San Onofre steam generator tubes leaking - why?

In summary: This is a manufacturing defect - the weld should have been a perfect fit. It's possible that the metal fatigue caused the crack.In summary, a manufacturing defect has been found in one of the replacement steam generators for the San Onofre nuclear plant. The cause of the defect is not yet known, but it is likely corrosion or cracking on the secondary side of the steam generator tubes. The SG tubes were made of Inconel, which is a tricky alloy system and is known to be susceptible to metal fatigue.
Engineering news on Phys.org
  • #2


Steam generator tubes are [STRIKE]leaking[/STRIKE] thinning .
This was a big problem stateside 1n 1970's.

Chemistry of water on secondary side is extremely important.
We measured impurities in parts-per--billion. One cup of tap water was enough to cause a shutdown to flush the steam generators, i know because one of our technicians used a cup of tapwater to top off a level instrument... once.

Metallurgy was important also. Copper contributed to corrosion.
We replaced the admiralty brass tubes in condenser with titanium, and feedwater heaters with stainless steel.

These lessons were learned almost forty years ago. Our replacement generators from Westinghouse-Tampa are doing fine.
So - what's going on now? That's REAL good question.
First question pops to mind is "Where did Mitsubishi procure the metal for the tubes in those replacement steam generators?"

Second is "How's the plant's water chemistry ?"

It'll be interesting to follow this one.

Edited first line. sorry.
 
Last edited:
  • #3
http://www.clipsyndicate.com/video/playlist/1510/3242180?cpt=8&title=ans_nuclear_clips&wpid=752

The San Onofre plant replaced its steam generators in 2010; it is almost certain this is a manufacturing defect.
 
  • #5
jensjakob said:
Hi,

San Onofre, menioned durring the Japan crisis, is leaking and corroding pipes fast:
http://www.scpr.org/blogs/news/2012/03/22/5205/san-onofre-reactors-down-indefinitely/

http://sciencedude.ocregister.com/2012/02/02/nuclear-leak-damage-to-both-reactor-units/167503/

What is going on- any ideas?

Take care

Jens Jakob
The fact that these are replacement SG's and the tubes failed during the first cycle of operation would implicate 1) a manufacturing defect or 2) a problem with installation.

SCE replaced SGs in SONGS2 during 2009.
24 February 2009

Two replacement steam generators have been delivered for the second unit of Southern California Edison's (SCE's) San Onofre Nuclear Generating Station (SONGS). Mitsubishi Heavy Industries (MHI) made the components.
http://www.world-nuclear-news.org/newsarticle.aspx?id=24719

04 October 2010

Mitsubishi Heavy Industries (MHI) has delivered two replacement steam generators for the third unit of Southern California Edison's (SCE's) San Onofre Nuclear Generating Station (SONGS).

MHI said that the replacement steam generators delivered for SONGS 3 are among the world's largest, each measuring approximately 20 metres in length, seven metres in diameter and weighing some 580 tonnes. Each of them contains about 10,000 heat transfer tubes.
http://www.world-nuclear-news.org/C-New_steam_generators_for_SONGS_3-0410105.html

'Why' is the question - indeed! There is an ongoing investigation. It must be determined if any corrosion (intergranular stress-corrosion) and/or cracking is occurring, and if it is initiated on the primary or secondary side.

I believe the material is Inconel 690, which is supposed to be superior to Inconel 600. However, Inconels are notoriously tricky alloy systems.
 
Last edited:
  • #6
"The company has received overseas orders for 31 units, mainly from North America and Europe. "

Hmmm - could be interesting to track those and compare corrosion.
 
  • #8
More info at the SONGS website:

http://www.songscommunity.com/news.asp

Confirms Unit 2 S/G replacement in 2009 and Unit 3 in 2010. The following link is good info on the testing methods being used:

http://www.songscommunity.com/docs/Test_Inspections.pdf
 
Last edited by a moderator:
  • #9
Picture of Steam Generator (MHI):

http://www.mhi.co.jp/en/products/detail/steam_generator.html

Unit 3 S/G had leakage before delivery. Note that this was not tube leakage so current problems may not be directly related.

http://mdn.mainichi.jp/mdnnews/news/20120229p2g00m0dm058000c.html

If there is a manufacturing problem, Mitsubishi has already delivered over 100 steam generators around the world.I am less worried about the 31 on order.
 
Last edited by a moderator:
  • #10
NUCENG said:
Unit 3 S/G had leakage before delivery. Note that this was not tube leakage so current problems may not be directly related.

http://mdn.mainichi.jp/mdnnews/news/20120229p2g00m0dm058000c.html
The article states
According to the NRC documents, Mitsubishi Heavy discovered a 5 inch (12.7 centimeter) long crack in the dissimilar metal weld between the divider plate and the channel head of the steam generator of the unit 3 reactor during its routine visual inspection in March 2009.
If that is the replacement S/G, that's rather troubling. In 2009, that SG would have been at the Mitsubishi shop - ostensibly before shipment. Or did Mitsubishi inspect the older in-service SG, which was replaced in 2010?

A crack in the divider plate is not bad as long as it doesn't propagate. A breach in the divider plate would allow leakage from the hot leg to the cold leg, thus by-passing the SG tube bundle. It is still within the primary system.

From the description, it sounds like the crack was at the edge of the divider plate where it joins the vessel (channel) head.
 
Last edited by a moderator:
  • #11
Astronuc said:
The article states If that is the replacement S/G, that's rather troubling. In 2009, that SG would have been at the Mitsubishi shop - ostensibly before shipment. Or did Mitsubishi inspect the older in-service SG, which was replaced in 2010?

A crack in the divider plate is not bad as long as it doesn't propagate. A breach in the divider plate would allow leakage from the hot leg to the cold leg, thus by-passing the SG tube bundle. It is still within the primary system.

From the description, it sounds like the crack was at the edge of the divider plate where it joins the vessel (channel) head.

I understand the crack was on the Unit 3 replacement S/G and was repaired before shipment and installation in 2010.
 
  • #12
NUCENG said:
I understand the crack was on the Unit 3 replacement S/G and was repaired before shipment and installation in 2010.
That would seem to be what the article implicates. I was hoping for confirmation.

Nevertheless, I'm puzzled about the crack, and also about welding dissimilar metals. I would expect the vessel shell to be line with stainless steel, and the divider plate to be made of the same stainless steel. Certainly if one welds a low carbon stainless steel to a high carbon steel, cracking can be an issue.

I'm curious about their process and procedures, since the procedures should be such that cracking is prevented/avoided.

I'd like to know if the Inconel tubes are cracking (which would imply either poor material and/or poor fabrication practice) or leaking about the fitup at the tube sheet (which would imply a poor process).

Nevertheless, it is very troubling that such failures occur in something that is designed to last 20 to 30 years.

Original SGs were supposed to last the life of the plant (40) years, and if possible now 60 years. They represent a substantial capital cost, and the economic models, which I studied at university, never included SG replacement. After I finished by undergrad, I learned about how Inconel 600 components (and certainly welding materials) were failing prematurely.

Primary water chemistry, and in some cases, secondary water chemistry are certainly factors.
 
  • #13
Astronuc said:
That would seem to be what the article implicates. I was hoping for confirmation.

Nevertheless, I'm puzzled about the crack, and also about welding dissimilar metals. I would expect the vessel shell to be line with stainless steel, and the divider plate to be made of the same stainless steel. Certainly if one welds a low carbon stainless steel to a high carbon steel, cracking can be an issue.

I'm curious about their process and procedures, since the procedures should be such that cracking is prevented/avoided.

I'd like to know if the Inconel tubes are cracking (which would imply either poor material and/or poor fabrication practice) or leaking about the fitup at the tube sheet (which would imply a poor process).

Nevertheless, it is very troubling that such failures occur in something that is designed to last 20 to 30 years.

Original SGs were supposed to last the life of the plant (40) years, and if possible now 60 years. They represent a substantial capital cost, and the economic models, which I studied at university, never included SG replacement. After I finished by undergrad, I learned about how Inconel 600 components (and certainly welding materials) were failing prematurely.

Primary water chemistry, and in some cases, secondary water chemistry are certainly factors.

And anytime you open a system there is a chance for loose parts or materials to enter a system. Flow induced vibration can cause wear and tear. A few years ago replacement condensate pumps for a nuclear plant were exposed to road grime and sludge during shipment because penetration seals were not properly installed. I too am concerned this has developed so soon after S/G replacement and that is likely why NRC sent the AIT. I am not a welding or S/G expert, but if I find additional information I will post it. We should learn more when the NRC AIT has their exit meeting.
 
  • #14
I'd like to know if the Inconel tubes are cracking (which would imply either poor material and/or poor fabrication practice) or leaking about the fitup at the tube sheet (which would imply a poor process).



From Nuceng's http://pbadupws.nrc.gov/docs/ML1207/ML12075A219.pdf

Continuing inspections of 100% of the steam generator tubes in both Unit 3
steam generators discovered unexpected wear, including tube to tube as well as tube to tube
support structural wear.
and http://www.nrc.gov/reading-rm/doc-collections/event-status/event/2012/20120319en.html
Steam generators do experience some wear during the first year of operation but the level of tube wear at Unit 3 is unusual.


Chemistry problems can deposit solids in the support to tube annulus and squeeze the tubes.
But tube-to-tube wear sounds more like a vibration issue arising from mechanical design.
It's hard to believe mechanical vibration trouble after this many years experience making Steam Generators.

As you said, it'll be interesting to see what they find.
 
  • #15
jim hardy said:
Chemistry problems can deposit solids in the support to tube annulus and squeeze the tubes.
But tube-to-tube wear sounds more like a vibration issue arising from mechanical design.
It's hard to believe mechanical vibration trouble after this many years experience making Steam Generators.

As you said, it'll be interesting to see what they find.
The chemistry practices are pretty standard these days. There could be an issue with commissioning a fresh surface.

Replacement generators may have higher flow rates. I can't remember if there was a plant uprate with the steam generator replacement.

Tube wear after one cycle of operation would be troubling. Despite experience, designer make 'improvements' that sometime may introduce performance problems. There was a case of two BWRs* in which new advanced turbines developed cracks in one of the late stages in the LP turbine. Subsequent CFD reveal a design flaw. The CFD analysis (which is very mature these days) should have been part of the initial design process.

*Hamaoka 5 and Shika 2 off line after turbine vane failures
http://www.neimagazine.com/story.asp?storyCode=2038314


High cycle fatigue (either mechanical (FIV) or thermo-mechanical) is a possibility if the frequency is in the acoustic range (10-1000s Hz) with 3.156E7 s/yr.
 
Last edited by a moderator:
  • #16
Astronuc said:
The chemistry practices are pretty standard these days. There could be an issue with commissioning a fresh surface.

Replacement generators may have higher flow rates. I can't remember if there was a plant uprate with the steam generator replacement.

Tube wear after one cycle of operation would be troubling. Despite experience, designer make 'improvements' that sometime may introduce performance problems. There was a case of two BWRs* in which new advanced turbines developed cracks in one of the late stages in the LP turbine. Subsequent CFD reveal a design flaw. The CFD analysis (which is very mature these days) should have been part of the initial design process.

*Hamaoka 5 and Shika 2 off line after turbine vane failures
http://www.neimagazine.com/story.asp?storyCode=2038314 High cycle fatigue (either mechanical (FIV) or thermo-mechanical) is a possibility if the frequency is in the acoustic range (10-1000s Hz) with 3.156E7 s/yr.

San Onofre Units 2 and 3 have both got approved Margin Uncertainty Recovery Power Uprates of 1.4% in 2001.
 
Last edited by a moderator:
  • #17
NUCENG said:
San Onofre Units 2 and 3 have both got approved Margin Uncertainty Recovery Power Uprates of 1.4% in 2001.
That's enough time to incorporate into the current replacement design. In terms of uprate, I was thinking more along the lines of an extended or stretch uprate with 5+% increase in reactor/plant output.

Mitsubishi is a Westinghouse licensee, and they have probably replaced more W-SG than CE SGs. The large CE plants (mostly 16x16 fueled) typically use 2 steam generators - with one hot leg and two cold legs. They are therefore typically larger than W-SGs. Could that be a factor?
 
  • #18
Hamaoka 5 and Shika 2 off line after turbine vane failures
http://www.neimagazine.com/story.asp?storyCode=2038314

Ahhh,, Turbine blades - another of those fascinating industry "niches" .
Rotor dynamics is fascinating.


High cycle fatigue (either mechanical (FIV) or thermo-mechanical) is a possibility if the frequency is in the acoustic range (10-1000s Hz)

Tubes will rattle.
I suppose it's quite a calculation to get the natural frequency and vibration modes of a long hollow tube that's pressurized with the fluid inside having considerable velocity.
When i read how a Coriolis Flowmeter works , i just felt like saluting the entire Mechanical Engineering community.
http://en.wikipedia.org/wiki/Mass_flow_meter#Operating_principle_of_a_coriolis_flow_meter
it somewhat resembles the u-tubes in steam generator, see this graphic
http://en.wikipedia.org/wiki/File:Coriolis_meter_vibrating_no-flow_512x512.gif

I'm admitting my abysmal ignorance here. I know just enough to not cast stones, and that wasn't intent of previous post...
 
Last edited by a moderator:
  • #19
NRC Region IV Administrator Elmo E. Collins said. “Until we are satisfied that has been done, the plant will not be permitted to restart.”

On Jan. 31, operators performed a rapid shutdown of the Unit 3 reactor after indications of a steam generator tube leak. Unit 2 has been shut down since Jan. 9 for a planned refueling and maintenance outage. Subsequent inspections at both units have identified unusual wear in many tubes of the steam generators, which were replaced in January 2010 at Unit 2 and January 2011 in Unit 3.

SCE has identified two causes of the unusual wear: tubes are vibrating and rubbing against adjacent tubes and against support structures inside the steam generators. They are still working to determine why this is occurring.

Only one tube required pressure testing on Unit 2. However, six other tubes required plugging, and 186 additional tubes were plugged as a precautionary measure. Eight tubes failed pressure testing at Unit 3, indicating that these tubes could have failed under some accident conditions. Evaluation for additional plugging or other corrective actions are continuing for Unit 2, based on ongoing evaluations of Unit 3 test results.
CAL 4-12-001 - http://www.nrc.gov/reading-rm/doc-collections/news/2012/12-011.iv.pdf
CONFIRMATORY ACTION LETTER – SAN ONOFRE NUCLEAR GENERATING STATION, UNITS 2 AND 3, COMMITMENTS TO ADDRESS STEAM GENERATOR TUBE DEGRADATION

For both Units 2 and 3, this was the first cycle of operation with new replacement steam generators. Unit 2 replaced its steam generators in January 2010, and Unit 3 in January 2011. Each steam generator has 9,727 steam generator tubes.
 
  • #20
Story on San Onofre Steam Generator Leakage.

http://www.power-eng.com/news/2012/04/02/expert-cites-reasons-for-san-onofre-troubles.html

Story above is based on Arnie Gunderson Report prepared for the Friends of the Earth environmental and anti-nuclear group.

http://fairewinds.com/content/foe-report-steam-generator-failures-san-onofre

Biggest error is that the quote from NRC chairman Jaczco that NRC approval is not required for restart is not current. The Confirmatory Action Letter issued to SCE requires NRC approval. I am still looking for a copy of the CAL itself. I haven’t found it on ADAMS yet.

I do like the list of changes implemented in the new steam generators. Arnie is correct that the increased number of tubes, change in tube alloy, changes in tube support structure (egg crate - implying fragility?) and increased coolant flow are potential causes.

I think pulling in the issue of BWR Dryer Cracking is a stretch though. Anyway this is a potential for a good discussion here on PF.
 
Last edited by a moderator:
  • #21
The CAL is in my previous post.

I don't think switching to 690 from 600 (?) is an issue, as that is the industry practice over the last two decades. Smaller tubes and increased flow rates might play a role, but I'd have expected a CFD analysis would have caught that - but maybe not if not done right.

BWR dryer is different material in a different environment over a longer period. That's mixing apples with oranges, but it is an interesting topic nevertheless.
 
  • #22
Astronuc said:
The CAL is in my previous post.

I don't think switching to 690 from 600 (?) is an issue, as that is the industry practice over the last two decades. Smaller tubes and increased flow rates might play a role, but I'd have expected a CFD analysis would have caught that - but maybe not if not done right.

BWR dryer is different material in a different environment over a longer period. That's mixing apples with oranges, but it is an interesting topic nevertheless.

Thanks, I was not paging down far enough to see the CAL - operator error!

You expressed interest in the divider plate defect in the rplacement S/G for unit 3. Here are the references I found. (I am not a weld engineer and won't even try to comment.)

IN 2010-07
http://pbadupws.nrc.gov/docs/ML1000/ML100070106.pdf

Slide Presentation on root cause:
http://pbadupws.nrc.gov/docs/ML0925/ML092590470.pdf

Non Proprietary Root Cause Report:

http://pbadupws.nrc.gov/docs/ML0926/ML092600513.pdf
http://pbadupws.nrc.gov/docs/ML0926/ML092600515.pdf
http://pbadupws.nrc.gov/docs/ML0926/ML092600516.pdf
 
  • #23
Somewhat relevant - Effects of Alloy Chemistry, Cold Work, and Water Chemistry on Corrosion Fatigue and Stress Corrosion Cracking of Nickel Alloys and Welds
http://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr6721/cr6721.pdf

They process used low carbon Alloy 152 in butter welds as expected.

They used gouging to remove the SS cladding in Unit 3 RSGs rather than the machining used in Unit 2 RSGs.
 
  • #24
I came across this news article that may be of interest.

It sounds as though anti-vibration supports were removed to increase the number of tubes in the generator. This has led to more vibration and mechanical wear as a result.
 
  • #25
Hologram0110 said:
I came across this news article that may be of interest.

It sounds as though anti-vibration supports were removed to increase the number of tubes in the generator. This has led to more vibration and mechanical wear as a result.

The source of that position is Arne Gunderson speculation on a cause. In previous posts we have listed other possible causes. Clearly the NRC recognizes the need to determine a root cause (as evidenced by the CAL). I urge you to reserve judgment until the facts are determined. I am certain that the root cause evaluation will be released (although some proprietary information may be withheld). I am also certain that NRC staff, ACRS, and every "nuclear watchdog" organization will subject the root cause to independent review. It is possible that Arnie is right, just not certain based on his previous record.
 
  • #26
You're absolutely right. From the article:

The report on San Onofre by Fairewinds Associates, a Vermont-based consultant that has worked with groups critical of nuclear power, suggests that "imprudent design and fabrication decisions" may be to blame for accelerated wear on generator steam tubes. Friends of the Earth commissioned the analysis.

The article just popped up in my news feed this morning and I remembered there was a thread about it on Physics Forums. I wasn't following the tread so I assumed that this was a 'new' analysis. Seems I missed that the tread was dead. Sorry about that.
 
  • #27
Hologram0110 said:
You're absolutely right. From the article:

The article just popped up in my news feed this morning and I remembered there was a thread about it on Physics Forums. I wasn't following the tread so I assumed that this was a 'new' analysis. Seems I missed that the tread was dead. Sorry about that.

Here is the latest - a Part 21 report from MHI.

http://www.nrc.gov/reading-rm/doc-collections/event-status/event/2012/20120416en.html

Part 21 Event Number: 47833
Rep Org: MITSUBISHI NUCLEAR ENERGY SYSTEMS
Licensee: MITSUBISHI HEAVY INDUSTRIES, LTD
Region: 1
City: ARLINGTON State: VA
County:
License #:
Agreement: Y
Docket:
NRC Notified By: EI KADOKAMI
HQ OPS Officer: JOHN KNOKE Notification Date: 04/13/2012
Notification Time: 15:58 [ET]
Event Date: 04/13/2012
Event Time: [EDT]
Last Update Date: 04/16/2012
Emergency Class: NON EMERGENCY
10 CFR Section:
21.21(a)(2) - INTERIM EVAL OF DEVIATION
Person (Organization):
BLAKE WELLING (R1DO)
KATHLEEN O'DONOHUE (R2DO)
DAVID HILLS (R3DO)
VINCENT GADDY (R4DO)
PART 21 GROUP (EMAI)


Event Text

PART 21 INTERIM REPORT - STEAM GENERATOR TUBE WEAR

This interim Part 21 is in regard to San Onofre Nuclear Generating Station, Unit 2, Steam Generator replacement.

"During the first refueling outage following steam generator replacement, eddy current testing identified ten total tubes with depths of 90 to 28 percent of the tube wall thickness. Some of the affected tubes were located adjacent to retainer bars. The retainer bars are part of the floating anti-vibration bar (AVB) structure that stabilizes the u-bend region of the tubes.

"Other tubes in the two steam generators had detectable wear associated with support points elsewhere in the AVB structure. Each steam generator has 9727 tubes with an 8 percent (778 tubes) design margin for tube plugging.

"Discovery Date: February 13, 2012

"Evaluation completion schedule date: May 31, 2012"

"Those Mitsubishi Heavy Industries customers potentially affected by this issue have been notified and will receive a copy of this interim report."

Reference Document: UET-20120089
Interim Report No: U21-018-IR (0)
 
  • #28
  • #30
i'm sure curious why the new tubes fail .

...limited to operate at a lower power rating to avoid the possibility that flow induced vibration is a cause.

Vibration can be excited from either inside or outside a tube.
Flow inside those tubes barely changes with power.
Seems to me a microphone on the steam generator could hear tubes clattering.
I'd instrument a steam generator and listen. If they clatter at zero power then excitation is from primary flow not secondary.
Most plants have loose parts monitors that are basically microphones at natural collection points like reactor vessel bottom and steam generator inlet side tubesheet. Move one up to vicinity of the tube wear region.

old jim
 
  • #31
jim hardy said:
i'm sure curious why the new tubes fail .

Vibration can be excited from either inside or outside a tube.
Flow inside those tubes barely changes with power.
If the tubes are contacting each other, then that's some relatively large amplitude vibration, which means the tubes are not sufficiently stiff, or there is some pretty substatial excitation mechanism.

Flow might have increased because of the reduced pressure drop, and perhaps flow was increased slightly, on the primary and/or secondary side in order to increase power output. Increased flow in the primary circuit always an issue when replacing steam generators.

I'm puzzled about what kind of analysis was performed concerning the design. In this day and age, we have pretty advanced CFD capability. I'm left wondering - what did they miss, or not consider, in the design and the analysis.

Seems to me a microphone on the steam generator could hear tubes clattering.
I'd instrument a steam generator and listen. If they clatter at zero power then excitation is from primary flow not secondary.

Most plants have loose parts monitors that are basically microphones at natural collection points like reactor vessel bottom and steam generator inlet side tubesheet. Move one up to vicinity of the tube wear region.
Acoustic emissions (noise) analysis would be appropriate, but I'm not sure it if is done on SGs.
 
  • #32
If the tubes are contacting each other, then that's some relatively large amplitude vibration, which means the tubes are not sufficiently stiff, or there is some pretty substatial excitation mechanism.

iirc the tube diameter was decreased and bending moment is in proportion to moment of inertia of cross section, i think 3rd or 4th power of diameter ?
http://en.wikipedia.org/wiki/List_of_area_moments_of_inertia
(Pardon me I'm no mechanical engineer) so reducing diameter will reduce stiffness ? Surely they calculated that. The tubes get additional stiffness due to internal-external Δp and i don't know how to calculate that. That Δp is not constant as main steam pressure changes from ~ 1000 psi to ~ 800 with power.
As you said surely they couldn't have missed that.

It gets curioser and curioser.
They'll figure it out. They have my genuine sympathy .

Acoustic emissions (noise) analysis would be appropriate, but I'm not sure it if is done on SGs.

we had loose parts sensors at entry point of feedwater line to steam generator. You could hear internals of check valve tinkling at low flow. That'd be the closest point i know of. Sound telegraphs pretty well through steel , so one might hear something at primary tube sheet.



old jim
 
  • #33
Installing such acoustic monitors might have been possible as part of the RSG startup, but I don't think they will be heating the unit up / running the RCPs just so they can listen in and try to identify the problem. Too late for that. They will have to figure it out with inspections, whatever operating data is available now, and some fancy analysis.
 
  • #35
Astronuc, How do you feel this affects MHI's reputation in the USA industry after this?
 

Hot Threads

Recent Insights

Back
Top