Grade 91 Steel Performance 3

[NCTHAI]

Has anyone come across issues with Grade 91 steel performance?

The grade 91 steel metullurgy changes dues to welding, stress, operating temperatures and other service parameters. Has any one come across this issue?

Has there been any in service failures experinced? What are methods for predictive / condition based maintenance to avoid this issue?

NC

 

[JohnBreen]

I believe you are asking about ASTM A-335, P91 piping material. This is a topic that has been discussed here many times.

To use the search feature to find previous discussions on this material do the following:

1. Go to the top of the page and find the window between the words "Google" and "search".
2. Type in "P91".
3. Click on the "search button.
4. Read at least 100 references to "P91"

Regards, John.

 

[davefitz]

Problems with P91 steel? Let me count the ways.

1. The ASME code minimum requirements for fabrication are not an adequate or complete guide to the required procedures needed to retain its high temperature creep resistance in the general case. The designer must have additonal specialized training or have the fabrication procedures reviewed and approved by a metallurgist familiar with P91 in order to ensure a relaible component has been fabricated.

2. hot bending- the entire hot bent piece must be N+T'd following hot bending . The industry is struggling with determining safe mitigation procedures for continued use of bends that were not N+T'd,

3.welding - The optimum weld procedure to prevent prompt cracking prior to PWHT is at odds with apprent ASME requirements. Raising the electrode Ni content to range of 0.4-0.8% is helpful in preventing prompt cracking in components that must withstand shop handling stresses or pipeline fab stresses, but this higher Ni content implies that the weld must be cooled to room temp for complete martensite transformation prior to PWHT. This technique to improve weld ductility may contradict ASME welding requirements , according to the way some people read the Code.

4.welding - low hydrogen procedures are neccesary, and all this implies to selection and handling of electrodes and protection of the welded piece prior to PWHT

5. N+T oven- the proper means of normalizing a P91 piece includes the requirement that the piece be cooled at rate faster than 9. F/min, yet many commercial furnaces do not recognize this requirement, with the result that excessive ferrite may reside in the piece.

6. dissimilar metal welds to lower chrome alloys- the huge disparity in allowable stress at elevated temperatures between P91 and P22 or 1.5 Cr- 1Mo- V forgings implies a 2:1 change in required wall thickness. This thickness transiton can lead to unacceptale fatigue stresses during fast startups or shutdowns or unacceptable creep strain redistirbution in steady state unless the detailed geomtry of teh part is designed to accomodate these issues. Generally, aP91 transition piece and / or a rate of change of thickness not greater than 0.33:1 should be respected in the weld zone.

 

[MJCronin]

Troubles.... with P91...!!??? No way....Naaaaaaaaa....!!!

http://www.steamforum.com/steamforum_display.asp?post=986

http://www.eng-tips.com/viewthread.cfm?qid=125940&page=3

http://www.psimedia.info/Industry Alert.pdf

Who has something more recent...???

Have there been any HRSGs coming apart as a result of Type IV cracking ?

Whhhheeeeee.......!!!!!

-MJC

 

[NCTHAI]

Thank you all for their responses!!

Great information!!

Regards,

NC

 

[unclesyd]

Here is a good paper by Walter E. Sperko of Sperko Engineering on the welding of P91 material. He has several very good articles on his website.

http://www.sperkoengineering.com/html/Grade 91 R 10-05.pdf

 

[stanweld]

In addition to the above, WP91 fittings are often mis-heat treated by the manufacturers. We've seen numerous fittings, which were annealed with 2 1/4-Cr-1Mo fittings or otherwise slow cooled after austenitizing.

P91 pipe has also been mis-heat treated by manufacturers; evidence of over tempering has been observed, also slow cooling during normalizing (ferrite microstructure seen at ID through 1/2 wall thickness, tempered martensite with hardness of 210 HV10 at OD surface to about 1/3 T. I am also concerned with very short hold times; e.g., 12 to 15 minutes for 5/8" wall thickness at the normalizing temperature shown on some Japanese manufactured pipe MTRs.