Thermal fatigue case studies
Mixture
Hot water in main pipe : 280oC
Cold water : 140oC
Striation pattern show evidence of
thermal fatigue
Case
347 stainless steel moderator circuit branch
piping in a pressurized hot water reactor
Cause of thermal fatigue
-Fluctuation in cool zone with stratified
temperature layers
- Mechanical fatigue and pressurized heavy
Circumferential crack at HAZ show water
high load in axial + residual stress
Cold spot Contract
16” 16” Under tension
47 bar 105 bar
400oC 14” 500oC
120oC
Hot spotExpand
under compression
Local cool cause stress
Schematic of jetting effect and crack location
Cause
Crack at desuperheater
Cause of thermal fatigue
Cyclic wet surface causing
thermal stress variation
brought about by BFW
Inside sprayer
Circumferential crack at welding seam Beachmark on fracture surface
(external) show high load in axial show evidence of fatigue
Distance from
midpoint
D
D
-The mixing of process streams
with large differences in
temperature, typically on the order
of 150oC or greater, can result in
cracking due to thermal fatigue
Stress as a function of initial mix point
- Location of pipe at 1D-3D
distance from is the highest stress
area
MOC case
Distance from crack to spray : 415 mm
(300+115)
ID : 230 mm
Ratio : 1.8 from midpoint
(in high stress zone)
Desuperheater crack
Stress corrosion Mechanical
Abnormal Thermal shock
cracking (SCC) damage
Material defect
Temperature
Beach change
Fatigue mark
Microstructure Hardness immediately
Intergrnular Cyclic pipe stress, Process Microstructure
Brittle fracture
corrosion constrain, vibration induced
Strain induced
Stainless steel Originated martensite
Ordinated
347H from external from internal
Residual stress
Crack size Crack size
Contaminant Internal > External
External > internal
Cl Cyclic stress
Pipe stress
O2 analysis Longitudinal thermal expansion (temperature fluctuation)
FEA shows low Improper atomization at
alkaline stress from nozzle
pressure and
temperature FEA support high stress
Stress
CFD support improper atomization
Vibration
Process pressure
Hoop stress induced
Longitudinal crack
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Pre-Analysis: Key Parameters
Nozzle Design Piping Design Operating Conditions
Hole diameter Pipe diameter Valve position
Holes arrangement Straight upstream Flow rate of BFW
length (Lup)
Spray angle (α)
Straight downstream
Evaporation length length (Ldown)
(Levap)
23
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Pre-Analysis: Key Parameters
Nozzle Design Piping Design Operating Conditions
Hole diameter Pipe diameter Valve position
Holes arrangement Straight upstream Flow rate of BFW
length (Lup)
Spray angle (α)
Straight downstream
Evaporation length length (Ldown)
(Levap)
Pipe Diameter
C B C Boiler Feed Water (BFW) BFW Droplet
Levap
DX D DX Lup Ldown
α
G F G
Superheated Desuperheated
Steam (SS) Steam (DSH)
24
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Pre-Analysis: Key Parameters
Nozzle Design Piping Design Operating Conditions
Hole diameter Pipe diameter Valve position
Holes arrangement Straight upstream Flow rate of BFW
length (Lup)
Spray angle (α)
Straight
Design downstream
Curve
Evaporation length length (Ldown)
(Levap)
25
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Results Interpretation and Analysis: Atomization Effects
1. Mass Fraction of BFW Along the Pipeline
Location of – Shows flow characteristics of BFW inside
Desuperheater Nozzle the pipeline.
2. Mass Fraction of BFW on Pipe Wall
– Indicates whether or not the pipe wall is
impacted by BFW.
3. Maximum Wall Temperature Difference
(between Left and Right Sides)
– Indicates homogeneity of mixing between
SS and BFW.
26
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Results Interpretation and Analysis: Atomization Effects
Operating Condition BFW = 3,814 kg/hr SS = 35,048 kg/hr Rws = 0.11
Lup (m/DDSL) 0.840 m/3.7DDSL
Nozzle Design/# Holes MOC/3 Holes MOC/6 Holes
Mass Fraction of BFW
- Along the pipeline
- Is pipe wall impacted YES NO
by BFW?
- Max. wall temp. diff. 67 K 40 K
Discussion:
• INCREASING NUMBER OF OPEN HOLES can PREVENT THE IMPACT ON PIPE WALL FROM BFW. The BFW velocity
decreases when the number of open holes increases which results in more evaporation time.
• INCREASING NUMBER OF OPEN HOLES results in MORE HOMOGENEOUS MIXING between SS and BFW according
27
to the decrease of maximum wall temperature difference.
Boiler Feed Water
External Crack
Internal Cracks
FEA CFD CFD
Stress Profile Wall Temperature Profile Wall Temperature Profile
Model Validation
Root Cause:
COMPUTATIONAL FLUID DYNAMICS (CFD) SIMULATION
Impact of BFW on the pipe wall leads to non-uniform
wall temperature profile and high stress areas.
By Tiyawut Tiyawongsakul
Conclusion ↓ Impact of BFW CFD Team
↑ Homogeneous Mixing
10 CFD case simulation
↑ Number
of Open
Holes ↑ Impact of BFW
↓ Impact of BFW
↓ Homogeneous Mixing
↑ Homogeneous Mixing
↑ Number
of Nozzles
↑ Rws
Flow
Characteristics
Low
injection ↑ Lup
angle
No Impact of BFW = Impact of BFW
↓ Homogeneous Mixing = Homogeneous Mixing
29
Desuperheater crack
Stress corrosion Mechanical
Abnormal Thermal shock
cracking (SCC) damage
Material defect
Temperature
Beach change
Fatigue mark
Microstructure Hardness immediately
Intergrnular Cyclic pipe stress, Process Microstructure
Brittle fracture
corrosion constrain, vibration induced
Strain induced
Stainless steel Originated martensite
Ordinated
347H from external from internal
Residual stress
Crack size Crack size
Contaminant Internal > External
External > internal
Cl Cyclic stress
Pipe stress
O2 analysis Longitudinal thermal expansion (temperature fluctuation)
FEA shows low Improper atomization at
alkaline stress from nozzle
pressure and
temperature FEA support high stress
Stress
CFD support improper atomization
Vibration
Process pressure
Hoop stress induced
Longitudinal crack
Further investigation of crack
morphology
1st crack fracture surface
Outer wall
Inner wall
Inner wall
Crack origin
Crack is originated from scratch
Another crack
Crack is originated from
scratch at internal surface
of pipe
Condition Endurance limit
Unnotch 315 1
V-notch 175 0.55
*reference from the material of nitriding
Inner wall
Desuperheater crack
Stress corrosion Mechanical
Abnormal Thermal shock
cracking (SCC) damage
Material defect
Stress Temperature
Beach change
concentration
(scratch) Fatigue mark
Microstructure Hardness immediately
Intergrnular Cyclic pipe stress, Process Microstructure
Brittle fracture
corrosion constrain, vibration induced
Strain induced
Stainless steel Originated martensite
Ordinated
347H from external from internal
Residual stress
Crack size Crack size
Contaminant Internal > External
External > internal
Cl Cyclic stress
Pipe stress
O2 analysis Longitudinal thermal expansion (temperature fluctuation)
FEA shows low Improper atomization at
alkaline stress from nozzle
pressure and
temperature FEA support high stress
Stress
CFD support improper atomization
Vibration
Process pressure
Hoop stress induced
Longitudinal crack
Lifetime extension by material Nov 10th,2015
Average life is about 6 years (through crack)
exiting
Option 0 1 2 3 4
Based material 347H 347H 347H 625 625
Polished x0.55 X 1.3 X 1.3
(scratch)
Crack initiation 1 1 1 X2 X2
(Thermal shock
resistance)
Crack 1 1 1 1.27 1.27
propagation
(after crack
initiation)
Life 0.55 1 1.3 2.52 3.3
Cost 200K 200K 230K 1000K 1030K
Cost / life 400K 200K 177K 397K 312K
Improve about 2.4 times
Improve about 6 times
Polish
Mirror polish is recommended for inside surface of pipe
to increase fatigue resistance
Crack initiation
(thermal shock parameter)
0.00007
0.00006
x 1.96
0.00005
0.00004
Fracture 347H
toughness 0.00003 625
base x 2.15
0.00002
0.00001
0
0 100 200 300 400
Tensile stress
base
347H 3.26233E-05 141.18486 Crack initiation is improved about 2 times
625 6.38181E-05 303.31927
Ratio 1.956211907 2.1483839
Crack propagation By Jatuporn
(after crack initiation)
Material SUS 347 Inconel 625 Margin (%)
Temperature limit (oC) 620 675 8.15
Yield Strength (MPa) 205 490 58.16
Thermal expansion
18.9 x 10-6 13.3 x 10-6 -30
cm / cm.oC
Endurance Limit (MPa) 197 314 37.26
Fatigue Life (106 Cycle) at
operating condition
2.2 3 26.67
Toughness KIC (Mpa*m^1/2) 119A 174 31.61
Minimum Extra life 1 1.27 27% Operation Condition
Pressure 12.5 MPa
Polish Surface finishing 30% Thickness 21.4 mm
OD 273 mm
Minimum Extra life 1.30 1.65
A : stainless 304 property Material Condition
C = 1.65E-8, m = 3
∆ = 2 MPa
Paris law
= ∆ Assumption
Crack growth rate of this
component is as Paris
+ =
1 model.
Initiation crack length is, ai
0.1 mm and the inspected
∆ crack length is, af 3.0 mm
Fatigue limit is defined at
38% of UTS
Life time improvement for desuperheater
Action
1) Change valve open range reduce the thermal fatigue and
increase life time substantly. Replace nozzle every 2 years
(100K/years)
2) Polishing 347H main line internal surface
(lifetime improvement about 2.36 times)
Next step for reliability improvement
- Study nozzle design from high spray angle to low spray angle
can reduce the thermal fatigue
(confirm operation effect)
-Improve material to 625 with polishing surface. This enhance at
least 6 times of life
- Maintain BFW/steam ratio, reduce continuous BFW feed
