-
Page 1 of 19
Fitness for Service of Cracked Valve on High Pressure Steam Line
Header
Sofyan PURBA
Inspection Section – Technical Department PT Badak NGL Bontang;
East Kalimantan, Indonesia
Phone : +62-548-55388; Fax : +62-548-552234; e-mail :
[email protected]
Abstract Cracks were found on 24” valve at steam header line.
This header connects high pressure steam header from Module I to
Module II, and the valve was installed in 1989. The cracks occurred
at valve body and bonnet with total of seven cracks were observed.
Due to risk of valve failure; fitness for service needs to be
conducted to ensure valve service ability. Alternating current
potential drop technique was used for measuring crack depth. The
crack dimensions and material properties are used as input for
finite element analysis modelling. Fatigue growth analysis and
fracture mechanic analysis were conducted also to obtain of valve
properties. The assessment shows that the valve is still in safe
region based on API 579 Level 3 assessment but need to monitor the
valve service temperature. Keywords: valve, steam, crack,
alternating current potential drop, finite element, fracture
mechanic 1. Introduction Badak LNG operates eight LNG Process
Trains, Train A-H, to produce LNG with maximum annual capacity of
22.5 MTPA. To support LNG production, a number of 11 water tube
type boilers are operated with steam production of 295
ton/hr/boiler at 62 kg/cm2.g and 450 0C in Utilities-I, and 10
water tube type boilers with steam production of 379 ton/hr/boiler
at 62 kg/cm2.g and 450 0C in Utilities-II. The steam from boiler is
distributed thru the piping header system to process area as driver
for compressor or pump, power generation and heating media. On
January 27, 2011 several cracks were detected on one block valve of
the high pressure steam header line, 31HS310-24”-JF2H, as showed in
figure 1. The steam line is an interconnecting line between
Utilities-I and Utilities-II. The crack found on the valve body had
a length of 170 mm. Six (6) crack indications were also found on
the valve bonnet. These had lengths ranging from 23 to 120 mm and
were characterized as branched cracking. The location of cracked
valve is showed on the Figure 2 below. Since the cracks exist,
there is the risk that the valve may fail therefore detail
assessment is required to ensure the cracked valve fit for the
service. To replace the cracked valve is not easy task it is
required well arrangement of Train E/F and Train C/D shutdown
because the valve located at superheated steam interconnecting line
of Utilities-I and Utilities-II. Besides that, the replacement
valve is not available at that time so the replacement cannot be
directly conducted. By doing the assessment, it is expected that
there is comprehensive analysis of the valve condition and the
remaining life prediction of the valve based on latest condition.
For this assessment, Level 3 Assessment As per API 579 [1] was
performed. Level 1 and 2 assessments because of complicated
geometry and/or loading conditions, expectation of crack growth or
has the potential to be active because of loading conditions, and
high gradients in stress on valve material.
11th European Conference on Non-Destructive Testing (ECNDT
2014), October 6-10, 2014, Prague, Czech Republic
-
Page 2 of 19
Figure 1. Crack observed on the valve
Figure 2. Location of cracked valve on Utilities area (Valve 6A)
2. Fitness for Service Approach The valve assessment was divided in
to several stages as described below. Before discussing the
assessment stage, the specification of the valve is described
first. 2.1 Cracked Valve Specification The 24”cracked valve is gate
type constructed using ANSI B16.34. This valve was installed in
1989 during construction of process train. Valve service is
superheated steam with service
-
Page 3 of 19
pressure of 61- 62 kg/cm2 and service temperature of 450 oC (842
oF). To meet with high temperature service, the valve material is
made from 1¼” Cr- ½” Mo material. Detail specification of the valve
can be seen on table 1.
Table 1. Technical specification of cracked valve
Design code ANSI B 16.34 Size / ANSI Class 24” / 600 Valve type
Gate flexible wedge type disc Year installed 1989 Material
specification ASTM A217 WC6 Service High pressure superheated steam
Design pressure 75 kg/cm2 (1066 Psig) Design temperature 450 o C
(842 o F) Operating pressure 61- 62 kg/cm2 (882 Psig) Operating
temperature 450 o C (842 o F) Insulation N/A For future operation
the surface temperature on the West and East side is 109 o C and
77.9oC
2.2 Visual Inspection and Flaw Sizing Visual inspection of the
cracked valve and obtain information about the crack appearance,
location and orientation. To perform detailed flaw sizing using
preferred method Alternating Current Potential Drop (ACPD) [2]. For
the ACPD flaw sizing on the block valve we are therefore
considering the thin “skin-effect”. For the flaw sizing the current
is injected to the component using two current injectors and the
surface potential over the crack and reference area is measured
using a two point contacting probe. For flaw sizing accuracy
determination it is important to use a reference block with cracks
of accurately know depths of the same material as in the component
where flaw sizing is to be performed. The basic principle for thin
skin flaw sizing is illustrated in Figure 3 below.
Figure 3. Basic principles of thin skin flaw sizing using the
ACPD technique
-
Page 4 of 19
2.2 Valve Material Testing and Characterization To obtain
material tensile and fracture toughness at the operating
temperature, it is not possible to do the test on the cracked
valve. As the solution, a series of tests on similar grade material
from a discarded valve originating from the same foundry as the
assessed valve was conducted. The material testing was to include
hot tensile, fracture toughness/fracture resistance and fatigue
crack testing at the operating temperature of the cracked valve in
view of the temperature gradients measured on the valve surface.
The testing programme was to be designed to facilitate an accurate
assessment of the remaining life of the valve based on applicable
procedures input data requirements of the applied FFS standard
ASME/API 579 [1]. 2.3 Finite Element Modelling The assessment
included detailed finite element analysis (FEA) of the valve using
the Abaqus software to develop a detailed finite element model of
the block valve and the materials data established from the
testing. The analysis was to be performed based on information
about the operating pressure, temperature, surface temperature
gradients, design stress/piping stress analysis and historic
records showing fluctuations over time associated with operation of
the steam system and the block valve, shut down history and changes
to the operation of the valve and steam line in question due to
commissioning of new trains. Further, for the remaining life
prediction information about the future operation of the valve i.e.
would it be kept open or closed and how would the pressure and
temperature be expected to vary based on the operation of the steam
turbines and the steam utility system is required. 2.4 Fitness for
Service Assessment The fitness-for-service and remaining life
assessment methodology applied for the analysis of the cracked
block valve conforms to the API 579 procedures for crack
assessment. As per API 579 Section 9.2.2.2 A Level 3 Assessment
should be performed due to the following condition:
a. Advanced stress analysis techniques are required to define
the state of stress at the location of the flaw because of
complicated geometry and/or loading conditions.
b. The flaw is determined or expected to be in an active
subcritical growth phase or has the potential to be active because
of loading conditions (e.g. cyclic stresses) and/or environmental
conditions, and a remaining life assessment or on-stream monitoring
of the component is required.
c. High gradients in stress (either primary or secondary),
material fracture toughness, or material yield and/or tensile
strength exist in the component at the location of the flaw (e.g.
mismatch between the weld and base metal).
For the block valve considered all above three conditions a), b)
and c) apply therefore detailed finite element modelling of the
valve to determine the operating stress in the valve bonnet and
body at the crack locations, the flaws detected are expected to be
in a sub-critical growth stage and on-line monitoring has been
performed, and is required for extension of the remaining life. The
block valve has been operating since 1989, and the maximum
operating temperature has not exceeded the design temperature of
450 oC i.e. the super critical steam temperature. This implies that
since the maximum valve body temperature is below ¼ of the steel’s
melting point i.e. Tm = 1530 oC, the block valve body and bonnet
should not have been subject to creep and creep fatigue. Further,
the sub-critical crack growth to be considered for the
-
Page 5 of 19
remaining life assessment is crack growth associated with cyclic
variations of the thermal and pressure stress i.e. fatigue crack
growth. The fatigue crack growth parameters of the ASTM A217 WC6
grade cast steel were determined by testing at the anticipated
future operating temperature of the valve, as described in above.
3. Test Result 3.1 Flaw Sizing of Crack in the Valve Body and
Bonnet The results of 5 ACPD measurements performed at each
location have been tabulated in Table 2 and Table 3 below. The
measurement locations are seen marked using a blue paint marker on
the crack body in Figure 4 and 5. The deepest flaw was found to
have a height of 31.8 mm at location A4 of the body and 13.3 mm at
location C6 of the bonnet. For assessing cracks in bonnet, the
network is idealized as a single planar predominant flaw in
accordance of API 579. Based maximum height of the branched network
from the ACPD measurements, the idealized flaw height determined as
1.2 x maximum crack height from measurement, i.e.: 16 mm and 130 mm
long. 3.2 Thickness Measurement Thickness measurements at the areas
adjacent to the valve body crack and bonnet cracks were carried out
at elevated temperature. The same reference block from the same
manufacturer, same year of manufacture and which had been cast from
the same material ASTM A217 grade WC6 as used for the crack height
measurement trials was used as reference for the thickness
measurements. The thickness of the reference block was measured at
room temperature and also at elevated temperature 77.9°C
(temperature measured at the time the thickness of the valve body
was taken), and 63°C (temperature measured at the time the
thickness of the bonnet) was taken. In Table 4 and 5 conservative
estimates of the remaining ligament thickness are presented as the
difference between the minimum wall thickness recorded for each
location, and the greatest crack height recorded by ACPD. The
numbering of the locations follows the same numbering convention as
used earlier for the crack height measurements.
Figure 4. Marking in the body
Figure 5. Marking in bonnet. Marking red location B, blue
location C, green location D and yellow location E
-
Page 6 of 19
Table 2. Crack height and length measurement result on the
body
Table 3. Crack height and length measurement result on the
bonnet
Table 4. Thickness measurement result on the valve body
3.3 Material Characterization Result 3.3.1 Chemical Analysis
Chemical analysis of the block valve material received for testing.
The chemical analysis was performed using the Optical Emission
Spectrography (OES). The detailed results of the analysis are
tabulated in table 6. The cast valve body material is seen to meet
the chemical requirements to ASTM Grade apart from silicon, which
exceeds the maximum of 0.60% weight. 3.3.2 Tensile Testing Hot
tensile testing of the block valve material was performed as per
ASTM E21 – 2009 using a
Location Height Maximum (mm) A1 16.2 A2 23.6 A3 28.2 A4 31.8 A5
30.4 A6 22.2 A7 6.2
Crack length : 165 mm
Location Height
Maximum (mm)
Location Height
Maximum (mm)
Location Height
Maximum (mm)
Location Height
Maximum (mm)
B1 B2 B3 B4
2. 4 3.7 7.8 8.7
C1 C2 C3 C4 C5 C6 C7 C8
4.1 6.7 9.0 7.2
11.7 13.3 7.5 4.1
D1 D2 D3 D4 D5
6.2 11.7 6.9 5.2 5.3
E1 E2
7.3 10.5
Crack length : 33 mm Crack length : 120 mm Crack length : 70 mm
Crack length : 23 mm
Location Average wall thickness (mm)
Max. Crack Height (mm)
Thickness of remaining ligament
A0 53 0 53 A1 55 16.2 38.8 A2 56 23.6 32.4 A3 59 28.2 30.8 A4 60
31.8 28.2 A5 63 30.4 32.6 A6 66 22.2 43.8 A7 69 6.2 62.8
-
Page 7 of 19
12.60 mm round bar test specimen with 50 mm gauge length. The
test temperature was chosen as 130 oC, i.e.: approximately twice
the temperature on the surface of the valve body and bonnet
measured as 77 oC and 65oC, respectively according to the latest
measurements. This test temperature conservatively, accounts for
the temperature gradient in the valve material. For reference and
for comparison with the ASTM A 217 WC6 standard requirements,
testing was also performed at room temperature. It is seen from
Table 7 that the cast valve meets the room temperature yield and
tensile strength requirements to ASTM A217 grade WC6.
Table 5. Thickness measurement result on the valve bonnet
* No back wall echo
Table 6. Chemical composition analysis result on valve body
Element Element % weight
ASTM A 217 - WC6 % weight
Carbon, C 0.20 0.05 to 0.20 Silicon, Si 0.70 0.60 max Manganese,
Mn 0.87 0.50 to 0.80 Phosphorous, P 0.033 0.04 max Sulphur, S 0.027
0.045 max Chromium, C 1.43 1.00 to 1.50 Molybdenum, Mo 0.94 0.45 to
0.65 Nickel, Ni 0.20 0.50 max* Iron, Fe balance Others 0.37
*Allowed as a residual element
Location Average wall thickness (mm)
Max. Crack Height (mm)
Thickness of remaining ligament
B1 53 2.4 50.6 B2 52.5 3.7 48.3 B3 52.6 7.8 44.7 B4 52 8.7 43.3
C1 35.75 4.1 31.6 C2 NBW* 6.7 - C3 56 9 47 C4 53 7.2 45.8 C5 52.55
11.7 40.8 C6 52 13.3 38.7 C7 51 7.5 43.5 C8 49.35 4.1 45.1 D1 49.7
6.2 43.2 D2 52 11.7 40.3 D3 53 6.9 45.1 D4 51.2 5.2 45.2 D5 52.25
5.3 46.7 E1 51.25 7.3 43.2 E2 50.6 10.5 39.9
-
Page 8 of 19
Table 7. Tensile testing result
YS (MPa) TS (MPa) El (%) RA (%) Room temp 23 oC 287 587 14 13
Elevated temp 130 oC 225 487 16 23 Room Temp Requirements to ASTM A
217 Grade WC6
Minimum 275 485–655
3.3.3 Fracture Toughness (CTOD) Testing Three crack tip opening
displacement (CTOD) test specimens were extracted from the
discarded valve body of cast steel ASTM A 217 WC6. The specimens
were machined to square section B x B = 40 mm x 40 mm specimens as
per BS 7448 Part 1 [4], and notched by mechanical notching and
pre-cracked by fatiguing in three point bending. The resulting
total crack height to specimen width was a/W = 0.33. Testing was
performed at the test temperature of 130oC i.e. the same test
temperature as for the hot tensile test. The test results obtained
have been summarized in Table 8. It is seen that the three CTOD
fracture toughness values are fairly similar, with a minimum value
of the three tests of 0.38 mm. Since two of the specimens i.e. Nos
PS24#1 and PS24#3 exhibited instability close to the maximum load
plateau, the minimum critical value of 0.38 mm was applied for the
fracture assessment.
Table 8. CTOD test results obtained by testing
Test specimen
No.
Notch & pre- crack height,
Ao (mm)
Plastic component of clip gauge opening,
Vp (mm)
Load P (kN)
CTOD (mm)
δ Comment
15600 PS 24 #1 14.88 0.82 75.72 0.38 Critical event, δc
15600 PS 24 #2 15.02 0.96 72.11 0.43
Max. load plateau, δm
15600 PS 24 #3 14.02 0.84 77.42 0.40 Critical event, δ
3.3.3 Fatigue Growth Testing Fatigue crack growth testing was
performed on a single edge notched tension specimen (SENT) with B =
9.6 mm and W = 29.6 mm. The initial notch and fatigue pre-crack
height was a =5.4 mm. The testing was performed at the same
temperature as the CTOD and hot tensile testing i.e. at 130 oC. The
fatigue crack growth testing was performed at 5 Hz varying the
stress range between 20 MPa, 40 MPa, 50 MPa, 75 MPa, 100 MPa and
150 MPa until the crack exhibited stable propagation. The crack
extension was monitored using ACPD and visual measurement of the
crack height on the side of the fatigue crack growth specimen. The
results of the fatigue crack growth testing of the block valve
material in terms of crack height versus number of stress cycles,
are showed in Figure 6. Also showed in this figure are the crack
growth parameters A = 3.1E-17 and m = 4.4 at 130 oC obtained from
the curve fitting. The fatigue threshold was found from the tests
to be approximately ∆Ko = 184 Nmm-3/2.
-
Page 9 of 19
Figure 6. Results of fatigue crack growth testing of the 24”
block valve material 3.4 Finite Element Modelling Result The
objective of the analysis was to study the stress distribution on
the 24” block valve in the closed condition, at the two crack
locations (i.e. West Side valve body and bonnet crack) and to
provide the stress values as input data to the fitness-for-service
and remaining life estimation. The 3D finite element model of the
block valve was developed based on the 2D engineering design
drawing for the valve and manufacturer specification brochure was
also reviewed for information purpose as well as thickness
measurement. In the current model, the dimensions of the bonnet
“cap” have been assumed since the curvature is anticipated to
affect the stress distribution at the bonnet “cap”. The model is
showed on the figure 7. The material properties input for the
finite element analysis was taken from the tensile test results
obtained from testing performed at high temperature (i.e. 130˚C).
In order to input the elastic-plastic material properties to
Abaqus, it is important to generate best fit smooth Ramberg-Osgood
stress strain curve since stress strain data obtained from the
testing will normally have a number of irregularities which may not
be acceptable for the FEA. The engineering stress-strain curve
generated does not give a true indication of the deformation
characteristics of a metal because it is based entirely on the
original dimensions of the specimen, and these dimensions change
continuously during the test. Therefore the true stress-strain
curve should be adopted in the finite element analysis. Fixed
boundary condition (BC) is assigned to West Side of the block valve
model while simply supported boundary condition is assigned to the
East Side. In simply supported BC, the model is free to move
axially. The BCs considered in the current analysis were based on
the piping isometric drawing as showed in Figure 8 and it is
anticipated that the 24” block valve is not rigidly fixed in all
directions at both end of the pipe. To simulate the operating
condition of the 24” block valve in closed condition, internal
pressure is applied to the West Side (62.25 bar = 6.225 MPa),
bonnet (61 bar = 6.1 MPa) and East Side (61 bar = 6.1 MPa) of the
24” block valve model.
-
Page 10 of 19
Two (2) different operating temperature distributions were
analysed, such as: 1. Low temperature case: measured at the block
valve surface.
a. West Side: 114.4˚C b. Bonnet: 73˚C c. East Side: 70.6˚C
It is noted that there is no information available regarding the
temperature distribution inside the block valve. Hence, in the
current analysis assumption has been made to increase the
temperature inside the block valve by 100˚C.
2. High temperature case: temperature distributions applied in
at the outer surface of the
model were based on updated information / surface temperature
measurement i.e. 415˚C for the West side, 412˚C for the East side,
402˚C for the bonnet cap and 375˚C for the valve body). For this
case, the operating temperature inside the block valve is set to
440˚C due to the flow of superheated steam inside the block
valve.
Figure 7. Isometric view (left) and front view (right) of finite
element model
Figure 8. Piping isometric drawing
-
Page 11 of 19
In Abaqus, to perform a thermal analysis an initial temperature
condition needs to be defined. In the current analysis, ambient
temperature of 25˚C was assumed as the initial condition at the
outer and inner surface of the 24” block valve. Based on the
summary of 3D finite element results listed in Table 8, it can be
seen that the increase in temperature between the low and high
temperature will further increase the stress istributions at the
valve body and the bonnet (i.e. ~72 MPa stress increase at valve
body and ~74 MPa stress increase at the bonnet). For the low
temperature case at the valve body crack location it is seen that
the stress does not decrease significantly if the temperature is
reduced by 10˚C (i.e. reduction of stress of 1 MPa). If the
temperature is increased by 20˚C, then the stress increases by 4.5
MPa. Hence, the stress range between low temperature -10˚C to low
temperature + 20˚C for the valve body is 4.5 MPa. Similarly we find
that for the valve bonnet the stress range between Low temperatures
-10˚C to Low temperature + 20˚C is 9 MPa. The cyclic stress and
number of cycles assuming temperature fluctuation once a day are
showed in Table 10.
Table 9. Summary of FE results
Case Location Von-Mises stress (MPa)
Average (MPa)
Low temperature Valve body 78 to 129 103.5 Bonnet 57 to 122
89.5
High temperature Valve body 157 to 194 175.5 Bonnet 132 to 195
163.5
Low temperature minus 10˚C
Valve body 77 to 128 102.5 Bonnet 54 to 116 85
Low temperature plus 20˚C
Valve body 84 to 130 107 Bonnet 61 to 127 94
Table 10. Cyclic stress and number of cycles assumed per
year
Case Location Stress range, ∆σ(MPa) Cycles per year Low
temperature – 10oC + 20oC
Valve body 4.5 365 Bonnet 9 365
Low temperature - High temperature
Valve body 72 365 Bonnet 74 365
3.5 Fitness for Service Assessment Result The failure criteria
at Level 3B is determined as per API 579 [1], from the failure
assessment diagram (FAD). For the block valve assessment the Level
3B i.e. the material specific FAD is established from the
engineering stress-strain curve obtained from the hot tensile test.
The material fracture toughness is established by fracture
mechanics testing of similar ASTM A 217 WC6 grade material from the
discarded valve of same rating, same manufacturer and year of
manufacture, and which has been operating in the same service in
the steam system. By implementing procedure in API 579 [1], the
failure assessment diagram (FAD) for 24” valve is showed on the
Figure 9.
-
Page 12 of 19
Figure 9. Material specific Failure Assessment Diagram for valve
body/bonnet at 130 oC
For the valve body, the assessment was conducted using API 579
Section 9 [1] procedures at Level 3B. It is seen from Figure 10
shows that the flaw is safe based on the maximum von-Mises stress
for the low temperature case +20 oC, derived from the finite
element analysis. Critical flaw sizes (heights versus lengths) were
determined for cracks in the valve body subjected to tensile stress
due the temperature gradient in the valve assuming different wall
thicknesses reported in the vicinity of the valve body crack. The
results presented in Figure 11 show that the 31.8 mm height and 170
mm long flaw is safe contained within the boundaries of the
FAD.
Figure 10. Location of failure assessment point for flaw in
valve body
-
Page 13 of 19
Figure 11. Location of failure assessment point for flaw in
valve body compared to the calculated
critical flaw heights versus flaw lengths
Figure 12. Valve body flaw profile and wall thickness profile
established from similar valve cross section and minimum wall
thickness derived from UT and measurement
In Figure 12 the valve body flaw has been plotted against the
minimum wall thickness measured in the area around the flaw. The
minimum wall thickness is estimated by UT thickness measurements in
taken in vicinity of the valve body crack, where back wall echo
could be detected. It is apparent that the thickness values derived
by the UT measurements are incorrect due to the problems of
obtaining the back wall echo at the exact crack location. Is
believed that the minimum thickness measurement is taken at the
location of the valve where there is a recess for the valve seals.
The ‘t2’ values showed in Figure 10.4 are based on the updated
information, which represents the wall thickness estimated by
sectioning a similar (same manufacturer and grade) discarded valve,
at the same position where the valve body crack was located and
also by correlating this with the thickness derived from the
manufacturer’s drawing. The measurement have some inherent
uncertainty, but are considered far more reliable than the
thicknesses derived from UT, which were associated
-
Page 14 of 19
with loss of back wall echo in the vicinity of the crack due to
the thickness transition and measurement at locations that are
associated with thinner wall, presumably due to the presence of the
recess for the valve seals. The valve is intended to operate the
valve in the closed condition, under the defined low temperature
case, with minimal temperature excursions i.e. Low temperature
-10oC and Low temperature + 20oC. The remaining life of the valve
can be estimated assuming the valve remains in the closed position,
and the operation of the steam system causes maximum temperature
swings between 10 oC below and 20 oC above the low temperature
case, with a frequency of once a day. Under these circumstances the
remaining life can be estimated by calculating the time for the
flaw to propagate to become critical or grow through the wall
thickness, eventually causing a leak before break (rupture) at
location A4. Calculations were performed using the stress range 4.5
MPa, one cycle per day and using the crack growth parameters showed
in Table 9. The results of this analysis showed that under these
conditions no crack extension would be expected, as showed in
Figure 13.
Figure 13. No crack growth is estimated for the valve body crack
for thermal stress cycles
Similar analysis was run for illustration, assuming the
temperature swings between the low and high temperature case once a
day. The analysis showed that if the valve experiences temperature
fluctuations between the low temperature case and the high
temperature case (inside 440 oC superheated steam) the body crack
would grow as showed in Figure 14, and the flaw would eventually
become unsafe. The expected failure mode is leak before
rupture.
For the crack in the bonnet, the same procedure is applied. It
is seen from Figure 15 that the 16 mm height and 130 mm equivalent
flaw is safe for the low temperature case i.e.: for a membrane
stress of 127 MPa as per the finite element analysis. The critical
flaw size was estimated for the low temperature case as showed in
Figure 16. Fatigue crack growth analysis was performed to check if
the flaw in the bonnet would grow to become critical. Based on the
results of the finite element analysis it was found that the
primary membrane stress range associated with fluctuating
temperature in the bonnet 10 oC
-
Page 15 of 19
below to 20 oC above the low temperature case was 9 MPa. It is
seen from Figure 17 that the flaw would not be expected to grow
under these conditions, if the frequency of the cyclic stress
change associated with the temperature fluctuation is once a
day.
Figure 14. Crack growth estimated for the valve body crack for
thermal stress cycles associated with temperature fluctuations
between the low and high temperature once a day
Figure 15. Location of failure assessment point for flaw in
valve bonnet
-
Page 16 of 19
Figure 16. Critical flaw sizes estimated for the valve bonnet
for the low temperature operation case
Figure 17. Results of fatigue crack growth calculations for the
re-categorized flaw for the low temperature case
The fatigue calculation was also performed for illustration
assuming a stress range of 74 MPa associated with a swing between
the low and high temperature case, once a day, as showed in Figure
18. The analysis shows that if the valve experiences temperature
fluctuations between the low temperature case and the high
temperature case (inside 440 oC superheated steam) the branched
bonnet cracks would grow and eventually become unsafe. The failure
mode is uncertain due to the nature of the branched cracking.
Potentially this could cause fracture before leak by interaction of
the branched cracking.
-
Page 17 of 19
Figure 18. Results of fatigue crack growth calculations for the
re-categorized flaw on the bonnet
for temperature fluctuations between the low and high
temperature cases
4. Discussion
ACPD crack height measurements were performed of the 24” block
valve. The results showed that the 170 mm flaw located at the valve
body on the West Side had a semi-elliptical shape with a maximum
height of 31.8 mm. ACPD measurements were also performed on the
valve bonnet and the greatest flaw height for the network of
branched cracks was determined to 13.3 mm. Applying the API579 [1]
procedure for re-categorisation of branched cracks the
characteristic idealized flaw or equivalent flaw size was
determined to be 16 mm in height and 130 mm long. Based on hot
tensile testing, crack tip opening displacement (CTOD) testing of a
discarded valve of the same material ASTM A217 grade WC6, and
finite element modelling of the valve in the closed position, the
subsequent fracture mechanics analysis, showed that both the body
crack and the equivalent bonnet flaw was safe for the low
temperature case considered. Fatigue crack growth analysis showed
that the flaw would not grow under the current low temperature
operating conditions, assuming temperature swings between 10 oC
below the low temperature and 20 oC above the low temperature case
for the closed condition. Crack growth analysis showed that
sub-critical crack growth could take place and the flaws could
become unsafe depending on the operation of the steam system. If
the valve experiences significant temperature changes (between the
low and high temperature case), once a day, a leak (no rupture)
could develop. The valve is intended in the closed condition and
under the defined low temperature case, with minimal temperature
excursions i.e. Low temperature -10 oC and Low temperature + 20 oC.
Under such strictly controlled conditions, the finite element
analysis showed that the valve body would experience a stress range
of 4.5 MPa at the crack location. Fatigue crack growth analysis was
performed imposing a stress range of 4.5 MPa once a day. The
results of the analysis showed that under these conditions no crack
extension would be expected.
-
Page 18 of 19
If the valve is continued operated under controlled conditions
it is recommended that crack monitoring of the body and bonnet
flaws is performed to confirm that no crack extension is taking
place as the current fatigue crack growth analysis indicates. Since
the analysis indicates that the main growth occurs in the thickness
direction hence, there may be difficult to detect any crack growth
on the valve surface. Therefore, the temperature monitoring shall
be conducted to avoid any high swing beyond lower temperature case.
5. Conclusion Based on the above the results of the ACPD flaw
height measurements, materials, CTOD, fatigue crack growth and
finite element modelling, Follows are the conclusion: a. The
results showed that the 170 mm long flaw located at the valve body
on the West Side
had a semi-elliptical shape with a maximum height of 31.8 mm. b.
Failure assessment based on the minimum thickness and the stress
derived from the finite
element analysis showed that the flaw is in the safe region of
the Level 3B failure assessment diagram.
c. Crack growth analysis showed that the crack in the valve body
is not likely to grow if the
valve experiences limited temperature cycling (between 10 oC
below to 20 oC above the low temperature case) but sub-critical
growth of the body crack will occur if the valve experiences
temperature cycling between the low and the high temperature case
(superheated steam at 440 oC inside), and the valve is likely to
develop a leak before break at the valve body crack location.
d. The largest flaw height associated with the network of
branched cracks was 13.3 mm and idealized or equivalent flaw size
was determined to be 16 mm in height and 130 mm long.
e. The fracture mechanics analysis performed based on the API
579 Level 3B procedures
showed that for the current low temperature case, the crack is
in the safe area of the failure assessment diagram.
f. Fatigue crack growth analysis performed based on the stress
range derived from the finite
element analysis assuming temperature fluctuation from 10 oC
below to 20 oC above the low temperature, showed that the
equivalent crack in the bonnet would not grow. But, if the
temperature swings between the low temperature and the high
temperature case (440 oC superheated steam inside) the analysis
showed that the equivalent bonnet flaw would experience
sub-critical crack extension.
g. Under the current low temperature case the valve can be
operated safely, it is
recommended to perform crack monitoring to confirm that there is
no crack extension using ACPD. Close monitoring of temperature
shall be conducted to prevent any temperature swing beyond lower
temperature case.
Acknowledgements We express our gratitude to DNV Singapore for
assistance in conducting this assessment.
-
Page 19 of 19
References 1. American Petroleum Institute, ‘Fitness for
service’, API 579-1/ASME FFS-1, JUNE 5,
2007. 2. Martin C Lugg, ‘An Introduction to ACPD’, TSC Technical
Bulletin TSC/MCL/1146
Rev. 18, February 20, 2002. 3. British Standards Institution,
‘Guide to methods for assessing the acceptability of flaws in
metallic structures’, BS7910: 2005. 4. British Standards
Institution, ‘Fracture mechanic toughness test. Method for
determination of KIc, critical COD and critical J values of
metallic material’, BS7448 Part 1: 1991.
5. DNV, ‘Fitness for service and remaining life assessment of
block valve in high pressure steam system PT Badak NGL Bontang ’,
Technical Report no CTC_R_2011029, October 2011.