Boiler superheater tubes in secondary creep stage

[robsalv]

Folks,

We have a medium sized, D style, marine boiler plonked in our plant for generation of steam. It's 46 yrs old.

In the Metallurgist's estimation, the metal replica surveys of the superheater tubes show that the tubes have entered secondary creep stage. Verbally I have been told that the pearlite has gone into solution and there's carbide formation in the grain boundaries. Haven't seen the micros yet.

Shield tubes are 1.25Cr 0.5Mo. Superheater element tubes are CS.

There are no obvious bulges, though there is some bowing in the superheater elements.

It looks like this could be the early stage of secondary creep, so I'm interested in opinions on how to manage, monitor or otherwise, with a view to determining when the budgeting cycle should allow for replacement.

At the very least I'm planning on instigating a yearly monitoring program to keep tabs on things.

Thoughts, opinions??

Thanks in advance.

Rob

 

[metengr]

Secondary creep rate of boiler tubing above 650 deg F is absolutely normal, and is no clause for alarm in the life of boiler tubing. Replicas in my opinion serve a limited useful purpose because they provide a view of only the tube OD surface. Degradation of pearlite is normal in these tube materials and is expected for service temperatures above 650 deg F. The best way to monitor the remaining service life of steam circuit boiler tubing is to perform ID oxide scale thickness testing in situ, and to remove periodic tube samples to corroborate the oxide scale thickness test results.

Typically, steam circuit boiler tubing should be monitored every 5 years or sooner (if, tube failures begin to occur). We have Cr-Mo superheater and reheater tube circuits along with low temperature primary superheater carbon steel tubes that have accumulated over 300,000 operating hours, and are still going strong. From my experience, I would expect properly designed boiler tubing to achieve at least 500,000 to 700,000 operating hours.

 

[sjrfc2]

Marine/shipboard boilers are designed for compactness (space and weight is a premium) so the heat flux through each tube is going to be higher. Where you might see 500K to 700K hours for properly designed plant boiler, it might not be realistic for a marine boiler.

I've seen some CRMO experience problems in less than 20K hours. Scale formation is the cause of the creep damage. Scale to prolonged overheat drives the tube material into a downward spiral: drives tube temp up and fostered additional scale growth, oxidation etc...

Part of the specific problem in our marine boiler was flow ditribution related (both steam and gas flow) and local to a specific area and selected tubes. We wound up managing the problem (extending the expected life of the effected tubes) by acid cleaning those tubes individually. The boiler was too old for a redesign.

I would have liked an NDT method to check the scale in-place but nothing was available for our application. B&W had a method to check scale but it wasn't applicable for the tube diameter and scale thickness in our application. Again since the heat flux was so high in the marine boiler, we were having problems with even 0.010-00.015" steamside oxide.

Again regarding compactness, periodic removal and sampling of tubes as tightly packed as those on a marine boiler often requires the removal of so many interference tubes, that it becomes more cost effective to just replace the superheater tubes rather than try to periodically remove a few for monitoring. We wanted to avoid that in our approach with acid cleaning individual tubes.

Individual acid cleaning was an effort though. You need access to the tube bores through header handholes and special adapters. Establishing flow through the tubes is also is crucial and may be tough if the design is complicated. We had bifurcated tubes and had to apply and maintain a vacuum prior to admitting the flow to ensure we had flow in all circuits. Otherwise we'd have pockets of air and no-flow circuits.

 

[metengr]

sjrfc2;
Periodic tube sampling along with oxide/scale thickness testing (if accessible) would achieve the desired results of monitoring tubes even in smaller industrial boilers.

 

[robsalv]

Thanks folks.

Interesting stuff. My boiler experience is a bit limited.

Cutting out a pup piece from both the shield tube and an accessible superheater tube would allow through section metallurgical micros and therefore truly establish material state.

One question. If you have really excellent water quality control and a top performing Demin plant, are you likely to build up an excessive oxide layer?

 

[robsalv]

Thanks folks.

Interesting stuff.

My boiler experience is a bit limited. Our B&W's have been well behaved beasties for as long as I've been in this area.

Good tip about the replica possibly only being representative of the surface. Due to space issues, cutting out a pup piece will be an issue, but I could probably arrange a diamond window cut out. This would allow through section micros and therefore truly establish the material state.

Our Demin plant output quality is excellent - which should reduce the propensity to steam side oxide scaling. But we haven't done an oxide check for years... Has this statement triggered anyone's alarm bells??

Cheers

Rob

 

[sjrfc2]

I think a steamside oxide layer will develop on CRMO tubes anyway. I remember this being the normal product of steam-ferrous metal reaction on the CRMO tube surface. The growth rate and thickness will be dependent on tube metal temperature. As a note, steamside oxide develops in steam piping as well (exfoliation f the scale leads to turbine blade erosion). When we checked a scale sample after a CRMO tube failed we found the scale to be greater than 95% iron confirming that it was a naturally occurring oxide rather than a scale from carryover.

 

[metengr]

sjrfc2;
Yes, oxide will develop. However, it is the thickness of the oxide over time that is the key to determining actual metal temperature that can be used to calculate the state of creep rupture damage. As the oxide becomes thicker due to exposure to higher than normal metal temperature in service, it will result in even higher heat flux (hotter tube metal) and increased oxidation (loss of tube wall). There have been empirical correlations developed on the family of Cr-Mo tube materials that show oxide thickness as a function of operating hours under various metal service temperatures. This information along with measured wall thickness were plotted using Larson Miller creep rupture data for the Cr-Mo tube materials, and an empirical correlation was developed and eventually made into a software program.

The in-situ testing consists of measured wall thickness (to establish service stresses in the tube wall), oxide thickness and known operating hours to date. This information is used in software to calculate a remaining service life based on a Cr-Mo creep rupture database. This is just a tool to be used along with removal of tube samples to evaluate the tube metal microstructure.

 

[robsalv]

Appreciate the additional comments sjrfc2 and metengr - I've been away so apologies for the delayed kudos.

Our thickness testing program showed that the tubes were essentially at nominal thicknesses, which bodes well.

I have a detailed (metal replica) metullurgical report which I'm yet to digest, but the indications are that there are quite a few more years left in our tubes - which will please the wallet holders of the organisation!

Thanks again.

Cheers

Rob

 

[Boiler67]

In addition to the evaluation method discussed by metengr, another point to consider is that prior to failure, Cr-Mo tubes will commonly swell. The amount of swelling is dependant on the failure cause; a tube that fails due to an internal blockage will swell much more (10% or more) than one that fails as a result of long term overheating (2%).

Swollen tubes are something that every boiler inspector should be on the lookout for.

 

[robsalv]

Hi there Boiler67, are we talking local bulges, or general overall swelling, that would take accurate measurements to detect?

Thanks.

Rob

 

[Boiler67]

Rob, typically the shorter term overheating (10% swelling) occurs locally, where the tube bulges. These types of failures can often be attributed to some immediate problem such as an internal blockage. The overall swelling is indicative of long term overheating, and typically doesn't occur within a short (1-dia.) length.

Since the original inquiry was about monitoring SH tubing that may be experiencing end of life issues, I would look for the overall swelling. Given the variations in tube wall, the most effective way I know of to measure swelling is to measure the inside diameter of a tube sample removed from the suspect section.

Also, using replication samples to evaluate tubing can be misleading because the outer tube surface is often decarburized during the tube manufacturing process. While more costly, I would recommend tube sampling before making any replacement recommendations.

 

[sjrfc2]

One thing that you might want to look at just for info is a fiber-optic scope on the inside of the tubes. When we had problems we noted you could see longitudinal lines on the inside of the tubes. It looked like longitudinal cracks in the tube surface.

The cracking wasn't in the tube but in the scale that had formed on the tube inside surface. The scale is brittle. As the ductile tube began to swell the brittle scale cracked.

Its by no means a fail-safe method and isn't quantifiable but if you see it you know something's going on. It might help identify where you want to pull samples if you go that route.

Are you having failures? We've had bowed tubes for years that have been OK. If they bowed apart to form a gas lane then it was a problem.

 

[robsalv]

Thanks for those replies Boiler67 and sjrfc2.

Fiberoptics sounds like it might be useful.

Yes, we have some bowing, which I've put down to creep - permanent elongation of tubes needs to go somewhere.

The poor results from the replica survey in conjunction with the bowing threw up the concerns raised in the OP. Discussions with the metallurgist [who has limited boiler experience] leaves me feeling a bit concerned though.

Basically, the CrMo shield tubes have some signs of microstructural degradation, put down to early signs of overheating - which I can live with. The remainder of the SH coil is carbon steel however, and the first and second rows are exhibiting bulk microstructural degradation, i.e., a bulk microstructure of ferrite, with the expected pearlite completely degraded into ferrite and spheroidal carbides. These spheroidal carbides had also precipitated to the grain boundaries. No voids or bulging observered at this stage.

What leaves me thinking that this is NOT a discountable surface replica type observation is that the micro structure was observed on the non flame exposed side, i.e., heat flux only.

We're instituting a yearly monitoring regime, and considering cutting out some diamond windows/pup pieces for more detailed analysis...

 

[metengr]

robsalv;
Tube or material sampling an excellent idea. Please be aware that microstructural degradation is typical for carbon steel boiler tubing in steam circuits that have been exposed to over 50,000 operating hours at metal temperatures above 750 deg F. This phenomenon is called spheroidization, and does not imply the tubes are in any imminent damager of failure from creep or stress rupture. The removal of samples followed by metallographic examination should provide you with good information.

PS; find yourself another competent metallurgical lab that can evaluate boiler tube microstructures for you.