creep-fatigue-oxidation interactions: predicting alloy ......2016/04/20 · creep-fatigue behavior...
TRANSCRIPT
Creep-Fatigue-Oxidation Interactions:Predicting Alloy Lifetimes under Fossil
Energy Service Conditions
Sebastien Dryepondt
Amit Shyam
Oak Ridge National Laboratory
2016 NETL Crosscutting Research Review Meeting
April 18-22 2016, Pittsburgh
2
Power Plants Will Need to be Capable of Flexible Operation
- Frequent (~daily) load cycling will result in significant creep-fatigue interaction- Project will focus on:
- Long term creep fatigue testing and lifetime modeling- Study of microstructurally small cracks under creep-fatigue loading- Interactions among creep fatigue and oxidation
week1 week2 week3 week4 week1 week2 week3 week4
* Ralf Mohrmann, Proceedings Liege conference 2014
*
3
Creep-Fatigue Behavior of Gr.91 Steel- Gr.91 creep performance is significantly affected by cyclic loading due to microstructure changes (Fournier et al. 2008)
- No significant effect of cyclic creep (unloading/loading) due to limited cyclic strain ( 3500h, ~50 cycles)
Creep-Fatigue Tests at 550ºC- sub grain coarsening- dislocation density decrease As Received
4
0
2
4
6
8
10
12
14
0 200 400 600 800 1000
Strain
(%)
Time (h)
Decrease of lifetime at 650ºC, 90MPa with 10h cycle
650ºC
Gr 91, ORNL heat 30176
0.0013%/h0.0029%/h
90MPa referenceNo cycling
9MPa
90MPa10h
5
0
2
4
6
8
10
12
14
0 500 1000 1500 2000 2500 3000 3500 4000
Strain
(%)
Time (h)
Increase of creep rate at 600ºC, 150MPa with 10h cycle
15MPa
150MPa10h
10min
Gr 91, ORNL heat 30176
0.00066%/h
0.0012%/h
6
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1000 2000 3000 4000 5000 6000 7000
Strain(%
)
Time(h)
Significant increase of creep rate at 550ºC, 210MPa with 10h cycle
21MPa
210MPa10h
10min
Gr 91, ORNL heat 30176
0.000041%/h0.00024%/h
Ongoing
7
Significant Increase of Creep Rate for Longer Creep Tests
- Systematic increase (~X2) of creep rate due to cycling- Need to focus on long term test / low creep rate
8
0
40
80
120
160
1.9
2
2.1
2.2
2.3
1344 1344.5 1345 1345.5 1346
Stress
(MPa
)
Strain
(%)
Time (h)
Recovery Observed During Unloading at Each Cycle
Recovery
Decrease of dislocation density?Coarsening of sub-grain structure?
Stress
Strain600ºC/150MPa
9
Creep-Fatigue Behavior of Gr.91 Steel- Decrease of cycle to rupture with hold time
- No cavity observed after testing
- Main effect of hold time is related to the formation of multi-layer thick scale vs thin scale on pure fatigue specimens
- Need for longer creep fatigue test duration
Fournier et al. (2008)
Fatigue Creep-Fatigue
10
Decrease of Nf with Hold Time for Creep-fatigue Tests at 625ºC, ±0.25%
Cycle 5No hold
Cycle 510 min
Gr.91, 625ºC10 min hold time
Stre
ss (M
PA)
−400
−300
−200
−100
0
100
200
300
Strain (%)−0.2 0 0.2
Faster softening due to 10min hold time (500h test)
11
Decrease of Nf with Hold Time for Creep-fatigue Tests at 625ºC, ±0.5%?
10min
10h
Gr.91, 625ºC
Cycle 5
No hold
Stre
ss (M
Pa)
−400
−300
−200
−100
0
100
200
300
Strain (%)−0.5 0 0.5
-Faster softening due to hold time but not as drastic as for ±0.25%-Increase of noise for 10h hold test (>2100h)
12
Significant Decrease of Stress During 10h Hold Time
- Close to a steady stress after ~2h- Creep lifetime at 625ºC, 100MPa ~ 5000h
10min
5 CyclesGr.91, 625ºC
10hStre
ss (M
Pa)
0
50
100
150
200
250
300
Time (s)0 100 200 300 400 500 600
5 CyclesGr.91, 625ºC
10h
10min
Stre
ss (M
Pa)
0
50
100
150
200
Time (s)0 10,000 20,000 30,000
2h
13
Slow Evolution of “Steady Stress” for 10h Hold Time Test
10hGr.91, 625ºC
5 Cycles
50 CyclesStre
ss (M
Pa)
0
50
100
150
200
Time (s)0 10,000 20,000 30,000
5 Cycles
10 min hold timeGr.91, 625ºC
50 cyclesStre
ss (M
Pa)
0
50
100
150
200
250
300
Time (s)0 100 200 300 400 500 600
14
Thicker Oxide Scale for Tests Performed at ±0.5%, 10min Hold Time
±0.25%,10min ±0.25%,No hold ±0.5%,10min ±0.5%,No hold
15
Higher Strain Leads to Oxide Scale Cracking and Re-oxidation
±0.25%,10min
±0.5%,10min
2 mm
2 mm
80μm
-40μm
30μm
-30μm
16
New Set up to Study MicrostructurallySmall Crack Growth at High TºC
• Sumit Bahl’s work (Indian Institute of Science)
• Slower propagation for small cracks
• Fournier et Al.(2008)
• - Initiation: Tanaka and Mura model
• - Propagation: Tomkins model
• Crack initiation at room temperature
• High cyclic Fatigue & Creep Fatigue Testing
• In Situ imaging of crack propagation
• Tests conducted at Room and 550ºC
17
Crack Growth Imaging at Room TºC
300μm
18
No Frequency Effect on Crack Propagation
19
No Effect of Hold Time on Crack Propagation
Test initiated with 10min hold time
Results consistent with decrease of dislocation
density at crack tip
20
BCS* Theory Crack Growth Characterization
−= 1)
2sec(r max
pmys
aσ
σπ
( )
−−=
−−= 1))1(
4sec(1)
2)1(
2sec(r maxmax
pcysys
RaRaσ
σπσ
σπ
Monotonic Plastic Zone Size*
Cyclic Plastic Zone Size
rpc
rpm
Monotonic and Cyclic Plasticity are Related by Load RatioSubstitute σmax σmax – σmin = σmax(1-R) and σys 2 σys
( )
−=
ys
ysm E
aσ
πσπ
νσφ
2secln
18 max2
R = -1, rpc = rpm; R > 0, rpc < rpm
Crack tip displacements (for distribution of edge dislocations at crack tip)**
Monotonic Crack Tip Displacement
( ) ( )
−−=
ys
ysc
RE
aσ
πσπ
νσφ
41secln
116 max2
Cyclic Crack Tip Displacement
* Bilby BA, Cottrell AH, Swinden KH. Proc Royal Soc A 1963;272:304.** Weertman J. “Dislocations Based Fracture Mechanics”
21
Small Fatigue Crack Growth Model
c
crN φ
φf
=Δ
ysmσαφ=Δa
ysσφμφ mc dNda =
wherecrφαμ f=
Number of cycles to failure
Crack extension at failure
Crack-extension force on dislocations in monotonic plastic zone
φc – cyclic crack-tip displacementφm – monotonic crack-tip displacementf – slip irreversibility factorφcr – critical opening displacementα – geometric factor
22
Results at room and 550ºC Consistent with Crack Growth Model
Room and 550ºC data
Tensile testing at different deformation ratesMicrostructure characterization to evaluate dislocation density
23
Creep-Fatigue-Oxidation Interaction in Steam?
In Air: From protective scale to non-protective scale due to cracking and re-oxidationIn Steam:Effect of strain on Non-protective scale?
±0.25%,10min
±0.5%,10min
2 mm
2 mm
80μm
-40μm
30μm
-30μm
24
Martensitic Gr. 91 and fully ferritic 9Cr alloys creep tested at 650ºC in air and
steam
Similar composition but different microstructures:-Gr.91-1 = standard commercial heat treatmentNormalization: 1040ºC and Tempering: 730-760ºC - Gr.91-2 = standard commercial heat treatmentNormalization: 1080ºC and Tempering: 760ºC - 9Cr = Non heat treated materialFully ferritic as-fabricated microstructure
Alloys Fe Cr Mo Si Mn V CGr.91-1 Bal. 8.31 0.9 0.13 0.34 0.26 0.08
Gr.91-2 (or 9Cr) Bal. 8.61 0.89 0.11 0.27 0.21 0.08
25
Effect of Load-Bearing Scale for Thin 9Cr Specimens
- Thinner specimen to increase the volume to surface ratio
Tim
e to
rupt
ure
(h)
9Cr
Oxide
F =σoxideSoxide
S +σ9Cr
S9CrS
9Cr
26
Stress Calculation Is Very Consistent With Load Bearing Scale
- Average oxide scale based on parabolic growthto calculate Soxide and S9Cr
− σ9Cr estimated from lifetime in steam
σ9Cr
Specimen (mm) Scale (μm) σ oxide (MPa)2 38 1271.5 62 1321 51 134
Thickness
27
No Effect or Small Increase of Lifetime in Steam vs Air for Gr.91 Alloys
- Again consistent with Load-Bearing Scale compensating for metal loss
Ongoing
28
Healed cracks lead to a Continuous Bearing Inner Oxide scale
Gr91-1, Steam, 350h,100MPa
50μmGr91-1
Inner Layer
Outer Layer
Healed Crack
29
Decrease of Lifetime in Steam due to Thermal Cycling
On going
650ºC
10h
250ºC
2h 3.5h
30
Similar Oxide Scale Microstructure for Specimens Thermally Cycled
- Local cracking of the scale leads to sudden increase of stress and rupture?- Ongoing cyclic air test to verify Gr.91 microstructure is not affect by thermal cycling
50μm
a) b)
Inner Layer
Outer Layer
Gr.91-1, 855h Gr.91-2, 2512h
31
Future Activities
- Conduct microstructure characterization of cyclic creep, creep fatigue and fatigue crack growth specimens
- Sub grain coarsening
- Decrease of dislocation density
- Cavity formation
- Continue long-term tests
- Conduct fatigue and creep-fatigue testing in steam.
- Design is Ready
- Effect of cyclic loading on a thick non protective scale?
32
Acknowledgements
- D. Erdman, C.S. Hawkins, T. Lowe, T. Jordan, J. Turan, J. Arnold and D. McClurg for assistance with the experimental work- B. Pint, P. Tortorelli, for exciting scientific discussions- This research was sponsored by the U.S. Department of Energy, Office of Fossil Energy under the supervision of Vito Cedro III