Creep-Fatigue-Oxidation Interactions:Predicting Alloy Lifetimes under Fossil

Energy Service Conditions

Sebastien Dryepondt

Amit Shyam

Oak Ridge National Laboratory

2015 NETL Crosscutting Research Review Meeting

April 27-30 2015, 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- Interactions among creep fatigue and oxidation- Study of microstructurally small cracks under creep-fatigue loading

week1 week2 week3 week4 week1 week2 week3 week4

* Ralf Mohrmann, Proceedings Liege conference 2014

*

3

FY15 Technical achievements

- Project initiated in FY15

- Developed two creep machines allowing automated loading/unloading sequences during testing

- Set up creep fatigue machine to conduct long term creep fatigue tests

- Continued work on interaction between oxidation in steam and creep testing

Creep testing in steam with additional thermal cycling

- Project focusing on ferritic/martensitic Gr91 (9Cr-1Mo) alloy

4

Very limited Gr.91 cyclic creep data

• Slight decrease of creep rate and increase of life due to cycling

• Beneficial effect of cycling due to dislocation recovery?

• Only short term data

Kim et al. MHT, 31, 249 (2014)

600ºC10min hold time

5

0

0.5

1

1.5

2

2.5

3

0 500 1000 1500

Strain

(%)

Time (h)

Decrease of creep rate at 600ºC, 150MPa with 10h cycle

creep rate: 0.0012%/h

Estimate from ORNL report without cycling:(several heats)

=0.0026%/htr = 1600h

standard creep testing was

initiated Increase of time to rupture

15MPa

150MPa10h

10min

Gr 91, ORNL heat 30176

6

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)

0.3

0.4

0.5

0.6

0.7

203 203.5 204 204.5 205

Strain

(%)

Time (h)

Recovery observed during unloading at each cycle

Recovery

Effect of cycling on dislocation network and subgrain?Effect of hold time during unloading?

Short primary stage

Stress

Strain

7

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 reference

9MPa

90MPa10h

Different microstructure evolution above and below 600ºC?

8

creep-fatigue 625ºC, ±0.5% results consistent with literature data

- Comparison with EPRI Round-Robin conducted at 625ºC“need longer hold time”- Significant softening of the alloy at the test beginning

Cycle 10

Cycle1Cycle 100

Mid-life

Gr.91, 625ºC10 min hold time

Stre

ss (M

Pa)

−400

−200

0

200

400

Strain (%)−0.5 0 0.5

9

Similar stress strain curves with 10min and 1h hold time up to ~50 cycles

- 10min hold time (tension) test “failed” after ~519 cycles- Buckling of 1h hold specimen. Specimen geometry has been modified to avoid buckling in future tests

Cycle 1 1h

Cycle 1 10min

Cycle 1010 min

Gr.91, 625ºC

Cycle 10 1h

Stre

ss (M

Pa)

−400

−200

0

200

400

Strain (%)−0.5 0 0.5

10

625ºC, ±0.5% 10 min hold time Fast stress relaxation during 1rst ~30s

- Stable relaxation curves after ~100 cycles

Cycle 10Cycle1

Cycle 100Mid-lifecycle 259

Gr.91, 625ºC10 min hold time

Cycle 500

Stre

ss (M

Pa)

0

50

100

150

200

250

300

350

Time (s)0 100 200 300 400 500 600

11

625ºC, ±0.5% Similar relaxation curves for 10min and 1h hold time

Stress plateau reached after ~30 min for 1hold time tests

12

625ºC, ±0.5%, Significant initial creep damage

- Initial high stress leads to high creep damage

N: number of cyclet0 hold time

Damage accumulation model

Df, fatigue damage Dc, creep damage

creep damage over 1 cycle

Cycle 1

Cycle 100

Dc+DF<1

13

625ºC, ±0.5% Similar creep damage for the 10 min and 1h tests

Total creep damage of ~0.3

Cum

ulat

ive

dam

age

14

625ºC, ±0.25% strain results in limited creep damage during first cycles

Lowering strain and increasing hold time to increase creep damage

625ºC, 10min hold time

15

Creep-fatigue modeling needs to integrate microstructure evolution

Damage accumulation modelDf, fatigue damage, Dc, creep damage

- Longer test duration needed to allow for microstructure change- Lower strain to limit initial high creep damage

- Subgrain and network dislocation evolution- Precipitate coarsening- Cavity formation- Depletion of solid solution elements

Microstructural creep damage

* Kitahara et al., acta met 54, 1279 (2006).

*

16

Ferritic/martensitic Gr. 91 and fully ferritic 9Cr alloys creep tested at 650ºC

in air and steam

Similar composition but different microstructures:-Gr.91 = standard commercial heat treatment (normalized and tempered)- 9Cr = Non heat treated material

Wt% Fe Cr Mo Si Mn V Cgr.91 base 8.31 0.9 0.13 0.34 0.26 0.089Cr base 8.61 0.89 0.11 0.27 0.21 0.08

50m50m

Strong Gr.91 Weak 9Cr-1Mo

17

Asses the effect of oxide scale and metal loss on gr.91 creep properties

Loss of metal = decrease of load bearing section?Composite behavior with scale bearing some load?

porous magnetite Fe3O4 layer

(Fe,Cr) spinel layer

50m

18

Lifetime X3 in steam for 9Cr alloy due to load-bearing scale

- Load-bearing oxide scale has a significant impact on weak 9Cr

Tim

e to

rupt

ure

(h)

9Cr

Oxide

F oxideSoxide

S 9Cr

S9CrS

9Cr Gr. 91

19

Higher effect of load-bearing scale for thin specimen

- Thinner specimen to increase the volume to surface ratio

Tim

e to

rupt

ure

(h)

9Cr

Oxide

F oxideSoxide

S 9Cr

S9CrS

9Cr

20

Cracking/debonding of the outer oxide layer but continuous inner layer

0.5mm

rupture

Cu plating

Gr91, Steam, 350h,100MPa

spallation cracks

21

“Similar” oxide scale with and without stress except for healed cracks

500h in steam at 650ºC no load

50m

Gr91, Steam, 350h,100MPa

50m center

rupture

22

Cracks in the inner oxide scale seems to heal by formation of new oxides

20m

oxidized crackcrack

Gr91, Steam, 350h,100MPa

5m

inner layer

inner layer

23

Similar continuous inner layer with oxidized cracks for 9Cr alloy

Small cracks close to the alloy/inner oxide interface

100m 50m

9Cr, Steam, 314h,65MPa

oxidized crack

crack crack

24

Addition of thermal cycling during creep testing in steam

- Cycling resulted in lifetime decrease in steam for 9Cr and gr91 alloys

Tim

e to

rupt

ure

(h)

9Cr,1.5mm

Gr. 91, 2mm

Cycle: - 10h at 650ºC- ~3.5h cooling down to 250ºC - 5min at 250ºC - 2h heating to 650ºC

25

No scale degradation due to thermal cycling for 9Cr alloy

- Thinner scale for thermally cycled specimen

9Cr, Steam, 314h,65MPa 9Cr, Steam, 256h,65MPaThermal cycle 650-250ºC 50m 50m

26

No scale degradation due to thermal cycling for Gr.91 alloy

- Difference in scale composition?- No clear explanation for lower lifetime with thermal cycling

Gr91, Steam, 827h,90MPaThermal cycle 650-250ºC 50m

center

50mGr91, Steam, 350h,100MPa

27

Localized effect of thermal cycling during creep testing in steam?

- Will assess the effect of cycling on Gr.91 and 9Cr microstructure

50m center50m

Gr91, Steam, 350h,100MPa

Gr91, Steam, 827h,90MPaThermal cycle 650-250ºC

50m

28

FY15 Milestones

- Perform five creep-fatigue tests in humid atmosphere

- need to conduct 3 more tests

- Initiate five long-term creep-fatigue tests

- done

- Submit an open-literature paper on creep-fatigue of ferritic-martensitic steel

- on track

29

Future Activities

- Start assembling a new creep-fatigue machine

- Design a set up for creep-fatigue testing in steam

- Compare the performance of different creep-fatigue models based on damage accumulation

- Focus on gr. 91 to develop a microstructure-based model

30

Acknowledgements

- D. Erdman, C.S. Hawkins, T. Lowe, T. Jordan, B. Thiesing, D. McClurg for assistance with the experimental work- B. Pint, P. Tortorelli, Phil Maziasz and R.C. Cooper for exciting scientific discussionsThis research was sponsored by the U.S. Department of Energy, Office of Fossil Energy under the supervision of Vito Cedro III