Unusual Failures in Hydrogen Production

1
Unusual Failures in Hydrogen Production

• Sheldon W. Dean
• Dean Corrosion Technology, Inc.
• Allentown, PA
• May 17,2005

2
Hydrogen is New Opportunity

• Ideal fuel for fuel cells.
• Exhaust only water.
• Energy conversion is efficient.
• Fuel for engines.
• No hydrocarbons or CO in exhaust.
• NOX can be minimized.
• Other uses
• - Removal of S, N, etc. from fuel.

3
Steam-Hydrocarbon Reforming

• Largest source of hydrogen today.
• Methane (natural gas) feed.
• Higher hydrocarbons can be used.
• High temperature-catalytic process.
• Used for more than 50 years.

4
Steam-Hydrocarbon Reforming

• Reforming reaction
• CH4 2H2O ? CO2 4H2
• Typical conditions 1650ºF (900 ºC) 600psi with
  excess H2O.
• If higher hydrocarbons are present
• CxHy (2X - ½Y)H2 ? XCH4
• This step is run before reforming.

5
Typical Reforming Process

• Pre-reforming (hydrogenation)
• Steam introduction
• Reform over catalyst
• High temperature shift reactor (HTS)
• Low temperature shift reactor (LTS)
• Cooling and water condensation

6
Materials Problems Involving Corrosion

• HTS inlet piping Cracks in bypass
• (KOH SCC ).
• LTS exit piping Cracks in mixing tee
  (corrosion fatigue).

7
HTS Inlet Bypass Cracking

• BACKGROUND
• Large steam hydrocarbon reformer.
• Feed CH4 and some higher HCs.
• Potassium promoted reforming catalyst.
• Start-ups run with nitrogen then steam.
• HC feed added when plant temp. OK.

8
Background Continued

• HTS designed with bypass for start-up
  acceleration.
• HTS piping 1¼ Cr, ½ Mo steel.
• Bypass 304 SS because of possible condensation
  during operation (carbonic acid corrosion
  concern).
• Operating temp 850 ºF (455 ºC).

9
HTS Cracking Incident

• Many start-ups in first 6 months (gt10).
• Then 2 months continuous operation.
• HTS by-pass split open and tore off suddenly!
• Shock wave and fire resulted.
• Explosion heard 30 miles away.
• Repairs required 3 months to finish.

10
Diagram of By-pass
To HTS
By-pass 304L
1¼Cr,1/2Mo
Location of cracks
11
Failure Analysis HTS By-pass

• Many cracks found on interior of 304L.
• Pipe interior has black oxide.
• Cracks show white halos.
• Samples of pipe with crack taken.
• Metallographic sections examined.

12
Photo of Pipe Interior
13
Metallographic Section
14
Failure Analysis Continued

• Cracked section broken open.
• SEM examination of fracture surface.
• Energy dispersive X-ray analyses of fracture
  surfaces.

15
Crack Fractograph
16
EDS Analysis
K
17
Results of Failure Analysis

• Cracks are transgranular and branched.
• Cracks are typical of caustic SCC.
• No evidence of sodium present but potassium is
  widespread.
• Residual stress in pipe from cold forming.

18
Conclusion

• Pipe failure caused by KOH SCC.
• Surprising because SCC usually causes leak
  before break in 304L Catastrophic SCC failures
  are rare!
• Questions Where did KOH come from?
• How did it survive in by-pass but did not crack
  1¼ Cr, ½ Mo steel pipe?

19
Investigation of Source of KOH

• Transfer line refractory not source.
• Waste heat boiler not source.
• HTS catalyst not source.
• Reforming catalyst is source
• Potassium promoted catalyst.
• Depletion of K found in catalyst.

20
Source of KOH

• Potassium carbonate added to catalyst.
• Carbonate decomposes in hot steam
• K2CO3 H20 ? 2KOH CO2
• KOH vaporizes during start-up when hot
• 1788ºF (976ºC) V.P. 40 torr
• KOH deposits on cool bypass.

21
Laboratory Study

• Purpose to define when KOH SCC occurs.
• Use ASTM G 129 Slow strain rate method (SSR).
• Because water partial pressure fixed only one
  variable (temperature) available.

22
KOH Lab Study

• Temperatures 370, 420, 550ºF (188, 216,
  288ºC).
• 304L, 1¼Cr, ½Mo steel specimens.
• KOH concentrations calculated for each
  temperature.
• Head space had H2 added in some tests.
• K2CO3 added in some tests to 90.

23
Results of Lab Study

• 304L specimens did not fail until 50 psi H2 or
  CO added to head space.
• 304L OK at 370ºF, failed at higher
  temperatures.
• 1¼ Cr, ½ Mo steel failed at 370
  º,420 º but not at 550 ºF.
• 90 K2CO3 caused no change.

24
(No Transcript)
25
Recommendations

• Redesign by-pass with free draining 1¼ Cr, ½
  Mo steel material.
• Questions for further research
• Why is hydrogen necessary for SCC ?
• Is there an upper temp limit for 304L?
• Are other alloys better for piping?

26
Mixing Tee Cracking

• Background
• Large hydrocarbon reforming plant.
• Feed refinery off gas (mainly methane).
• Condenser used as boiler water pre-heater.
• By-pass used to prevent boiling in BWPH.

27
Mixing Tee Incident

• Plant operated for about 3 years.
• Leak noted in mixing tee at exit of BWPH.
• Plant shut down and inspected internally.
• Extensive cracking found in tee and down
  stream, also in expander.

28
Concentric Cone Heads
315F
220ºF
BOILER WATER PREHEATER
Location of cracks
Control Valve
Expander
By-pass Line
29
Failure Analysis

• Most cracks associated with welds.
• Metallography some cracks straight, others
  branched.
• Some cracks showed beach marks.
• Fracture surface quasi-cleavage.
• EDS shows faint Cl and S on fracture surface.

30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
Investigation

• Several older plants had similar cracks.
• Found in hydrocarbon reformers (not natural gas
  feed plants).
• Upgrade to 316L or Alloy 20 not successful.
• Upgrade to 625 successful.

36
Laboratory Study

• ASTM G 129 SSR tests.
• Cyclic straining tests
• 0.3 offset pre-strain,40.7 to 13.7ksi stress
  range.
• 2000 cycles _at_ 1 Hz.
• Potential control, Ag/AgCl ref. electrode.
• 304L, 2205, 625 specimens.

37
Results 304L

• No cracking with SSR tests.
• No cracking with cyclic straining w/o
  electrochemical polarization.
• Cracking gt200mV, 10ppm Cl-, 150ºC.
• Cracking gt200mV, 1ppm Cl-, 1ppm NaCNS, 150ºC.

38
Results 625 2205

• No cracking in cyclic straining tests up to
  600mV.
• No cracking in SSR tests.

39
Conclusions

• 304L shows synergistic cracking in water with
  Cl- and potential gt 200mV.
• SCN- aggravates chloride cracking.
• Both 625 and 2205 resist cracking.

40
Recommendations for Plant Design

• Minimize thermal cycling of equipment.
• Use resistant materials (2205 or 625) where
  thermal cycling occurs.
• Avoid concentric cone heads on heat exchangers
  where condensation occurs.

41
Issues for Further Study

• What are the effects of other impurities in hot
  water on corrosion fatigue?
• What role does corrosion play in the cracking
  process?
• What other alloys resist corrosion fatigue?
• Can we predict cracking susceptibility?

42
THANK YOU!

• Acknowledgement
• Air Products and Chemicals, Inc.
• Intercorr International Inc.
• W. R. Watkins
• K.L. Baumert
• J. W. Slusser