Davefitz - Creep Information 2

[metengr]

Davefitz;
I have reviewed the recent conference proceedings and offer the following technical papers concerning the situation of your P91 induction pipe bends.

The first technical paper of interest provides an in-depth explanation of the background and creep behavior of Grade 91 steel. The title of the paper is
"Long term Creep Behavior and Microstructural Evolution of ASTM Grade 91 Steel" . The authors are S. Caminada and G. Cumino they are affiliated with TenarisDalmine, in Dalmine, Italy.

The title of the second paper is;
"Creep Strength Evaluation of Serviced and Rejuvenated T91 Using the Stress Relaxation Method" . The authors are D. Woodford and R. Swindeman, and are affiliated with Oak Ridge National laboratory.

The Oak Ridge paper was indeed interesting regarding the stress relaxation test (SRT) method for evaluating creep behavior in Grade 91 steel. Specifically, the SRT method can simulate five (5) decades of creep exposure in one 20 hour test. The specimens are loaded up to 1% creep stain and are held at this strain level while the material begins to relax and elastic strain is replaced by plastic strain. They have provided SRT data that shows good correlation with conventional constant displacement rate creep testing of T91 material.

Now for your situation, since you have confirmed the presence of ferrite in the matrix of tempered martensite, it may be worth while to consider the SRT tests to evaluate the creep behavior of this microstructure. My recommendation is to obtain some Grade 91 plate and simulate the induction bending heat treatment to obtain a similar microstructure.


Once this is completed, I would have SRT creep tests conducted to compare the creep and stress rupture behavior of this mixed structure with the creep behavior of N&T Grade 91. The SRT is performed using standard creep specimens.

What I like about the SRT is the method is logical and provides creep behavior that is representative of five decades of creep exposure seen with conventional creep testing. With the relatively quick turn-around in creep behavior data you can discuss the results with your clients to decide on a periodic inspection program to monitor creep damage or to optimize replacement of the elbows.

 

[davefitz]

Thnaks Meteng.

If we find the SRT method to be applicable, we will probably remove and replace the worst bed ( based on hardness survey and replication) and test it .

Out of curiousity, do the referenced papers discuss the case of creep testing of ferrite laced P91?

 

[metengr]

No. I was specifically looking for this information. I thought that the key might be creep testing that has been published for Type IV, intercritical HAZ (ICHAZ) failures in pipe or tube butt welds, but I could not find anything specific related to ferrite/martensite microstructures. I believe what you propose is the best approach, and is something you can defend. You will be blazing the way with your testing. Good Luck.

 

[DSB123]

davefitz,
I enjoy reading your thoughts/information on P91 material. We have just installed a P91 piping system on our Plant and I am concerned on several matters;
1. Whether the PWHT is correct.
2. The use of high allowable stresses.

However the worst situationis the fact that some of the piping has been highly stressed during the first couple of months (due to incorrect installation)with calculated stresses well above the Code allowables (using elastic analysis) where I beleive permanent damage could have resulted. The effect on the long term life of the piping due to this overstress for hundreds of hours operation.

Reading all of the articles around and many of the papers on P91 means it must be treat with the deepest respect as I see it.

 

[davefitz]

DSB123:

I agree with your assessment. I finally found a copy of the CD of the conference proceedings that Meteng is referring to, and there are several authors that have claimed that the initial service pipeline stresses are highly damaging to P91 weldments, espescially if cold springing was not utilized and if the operating temp is below 1050 F, because the P91 will not sufficiently creep relieve pipeline stresses at low operating temps.

In the last year we have experienced several "failures" of large P91 piping elements, including :

a) a P91 HP main steam pipe to 1.25 Cr 1 Mo V turbine stop valve through wall crack , at 3 duplicate stations. The root cause was , basically, technical incompetence of the EPC vendor incorrectly ommitting the required "transition piece", after only 1 yr service

b) 8 Hrsg to steam turbine main steam and hot reheat steam lines which have incorrectly fabricated hot bends ( foundry neglected to N+T following hot bending with local temps to 2000F ), resulting in soft zones ( Hb < 180 )and ferrite bands , with an unknown creep life of perhaps 150 elbows.

c) during a review of the fab procedures for the same foundry , discovered that they had never N+T'd any hot bent P91 elbow in the last 10 yrs, and when our spec required a N+T, they cooled it at the incorrect rate ( must be cooled faster than 9 F /min) the resulting component was all ferrite, and it requiered a second N+T to provide the correct tempered martensite crystal structure.

d) a review of the field weld procedure to be used on a new unit revealed that the weld procedures that had been used in the past by 2 different EPC contractors had neglected to use "low hydrogen" protocols for such P91 welding ( too many erros to list here)

Rather than blather on about every little fault, it seems the Code leaves the responsibility for proper welding and fabircatkion proceudres on the "designer", which is normally an EPC contractor. The foundry is not responsible for incorrect procedures. Of the 3 EPC vendors I am familiar with , none of them employ a staff metallurgist nor contact with one prior to specifying the P91 welds and fabrication techniques. Basically, there are a lot of incorrectly fabricated P91 pipelines out there.

 

[hacksaw]

Part of the problem is that the EPC is what it states: "Engineering Procurement Contractor". Typically it is unsupervised procurement with engineering is just thrown in as window dressing. All to common and more is to come.

 

[unclesyd]

metengr,

This might be old hat but might be worth a look.

http://www.twi.co.uk/j32k/unprotected/band_1/hightemp_index.html

 

[athomas236]

davefitz

Thanks for your useful information.

I always found that code writing bodies answered the easy questions in which case who needs them but they always left the rest to experts so who are they and why are they there

athomas236