting pots. Grain refinement was observed due to the electromagnetic stirring of the liquid

melt.

Other publications report on the use of magnetic fields to modify the arc during plasma

welding [10], or to influence the laser weldung process for high-alloyed steels [11] or

aluminum alloy [12].

1.2.2 Impact of Ultrasound

High power ultrasonic waves applied to a liquid metal have been described as grain-refining

[13]. In [14] ultrasonic treatment of solidified weld layers of an austenitic low carbon steel was

shown to produce a finer microstructure as well as improved mechanical properties (yield

strength, micro hardness).

1.2.3 Impact of Modifications of the Welding Process

Another method of influencing the microstructure of a weld is to modify the welding process

itself. Leading manufacturers of welding equipment experiment use approaches based on a

reduction of heat input into the component during welding, such as EWM’s “coldArc”

Technology [15] and the CMT-Technology developed by Fronius. With these “cold” welding

processes, significantly shorter cooling times of the weld and modified microstructures can

be achieved.

2 Full Description of the current study

The survey of existing literature has shown that research on methods for modifying the grain

structure of austenitic welds by influencing the welding process is rather limited. Application

of magnetic fields near the arc appears to be the most promising tool to influence the

microstructure. These magnetic fields can be static or fluctuating. In this investigation,

magnetic fields, as well as a modification of parameters for the welding process are applied.

479

2.1 E

Electrom

growth

melt and

detachm

of grain

improve

For the

automa

impleme

with the

(Figure

With thi

1.4432)

investig

The mo

100 Hz

metallog

pass. T

significa

Figu

varia

Experimen

magnetic fie

of dendritic

d causes lo

ment of den

n growth is

ement of we

e applicatio

ted TIG-w

ented. The

e TIG-elect

5).

is experime

) were crea

gations the w

ost significan

z for the

graphic sec

The dendrite

antly reduce

ure 3: Schem

ations in chem

nts using

elds result i

c structures.

ocal variatio

ndrite tips, c

influenced

elding seam

on of a pu

welding sys

electromag

rode at con

ental set-up

ated, applyi

welding fille

nt changes

applied ma

ction of the

es appear

ed.

matic represe

mical compo

magneti

n a modified

. The magn

n in the con

creating new

by the mag

m quality can

ulsed mag

stem equi

gnet was loc

nstant spee

, several m

ing pulsed

er material S

in the struc

agnetic fiel

test weld.

finer and s

ntation of the

sition within

ic fields

d heat trans

neto-hydrod

ncentration

w sources

gnetic field

n be observ

netic field

ipped with

cated below

ed and a c

multi-pass b

magnetic f

SAS 4-IG w

cture of den

ld. Figure

The orienta

shorter. Gro

e detaching o

the weld cau

sfer in the li

ynamic effe

of the alloy

for nucleat

and can the

ved [12], [16

under rea

h an elec

w the weld ro

constant dis

utt welds o

fields at dif

as used.

ndrites were

6 shows

ation of den

owth over s

of dendrites

used by mag

quid melt w

ect creates

ying elemen

ion (Figure

erefore be

6].

alistic weld

ctromagnet

oot and mo

stance of 4

f 16 mm th

fferent puls

e achieved

a photogra

ndrites chan

several wel

tips due to c

gnetic fields (

which influen

a current w

nts. This can

3). The or

influenced.

ing conditi

(Figure

oved simulta

mm from

hick plates (

se rates. Fo

with a puls

aph of an

nges for ea

d passes c

currents and

(S: solid; L: li

nces the

within the

n lead to

ientation

Also an

ons, an

4) was

aneously

the root

(material

or these

e rate of

etched

ach weld

could be

local

iquid)

480

2.2 Ex periments with Influencing the TIG arc

The second technique for influencing the solidification process of the weld used in this study

was the application of a pulsed TIG arc. Figure 7 shows the weld seam which shows the

most significant changes in microstructure obtained with a pulsed TIG arc.

Figure 6: Austenitic weld, welded under influence of an

electromagnetic field, pulse rate of f = 100 Hz

Figure 4: Automated TIG-welder with TIG-

electrode, filler wire feeder, electromagnet, and

test plate with weld seam preparation

Figure 5: Detail of electrode and electromagnet

481

In order to exclude a possible influence of grain structure in the root passes on the formation

of dendrites in the filler weld passes, the geometry of the weld seam preparation was carved

out of the test plate using spark erosion. The photo of the etched metallographic section

(Figure 7) shows a more fine-grained structure of the first filler weld pass compared to a weld

with hand welded root pass (see Figure 6).

2.3 Ex periments with different interpass temperatures

In order to investigate the influence of interpass temperature Tip on the formation of

dendrites, experiments with minimum interpass temperature (room temperature RT), with the

maximum recommended interpass temperature, as well as significantly excessive interpass

temperatures (210 °C – 245 °C) have been performed. For the austenitic Cr-Ni(Mo)-steel

used here, it is recommended not to exceed Tip = 150°C [17].

For the weld seam welded at minimum Tip (RT, Figure 8), the grain structure is almost

identical to the structure for the weld without modifications in the welding process (welded at

Tip ≦ 150°C). For the maximum interpass temperatures, growth of dendrites through several

weld passes can be observed, Figure 9. This can be explained by the higher content of heat

stored in the material due the high interpass temperatures, allowing an extended grain

growth in one preferred direction, and resulting in larger grains of one particular orientation.

Figure 7: Manual TIG-weld, pulsed TIG-arc at 100 Hz, weld seam

preparation by spark erosion, no root gap

482

2.4 Ultrasonic Testing of Influenced Welds

For the first trial of ultrasonic testing of the different welds, test blocks were manufactured

from one non-modified weld (SN17, welded with standard parameters), three pulsed arc

welds (SN 18 - 20), as well as the weld with minimum interpass temperature (SN22). A side

drilled hole (SDH) with a diameter of 1,5 mm was drilled in the center of the weld as a test

reflector (Figure 10). All test blocks have identical geometries.

All test blocks were inspected by mechanized ultrasonic phased array testing, using shear

waves at a frequency of 2.25 MHz, at different angles of incidence. The aim of the

inspections was to analyze the influence of the microstructure on the sound propagation

through the weld. For a better sound propagation a better flaw detectability can be expected.

The influences of the microstructural modifications (SN18 - SN22) are compared with

acoustic characteristics of the non-modified weld (SN17).

Figure 8: Manual TIG weld with minimum Tip. (RT) Figure 9: Manual TIG weld with too high Tip.of 245°

Figure 10: US-test block with side drilled hole in the center of the weld (SN20)

483

Table 1 shows the B-scan projection images (side view) of the phased array inspection for

the 2.25 MHz shear wave with an angle of incidence of 70 °. For a better comparability only

the manual TIG welds with the weld seam preparation by spark erosion are shown.

Table 1: Phased-Array B-scans of test blocks

Type of weld B-scan for 70° angle Macrosection

SN17:

No modification

SN18:

Pulse of light arc

100Hz

SN19:

Pulse of light arc

50 Hz

SN20:

TIG 500 Hz,

SN22:

Min. Interpass-

Temperature

484

For the test blocks SN18 – SN20, obtained with different pulse frequencies of the TIG arc,

the evaluation of the B-scans shows variations in the acoustic characteristics of the welds. It

appears that for low frequencies a better detectability of the defect is achieved. For higher

frequencies the signal-to-noise ratio decreases and a higher noise level due to scattering

from the microstructure of the weld can be observed.

Test block SN22 was welded with the minimum interpass temperature, allowing the weld

beads to cool down to room temperature. It was expected to obtain an improved

microstructure, resulting in improvements in the acoustic properties. This expectation was

confirmed neither by ultrasonic inspection nor by metallography.

SN17, the non-modified test block, shows the characteristic grain structure of austenitic

welds. According to the geometric characteristics of the sound beam the B-scan shows a

good correlation with the theoretical image of the scan, except for the localization of the

defect.

Further ultrasonic inspections are currently carried out in order to verify these first

experimental results and interpretations.

3 Summary and conclusion Metallographic examinations show an influence on the growth of dendrites due to

electromagnetic fields and pulsation of the TIG arc with varying frequencies during the

welding process. Evaluation of the microstructure of electromagnetically influenced welds

leads to the assumption that for low frequencies (approx. 100 Hz), the strongest

modifications of the dendritic structure can be achieved: Significant changes in the

orientation of dendrites, more fine-acicular grains are observed, and a growth through

several weld passes could be avoided. The assumption that this also leads towards

improved acoustic characteristics for ultrasonic inspection has yet to be verified.

For the modification of the light arc the following results can be observed: For pulsed arcs

with low frequencies, microstructures with short dendrites with different orientations can be

produced. The experiments show a fine-grained area in the first weld pass. The influence of

nucleation from the root pass structure could be avoided by carving the weld geometry out of

a plate, thus creating a weld without a root.

For optimization and evaluation of influencing parameters, ultrasound propagation

experiments were performed using ultrasound test blocks with reference flaws (1,5 mm

SDH). By studying ultrasonic propagation through the weld, influences on the acoustic

properties of the usually anisotropic welds can be observed. The first results of ultrasonic

inspection lead towards the assumption, that for lower pulse frequencies an improved sound

transmission can be obtained.

485

An impact of a reduced interpass temperature on the growth of dendrites could not be

observed. Also, the maximum applied interpass temperature in the range of 210 °C – 245 °C

did not show significant changes in the microstructure. Ultrasonic inspections of these welds

are currently under way.

Further studies considering the effects of high power ultrasound on the solidification and

grain growth in the weld will also be conducted as within this research project.

References

[1] Williams, P. (1976). Ultrasonic testing of austenitic steels (a selectiv bibliography

1958-1975). (Bd. Bibliography 252). London: Central Electicity Generating Board.

[2] Neumann, E. et al. (1995). Ultraschallprüfung von austenitischen Plattierungen,

Mischnähten und austenitischen Schweißnähten (Kontakt & Studium), Bd. 377.

Expert-Verlag.

[3] Just, T., Kuhlow, D., Matthies, K., Römer, M. (1978). Über den Stand der

Entwicklungen einer Ultraschallprüpftechnik für austenitsiche Schweißverbindungen

(Schweißen in der Kerntechnik Ausg. Nr. 52). Hamburg: DVS-Berichte .

[4] Schmid, R. (1995), Ultraschallprüfung von austenitischen Plattierungen, Mischnähten

und austenitischen Schweißnähten, Eberhard Neumann (Hrsg), Expert-Verlag,

(Kontakt und Studium) Bd. 377, Kapitel 4

[5] „Prüfbarkeit von dickwandigen Bauteilen aus Nickellegierungen und

Schweißverbindungen mit zerstörungsfreien Prüfmethoden“ gemeinsamer

Abschlussbericht der Forschungsstellen BAM Berlin, IZfP Saarbrücken, MPA

Stuttgart zum Vorhaben COORETEC TD-1 (unveröffentlicht)

[6] Lu, D. (2007). Refinement of primary Si in hypereutectic Al–Si alloy by

electromagnetic stirring. Journal of Materials Processing Technology, p. 13 - 18.

[7] Gao, Y.-L. (2004). Comparative study on structural transformation of low-melting pure

Al and high-melting stainless steel under external pulsed magnetic field.

Materialwissenschaft und Werkstofftechnologie, p. 385 - 395.

[8] Wang, X. D. (2005). Two kinds of magnetic fields induced by one pair of rotating

permanent magnets and their application in stirring and controlling molten metal

flows. Journal of Crystal Growth, p. e1473 - e1479.

[9] Jin, W. (2008). Grain refinement of superalloy IN100 under the action of rotary

magnetic fields and inoculants. Materials Letters, p. 1585 - 1588.

486

[10] Cheng, J. (August 2008). Effect of electromagnetic stirring on the microstructure and

wear behavior of iron-based composite coatings. Journal of University of Science and

Technology Beijing, Number 4, p. 451.

[11] Berger, P., & Graf, T. (2008). Gestaltung und Kontrolle des Nahtdurchhangs beim

Schweißen. Forschungsvorhaben IFSW Universität Stuttgart, IGF-Vorhaben

14.959N.

[12] Lindenau, D. (2006). Magnetisch beeinflusstes Laserstrahlschweißen (Bd.

Dissertation IFSW Universität Stuttgart). München: Herbert Utz Verlag GmbH.

[13] Zhang, Z. (Feb. 2009). influence of high-intensity ultrasonic treatment on the phase

morphology of Mg-9.0wt%Al binary alloy. Rare Metals, p. 86.

[14] Gust, W. (1999). ultrasonic shock treatment of welded joints. Materials Science, p.

678 - 683.

[15] Kocab, H. (2010). Neue Lichtbogenform für hohe Wirtschaftlichkeit und Nahtqualität.

Schweißen im Anlagen- und Behälterbau, Sondertagung München, S. 40 - 46.

[16] Kern, M., Berger, P., Hügel, H. (2000) Magneto-Fluid Dynamic Control of Seam

Quality in CO2 Laser Beam Welding. Welding Journal, p. 72-s – 78-s

[17] Böhler-Handbuch “Wissenswertes für den Schweißer” Ausgabe 09/ 2010, Böhler

Schweißtechnik Austria GmbH

487