P91 hardness testing 2

[tigwired]

Am wondering if anyone can explain why the hardness numbers on a fillet weld of P91 materials would be in the 350 to 400 range(vickers)while the valve is around 200 and the pipe is in the 160 range?
The welding procedure was strictly adhered to including the monitoring if interpass temps, PWHT cooling was at 250F for the cooldown. Filler material was of the 9018-B9 TIG with checks done using bruscato formula.....also PMI of the weld showed it to be P91

 

[0707]

Maybe PWHT holding time was not long enough.


lm

 

[metengr]

tigwired:
Hardness testing of Grade 91 must be performed using a qualified procedure for both surface preparation, and hardness testing itself. Also, the type of portable hardness test is also important. At this point, I would not make any comment unless you can provide more information as to hardness testing details. I have seen many strange results due to inadequate surface preparation, wrong portable hardness tester, etc.

 

[stanweld]

The hardness of deposited weld metal and base materials cannot be equated with normally applied PWHT. If your stated hardness values have been obtained through proper testing methodology, you have a major problem. Of the thousands of hardness tests that we have made, those base materials having hardness of 160 BHN never came up after grinding as much as .250" below the surface. Replicative microscopy of these low hardness zones always revealed a feritic or mixed ferritic microstructure. There is a good chance that the pipe has been improperly heat treated or oxyfuel gas heating has been improperly administered or PWHT was too high. The high hardness of the deposited weld metal is generally indicative of low PWHT temperature or short hold times.

Advise that you make additional checks of pipe hardness out or the area subject to PWHT, assuming local PWHT was performed. You should also perform replicative microscopy to determine apprpriate microstructure. You may also wish to remove a section of pipe for through wall microscopic examination.

As advised by metengr, please provide greater detail as to the hardness test methods employed.

 

[tigwired]

Thank you for your prompt replies.
Here is more info.Hope it is of some help.
The lines were a 2” shd 160, and a 1” shd 160.
PWHT ramp up at 200F per hour, hold for 2 hours and ramp down at 250 F per hour, had the wrap on until it cooled to 200 F .
Microdur was used for hardness test.
The de-carburized zone was removed completely before hardness test was done.
Had two separate nde companies do the test with varying results.
Replication of the existing pipe has not been done. Although replication of other existing pipe showed some course grain structure that appeared to be largely ferretic in nature, with course grain boundary carbides. Random disperesed patches showed a martensitic appearance (again this was not on the particuliar piece that was welded but further up stream of this.)
Persons doing the testing were qualified and the testing was done at 4 places on each weld.
Several welds were done at the same time on each of these lines with 2 out of three being found acceptable on the 2 inch and three out of four acceptable on the 1 inch.
No evidence of signifigant creep damage was noted.
Am beginning to believe as someone in the forum suggested that this is indeed the Devils Material.

 

[metengr]

PWHT ramp up at 200F per hour, hold for 2 hours and ramp down at 250 F per hour, had the wrap on until it cooled to 200 F .

What was the actual PWHT temperature? This is critical because if your hardness testing of the fillet weld is correct, this implies an improper PWHT temperature.

Although replication of other existing pipe showed some course grain structure that appeared to be largely ferretic in nature, with course grain boundary carbides. Random disperesed patches showed a martensitic appearance (again this was not on the particuliar piece that was welded but further up stream of this.)

Your description of the microstructure is not what I would expect for properly heat treated Grade 91.

 

[tigwired]

Metengr
hold time for PWHT was 1400 F for 2 hours
interpass temps did not exceed 600 F.
Agreed that the appearance of the microstructure suggests that it had not been subjected to a proper heat treatment during manufature.

 

[metengr]

tigwired:
One last question is the 1400 deg F for 2 hours the actual recorded PWHT temperature on a chart recorder or is this a value in some WPS or PWHT specification?

 

[tigwired]

Thanks Metengr
Actual holding time as recorded on the chart, also a separate test was performed to verify that PWHT equipment was indeed at this temp. results showed it to be within 1 degree F at this temp.

 

[stanweld]

Apparantly you have determined that the P91 piping in question has been poorly manufactured. Or were the micros taken only near welds? If the microstructures observed were taken in areas away from the welds (not subject to PWHT), consideration of replacement must be given. I would mandate it.

Our industry has seen considerable problems resulting from the use of oxy-fuel gas heating for preheat or straightening creating localized ferritic areas in the microstructure. We had one case where an overzealous, uncontrolled welder heat straigtened some 3/4" shedule 160, P91 pipe; it failed (burst) on steam blow. We have long ago banned the use of oxy-fuel gas torches on P91 systems. There have been numerous reports of premature failure in P91 tubes at bends where such practices were employed.

I would tend to question the high hardness in the socket weld as resulting from improper testing assuming your PWHT was correct. Did you control the %Ni + %Mn in the weld filler metal?





.

 

[tigwired]

Thanks Stanweld
I stand in error on the filler material
It was ER90S-B9.
The Ni and Mn were shown to be < 1.5%
The Mn/S were shown to be > 50
The X factor as calcuated from the Bruscato Formula
X=(10P+5Sb+4Sn+As) x 100
was shown to be 9.3
The microstructure was taken at weld, HAZ and well away from the weld (as stated previously this was done on another section of the run (an elbow) not on the weld in question.)
Preheat/ PWHT was done thru ceramic heaters
am very grateful for your comments and metengr's also
thanks again

 

[metengr]

tigwired
Regarding the excessive hardness of the fillet weld deposit (and provided the hardness is accurate), I believe you may not have cooled to allow for complete austenite transformation after welding. If you have excessive hardness, this implies fresh martensite formation that occurred from retained austenite during cooling after tempering.

 

[tigwired]

Metengr
we cooled the weld to just under 200 F for 45 min after welding sequence and prior to PWHT, the cool down from the preheat of 400 F to 200 F took 1 hour.
am wondering if this was adequate in light of your comment about martinsite formation.
Is there a recommended cooling time and or temperature that would hopefully give us better results?
I guess the thing that throws me is the fact that the other welds all came up with hardness numbers which were acceptable.
I do appreciate your input.

 

[stanweld]

Was the elbow a bend in the small bore pipe or a socket weld fitting? Did the small bore pipe of the weld in question have bends in it?

 

[tigwired]

stanweld
the elbow replications were done on an 8" and 16".
the fillet welds (one inch and two inch)which have the high hardness readings has elbows, valves and risers.
your statement regarding replacement is the same as the engineering dept has come up with. my main concern at this point in time is the abnormal hardness readings and the fact that most are acceptable while a couple are not...seeing all were subjected to the same welding and pwht procedures. I am sure that in house discussions will range from changing out with similiar material, change out with P22 material (depending on weight loading of hangers) or perhaps even a derating scenario.
thanks again for your valueable input.

 

[stanweld]

tigwired,
You are mixing too many variables. The low hardness issues plaguing WP91 fittings have been well documented. Based on the testing on our jobs, we expect to find at least 10% of WP91 fittings to be mismanufactured. Some jobs (not ours)have seen over a third of fittings installed to have been mismanufactured. I assume that if you found at least one fitting that exhibited low hardness and an inappropriate microstructure that you hardness tested all fittings. Do not equate 8" and 16" fittings with the problems encountered on your small bore pipe weld. You also need to determine the extent of low hardness in your small bore pipe.

We had similar problems with high hardness deposited weld metal several years ago regarding socket welds to valves. Essentially it was attributed to a "chimney effect" of the opened valve connected to a long vertical run during PWHT in addition to the larger mass of the valve body. These issues are much more common in welding large bore pipe to valves. In effect the controlling T/C was measuring the temperature of the pipe. It was resolved by placing the T/C on the valve at the weld toe and extending the insulated area.