Creep measurement 2

[MalcolmA]

I'm looking for advice on techniques to measure the amount of creep that has been sustained by furnace tubes.
Also, what is considered an acceptable amount of creep?
Are there any codes or standards that provide guidelines?

 

[metengr]

Typically, for boiler tube samples one option to have isostress creep testing performed on test specimens removed from boiler tube samples. Because of the wall thickness used in standard reheater and superheater boiler tubing, you are faced with having to remove sub-size tensile specimens from the longitudinal (along the length of the tube) direction and have them creep tested in an inert atmosphere to avoid surface oxidation.

I have found that isostress creep testing (where the test specimens are loaded to actual service stress levels and the test temperature is increased above the typical service temperature to provide various creep deformation rates and creep rupture lives), provides the best estimate to measure remaining creep life. You will need to have the boiler tube samples sent to a reputable metallurgical laboratory that can conduct creep testing. In addition, you will need to provide the lab with boiler tube operating data - typical working pressure and temperature conditions applicable to the tube circuit. I want to warn you that over the years that I performed failure analysis on boiler tubes, I have fond considerable scatter in attempting to correlate remaining creep life to actual remaining service life because of the numerous factors in a boiler environment that can effect the service life of boiler tubes - local metal temperature variations in a steam circuit, surface oxidation effects and thermal/mechanical stresses on boiler tubes.

A second option to isostress creep testing is to perform in-situ oxide thickness testing on boiler tubes. This is a proven method of determining the remaining creep life of steam-touched ferritic boiler tubing by measuring the thickness of the oxide scale on the tube ID surface, and the actual wall thickness. There is an equation that was developed that relates measured wall thickness and measured oxide thickness scale of boiler tubes to a Larson Miller creep parameter. This parameter is plotted on a graph of stress versus Larson Miller parameter for new boiler tube alloys to determine remaining creep rupture life. The oxide scale thickness testing method is offered by most boiler OEM's and is a valuable tool to assess the condition of a reheater or superheater steam circuit. It also avoids the cost of having to perform isostress creep testing on individual tube samples.

The original design assumption used by the ASME Boiler and Pressure Vessel code committee regarding the use of boiler tube materials exposed to time dependent damage (ie., creep or stress rupture) is mentioned in ASME Section II, Part D Appendix 1. In general, the selection of boiler tubing for a targeted service temperature and stress level is based on the lowest of the following;
a creep deformation rate of 0.01%/1000 hours or
80% of the minimum stress to cause rupture after 100,000 hours or
100% of the average stress to cause rupture after 100,000 hours.

 

[davefitz]

Meteng:

The above reply was exceelent, and seems applicable to ferritic tubes and piping that was originally supplied with the correct metallurgical structure.

We have a slightly different problem: we have 8 units which have P91 HP main steam lines which were provided incorrectly heat treated hot-bends . The foundry simply placed the piping in a hot bending machine, locally heated ( via induction) the bend section to 2000F + , bent the pipe, then only provided a tempering treatment tot he entire pipe section. This process led to formation of ferrite in the areas near the hot bends, and the lack of a correct N+T prevented formation of uniform tempered martensite. This problem was detected after plant startup by later in-situ hardness testing and review of the foundry records.

As a result , we have 8 plants with an unknown creep life of the bends, since we have been unable to find published data on the creep life of ferritic modified 9 crome .

 

[davefitz]

also, the ferrite microstructure was confirmed via plastic film replication, and the foundry has provided the same incorrect hot bending treatment of P91 piping on about 50 other plants over the last decade.

 

[metengr]

davefitz;
I will try to do some investigating on my end in terms of published technical papers. I recently attended the 4th International Conference on Advances in Materials for Fossil Power Plants. I have a lot of technical papers on P91 material, specifically with long term creep testing and microstructure correlations. I will report back to you in this forum what I find on this subject.

 

[davefitz]

Thanks, Meteng. I missed than conference, but I heard the papers will be published by ASM in 2005.

 

[NickelMet]

Fellas, what about using the calculations (methodology) in API 530 (Calculation of Heater Tube Thickness in Petroleum Refineries) and/or API 579 (Fitness for Service)?

We have used both concurrently and coupled random field metallography & replication work with the analysis to give us our "warm and fuzzy". Comments?

~NiM

 

[metengr]

NickelMet;
I believe the original question was related to determining the remaining creep life of furnace tubes. Other than actual creep testing of tube samples or the ID oxide/wall thickness testing (as mentioned above) coupled with metallographic examination of the tube metal microstructure, you would not be able to estimate the remaining useful service life.

Field metallography or replication are also useful tools that will provide some subjective level of in-service damage assessment, but it cannot be used to predict actual remaining service life.

 

[NickelMet]

Thanks for the information Metengr.

~NiM

 

[MalcolmA]

I would like to add some clarification to my situation.
The particular tubes that I am involved with are a horizontal steam superheat coil in the convection section of a furnace. The flue gas is entering the covection section at a temperature of 2000 deg F. The tubes are incoloy 800H. The flue gas has been running hotter than the original design parameters and the tubes have exceeded the life predicted by API RP 530.

I am at a point where I would like to know what measurements are available to determine the amount of creep the tubes have sustained and what is an acceptable amount of creep so I can make a determination of when to replace the tubes.

Is ID oxide test applicable to 800H material?

I appreciate any advice I can get.

 

[metengr]

MalcomA;
The oxide scale thickness test is not applicable to Incoloy 800H material, it is only used on ferritic tube materials that contain less than 9% Cr.

For your situation, you need to conduct actual isostress creep testing to evaluate the remaining creep life. I would suggest removing several representative tube samples from 1/4 point locations across the width of the furnace.

The creep testing will consist of evaluating creep deformation rates and creep rupture times using a matrix of temperature and stress conditions. The data will be converted to a Larson Miller parameter and compared with creep curves for new Incoloy 800H material.

Have you experienced any tube failures?

 

[metengr]

To give you some concept of the information obtained from creep rupture testing as I mentioned above, the following web site is from Special Metals with a Tech Bulletin on Incoloy 800H. Pages 5 and 6 of the bulletin show creep rate and rupture data that would be compared with your actual tube sample creep test results. Your data would be plotted on these curves to determine remaining useful service life.

http://www.specialmetals.com/products/incoloyalloy800h&ht.htm

 

[MalcolmA]

metengr

Thank you for the info.
We have not experienced any ruptures yet. We would like to replace the tubes before there is a failure. The steam is at a pressure of 1500 psi and would cause some serious damage if it ruptured.