Beruflich Dokumente
Kultur Dokumente
\
|
> (8.1)
This can be re-arranged to give:
3 / 2
mm
b
7.5d b SDD |
.
|
\
|
+ > (8.2)
Formula 8.2 can be used for both the standard and higher qualities of wall loss inspection.
27
HOIS(09)RP1 Issue 1
For larger diameter pipes, formula (8.2) allows the source to be placed adjacent to the pipe
wall opposite the detector (as is conventional practice for DWSI).
This may not be true for pipes with diameters of 4" or less, or those covered by insulation, so
that the detector could not be placed in close contact with the pipe wall. The SDD should
then be calculated according to formula (8.2) above, and the source positioned accordingly.
For pipes with diameters of less than about 3" to 4", the DWDI technique (see below) may be
preferable in some cases as greater axial coverage can be achieved in a single exposure.
8.2 Double wall double image (DWDI) inspection
For the DWDI technique, the SDD is increased compared with DWSI inspection, allowing
inspection of both pipe walls, as illustrated in Figure 8.2, but with a significant increase in
exposure time. The detector is then usually flat, and not wrapped around the pipe wall. In this
case, the relevant object plane is the external surface of the pipe closest to the source.
SDD
b
Source, size
d
Detector
f
Figure 8.2 Distances for DWDI
In general, a large source to detector distance (SDD) will minimise the unsharpness in the
radiograph caused by the size of radiation source (known as geometric unsharpness).
However, large source to detector distances can lead to very long exposure times, and
increased shielding difficulties. Thus trade-offs must be made, whilst ensuring acceptable
image quality.
There is no universally accepted method for the choice of the source to detector distance
which will provide a satisfactory radiographic technique [3] and existing standards from
various countries differ widely.
For wall loss inspection, it is recommended that a geometric unsharpness criterion is used to
provide a basis for setting source to detector distances, following the HOIS Recommended
Practice on radiographic inspection of small bore piping [4]. In [4], a specified unsharpness
value in the plane of the object was used as opposed to the unsharpness in the detector
plane. The unsharpness in the plane of the object was considered to be more physically
meaningful than the unsharpness in the detector plane, as it is the unsharpness in the object
plane that principally affects the minimum discernable detail size in the component under
inspection.
28
HOIS(09)RP1 Issue 1
29
In [4], it was shown that the source to detector distance, SDD, needed to achieve a specified
geometric unsharpness in the plane of the object (Ug') is given by:
SDD = (d . b) / Ug' (8.3)
Where d is the effective source size for geometric unsharpness calculations and b is the
distance between the detector and the source side of the external diameter of the pipe (OD).
For small bore connector inspection, [4] recommended a value of Ug' of 0.3mm, and it is
recommended that this value is used for the higher quality of wall loss inspection as defined
in the present document.
Thus, for higher quality wall loss inspection, the following equation should be used for SDD:
SDD
min
= (d . b)/0.3 (8.4)
For the standard quality wall loss inspection, a higher unsharpness value of 0.6 mm is
acceptable, i.e.
SDD
min
= (d . b)/0.6 (8.5)
Where b is the distance between the source side of the pipe and the detector.
The recommended SDDs derived using equations (8.4) and (8.5) are shown in Figure 8.3,
for an assumed source size of 2.3 mm. Tabulated values are given in Table 8.1.
These distances will be increased for larger source sizes, and reduced for smaller source
sizes, as given by equations (8.4) and (8.5).
HOIS(09)RP1 Issue 1
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 20 40 60 80 100 120 140 160 180 200 220
Distance from Detector to Source side of pipe OD, b (mm)
S
D
D
(
m
m
)
Standard Quality
Higher Quality
Approximate pipe nominal bore for detector in contact with pipe
1 2 4 3 5 6 8
Figure 8.3 Recommended Source to Detector distances (SDD) for DWDI, for an
assumed effective source size of 2.3mm.
Table 8.1 Recommended source to detector distances (SDD) for DWDI, for an
assumed effective source size of 2.3mm and for the detector in contact with the pipe
wall.
SDD
mm b
mm
Pipe Nominal Bore
(if detector in contact with
pipe wall)
in
Standard
Quality
Higher
Quality
33.4 1 128 256
60.3 2 231 462
88.9 3 341 682
114.3 4 438 876
141.3 5 542 1083
168.3 6 645 1290
219.1 8 840 1680
It should be noted that the SDD values given in equations (8.4) and (8.5) are constant factors
of the pipe diameter, for detectors in contact with the pipe wall. For a 2.3 mm source size, the
SDD for the standard quality is 3.8b, or 3.8 times the pipe diameter, and for the higher quality
the SDD is 7.7b, or 7.7 times the pipe diameter.
30
HOIS(09)RP1 Issue 1
8.3 Tangential Inspection
8.3.1 Source on pipe centre line
For tangential radiography with the source on the pipe centre line, the dimensions and
distances are shown in Figure 8.4.
SDD
b'
Source, size
d
Film/Detector
SPD
Figure 8.4 Dimensions and distances for tangential radiography (source on pipe
centre line)
For tangential radiography, there are two main factors which affect the accuracy of wall
thickness measurements and hence the recommended SDDs.
- Dimensions measured in the CR image are progressively distorted away from the
source axis due to the finite source to detector distance. This has differing effects on
the accuracy of the various methods used for calibration of dimensions in the CR
images.
- Geometric unsharpness.
Consideration of the accuracy of dimensional measurements is given below. It is likely that
this will lead to a greater value for SDD than those based on geometric unsharpness
considerations (see equations (8.7) and (8.8) below. If this is the case, the larger of the two
values for SDD should always be used, where practicable.
Figure 8.4 shows calculations of the errors on the measured wall thickness as a function of
the ratio (source to pipe centre, SPD)/OD, for four different dimensional calibration
techniques, as described further in Section 14. Note that these calculations apply to
conditions in which the pipe may not be in contact with the detector, due for example to the
presence of pipe insulation, as well as cases in which the pipe OD is in contact with the
detector.
The centre line magnification calculations are based on the approximate equation (14.1). The
ball bearing comparator calculations are based on the equations (14.3) and (14.4). The fan-
31
HOIS(09)RP1 Issue 1
beam geometry calculations assume that the radiation beam diverges from a point source in
a thin, two-dimensional fan beam, as shown in Figure 14.3, and that the actual wall thickness
WT is related to the measured value WT by equation (14.2).
In reality, for all usual forms of radiography, the radiation beam diverges as a three-
dimensional cone from the source. However, if the line joining the source to the centre of the
detector is perpendicular to the detector plane, then the fan-beam geometry applies to any
line drawn through the centre point of the CR image.
For the purposes of the calculations shown in Figure 8.5, the 3-D cone beam complication is
ignored, and it is assumed that the fan-beam geometry calculation method gives no errors
in the derived wall thicknesses (provided the distances are measured accurately). For the
other calibration methods, the wall thickness measurement accuracy decreases with
decreasing SDD.
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9
SPD / OD
E
r
r
o
r
i
n
W
T
m
e
a
s
u
r
e
m
e
n
t
(
%
)
10
Centre line Magnification
Pipe Diameter measurement
Ball bearing comparator
Fan-beam geometry calculations
Figure 8.5 Calculations of effects of different dimensional calibration techniques on
accuracy of wall thickness measurements using the tangential
technique.
From Figure 8.5, it can be seen that the accuracy of all the calibration techniques decreases
rapidly for SPD /OD less than about 3.5 (with the exception of the exact fan beam geometry
calculation method).
It is therefore recommended that for tangential radiography with the source on the
pipe centre line, the minimum source to pipe centre distance, SPD should be at least
3.5 times the pipe OD.
Using the distances given in Figure 8.4, this corresponds to:
SDD = 3.5 OD + b (8.6)
Where b is the distance from the pipe centre to the film/detector.
32
HOIS(09)RP1 Issue 1
33
It should also be verified that the SDD is not smaller than the values given below. These are
based on geometric unsharpness considerations similar to those used for DWDI (see Section
8.2), but adapted for the different plane of interest on the pipe (tangent position instead of the
source side of the pipe):
Standard quality:
SDD
min
= (d . b)/0.6 (8.7)
Higher quality:
SDD
min
= (d . b)/0.3 (8.8)
Where b is the distance from the pipe centre to the film/detector.
8.3.2 Combined tangential/DWDI image radiography
For smaller diameter pipes (< 4" - 6" diameter) , the double-wall double image (DWDI)
technique is often combined with the tangential technique with the source on the pipe centre
line.
The SDDs for this combined technique should be those for DWDI, i.e. as given in equations
(8.4) and (8.5), unless the value given by equation (8.6) is larger, in which case that should
be used where practicable.
8.3.3 Source offset from pipe centre line
For tangential radiography with the source offset from the pipe centre line, the effects of the
different dimensional calibration techniques on measured wall thickness accuracy are much
less significant than with the source on the pipe centre line.
It is therefore appropriate to use recommended SDDs based solely on an unsharpness
criterion. In this case, the plane of the object of interest is the pipe tangent position, as
illustrated in Figure 8.6 below.
HOIS(09)RP1 Issue 1
SDD
b
Source, size
d
Detector
SPD
Figure 8.6 Distances for offset source tangential radiography
The recommended SDD is then given by equations (8.7) and (8.8) for the higher and
standard qualities of inspection.
These recommended SDD values for offset tangential radiography are plotted in Figure 8.7.
0
200
400
600
800
1000
1200
0 20 40 60 80 100 120 140
Distance from detector to centre of pipe, b (mm)
S
D
D
(
m
m
)
Standard Quality
Higher Quality
Approximate pipe nominal bore for detector in contact with pipe (in)
1 2 4 3 5 6
10 8
Figure 8.7 Recommended Source to Detector distances (SDD) for offset tangential
radiography, for an assumed effective source size of 2.3mm.
34
HOIS(09)RP1 Issue 1
9 Circumferential and axial coverage and overlap
9.1 Circumferential coverage
When using the DWDI and DWSI techniques, then full circumferential coverage of a pipe is
achieved by taking a number of different exposures around the pipe circumference.
9.1.1 DWSI
For DWSI, the approach used in EN 1435 : 1997, is for the number of circumferential
exposures to be calculated on the basis of a maximum permissible increase in penetrated
thickness due to inclined penetration at the edges of the diagnostic area. This increase in
penetrated thickness is a function of:
- SDD
- Wall thickness, WT
- Pipe OD
Following EN1435 class A for DWSI, with the standard and higher quality classes of CR wall
loss inspection, it is recommended that for wall loss CR, the maximum permissible increase
in penetrated thickness should be 20%. The minimum number of exposures is then given by
Figure A.4 of EN 1435 :1997 (page 17).
For the source positioned outside the pipe, Figure 9.1 shows the number of exposures
needed, as a function of two dimensionless variables WT/OD and OD/SDD.
Minimum number of exposures DWSI
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
1.05
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3
WT/OD
O
D
/
S
D
D
> 3
> 4
> 5
> 6
> 7
> 8
4
5
7
6
8
9
3
Figure 9.1 Minimum number of DWSI exposures circumferentially around a pipe as
a function of the ratios WT/OD and OD/SDD. This graph shows the same
information as Figure A.4 from EN 1435: 1997, for source positions
outside the pipe.
35
HOIS(09)RP1 Issue 1
If the detector is offset from the pipe wall due for example to the presence of insulation, then
Figure 9.1 is not applicable, and the values given in Figure 9.2 should be used instead. In
Figure 9.2 note that the vertical axis is the pipe OD divided by the distance from the source
to the pipe axis (SPD).
To obtain the circumferential angular difference (in degrees) between exposures, the
following formula should be used for DWSI:
Angular difference = 360 / (Number of exposures)
Minimum number of exposures DWSI
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
2.1
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3
WT/OD
O
D
/
S
P
D
> 3
> 4
> 5
> 6
> 7
> 8
4
5
7
6
8
9
3
Figure 9.2 Minimum number of DWSI exposures circumferentially around a pipe, as
a function of the ratios WT/OD and OD/SPD, where SPD is the distance
from the source to the pipe axis (centre). This figure should be used
instead of Figure 9.1 if the detector is not in contact with the pipe wall
(e.g. due to the presence of insulation).
For DWDI (see Section 9.1.2), recent trials [5] have shown that the maximum permissible
increase in penetrated thickness can be greater than 20%, without significant loss of
sensitivity for pitting type flaws in pipes.
Thus for DWSI, it may also be possible to reduce the number of exposures below the values
shown in Figure 9.2. However further experimental trials would be needed to investigate this.
For DWSI large variations in source to detector distance occur across typical images, which
result in significant changes in background image brightness. This effect should also be
considered if new criteria are to be established for DWSI circumferential coverage.
9.1.2 DWDI
For DWDI, the detectability of flaws at the tangential position represents the most demanding
position for circumferential coverage. Recent trial results [5] showed that the detectability of
36
HOIS(09)RP1 Issue 1
37
pitting flaws at or very close to the tangent position was generally reduced compared with
that obtained when the flaws were on the pipe centre line, especially with pipes having larger
maximum penetrated thickness values, w
max
. However rotation of the source/detector so that
the flaws were 15-20 from the tangent position then gave detectability similar to that
achieved with the flaw on the pipe centre line.
In general, at least two exposures are recommended for DWDI, separated in circumferential
angle by at least 30-40 (for best results the angle would be 90).
However, for those pipes having maximum tangential paths, w
max
, less than about 50% of the
maximum recommended values for the standard tangential image quality (see Table 7.2),
adequate sensitivity may be achieved using a single exposure.
9.1.3 Penetrated Thickness measurement
The number of circumferential exposures derived from Figures 9.1 9.3 can be used for
general DWDI and DWSI techniques.
However, if quantitative analysis of the image grey level information is being used to estimate
penetrated thickness (Section 12), it is important to ensure that the feature being measured
on the CR image is as close as possible to the centre of CR image. This is likely to require
additional circumferential exposures to align the flaw of interest more closely with the source
axis.
9.2 Axial coverage
The maximum axial coverage for a single CR image can be determined in a similar manner
to that for DWSI circumferential coverage (see Section 9.1.1). For axial distances, the
geometry is then simpler, as illustrated in Figure 9.4. For a 20% increase in penetrated
thickness at the edge of the area to be inspected, then it can be shown the total extent, L
d
, of
this area on the detector is
L
d
= 1.32 SDD
The corresponding axial coverage on the source side of the pipe, L
p
, is then:
L
p
= 1.32 f
where f is the distance between the source and the source side of the pipe. Lp should be
used when determining the interval along the pipe of exposures.
Note that for DWDI, typical CR image plates are likely to be significantly less than L
p
, so this
is unlikely to be a constraint found in practice, and the axial interval between exposures will
be determined by the dimension of the CR image plate.
However, for DWSI, especially on smaller diameter pipes, the value for L
p
needs to be
calculated and used when determining the axial interval between exposures.
As for circumferential coverage, if quantitative analysis of the image grey level information is
being used to measure penetrated thickness (Section 13), it is important to ensure that the
feature being measured on the CR image is as close as possible to the centre of CR image.
HOIS(09)RP1 Issue 1
SDD
Source, size
d
Detector
f
L
d
L
p
Figure 9.4 Axial cross section for DWDI inspection (also applicable to DWSI),
showing the maximum permissible axial length, L
d
, of the evaluated area
for a single source position, on the detector and along the pipe, L
p
.
9.3 Overlap of images
The separate CR images shall overlap sufficiently to ensure that no portion of the component
remains un-examined. Unless otherwise specified, the minimum overlap shall be 25 mm
axially either side of the diagnostic area, measured on the source side.
38
HOIS(09)RP1 Issue 1
39
10 CR image quality indicators
10.1 Background
Experience has shown that some form of objective check on image quality is important for in-
service computed radiography. Operators should be aware that very short exposure times
with computed radiography can give CR images which may appear to the eye, subjectively at
least, to have adequate quality, when displayed on a computer monitor. However,
quantitative analysis can show that such images may have low signal to noise ratios, and
hence poor sensitivity for wall loss detection by image grey level variations.
It is strongly recommended that some form of image quality indicator, or objective
check on image quality is used on all CR exposures, as described below.
10.2 Image Quality Measures for Double Wall Techniques (DWDI &
DWSI)
10.2.1 Target Grey level range
For the standard quality class of wall loss inspection CR, a minimum check on CR image
quality is to ensure that a specified (target) grey level range is achieved in the area of interest
of the CR image. The grey level range achieved for a particular radiation exposure also
depends on the IP type and the scanner gain (or sensitivity) used, so it is also necessary to
specify the IP type and all user selectable scanner parameters when setting a target grey
level range (pixel size, gain/sensitivity etc).
In setting a target grey level range for a particular CR scanner, IP type and scanner gain,
measurements of image normalised signal to noise ratio (see Section 10.2.2 below) should
first be made for a range of exposures. The results should be plotted as a function of grey
level, as illustrated in Figure 10.1. This allows the grey level corresponding to a normalised
signal to noise ratio of 50 to be determined, for a specified scanner/IP combination and the
specified scanner user settings including gain/sensitivity and pixel size.
In the example shown in Figure 10.1, for low gain, SNR_N > 50 can be achieved for grey
levels >3700, whereas for medium gain grey levels must be >24000. These target grey level
ranges are specific to the scanner, the IP and all the scanner settings and the CR analysis
software. If changes are made to any of these variables, a repeat calibration is required to
establish the target grey level for which SNR_N = 50.
HOIS(09)RP1 Issue 1
0
10
20
30
40
50
60
70
80
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000
Grey Level
S
N
R
_
N
CR50P
Low gain
CR50P
Medium gain
Figure 10.1 Measurements of normalised signal to noise ratio as a function of local
image grey level, for two different gain/sensitivity settings with Ir 192
and a specific CR scanner/IP and pixel size.
For the higher quality class of wall loss inspection CR, it is required that one of the more
rigorous measures of image quality described below is also applied to ensure the required
image quality has been achieved.
10.2.2 Signal to Noise Ratio (SNR)
For the standard and higher quality classes, it is recommended that CR image quality is
assessed by measuring the normalised signal to noise ratio (SNR_N) in the area of interest,
using appropriate software if available.
Note that it is important when making SNR measurements for the images to have grey levels
directly proportional to radiation intensity (i.e. linear response). For non-linear response CR
scanners it is necessary to select an appropriate look-up table (LUT), to achieve a linear
relation between radiation intensity and CR image grey level. Different CR scanners can
have different characteristics (e.g. logarithmic or square-root, as well as linear), and the
correct LUT must be selected for the scanner used. If the CR images are not linearised
correctly, the values measured for SNR_N can be misleadingly high. For typical CR images,
SNR_N values will typically be in the range ~ 50 200. Values in excess of this may be due
to use of an incorrect LUT, and should be checked carefully.
When making this measurement, it is important to ensure the analysis area does not include
any significant variations in grey level due to changes in penetrated thickness. Thus the SNR
area should not include, for example, component edges, areas of wall loss, or any other
image areas where there are significant changes in penetrated thickness.
The size of the SNR area should be 20 pixels horizontally x at least 55 pixels vertically to
provide a dataset of at least 1100 values, as required by EN 14784-1.
40
HOIS(09)RP1 Issue 1
41
In all cases, SNR_N measurements should be made at several locations (minimum of four)
within the area of interest, and a mean value derived.
Note that for some components it may not be possible to find a homogeneous area of the
image suitable for SNR_N measurements. An example would be a pipe containing
generalised irregular corrosion over the whole area of the image. In such cases, the target
grey level method described in the Section above should be used instead.
The SNR measured on a CR image needs to be normalised using a factor which depends on
the basic spatial resolution (BSR) of the CR system (see EN 14784-1, equations 2 and 3).
The formula to be applied to calculate the normalised signal to noise ratio is:
SNR_N = SNR
meas
(88.6/BSR) (10.1)
Where
SNR
norm
is the normalised signal to noise ratio
SNR
meas
is the signal to noise ratio measured on the CR image
BSR is the basic spatial resolution of the CR system in microns (depends on the
scanner pixel size and the model of CR plate)
88.6 is the length of the side of a square having the same area as a circle with
diameter 100 microns (see EN14784-1, p11 for explanation)
For double wall techniques (DWSI and DWDI), and the standard quality class, the
normalised signal-to-noise ratio (SNR_N) as calculated from equation (10.1) in the pipe
centre should be at least 50.
For double wall techniques (DWSI and DWDI), and the higher quality class, the
normalised signal-to-noise ratio (SNR_N) as calculated from equation (10.1) in the pipe
centre should be at least 80.
10.2.3 BSR Measurement
The basic spatial resolution (BSR) can be measured according to EN 14784-1, which
involves use of orthogonal duplex wire IQIs. An improved BSR measurement technique for
isotope sources, to minimise the effects of noise, is described in [5]. This method involves
the analysis of the responses obtained with a Duplex wire IQI, as shown in Figure 10.2.
HOIS(09)RP1 Issue 1
Figure 10.2 CR image of the Duplex wire IQI, rotated so that the wires are as
accurately vertical as possible.
The modulation of the different wire pairs can then be plotted as a function of wire diameter.
The variation of modulation with wire diameter is linear on a log log plot, which allows
ready derivation of the wire diameter for which the modulation is 0.2 (or 20%), as illustrated
in Figure 10.3.
Derivation of BSR based on measured wire pair modulations
0.1
1
0.1 1
Duplex wire diameter (mm)
M
o
d
u
l
a
t
i
o
n
BSR
Figure 10.3 Measured Duplex wire pair modulation plotted as function of the wire
diameter. The BSR is derived from the fitted line for a modulation of 0.2.
42
HOIS(09)RP1 Issue 1
43
The advantages of this more quantitative method for BSR determination include improved
repeatability, better noise tolerance, less subjectivity, and determination of values to a
precision of better than the spacing between available wire pair diameters.
In some cases, the measured BSR values vary with direction in the CR image. In this case, it
is important to ensure consistency between the directions of the SNR_N and BSR
measurements. For the default SNR measurement direction of horizontal within a CR image
(along the fast scan direction), the wires themselves within the duplex wire IQI should be
aligned orthogonal to this direction (i.e. vertical in the displayed image). This is achieved
when the long axis of the duplex wire IQI assembly is horizontal in the image, i.e. aligned
with the SNR measurement direction.
Measured values for BSR for some current CR systems are given in Table 10.1, taken from
[5]. Note that the values depend on the radiation source as well as the imaging plate, pixel
size and scanner model.
If a value for the BSR for the screen/scanner combination in use is not available, a
conservative estimate of 200 microns can be used, provided the screen resolution is known
to be similar to those screens given in Table 10.1, and the scanner pixel size does not
exceed 100 microns.
HOIS(09)RP1 Issue 1
44
Table 10.1 Basic spatial resolution (BSR) of some CR scanners and imaging plates,
measured using Duplex wire IQI
CR System
Basic spatial resolution,
BSR*
(microns)
Scanner IP Pixel
size
(m)
Radiation
source
According
to
EN14784-1
More
quantitative
method
[see ref 5]
HD CR 35
NDT
White, pre April
2008
100 Iridium 192 160
HD CR 35
NDT
White, pre April
2008
100 Selenium 75 160
HD CR 35
NDT
White, post
April 2008
100 Iridium 192 145 163
HD CR 35
NDT
White, post
April 2008
50 Iridium 192 115 133
HD CR 35
NDT
White, post
April 2008
100 Selenium 192 145
HD CR 35
NDT
White, post
April 2008
50 Selenium 192 100
HD CR 35
NDT
Blue 50 Iridium 192 100
HD CR 43
NDT
White, post
April 2008
100 Iridium 192 145
HD CR 43
NDT
Blue 50 Iridium 192 90
Fuji system
DynamiX
ST-VI 100 Iridium 192 160 176
Fuji system
DynamiX
ST-VI 100 Selenium 75 160
GE CR50P IPC2 100 Iridium 192 200 236
GE CR50P IPS 50 Iridium 192 130
GE CR50XP IPC2 100 Iridium 192 200
GE CR50XP IPS 50 Iridium 192 145 165
GE CR50XP IPS 50 Selenium 75 130
GE CR100 IPC2 100 Iridium 192 200 230
GE CR100 IPC2 100 Selenium 75 200
GE CR100 IPS 100 Iridium 192 160
GE CRXTower IPC2 100 Iridium 192 160
GE CRXTower IPC2 100 Selenium 75 160
GE CRXTower IPS 50 Iridium 192 130
GE CRxFlex IPC2 100 Iridium 192 160 222
GE CRxFlex IPS 100 Iridium 192 145
GE CRxFlex IPS 50 Iridium 192 100 100
* The effective spatial resolution or unsharpness of the CR plate/scanner is
approximately twice the BSR [see equation (4) of EN14784-1].
HOIS(09)RP1 Issue 1
10.2.4 Wire IQIs
Use of physical IQIs (wire or step/hole) is recommended for CR inspection for wall loss to
provide a quantifiable measure of image quality. In general, the minimum IQI wire diameter
visible on a CR image will be limited by the total unsharpness of the image, which is made up
from contributions from the screen unsharpness and geometric unsharpness. The CR image
signal to noise ratio and radiographic contrast will also affect the minimum visible IQI wire
diameter.
For DWDI, the IQIs should be placed on the source side of the pipe, where possible. If the
pipe is insulated, the IQI will need to be placed on the detector side. For DWSI, the IQIs need
to be placed on the detector side (between pipe wall and detector).
For both DWDI and DWSI, the IQIs should be positioned close to the centre of the resulting
CR image. If the IQIs are close to the edges of the images, a smaller number of wires may
be detected than for centrally placed IQIs.
For DWDI inspection using Ir 192, source-side measured and target values for IQI wire
numbers (as defined in EN 462-1) are shown in Figure 10.4. These measurements cover
pipes with wall thicknesses in the range c. 3 mm to 22 mm. Separate values are shown for
the standard and higher image quality classes (SNR_N 50 and 80 on the pipe centre line,
respectively). For product filled pipes, the total equivalent steel penetrated thickness should
be calculated using:
w
t
= w
s
+ ID/ f (10.2)
where w
s
is the steel penetrated thickness, ID is the pipe internal diameter and f is a factor
representing the lower attenuation of the product compared with steel. Estimates for f are ~9
for water and ~11 for oil (no measured value is available for oil).
DWDI - Ir 192
5
6
7
8
9
10
11
12
13
14
15
0 10 20 30 40 50 6
Equivalent steel total penetrated thickness (mm)
I
Q
I
v
a
l
u
e
W
0
DWDI Bergen Oct 08 IPC2/CR100
DWDI IPC2/CR50P NDT Services
DWDI Bergen Feb 08 IPC2/CRX Tower
DWDI Bergen May 09 ST6/Fuji
DWDI Water filled Bergen Feb 08 IPC2/CRX Tower
DWDI Water filled Bergen Oct 08 IPC2/CR100
DWDI Water Filled Bergen May 09 ST6/Fuji & HD CR 35/White
DWDI Ir 192 Target values
DWDI HIGHER Bergen Feb 08 IPS/CRX Tower
DWDI HIGH Bergen Oct 08 IPS/CR50XP
DWDI HIGHER Bergen May 09 HD CR 35 White IP
DWDI Water filled HIGHER Bergen Feb 08 IPS/CRX Tower
DWDI Water filled HIGH Bergen Oct 08 IPS/CR50XP
DWDI Ir 192 High Target values
EN1435 Class A (DWDI)
Figure 10.4 Measurements of smallest IQI wires visible on DWDI Ir192 CR images, as
a function of equivalent steel total penetrated thickness (on pipe centre
line). The values from EN1435 for film radiography weld inspection are
shown for comparison purposes. The IQI wires were on the source side
of the pipe.
45
HOIS(09)RP1 Issue 1
Corresponding measured IQI values and target values for DWSI inspection using Iridium 192
are given in Figure 10.5.
DWSI - Ir 192
5
6
7
8
9
10
11
12
13
14
15
16
0 10 20 30 40 50 6
Equivalent steel total penetrated thickness (mm)
I
Q
I
v
a
l
u
e
W
0
DWSI Bergen Oct 08 IPC2/CR100
DWSI IPC2/CR50P NDT Services
DWSI Bergen Feb 08 IPC2/CRX Tower
DWSI Bergen May 09 ST6/Fuji
DWSI Water filled Bergen Feb 08 IPC2/CRX Tower
DWSI Water Filled Bergen May 09 HD CR 35/White
DWSI Ir 192 Target values
DWSI Water filled Bergen Oct 08 IPC2/CR100
DWSI HIGHER Bergen Feb 08 IPS/CRX Tower
DWSI HIGHER Bergen Oct 08 IPS/CR50XP
DWSI HIGHER Bergen May 09 HD CR 35/White IP
DWSI Water filled HIGHER Bergen Feb 08 IPS/CRX Tower
DWSI Water filled HIGHER Bergen Oct 08 IPS/CR50XP
DWSI HIGHER Water Filled Bergen May 09 HD CR 35/White IP
DWSI Ir 192 High Target values
EN1435 Class A (DWSI)
Figure 10.5 Measurements of smallest IQI wires visible on DWSI Ir192 CR images, as
a function of equivalent steel total penetrated thickness (on pipe centre
line). The values from EN1435 for film radiography weld inspection are
shown for comparison purposes.
With Selenium 75, increased radiographic contrast is generally obtainable compared with
Iridium 192, within the range of applicability of the source. This results in higher IQI values
with Selenium 75 for images having the same SNR_N values. The available measurements
for Selenium 75 DWDI and DWSI inspections are given in Figures 10.6 and 10.7
respectively.
DWDI : Se 75
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25 30 35 40
Equivalent steel total penetrated thickness (mm)
I
Q
I
v
a
l
u
e
W
DWDI SELENIUM Bergen Feb 08 IPC2/CRX Tower
DWDI Bergen Oct 08 IPC2/CR100
DWDI Bergen May 09 HD CR 35/White
DWDI Water Filled Bergen Oct 08 IPC2/CR100
DWDI Water Filled Bergen May 09 HD CR 35/White
DWDI Ir 192 Target values
DWDI HIGHER Bergen Oct 08 IPS/CR50XP
DWDI HIGHER Bergen May 09 HD CR 35/White
DWDI Ir 192 High Target values
EN1435 Class A (DWDI)
Figure 10.6 Measurements of smallest IQI wires visible on DWDI Se 75 CR images,
as a function of equivalent steel total penetrated thickness (on pipe
centre line). The IQI wires were on the source side of the pipe.
46
HOIS(09)RP1 Issue 1
DWSI : Se 75
6
7
8
9
10
11
12
13
14
15
0 10 20 30 40 50 60
Equivalent steel total penetrated thickness (mm)
I
Q
I
v
a
l
u
e
W
DWSI SELENIUM Bergen Feb 08 IPC2/CRX Tower
DWSI Bergen Oct 08 IPC2/CR100
DWSI Water Filled Bergen Oct 08 IPC2/CR100
DWSI Bergen May 09 HD CR 35/White
DWSI Water Filled Bergen May 09 HD CR 35/White
DWSI Se 75 Target values
DWSI SELENIUM HIGHER Bergen Feb 08 IPS/CRX Tower
DWSI HIGHER Bergen Oct 08 IPS/CR50XP
DWSI HIGHER Bergen May 09 HD CR 35/White IP
DWSI Higher Water Filled Bergen Oct 08 IPC2/CR100
DWSI HIGHER Water Filled Bergen May 09 HD CR 35/White IP
DWSI Se75 High Target values
EN1435 Class A (DWSI)
Figure 10.7 Measurements of smallest IQI wires visible on DWSI Se 75 CR images, as
a function of equivalent steel total penetrated thickness (on pipe centre
line).
The tables below summarise the target IQI values derived from the experimental CR images
collected during the HOIS trials, as given in Figures 10.4 to 10.7 above. Note that the
thickness t is the equivalent steel total penetrated thickness on the pipe centre line, w
t
, as
given by equation 10.2.
DWDI Iridium 192 (source side IQIs)
Standard Image quality Higher Image quality
Thickness
range (mm)
IQI value, W Thickness
range (mm)
IQI value, W
5 s w
t
< 15 9 5 s w
t
< 8 12
15 s w
t
< 25 8 8 s w
t
< 12 11
25 s w
t
< 40 7 12 s w
t
< 15 10
40 s w
t
< 60 6 15 s w
t
< 20 9
DWSI Iridium 192 (detector side IQIs)
Standard Image quality Higher Image quality
Thickness
range (mm)
IQI value, W Thickness
range (mm)
IQI value, W
5 s w
t
< 10 11 15 s w
t
< 20 10
10 s w
t
< 15 10 20 s w
t
< 35 9
15 s w
t
< 25 9 35 s w
t
< 60 8
25 s w
t
< 30 8
30 s w
t
< 60 7
47
HOIS(09)RP1 Issue 1
48
DWDI Selenium 75 (source side IQIs)
Standard Image quality Higher Image quality
Thickness
range (mm)
IQI value, W Thickness
range (mm)
IQI value, W
5 s w
t
< 12 11 5 s w
t
< 10 12
12 s w
t
< 18 10 10 s w
t
< 20 11
18 s w
t
< 25 9
25 s w
t
< 35 8
DWSI Selenium 75 (detector side IQIs)
Standard Image quality Higher Image quality
Thickness
range (mm)
IQI value, W Thickness
range (mm)
IQI value, W
5 s w
t
< 15 11 5 s w
t
< 25 11
15 s w
t
< 25 10 25 s w
t
< 40 10
25 s w
t
< 30 9 40 s w
t
< 45 9
60 s w
t
< 60 8
10.2.5 Image quality for tangential techniques
For tangential radiography, conventional wire or step/hole IQIs are not directly applicable,
because they cannot be positioned near to the tangential pipe position, and the rapid
changes in penetrated thickness in this part of a radiographic image makes it impossible to
assess IQI visibilities in any meaningful way.
However, very noisy CR images will give lower wall thickness measurement accuracies than
less noisy CR images. In addition, the unsharpness of the CR image will influence the WT
measurement accuracy.
Hence, some form of quality control for tangential CR radiography is considered necessary,
as follows.
- Ensure that the SDDs recommended in Section 8.3 are used.
- The exposure time should be adjusted so that the un-impeded radiation beam outside
the pipe wall does not exceed 80-90% of the CR imaging systems saturation value
(see Sections 12 and 14.4 for further information on exposure times and burn-off
effects).
- The resulting exposure time depends on the gain setting of the CR system. It is
important that a sufficiently low system gain is used to allow the SNR_N values given
below to be achieved.
- If the available software provides this function, the normalised signal to noise ratio
(SNR_N) should be measured with appropriate image analysis software. The average
SNR_N values obtained in the free beam outside the pipe should be at least 70 for
the standard quality and 110 for the higher quality classes of wall loss inspection.
HOIS(09)RP1 Issue 1
49
- If the pipe centre line is available for measurement of SNR_N, then the DWDI SNR_N
values of 50 and 80 for standard and higher qualities respectively can be used as an
alternative to the above free beam SNR_N values.
Note that in all cases when measuring SNR_N, it is important that the image is in a form
having the image grey levels directly proportional to radiation intensity, otherwise the values
can be misleadingly high.
HOIS(09)RP1 Issue 1
50
11 Screens/filters
The filmless plates used in CR are more sensitive than conventional film to scattered
radiation, which generally has a lower average photon energy than that of the primary
radiation beam. It is therefore important to reduce scattered radiation by means of lead
screens.
The recommended values given in EN 14784-2 are reproduced for convenience in
Table 11.1. These screen thickness values are generally greater than those recommended
for film radiography and are thicker than standard.
Thinner screens with standard thickness can be used provided the specified image quality
values are achieved (see Section 10).
Table 11.1 Recommended metal screens for computed radiography taken from
Table 2 of EN 14784-2. Thinner (standard) screens can be used provided the specified
image quality values are achieved (see Section 10).
Type and minimum thickness in mm of metal
screens
Radiation Source Penetrated WT, w
mm
Front Back
w < 5 Pb 0.1 Pb 0.1 Yb 169, Tm 170
w > 5 Pb 0.1 Pb 0.1
w < 50 Pb 0.3 Pb 0.3 Ir 192, Se 75
w > 50 Pb 0.4 Pb 0.4
w < 100 Fe 0.5/Pb 1.5 Fe 0.5/Pb 1.0 Co 60
w > 100 Fe 0.5/Pb 2.0 Fe 0.5/Pb 1.0
For the corresponding values for X-ray sources, see EN 14784-2, Table 2.
For tangential radiography, the values given in Table 11.1 should be regarded as minimum
values, and thicker screens should be used, where possible, to reduce burn-off effects at the
outer edges of the pipes, and lower the overall image contrast.
For pipes containing product, increased levels of scatter may be obtained, depending on the
pipe wall thickness, and the penetrated thickness of product. Increased levels of scatter are
likely to be associated with greater penetrated thicknesses of product, and reduced pipe wall
thicknesses.
Front and back screens with increased thickness (e.g. up to 1.5 mm thick front screens) may
then be beneficial in some cases. Note however that use of thicker lead screens will then
require increased exposure times to achieve acceptable SNR_N values, as defined in
Section 10.
For high strength sources, significantly thicker back screens can be needed to compensate
for increased scatter from the plant environment in the vicinity of the detector.
HOIS(09)RP1 Issue 1
12 Exposure time
CR plates have a wide tolerance or latitude to variations in exposure times. However,
experience has shown that using very short exposure times will result in poor image quality,
and hence poor sensitivity to wall loss with the DWDI and DWSI techniques. Correspondingly
lower wall thickness measurement accuracies can be anticipated with the tangential method
for short exposure time images.
To avoid these issues, the exposures times for computed radiography should be sufficient to
give CR images with the required quality, as described in Section 10.
In all cases, for isotope sources, it is recommended that the time taken to move the source
out of the container into the exposure position and back again should be less than 10% of
the exposure time, to avoid effects connected with movement unsharpness.
12.1 DWSI and DWDI
To calculate exposure times, the following approximate formula may be used:
( )
S
SDD
2
tot 0
4 -
w
w exp E 2.4x10 E + = (12.1)
Where
E
w
is the exposure time in sec for penetrated thickness w.
E
0
is the exposure needed to achieve the required SNR_N value for zero penetrated
thickness in units of Ci.min @ 500mm.
S is the source strength in Ci.
SDD is the source to detector distance in mm.
is the measured effective material attenuation coefficient (0.04 /mm for Ir 192 and
0.08 /mm for Se 75).
w
tot
is the total steel equivalent penetrated thickness, including any product in the pipe.
Note that if the exposure, E
0
, is expressed in units of gBq. sec @ 1000mm, then the constant
in equation (12.1) is 2.7 x 10
-8
.
For pipes fully filled by product, the steel penetrated thickness (2 WT) on the pipe centre line
should be increased by approximately ID/9 for water and c. ID/11 for oil (density of 0.8
gm/cm
3
) to give the total steel equivalent penetrated thickness, w
tot
.
The exposure times required to obtained the recommended SNR_N values on the pipe
centre line depend on the imaging plate and scanner.
For the standard image quality class (SNR_N 50 on the pipe centre line), exposures, E
0
in
the range ~3 - 15 Ci.Min @ 500mm have been measured in practical trials [5], depending on
the CR scanner and imaging plate being used.
For the higher image quality class (SNR_N 80 on the pipe centre line), exposures, E
0
in the
range ~10 - 30 Ci.Min @ 500mm have been measured in practical trials [5], depending on
the CR scanner and imaging plate being used.
51
HOIS(09)RP1 Issue 1
52
Significantly longer exposures may be needed if high-resolution imaging plates intended for
weld inspection are used for in-service examination (E
0
~ 30 Ci.Min @ 500mm for standard
image quality and E
0
~ 70 Ci.Min @ 500mm for higher image quality).
12.2 Tangential
For tangential radiography, the times given above for DWDI can generally be used as a
guide.
It is also important to ensure that the scanner gain/sensitivity setting is adjusted so that the
unimpeded radiation beam outside the pipe wall is not saturated, which will cause burn-off
and errors in wall thickness measurement.
HOIS(09)RP1 Issue 1
53
13 Penetrated Thickness measurements
13.1 Introduction
For the DWDI and DWSI techniques, computer analysis of the CR image grey levels can, in
certain conditions, be used to estimate wall thickness changes. It is important to note
however that this a relatively new method which has not yet been fully validated.
This method should therefore be used with caution, and validated in advance for the
component under inspection using test components with closely similar wall thickness and
pipe OD. These test components should contain areas of wall loss of known through-wall
extent, to allow validation of the penetrated thickness analysis method, described below.
It should be emphasised that this is a relative method for measurement of wall loss, and
unlike the tangential method (Section 14), the penetrated thickness method does not provide
direct measurements of remaining wall thickness.
13.2 Principle of method
The software used for this purpose generally assumes that the detected radiation intensity is
related to penetrated thickness by:
I(w) = I
0
exp(- w) (13.1)
Where
I(w) is the intensity for penetrated thickness w
I
0
is the unimpeded beam intensity
is the effective linear attenuation coefficient of the material
13.3 Effects of scattered radiation
In practice, it is important to appreciate that a number of effects, in particular the presence of
scattered radiation, complicates the analysis. Thus equation 13.1 is often only approximate,
and generally only valid for relatively small percentage changes in wall thickness.
The presence of significant levels of scatter on CR images also makes nominal attenuation
values unreliable. For example, a theoretical value (with no scatter) for the linear attenuation
coefficient, , for Ir 192 and steel is 0.084 /mm. However, measured values can vary widely
between at least 0.02 /mm to 0.05 /mm, depending on scatter levels.
13.4 Calibration using step wedge
For this method, it is important to calibrate using a step wedge placed on the component
under inspection, as close to the area of interest as possible. For the DWDI method, the step
wedge can be positioned on the source side of the pipe. For DWSI, the step wedge needs to
be positioned between the pipe wall and detector, which leads to certain practical issues (see
Section 13.6).
HOIS(09)RP1 Issue 1
54
For both DWSI and DWDI methods, the penetrated thickness increases away from the pipe
centre-line, so it is important that the step wedge is positioned so that the penetrated
thickness for the step wedge is similar to that for the wall loss being measured.
The effective attenuation coefficient should be measured from at least the first and second
steps on the step wedge. If the attenuation values derived from the different steps are in
good mutual agreement, an average value can be used, and this provides some confidence
in the accuracy of the subsequent measurements of wall loss. However, if the attenuation
coefficients derived from the different steps are not in good mutual agreement, the value
derived from the first step (smallest increase in penetrated thickness) should be used when
estimating wall loss. Such differences can also indicate the presence of significant levels of
scatter on the image, which lead to a loss of accuracy in the wall loss measurements,
especially for larger percentage wall losses.
13.5 Key Points
The key points for this technique are:
- CR image grey levels must be linearised using the correct look-up table for the CR
scanner, if a non linear amplifier is built into the CR scanner.
- A step wedge must be used to measure the effective attenuation coefficient, and
located such that the local penetrated thickness is close to that for the wall loss being
measured.
- The reference and measurement areas should be as large as possible given the size
of features in the image to measure. (Small areas are more affected by noise).
- The reference and measurement areas should be as close together as possible,
given the limitations of the image content.
With this technique, initial results suggest that accurate (1mm) results can only obtained for
relatively small wall losses of c. 30 50%.
If the actual total penetrated thickness is not known accurately, differences from an assumed
value can be used to measure small percentage wall losses with reasonable accuracy.
Use of substantially thicker lead screens may preferentially absorb the lower energy
scattered radiation and hence increase accuracy in some cases.
13.6 Limitations
If this method is used to estimate wall loss, it is important to appreciate its limitations and to
note that it is a relatively new method which has not been fully validated.
Practical issues can include:
1. For the DWDI method, if wall loss occurs on both pipe walls (those near to and far from
the source) the corresponding radiation intensity changes can be superimposed on the
CR image, leading to an overestimation of the wall loss, if it is assumed to be present in a
single pipe wall only.
2. For DWSI it can be difficult to place a step wedge between the pipe wall and detector,
without distorting the detector and increasing the object to detector distance significantly
(which can lead to a need to increase the SDD see Section 8.1).
HOIS(09)RP1 Issue 1
55
3. For DWSI, and some DWDI images, the CR images can contain significant variations in
background grey level, due the rapidly changing SDD and penetrated pipe wall thickness
values across the image. These background variations introduce uncertainties into the
measurement of an appropriate reference grey level for the area of wall loss, especially if
a single measurement area on only one side of the indication is used.
In the presence of significant variations in background grey level, analysis methods
based on extrapolation of grey-scale profiles extracted from the CR image are
recommended to estimate the background image grey level underneath the area of wall
loss.
HOIS(09)RP1 Issue 1
14 Tangential radiography
This Section contains important recommendations regarding tangential radiography.
14.1 Recommended SDD
The recommended source to detector distances (SDD) for tangential radiography have
already been covered in Section 8.3.
14.2 Source location relative to pipe centre line
For small diameter insulated or non insulated pipes, the radiation source for tangential
radiography is generally positioned on the pipe central axis, as illustrated in Figure 14.1. This
allows both walls, on either side of the pipe, to be inspected on a single CR image.
CR plate
Figure 14.1 Tangential radiography with the radiation source on the pipe centre line.
For larger diameter pipes (both insulated and non insulated), the radiation source should be
offset from the pipe centre line, to be instead in-line with the tangential position on the pipe,
as shown in Figure 14.2.
In all cases, it is important to check that the tangential path for the pipe is within the source
penetration limits given in Table 6.2.
56
HOIS(09)RP1 Issue 1
CR plate
Figure 14.2 Tangential radiography with the radiation source offset from the pipe
centre line by half the pipe diameter, so it is in-line with the tangential
position on the pipe.
There are no specific recommendations concerning the maximum pipe diameter which
should be inspected using the radiation source on the pipe axis. This will in practice depend
on the size of the CR plate available, as well as the method used for dimensional calibration
(see Section 14.3 below).
14.3 Dimensional calibration
For tangential radiography, when making dimensional on-screen measurements of wall
thickness, it is important to calibrate the distances involved in the radiography, to allow for
the image enlargement or blow-up. The geometric magnification effect for tangential
radiography is shown in Figure 14.3.
57
HOIS(09)RP1 Issue 1
Detector
PDD
r
WT
u
SPD
x
SDD
Figure 14.3 Geometric magnification for tangential radiography showing the
measured wall thickness WT.
Two methods can be used for dimensional calibration, to derive the actual wall thickness WT
from the measured value WT as follows.
14.3.1 Measurement of distances
This method involves direct physical measurement of the key distances involved in the
radiography.
Note that reliable and accurate physical measurements of distances may be difficult to
achieve in a plant environment, and the measurements may not be available at the time of
the analysis. If this is considered to be the case, the alternative comparator method (Section
14.3.2) is recommended.
For calibration by the distances method, two of the following distances need to be measured
accurately (to within a few percent):
1. Source to detector distance SDD
2. Distance from source to pipe centre line, SPD.
3. Distance from detector to pipe centre line, PDD.
In addition, the following distance must be recorded:
4. Lateral offset (if any) of source from pipe centre line, x.
This method also assumes that the CR system software already contains correct information
on the scanner resolution (pixel size) used when acquiring the CR image.
58
HOIS(09)RP1 Issue 1
For offset tangential radiography (with x ~ r) , the true wall thickness WT at the tangential
pipe position as shown in Figure 14.2 can be calculated from the projected (measured) wall
thickness WT on the CR plate using the approximate equation (14.1):
' WT ) 1 ( WT =
SDD
PDD
(14.1)
For tangential radiography on the pipe centre line (x =0), equation (14.1) is an approximation
which becomes less accurate as the angle, u shown in Figure 14.3 increases. The exact
calculation then requires application of the more involved equation (14.2) [6].
|
|
.
|
\
|
+
|
|
.
|
\
|
=
SDD
WT'
r
r
1
SDD
WT'
r
r
r WT
2 2
2 2
SPD
SPD
SPD
(14.2)
Where r is half the pipe outside diameter (=OD/2).
Provided all the distances involved can be measured sufficiently accurately, and that the CR
image pixel size is known reliably, then software which applies equation (14.2) to derive the
actual wall thickness WT from the measured value WT is capable of high wall thickness
measurement accuracy, even for values of SDD/OD < 3.
Note however, that equation (14.2) is only valid for fan-beam geometries, and will be only
approximate for three-dimensional cone-beam geometries, i.e. for sections through the
pipe, either above or below the source position in the CR image, for which the angle | shown
in Figure 14.4 is non zero (see also Section 8.3.1).
59
HOIS(09)RP1 Issue 1
Detector
R
|
f
Pipe walls
SDD
Figure 14.4 Section along the pipe axis showing the three-dimensional cone beam
effect (for non zero values of the angle |).
60
HOIS(09)RP1 Issue 1
14.3.2 Dimensional comparator
An alternative method for dimensional calibration is the use of ball bearing or similar
dimensional comparator. This is an effectively radiation opaque object (usually spherical)
with a known diameter, which is placed close to the pipe, and in the same plane as the
tangent position on the pipe wall, as illustrated in Figure 14.5.
Detector
Comparator,
diameter c
c'
Figure 14.5 Tangential radiography showing use of comparator for dimensional
calibration. The comparator should be placed as close to the pipe wall as
possible, without overlapping it.
On-screen measurements of the imaged size of the comparator, using the CR system
software, then allow the pipe wall thickness measurement to be calibrated in mm, hence
allowing for the radiographic magnification or blow-up. The actual wall thickness WT is then
given by:
c'
c
WT' WT = (14.3)
Note that if the comparator cannot be placed adjacent to the pipe tangent position, due for
example to the presence of external insulation, it is recommended that the source is offset
from the pipe centre-line to be aligned with the pipe wall as shown in Figure 14.6.
As an alternative to the use of a dimensional comparator, if the outside diameter of the pipe
is known reliably, then the measured pipe diameter on the CR image can be used for
dimensional calibration. This method does not then require an additional calibration object,
and can have advantages in terms of accuracy provided the pipe outside diameter is known
accurately.
61
HOIS(09)RP1 Issue 1
Detector
Comparator
Figure 14.6 Tangential radiography showing use of offset source position with
comparator for dimensional calibration, for insulated pipes, where the
comparator should be placed as close to the outside of the insulation as
possible.
14.4 Use of lead strips to avoid burn-off
Some CR procedures specify the use of lead strips close to, or coincident with edge of the
pipe to avoid image burn-off effects at the pipe OD.
However, the use of these strips is not recommended for computed radiography, as they
would adversely affect the accuracy of the preferred wall-thickness measurement methods,
which involve analysis of the image values (profiles) extracted along lines orthogonal to the
pipe wall, see Section 14.5 below.
Burn-off should instead be minimised by ensuring the exposure time is adjusted so that the
unimpeded beam beyond the pipe wall is not greater than 90% of the dynamic range of the
CR plate (not too close to saturation).
In addition, the use of lead screens, thicker than the minimum values given in Table 10.1,
can be useful in avoiding burn-off, by reduction of scattered radiation.
14.5 Wall thickness measurement
14.5.1 Interactive on-screen measurements
Many CR systems contain software options which allow on-screen interactive dimensional
measurements using a cursor overlaid on the CR images. The user then judges by eye the
locations in the image of the inner and outer edges of the pipe wall.
This method can however be prone to error, especially if the penetrated thickness of the pipe
wall at the tangent position (w
max
) is approaching the maximum possible (see Section 7.2),
given the radiation source in use. This makes the exact positions of the outer and especially
62
HOIS(09)RP1 Issue 1
63
inner walls difficult to determine by eye. Furthermore, the apparent locations can be affected
significantly by the contrast and brightness settings in use on the CR image at the time.
Thus, the WT measured using this technique can change appreciably as the contrast and
brightness settings are varied on the same image.
If the on-screen measurement method is used, it should, if at all possible, first be checked for
accuracy using the current contrast and brightness settings of the displayed image, by
application to a section of the pipe with known wall thickness (e.g. known to be uncorroded
or not eroded).
For pipes having maximum tangential paths which approach the maximum permissible for
the source in use (see Table 6.2), the interactive on-screen measurement method is best
applied to CR images having logarithmic response functions to radiation exposure/dose,
which reduces the overall dynamic range of the image. This improves the visibility of the
position of internal pipe wall (ID).
Logarithmic CR images can be obtained from logarithmic-response CR scanners (e.g. the
CR100) or by application of an appropriate logarithmic look-up table (LUT) to CR images
resulting from scanners with linear response functions.
In addition, use of high-frequency spatial filtering (sharpening) is recommended since this
improves the accuracy of this measurement method, by emphasizing the positions of the
edges of the pipe wall in the CR images, and reducing any dependence on the contrast and
brightness settings on the image.
It is however recommended that for improved accuracy interactive on-screen measurements
are made in combination with grey-level profile analysis methods as described in the
following section.
14.5.2 Grey-level profile analysis methods
Many CR systems have software which allows the user to mark lines on the CR image
orthogonal to the pipe wall axis. The software extracts a grey-level profile along this line,
which is then generally presented on-screen, superimposed on the image. Measurements of
wall thickness can be obtained by either interactive or automated analysis of these grey-level
profiles.
Automated routines
Automated analysis routines can increase the reliability of the measured wall thickness
values, unless the maximum tangential penetrated thickness (w
max
) is approaching the
maximum possible, given the radiation source in use (see Section 7.2). In addition, other
factors such as the presence of external scale, corrosion products or irregular
internal/external corrosion may affect the accuracy of these automated routines.
In these cases, the automated routines are subject to uncertainties, and the operator should
check the consistency of the derived values with the density profile.
Interactive methods
In the absence of automated routines for wall thickness analysis, the operator should use the
available interactive facilities for analysis of the image profiles. Accuracy is likely to be
improved, especially for pipes having larger w
max
values, if the CR images have a logarithmic
response and are high-pass filtered.
HOIS(09)RP1 Issue 1
Figure 14.7 shows an example of interactive measurement of wall thickness, using cursors
on a grey level profile across the pipe wall, after applying a logarithmic look-up table to the
CR image, and high-pass filtering to enhance details. The position of the outer diameter
corresponds to a clear peak in the profile, and the location of the inner diameter is given by
the minimum and pronounced change in gradient of the profile.
This method, combined with a visual assessment of the image, is recommended.
Figure 14.7 Example of interactive wall thickness measurement using cursors
superimposed on a grey level profile taken across the pipe wall.
It should be noted that the accuracy of all measurement methods decrease as the tangential
penetrated thickness, w
max
, approaches the maximum value recommended for the isotope in
use (see Section 7.2), since the location of the inner wall becomes increasingly difficult to
determine with any reliability due to lack of contrast and increased noise.
64
HOIS(09)RP1 Issue 1
65
15 Scanner parameters, image recording and
processing
15.1 Scanner parameters
For the standard image quality, the scanner pixel size should be 100 microns or less.
For the higher image quality, the scanner pixel size should be 50 microns or less.
For a given radiographic exposure, it is important to note that increasing the scanner gain or
sensitivity increases the grey levels, but has negligible effect on the image quality, as
measured by the normalised signal noise ratio (SNR_N). To increase SNR_N, the exposure
must be increased, not the scanner gain.
For scanners with linear responses between radiation dose and grey level, use of low
gain/sensitivity reduces the risk of image saturation. For higher scanner gains, image
saturation may occur, especially in the free beam areas, for relatively short exposures, which
do not give sufficiently high image SNR_N values to meet the quality criteria given in Section
10.
15.2 Image recording and storage
The CR images from the scanner should be stored in a file format which supports a minimum
of 12-bits/pixel.
The images should be stored at full resolution and full dynamic range, as delivered by the
detector system. Only image processing required for detector calibration should be applied to
the images prior to storage of these raw data.
Any data compression techniques used in the storage of these files should be loss-less.
The following information should be recorded as header information, attached to each CR
image, to be used for reporting purposes.
(a) Name of company carrying out the inspection
(b) Test report number
(c) Component under test
(d) Date of test
(e) Material tested
(f) Nominal pipe diameter and wall thickness
(g) Details of any lagging/insulation, including thickness
(h) Radiographic technique and class
(i) Radiation source: type, size of focal spot & equipment used
(j) CR plate used (IP), screens and filters
(k) X-ray tube voltage & current or gamma ray source strength
(l) Exposure time
(m) Source to detector distance (SDD)
(n) Component to detector distance and/or source to component distance
(o) Type and location of IQIs used
(p) Any deviations from this procedure
HOIS(09)RP1 Issue 1
66
(q) Name and certification of responsible person
(r) Average pixel values (grey levels) in the area of interest
(s) Measured SNR, if software permits derivation
15.3 Image processing
All measures of image quality (mean image grey level in area of interest, signal to noise ratio
if available, IQI wire visibility etc) must be made on images which have not been subject to
any image processing routines which use spatial filtering.
Spatial filtering is defined as a process in which the value of the current image pixel is altered
by an amount which depends on other (usually adjacent) image pixel values. Examples of
spatial filtering or spatial image processing include noise smoothing/filtering, edge
sharpening, edge contrast, un-sharp masking etc. Spatial filtering such as edge sharpening
may however be beneficial for the interactive measurement of wall thickness using the
tangential method (see Section 14.5).
Image processing which changes contrast and brightness of the displayed CR image
(sometimes referred to as level control) is however permissible when evaluating image
quality, and strongly recommended for image evaluation at the monitor.
It is recommended that the results of any image processing are saved as a separate file, and
that the raw data derived from the CR scanner should not be altered by any subsequent
processing.
15.4 Monitor viewing conditions
The computed radiographs shall be examined in a darkened room using a monitor, or film
hardcopy, as specified in Section 7.10 of EN 14784-2.
HOIS(09)RP1 Issue 1
67
16 Acknowledgments
Several members of the HOIS JIP provided useful comments on earlier drafts of this
document.
17 References
1. Burch S F (2009). Results from HOIS CR Trials in Bergen October 2008. HOIS(08)R8
Issue 1. January 2009
2. Development of protocols for corrosion and deposit evaluation in large diameter pipes by
radiography IAEA report, Vienna, 2008.
3. Halmshaw, R (1995) Industrial Radiology Theory and Practice Second Edition, Chapman
& Hall, p125
4. Burch, S F and Collett N J, (2005) Recommended Practice for the rapid inspection of
small bore connectors using radiography, HSE Research Report, RR 294.
5. Burch S F (2009). Results from HOIS CR Trials in Bergen May 2009. HOIS Report
HOIS(09)R5 Issue 1. August 2009
6. Willems P, Vaessen B, Hueck W and U Ewert (1999) Applicability of computer
radiography for corrosion and wall thickness measurements. Insight, Vol 41, No 10, 635-
637.
HOIS(09)RP1 Issue 1