Professional Documents
Culture Documents
06 March 2004
Methods for
Ultrasonic
Examination
of Welds
Part 2: Automatic examination of fusion
welded butt joints in ferritic steels
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
Co-operating Organizations
Contents
Page
Co-operating organizations Inside front cover
Foreword ii
1 Scope 1
2 Definitions 1
3 Operators 1
4 Equipment 1
5 Surface condition 1
6 Parent metal examination 1
7 Weld examination 3
8 Evaluation of imperfections 4
9 Test plates 4
10 Presentation of results 4
Appendix A Guidance on the determination of probe characteristics 5
Appendix B Notes for guidance on the D.G.S. diagram 5
Appendix C Method for setting sensitivities where maximum
sensitivity is required 12
Figure 1 Relationship between ratio of diameter to
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
BSI 01-1999 i
BS 3923-2:1972
Foreword
This British Standard forms part of a series dealing with methods for the
non-destructive testing of welds. It does not state when this particular type of
testing should be employed not does it give standards of acceptance, as both of
these aspects should be covered in the appropriate application standard or be
agreed between the contracting parties.
This standard was first published in 1965 as a result of ultrasonic examination
procedures being agreed between organizations interested in the manufacture
and use of boilers and pressure vessels. Part 2 has now been revised to cover
automatic examination of butt joints in a wide range of shapes and forms of
ferritic steels.
The use of a D.G.S. (Distance, Gain, Size) diagram for obtaining the appropriate
sensitivity setting only has been introduced although it is not the sole method.
Guidance on this method is given in Appendix B.
The revised standard is not associated with any particular type of fabrication, but
has been prepared to cover a wide range of products and as such lays down the
broad principles of automatic ultrasonic examination. It is emphasized that a
satisfactory technique can only be determined after taking into account all the
relevant factors regarding the equipment to be used and the characteristics of the
weld to be examined.
Prior to ultrasonic examination the weld should be visually inspected and any
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
Summary of pages
This document comprises a front cover, an inside front cover, pages i and ii,
pages 1 to 14, an inside back cover and a back cover.
This standard has been updated (see copyright date) and may have had
amendments incorporated. This will be indicated in the amendment table on
the inside front cover.
ii BSI 01-1999
BS 3923-2:1972
1)
For the examination of welds used in the manufacture of pipes and tubes on a continuous basis see BS 3889-1A.
2)
Guidance on the determination of probe characteristics is given in Appendix A.
BSI 01-1999 1
BS 3923-2:1972
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
2 BSI 01-1999
BS 3923-2:1972
direction of welding the weld shall be scanned along with the recommendations of BS 2704.
the weld axis.
If a variable angle probe is used, the upper and
NOTE The automatic system may not facilitate the satisfactory lower limits of the range of angle shall be
detection of transverse imperfections. In such cases it will be
necessary for an alternative automatic or manual method to be determined in a similar manner to that for a fixed
agreed between the contracting parties before testing is angle probe.
commenced.
7.5 Sensitivity. The method of setting the
7.2 Test frequency. The test frequency used shall sensitivity shall be agreed between the contracting
be compatible with the requirements for sensitivity parties.
in 7.5.
Methods of achieving the gain settings are described
NOTE For guidance, the following frequencies are suitable for
the beam path lengths quoted:
in Appendices B and C.
4 MHz to 5 MHz for up to 200 mm The sensitivity of the flaw detector/probe
2 MHz to 2.5 MHz for up to 400 mm combination shall be such that a clear signal will be
1 MHz to 1.25 MHz for over 300 mm obtained from the smallest defect to be detected
7.3 Scanning. Wherever possible the scanning of throughout the scanning distance.
the weld shall be carried out from both sides of the The size, type and orientation of the smallest flaw to
weld on the same surface. Where the configuration be detected shall be agreed between the contracting
is such that examination from both sides is not parties before testing is commenced.
possible, this fact shall be included in the report
(see Clause 10). With variable angle probes a marked loss in
sensitivity may occur at the higher probe angles and
The full cross section of the weld shall be scanned at therefore the sensitivity shall be determined using
the agreed sensitivity. This may be achieved, for that angle which gives the weakest echo.
example, by the use of
7.6 Recording. If the recording system is of
(1) a variable angle probe; the GO/NO-GO type, the echo height shall be
or (2) a to and fro movement of a fixed angle measured at which
probe; 1) the recorded signal appears,
or (3) some special configuration of probes. and
The distance between successive scans of the weld 2) the recorded signal disappears,
shall not exceed the breadth of the transmitting
crystal. and there shall not be a difference of more
than 2 dB.
The speed of movement of the probe(s) shall be such
that the optimum response of the recording With a proportional type recording system, the
apparatus is obtained. recording indication shall be compared over its full
range, with the echo height on the flaw detector.
BSI 01-1999 3
BS 3923-2:1972
test.
4 BSI 01-1999
BS 3923-2:1972
derived from the formula: would not restrict the beam. The curve should be
plotted on logarithmic paper and after
velocity
Frequency = -------------------------------- approximately three near field lengths, it should
wavelength
form a straight line. If not, attenuation has to be
For boiler plate the velocity in compression is allowed for (see B.1.1), and when taken into account
usually in the order of 5.93 103 m/s and the for the back wall echo a straight line should result
velocity in shear is in the order of 3.23 103 m/s. in the far field past three near field lengths. If the
The velocity is known for the material and the straight line is then projected up to the zero decibel
wavelength is read from the trace as the distance N
line, the intersection point will be -------- or 1.57 N,
covered by two consecutive half cycles for a half 2
where N is the near field length.
wave rectified display. For flaw detectors which
show a full wave rectified display, it would be the Appendix B Notes for guidance on the
distance covered by four half cycles. From this it
would appear that two wavelengths were being
D.G.S. diagram
measured but in fact normal time-base calibration
B.1 Normal probes. (See Figure 2.)
for flaw detection purposes is always half of the total
distance travelled; the half calibration therefore B.1.1 General. The D.G.S. diagram for single crystal
cancels out the double wavelength. The half cycles normal probes was drawn by plotting amplitude in
should be regularly spaced, uneven spacing decibels from a series of disc shaped reflectors with
indicating a distorted pulse, and probes exhibiting increasing distance from the probe in water. The
this feature should not be used with the D.G.S. loss due to water attenuation was allowed for in
system. each case and therefore the graph shows the
reflection conditions for any material assuming no
attenuation.
The distance is given in near fields and is on a log
base. If the near field length N is not known for the
probe in use it may be calculated from the formula
2
D
N = -----------------------------------------
4 wavelength
where D is the crystal diameter.
BSI 01-1999 5
BS 3923-2:1972
The backwall echo line indicates the maximum The term small reflectors is related to the beam
reflection from a large reflector with increasing width and not to the crystal diameter, since
distance from the probe, which becomes a straight reflectors greater than the probe diameter behave
line after three near field lengths. This confirms the as small reflectors when far enough away to be
radiation law for large reflectors in the far field, i.e. encompassed within the spreading beam width.
the amplitude is inversely proportional to the This is shown in Figure 2 by the lines with S values
distance, so that if the distance is doubled the of 1 and 2 times the crystal diameter.
amplitude is halved: a 6 dB reduction. This law B.1.2 Setting the sensitivity. In order to use Figure 2
provides a simple method of measuring attenuation. to determine the sensitivity setting, the operator
If a difference of more than 6 dB is measured, then first needs to know the smallest disc shaped flaw or
this will be due to attenuation in the material, equivalent reflector size that has to be detected and,
leaving a simple calculation for the attenuation secondly, at what amplitude he wishes to record it.
factor. The following plate testing example illustrates the
Example 1 method:
Plate thickness 30 mm Example 2
Probe frequency 4 MHz Plate thickness 100 mm
Probe diameter 10 mm Probe frequency 4 MHz
Near field length 17 mm Probe diameter 25 mm
The second back echo at 60 mm distance is greater Near field length 104 mm
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
than 3N (51 mm), therefore, the second and fourth Smallest flaw to be detected equivalent to
echoes are on the 6 dB slope. flaw 5 mm diameter (0.2S). Defect echo height 2
Suppose the difference in echo amplitude measured screen divisions.
between second and fourth echo = 10 dB. Total beam The first back echo will be at a distance of
path between second and fourth echo = 120 mm. approximately 1 near field. It can be seen from
Therefore, the attenuation of the plate material Figure 2 that close to the surface an 0.2S flaw will
at 4 MHz is give an echo approximately 26 dB below the back
10 6 1 echo set to full screen height.
---------------- = ------ dB/mm or 33 dB/m
120 30 Near the bottom of the specimen such a flaw will
Alternatively, back echoes in the near field may be only be 17 dB below but at the worst position, 0.5N
used. Any difference between their amplitude in distance (52 mm), it will be 29 dB below.
excess of what is shown in Figure 2 indicates the In order to find the flaw at 2 screen divisions at the
amount of attenuation over the total distance worst point (29 dB below), the back echo has first to
travelled between the two echoes. be set to 2 screen divisions and then increased
NOTE The attenuation factor in the example 1/30 dB/mm by 29 dB. This calibration ensures that all flaws
means that the sound is attenuated by 1/30 dB for every that have to be detected will give a signal of not less
millimetre of sound path travelled, bearing in mind that the than 2 screen divisions height, whether or not there
sound in a reflection technique has to travel there and back.
is attenuation present.
Since the flaw detector calibration is always for the
It is obvious when the minimum record level is
distance in front of the probe and not the total
drawn on the diagram that reflections from certain
distance travelled, it may be more convenient to
depths could give echoes greater than that level and
express the factor in the same way, i.e. 1/15 dB/mm
yet still be acceptable. To determine therefore
or 66 dB/m. It is important to indicate in which form
whether or not the echo is greater or less than the
the factor is being used.
minimum size or in fact to determine its actual
The lines below the back echo line show the equivalent flaw size, the gain should be reduced by
amplitude of reflectors which are smaller than the the decibel difference between the 0.2S line and the
beam width. These values are shown in relation to record level at the defect depth. The 2 screen
the crystal diameter and, for small reflectors within divisions is then the minimum record level at the
the near field, the nodes and anti-nodes of distance.
sensitivity are clearly indicated. They do not arise
with a large reflector since all the incident energy is
totally reflected.
Well into the far field all these lines indicate that
reflections from small reflectors follow the inverse
square law, so that if the distance is doubled then
the amplitude is reduced by a factor of 4 (12 dB).
6 BSI 01-1999
BS 3923-2:1972
This example assumes no attenuation in the B.2 Angle probes. (See Figure 3 and Figure 4.)
material at the frequency in use, although by taking
B.2.1 General. The D.G.S. diagram for angle probes
the back echo as a reference any attenuation
is essentially the same as for normal probes, except
present would only make the sensitivity too high that distance is measured from the probe index and
rather than too low. If the attenuation has been therefore most of the near field is fortunately lost in
measured it may turn out to be insignificant
the probe. The horizontal axis as before gives the
1 - dB/mm (5 dB/m) in view of the thickness
e.g. --------- distance along the beam path; corresponding
200 equivalent surface distances are shown for the
involved, If the attenuation was significant, different angle probes. Note that in the case of the
however, as shown in Example 1 (1/30 dB/mm) then two D.G.S. diagrams shown in Figure 3 and
this should be taken into account as follows: Figure 4, the probe size is stated and therefore the
Draw in the attenuation curve as shown. In this equivalent flat-bottomed hole sizes and the
case, since the back echo has been taken as the horizontal axis have been calibrated directly in
reference reflector, the allowance for attenuation millimetres.
is greater the nearer the reflector is to the In order to use these diagrams to determine the
surface. The difference at the surface in this case sensitivity setting the operator first needs to know
is 6.7 dB i.e. (100 mm 2 1/30). the smallest disc-shaped flaw or equivalent reflector
Assuming now that a reflector in excess of the record size that has to be detected and secondly, at what
level was found at a depth of 0.3N, two corrections amplitude he wishes to record it.
should be made to the gain: 4 dB for the difference To set the sensitivity for weld testing with an angle
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
between the 0.2S line and the record level, and 5 dB probe however, is more complex than using a simple
for the record level to attenuation line, i.e. the total normal probe. Because there is no backwall echo to
correction to be made is a reduction in echo height use as a standard reference, a separate reference
of 9 dB after which the decision as to whether or not echo has to be used, such as the 100 mm radius of
the flaw is rejectable is directly related to the record the A2 calibration block (see BS 2704). This means
level of 2 screen divisions. that an allowance has to be made for the different
If it is necessary to determine the actual equivalent surface conditions (transfer loss) between the test
flaw size for the flaw echo at 0.3N, its amplitude in block and the plate, the different attenuation factors
decibels above the 2 screen divisions should be between the two materials and also the difference in
measured and reduced by the decibel value, record distance between the 100 mm radius reflector and
level to attenuation line. For example: the farthest defect distance in the material.
Flaw depth 62 mm = 0.6N The procedure to arrive at the correct gain setting to
Echo height above 2 screen division = 19 dB ensure that the smallest defect to be detected
throughout the scanning distance will at least give
Record level to attenuation line = 3 dB an echo at the appropriate screen height, taking into
Therefore, final position above record account the factors mentioned above, is as follows:
level = 16 dB (on 0.4S line) 1) Measure attenuation of material (see B.2.2).
From Figure 2, S crystal diameter = equivalent 2) Measure transfer loss between A2 test block
flat-bottomed hole diameter and plate (see B.2.3).
Therefore: 0.4 25 = 10 mm diameter 3) Read difference in decibels between back echo
NOTE 1 In the example used the back echo happened to be from 100 mm radius on A2 block to minimum
at 0.96N but it could have been 0.4N or 5.5N or any other value,
depending on the material thickness and the probe near field
defect size at worst position on D.G.S. diagram
length. Furthermore, the S values and the distance values in which should be available for the probe in use.
near fields could both be given directly in millimetres once the 4) Set back echo from 100 mm radius to 2
probe crystal diameter and near field length are known.
NOTE 2 The D.G.S. diagram shown in Figure 2 is based on the divisions, plus difference in decibels from (3) plus
established sound radiation laws. It is essential, therefore, that transfer loss from (2) plus difference in
the amplifier and probe characteristics in use do not give attenuation from 100 mm radius on A2 block to
conflicting results from those shown on the diagram. In other maximum testing distance in plate.
words, the flaw detector and probe combination should be
entirely compatible with the D.G.S. diagram in use.
BSI 01-1999 7
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
BSI 01-1999 8
BS 3923-2:1972
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
9 BSI 01-1999
BS 3923-2:1972
10
BS 3923-2:1972
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
BSI 01-1999
Figure 4 Typical D.G.S. diagram for angle probe of 4 MHz, 8 mm 9 mm crystal size
BS 3923-2:1972
50 mm
size to be recorded will at least reach 2 screen
divisions, wherever it appears throughout the beam path on screen.
scanning distance. It is possible as with the normal B.2.3 Measurement of transfer loss. The transfer
probe test that echoes in excess of 2 screen divisions loss between the A2 block and the plate under test
could be acceptable. To determine whether or not can be measured as follows:
this is so, and at the same time to evaluate the
1) Use one probe of the same angle and frequency
equivalent flaw size for any such echo, proceed as
as will be used for the test and calibrate the
follows:
screen for beam path.
a) Draw on the diagram the attenuation curve as
2) Use two probes, separate transmitter and
shown.
receiver, and locate as shown in Figure 6.
b) Locate the defect depth on the diagram in
3) Measure difference between through
distance and amplitude above the record level.
transmission signal height of A2 block and plate.
c) Reduce the defect position by the difference in
4) Subtract the difference indicated on the D.G.S.
decibels between the record level and the
diagram, the remainder being transfer loss.
attenuation line.
Example 6
Example 4
Defect echo 50 mm beam path from screen Probes of 4 MHz, 8 mm 9 mm crystal size
Echo height 34 dB greater than record level Through transmission signal
of 2 screen divisions
( 1-- skip on screen), A2 block 56 dB
Record level to attenuation line 8 dB 2
Therefore, new position on Through transmission signal
diagram = 34 8 = 26 dB above record level and
hence equivalent flaw size is 4 mm diameter. ( 1-- skip on screen), plate 68 dB
------
2
12 dB
B.2.2 Measurement of shear wave attenuation. The
shear wave attenuation of the plate can be Difference according to D.G.S.
measured as follows: diagram, 3 dB (50 mm to 80 mm)
1) Use one probe of the same angle and frequency Attenuation difference according to attenuation
as will be used for the test and calibrate the curve for A2 block at 50 mm and attenuation
screen for beam path. curve for plate at 80 mm, 5 dB.
2) Use two probes, separate transmitter and Therefore, transfer loss = 12 3 5 = 4 dB.
receiver, and locate as shown in Figure 5 so that
signal A. B1 appears at 50 mm and signal A. B2
at 100 mm on the screen.
BSI 01-1999 11
BS 3923-2:1972
Appendix C Method for setting 2) The probe is next placed on the A2 calibration
sensitivities where maximum block (see BS 2704) and a record made of the
difference in the attenuator reading required to
sensitivity is required bring the echo from the 100 mm radius for angle
After following the requirements of the standard for probes, or the 100 mm parallel section for normal
a particular probe selection and time-base probes, down to full screen height.
calibration the sensitivity settings should be made 3) The attenuator difference between the 100 mm
as follows: radius for angle probes, or the 100 mm parallel
1) The probe is applied to the test surface and the section for normal probes, of the A2 block and
gain controls are adjusted so that the grass that required to raise an echo from a 1.5 mm
level of the material grain structure is drilled hole at an appropriate beam path
raised 2 mm high above the base-line at the distance, is recorded.
appropriate testing distance. Any reduction or It should be stated whether this drilled hole is
increase in the sensitivity to facilitate the parallel to or at right angles to the test surface.
exploration of defects is to be noted in decibels.
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
12 BSI 01-1999
BS 3923-2:1972
BSI 01-1999 13
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI
14
blank
BS 3923-2:1972
BSI 01-1999
|
|
|
|
|
|
|
|
|
BSI British Standards Institution
|
|
|
|
|
|
| BSI is the independent national body responsible for preparing British Standards. It
|
| presents the UK view on standards in Europe and at the international level. It is
| incorporated by Royal Charter.
|
|
| Revisions
|
|
| British Standards are updated by amendment or revision. Users of British Standards
|
| should make sure that they possess the latest amendments or editions.
|
|
| It is the constant aim of BSI to improve the quality of our products and services. We
|
| would be grateful if anyone finding an inaccuracy or ambiguity while using this
| British Standard would inform the Secretary of the technical committee responsible,
|
| the identity of which can be found on the inside front cover. Tel: 020 8996 9000.
|
| Fax: 020 8996 7400.
|
|
| BSI offers members an individual updating service called PLUS which ensures that
|
| subscribers automatically receive the latest editions of standards.
|
|
| Buying standards
|
| Orders for all BSI, international and foreign standards publications should be
|
| addressed to Customer Services. Tel: 020 8996 9001. Fax: 020 8996 7001.
|
|
| In response to orders for international standards, it is BSI policy to supply the BSI
|
Licensed Copy: Tom Magee, Howden Power, 06 March 2004, Uncontrolled Copy, (c) BSI