Вы находитесь на странице: 1из 7

722 kir.


Structural Integrity Assessment Applying Ultrasonic Testing

dr Miodrag Kiri dipl. ing.
The Institute for Materials Testing of Republic Serbia
Boul. V. Miia 43, Belgrade

European standards for ultrasonic examination of welded joints are discussed with regard to the methods for sensitivity setting.
It is especialy important the frequently used method 1 based on side drilled cylindrical holes which is compared with ASME
standard. For methods 1 and 2 are given expressions for received acoustical signals reflected from corresponding reflectors in
far field of the normal and angle probes. The possibility of application of notch beyond the standard limits is considered, as well
two concepts of ultrasonic testing of welded joints and the characterization of discontinuities as planar.
The experiences gained in pressure vessels testing have shown that manual ultrasonic testing (UT) by using pulse-echo
technique compared to other nondestructive methods is the best method for discontinuities characterization and revealing because of
its versatility and possibility of quantification and memorization of results enabled by digital ultrasonic flaw detectors.
The aim of UT determines the technique to be applied, also its sensitivity and testing levels [1,4-6]. The determination of crack
features may include: basic ultrasonic parameters, crack parameters and its proximity to the surface or to other discontinuities as
well its basic shape and orientation.
The paper considers UT of welded joints from the European standards point of view. It is a multitask testing because it is
aimed at detection of discontinuities, characterization of discontinuities and evaluation of discontinuities, by taking into account the
geometry and welding technology. When a discontinuity is detected, the determination of its features is necessary for its
evaluation with respect to known acceptable criteria. The differentiation between the quality control concept and the fitness-forpurpose concept should be made, [1]. In order to correctly identify the discontinuity types specified in the acceptance criteria, or to
make a correct fitness-for-purpose evaluation, it may be necessary to make a more detailed sizing and assessment of the shape
of the discontinuity, [2,7]. Echodynamic patterns obtained as transient echo shapes during probe movement, are used as one of
criteria for the classification, as well for surface discontinuity sizing. Surface notches of equal lengths and different depths give
echo envelopes whose widths can be related to actual notch depths, [3], thus it is possible to evaluate the discontinuity depth.
Evaluation of indications detected by testing of welded joints, is based on reference levels, established by standard EN
1712 [4]. The setting of sensitivity and reference level for method 1 is performed by 3 mm diameter side drilled hole (SDH)
perpendicular to ultrasonic beam axis. The diameter is the same for different frequencies and parent material thickness (PMT),
while according to ASME Section V it is dependent on PMT. For method 2 of EN 1712 (the distance gain size system, or
DGS), ultrasonic probe frequencies (f) and reference levels depend on PMT, as given in Table 1 and Table 2 for longitudinal
(L) and transverse (T) waves. Reference levels for method 2 are defined by the diameter d of equivalent plane circular reflector
(FBH), perpendicular to ultrasonic beam axis. Evaluation level, related to reference level, is important from two reasons: all
indications equal to or exceeding evaluation level shall be evaluated and the other is that it is used to evaluate the lateral
dimension of an indication [4,5].
TABLE 1. Reference levels for L-waves and normal probes for method 2 (DGS), [4, 6]
Nominal probe frequency
1,5 to 2,5
3 to 5

Thickness of parent material, t (mm)


t <15


t < 40



d=2 mm

d=3 mm

d=2 mm

d=2 mm

d=3 mm

TABLE 2. Reference levels for T-waves and angle probes for method 2 (DGS), [4, 6]
Nominal probe frequency
1,5 to 2,5
3 to 5

Thickness of parent material, t (mm)


t <15


t < 40



d=2 mm

d=3 mm

D=1 mm

d=1,5 mm

One remark should be given and it is that sensitivity required for T-waves is higher than for L-waves if f is from 3 to 5 MHz,
because shorter wave length of T-waves compared to wave length of L-waves at the same frequency allows to detect the reflector
of smaller diameter d. However, at higher frequencies of L-waves, reference levels are the same as for lower frequencies.
Similar question can be put regard to the reference level for tandem examination; it is a distance amplitude curve (DACcurve) for flat bottom hole of diameter 6 mm, which is the same when PMT or probe frequency changes. DAC curve is at the
same time the evaluation level and recording (acceptance) level for discontinuities perpendicular to the surface.
Three acceptance levels for method 2 are illustrated in Fig. 1 according to EN 1712 (the first of them should be the level 1). It
is to be noted that EN 1714 [6] specifies methods for manual UT of fusion welded joints in metallic materials equal to and above 8
mm thick and EN 1713 requires that the examined depth should be greater than 5 mm, so that the surface layer of material up to
5 mm should not be tested [7]. However, according to ASTM E 169, UT is performed if metallic material thickness is above 6 mm.
Relative echo amplitude
level 3


Reference level



0,5 t

1,5 t


Length of indications

Acceptance level
Evaluation level

Reference level: 0 dB
Recording levels: Acceptance level 2 = Reference level
Acceptance level 3 = Reference level + 4 dB
FIGURE 1. EN 1712 acceptance levels for method 2 and 15 mm t < 100 mm
ASME standard Section XI gives acceptance criteria for internal as well for surface flaws, based on linear fracture mechanics
relations of Newman and Raju [8]. The criteria for surface planar flaws, perpendicular to the surface, are illustrated here by the
diagram in Fig. 2. The acceptable flaw sizes, a and l correspond to points below the curve for given PMT. It is evident the

resemblance of Fig. 1 with discrete levels and Fig. 2 with continual levels. For thick materials, the influence of
value of a is small if

on allowable

> 25 mm according to Fig. 2, but according to Fig. 1 it is neglected in three intervals of flaw length

dependent on PMT, with transitions between them described by step functions of length


Complementary examination should be carried out in the case of any doubt by using additional probes, by analysis of
echodynamic pattern in the lateral movement or using other NDT as radiography, or time of flight ultrasound diffraction technique,
TOFD, which is more appropriate for defect sizing in the material thickness direction. TOFD is the approach whose significance is
that it is not amplitude dependent and the principles involved (energy is dissipated only from the extremities of discontinuity),
enable precise through-wall size and location measurement of detectable defects. It is standardized by EN 583-6 [9].

Flaw height
a, mm

Surface planar flaws

D=100 mm
D= 65 mm
D= 13 mm
The boundary between significant
and non-significant flaws

Unacceptable flaws

Acceptable flaws






Length l, mm

FIGURE 2. The dependence of allowable height a on length

for surface planar flaws according to ASME Section XI


Ultrasonic frequency is among the most important probe parameters, because it influences beam shape, resolution, wave
attenuation and, as it will be shown by taking in mind the data given in Tables 1 and 2, the ratio of received and emitted acoustical
signal P/P0. UT is greatly determined by area of piezoelectric plate, Sa , the distance of a reflector, r, from the entrance point of
beam axis on test surface and the near field length

N = S a = S a f v


where is wave length and v is wave velocity. Half beam angle defined by the first wave amplitude minimum is

= arc sin(0,61 / a ) = arc sin(0,61v / af )


It is assumed that artificial reflector size 2b fulfils the condition for proportionality between echo amplitude and reflector area Sb,
oriented normally to the beam axis:

< 2b < 0, 4 W


where the half width of the main leaf of the ultrasonic field for a normal probe, W, depends on distance r and angle in

W = r


Thus the condition (3) at small angle becomes

< 2b < 0,244



and at the end of the near field it is approximately

< 2b < 0, 2a


If the reflector is flat bottom hole (FBH) in a normal probe far field, the ratio of received and emitted acoustical signal is [10]

Pfbh P0 =

Sa Sb

2 r 2



where is attenuation coefficient for L-waves in the material.

For a long side drilled hole (SDH) in far field, perpendicular to beam axis, the ratio is also dependent on frequency:

Pcyl P0 =


b 2r


For a FBH perpendicular to beam axis in far field of an angle probe, holds an approximative formula for Pfbh P

Pfbh P0 =

Dlt S a S b cos
2 ( 1 r1
t2 ( r + r2 ) 2 cos

+ t r)


where t is wave length of T-waves, Dlt is the transmission coefficient for incident L-waves at the interface between an angle
beam probe and material under testing, 1 and r1 are the attenuation coefficient and the path length for L-waves in refracting
prism, respectively, r2 is the path length for L-waves in the prism emitted by the equivalent piezoelectric plate and is incident
angle in probe and is refracted angle in the material [10].
For a SDH perpendicular to beam axis in far field of an angle beam probe, the ratio of received and emitted acoustical
signal is

Pcyl P0 =

Dlt S a cos
2t cos

2 (1 r1 + t r )
( r + r2 ) 3


The ratio of received and emitted acoustical signal for a FBH in far field is proportional to the second power of frequency
for a normal probe as well for an angle probe, eq. (7) and eq. (9).
The ratio of received and emitted acoustical signal for a SDH in far field is proportional to frequency for a normal probe as
well for an angle probe, eq. (8) and eq. (10).
The interpretation of these relations is that if frequency increases, the ratio of received and emitted acoustical signal for
both probes should increase proportional to f for SDH and proportional to f for FBH. If other parameters and reflector to probe
distance are the same and if attenuation change is not considered, the effect of increased frequency is the same as it was the
effect of increased b , or Sb for SDH. The effect when FBH is used is more pronounced, because it is proportional to f . It is
shown that the change of Dlt is not significant if is between 38 and 53 , corresponding to from about 45 to 70 [10,11].
It is a simple task to find out the relation between sizes of equivalent FBH and SDH at the same distances in far field. For
an angle beam probe it is shown in [12] that the diameter of an SDH, equivalent to the FBH respect to the acoustical signal is

2b =

8 Sb


t (r + r2 )

The diameter 2b of equivalent SDH depends on Sb, the surface of FBH, but also on f . Similar relation holds for a normal
probe and L-waves, too. It can be concluded that if different sizes of FBH, dependent on frequency, are used for sensitivity
setting, according to the relation (11), the diameter 2b of equivalent SDH should change with Sb and independently, also with
frequency as f . However, according to EN 1712 ultrasonic probe frequencies and reference levels (diameters of FBH) for
method 2 depend on PMT, but for method 1 diameter of SDH does not.
Standard EN 1712 method 3 for sensitivity setting if a probe angle 70 is used for the thickness range 8 mm t<15 mm is
based on a rectangular notch with a depth of 1 mm.The reference level is equal to a DAC curve for the notch. Notch as calibration
reflector is defined also by ASME standard. Here will be discussed the use of notch for UT with regard to EN 1712 and EN 1713.

The research of notches as reflectors for distance scale and sensitivity calibrations proved that they are well defined and
easy applicable reflectors. This can be illustrated by nothces made using electric-discharge machine in a 30 mm thick plate,
perpendicular to its surface, made of low-carbon HSLA steel SN 11484.1 for a spherical tank and cutted from its wall. The
notch depths are 1, 2 and 3 mm, lengths are 12 mm and the notch width is 0,3 mm. The A-scan and echodynamic pattern from
1 mm notch depth are given in Figs. 3 and 4 respectively. The echoes are obtained using the double traverse technique and a
miniature 70, 4 MHz probe. Echo height is given in percents of the full screen height (FSH) at the top of the screen. Thus, UT
can detect the notch of depth 1 mm from the distance 175,8 mm, simulating a shallow surface crack, when the material
thickness is 30 mm. The A-scan and echodynamic pattern from 2 mm notch depth and the same angle beam probe are also
given in Figs. 5 and 6 respectively.

Figure 3. Echo from the 1 mm notch depth

Figure 4. Dynamic echo from the 1 mm notch depth

(70, 4 MHz probe, distance 175,8 mm)

(70, 4 MHz probe, echo beginning at 172,6 mm)

Figure 5. Echo from the 2 mm notch depth

Figure 6. Dynamic echo from the 2 mm notch depth

(70, 4 MHz probe, distance 175,8 mm)

(70, 4 MHz probe, echo beginning at 178,9 mm)

The notches are compared to SDH of 3,2 mm diameter by comparing their echos, Figs. 7 and 8. According to the ASME
Section V, the reference calibration reflector for material thickness equal to 30 mm is the SDH 3,2 mm in diameter. The DAC
curve is constructed for the same miniature angle probe. The test instrument calculated and displayed the projection distance
PR=71,8 mm and the notch depth D=29,1 mm in Fig. 7 because the direct scan technique is used. From Figs. 7 and 8 it is seen
that notch echos are higher than DAC curve for 2 dB for 1 mm notch and for 6,5 dB for 2 mm notch, thus the SDH 3,2 mm
echo at the same distance is the smallest one. It should be noted that notch lengths are 12 mm.

Figure 7. DAC curve and the echo from the 1 mm notch

Figure 8. DAC curve and the echo from the 2 mm notch

By regression analysis it was obtained the dependence of ultrasonically determined notch depth from known notch depth.
According to ASME standard, as the measure of the notch depth was taken the dynamic echo width at the half of the echo
amplitude. For small depths, up to 3 mm, it was obtained a linear regression given in Fig. 9. The linear regression is explained
so that the echo amplitude is a linear function of notch size that reflects the beam, because notch lengths are equal. It should
be noted that correlation coefficient is high (R=1) and it supports linear character of the relationship between measured and
known notch depths.
The conclusion may be that ultrasonic examination can reveal and evaluate surface notches of limited length even at PMT
equal to 30 mm. Thus the restriction of EN 1713 standard concerning the surface layer which is not tested is too conservative.


The depth of the notch, mm





Y = - 0,625+0,25 X
R=1,000 SD=0







The width of dynamic echo, mm

Figure 9. The linear relationship between the echo width and notch depth [3]
UT of welds for structural integrity assessment is a multitask examination because it is aimed at detection of discontinuities,
characterization of discontinuities and evaluation of discontinuities. UT of welded joints has one of two functions:
1. The control of weld quality based on flaw acceptance criteria. This is the quality control concept.
2. To give a reasonable guarantee that the weld under examination contains no discontinuities which can bring into danger its
structural integrity in normal exploitation. This is the fitness-for-purpose concept.
The acceptability of a welded joint is assessed from one of two concepts. Standards for acceptance or rejection must take
into account the safety reasons, the engineering aspects of the structure, its cost and required service performance. The
second concept allows greater discontinuities then the quality control concept. Cracks can be allowed under some conditions,
too [13]. Some details of two concepts are compared in Table 2.

Weld quality
Quality control



TABLE 3. Comparison of two concepts of UT [1]

Structural integrity
on engineering Fitness-for-purpose criteria
mechanics calculations

1. Defect evaluation
Measurable defect parameters:
1.1 Maximum echo indication
1.2 Probe movement along and perpendicular to weld
1.3 Number of echos when scanning
1.4 Discontinuity orientation and nature
2. Procedure characteristic
2.1 Directed towards most serious discontinuities and
expected (manufactured) discontinuities
2.2 Sensitivity setting which determines the ability to
detect certain (the smallest allowable) discontinuities
2.3 Ussually limited number of probes (it increases with
quality class)
2.4 Thick welds are usually considered divided into
thickness zones and examined independently
2.5 Detection probability depends on quality class;
modest operators skill is necessary




Geometrical defect characteristics:

1.1 Height (through thickness size)
1.2 Length
1.3 Proximity to scanning surface
1.4 Orientation and nature
2.1 Directed towards most serious discontinuities and
expected (service-induced) discontinuities
2.2 Sufficiently high sensitivity
2.3 Several probes and beam directions; the use of
additional creeping wave probes.
2.4 In some cases it may be sufficient to examine outer
(surface) layers to demonstrate fitness-for-purpose
2.5 Detection probability is high; good operators skill is

The criteria for a discontinuity to classify as planar or non-planar are of primary importance. The criteria according to EN
1713 are :
- welding technique,
- location of the discontinuity,
- maximum echo height,
- directional reflectivity,
- Ascan,
- echodynamic pattern.
The classification requires an examination each of the parameters against all the others. The procedure of the
classification is stopped as soon as one of the above criteria is fulfilled and can be given by a flow chart diagram. High echo
amplitude is characteristic for a planar discontinuity. It is compared to the reference level increased by 6 dB. A discontinuity is
classified as planar if there is high directional reflectivity and a high echo in the A-scan. It means that there is a diference in
echo amplitudes of at least 9 dB between two angles of incidence of the discontinuity examination using shear waves. If
testing is carried out with shear waves from one direction and with longitudinal waves from the other direction, the signal
difference of at least 15 dB is indicative of a planar discontinuity. If directional reflectivity is low, the next step is the
examination of A-scan. If it doesn`t show a single smooth echo, the echodynamic pattern is examined.
Evaluation of discontinuities may include discrimination between planar and non-planar discontinuities as the primary
discrimination of an acceptable or rejectable indication. In this case all discontinuities above the evaluation level shall be
characterized and if characterized as planar, shall be rejected [4].
From relations and illustrations given in this paper, it is seen that UT is an efficient and useful method for structural integrity
assessment, because it provides high sensitivity and versatility in manual technique, but, additionally it is, with much greater
investments, very useful when automated. The reliability of UT depends on acceptance criteria and testing levels, but also on
operators skill and attention. The application of European standards criteria requires their good knowledge and training. They
allow the application of two concepts, one based on criteria for quality assurance and the other based on fracture mechanics.
The discussable is the setting of sensitivity for method 1 of EN 1712 by 3 mm diameter SDH, because received acoustical
signal depends among others on SDH diameter, frequency and refracted angle chosen regarding PMT. In addition, there is a
simple relation derived between diameter of FBH in far field and the diameter of an equivalent SDH at the same distance. Also,
the echograms with A-scans and echodynamic patterns given here illustrate the possibility for the use of rectangular notch as
calibration reflector beyond the limits determined by EN 1712.
1. Kiri, B.M. and Fertilio M.A., Ultrasonic Assessment of Welded Joints Against Quality Control and Fracture Mechanics
Criteria (in serbian and english), Zavariva, Vol. 40, pp. 209-223, 1995.
2. Kiri, B.M., The approaches of European standards for ultrasonic testing of welded joints of ferritic steels Part II:
Characterization of indications (in serbian), Zavarivanje i zavarene konstrukcije, Vol. 47, 239-245, 2002.
3. Kiri, B.M., The Ph.D. thesis (in serbian), The Faculty for mechanical engineering, Belgrade, 1-251, 2000.
4. EN 1712 Non destructive examination of welds Ultrasonic examination of welded joints Acceptable levels, CEN,
Brussels, aug. 1997.
5. Kiri, B.M., Zavarivanje i zavarene konstrukcije, The approaches of European standards for ultrasonic testing of welded
joints of ferritic steels Part I: Acceptance levels (in serbian), Vol. 47, 167-174, 2002.
6. EN 1714 Non destructive examination of welds Ultrasonic examination of welded joints, CEN, Brussels, aug. 1997/ nov. 2002.
7. EN 1713 Non destructive examination of welds Ultrasonic examination of welded joints Characterization of indications, CEN,
Brussels, may 1998.
8. ASME Boiler and Pressure Vessel Code, Section XI Division 1 Rules for Inservice Inspection of Nuclear Power Plant
Components, 1980 Edition, SI Edition 1983 and later editions.
9. EN 583-6 Non destructive testingUltrasonic examinationPart 6: Time-of-flight diffraction technique as a method for detection
and sizing of discontinuities, CEN, Brussels.
10. Gurvi, A.K., Ermolov, I.N., Ultrazvukavoj kantrolj svarnih vov (in russian), Tehnika, Kijev, 1972.
11. Ermolov, I.N., Progress in the Theory of Ultrasonic Flaw Detection. Problems and Prospects, Russian Journal of
Nondestructive Testing, Vol. 40, No. 10, 2004, pp. 655678. Translated from Defektoskopiya, Vol. 40, No. 10, 2004, pp. 1348.
12. Kiri, M., Analysis of ultrasonic testing of smaller thicknesses butt welded joints with regard to the possibilty of defects
revealing (in serbian), International Symposium - Nondestructive testing in insurance of quality of materials and equipment for
chemical, petrochemical and power supply plants, Sarajevo, 1978. Proceedings, Book II, pp. 5-13.
13. Kiri, M., Testing for detection and evaluation of cracks, International Fracture Mechanics Summer Schools - New Trends in
Fracture Mechanics Applications (IFMASS 9), editor S. Sedmak, Varna (Bulgaria), 19-24. sept. 2005. (CD)