Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations IDS on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

91 headers seamless/welded 2

Status
Not open for further replies.

GBENARD

Specifier/Regulator
Dec 12, 2003
13
We have some problem. ASME code seems to allow the use of welded headers (A387 91) as well as seamless pipe (A335 P91) for the manufacturing of headers for boilers. Even if it seems, from some studies, that we can expect a more significant creep-fatigue interaction for 91 in HAZ than from 22 material, as well as some reduce impact strength for welded pipe in from of seamless, we have no definite statement for the advantage of the use of seamless headers against welded pipe headers. I should add that this is concerning HRSG designed for a baseload service, something like 20 cold start up per year for a 25 years design life, the headers working at 565°C/130 and 30 barg.

Is there anybody knowing of a serious study on that problem or having experience on that matter (in that case, for how long because my concern is the long term)?
 
Replies continue below

Recommended for you

Here are three websites that may have information. I’ve seen a paper, I believe from EPRI, that went into a detail discussion about the use of different materials, pros and cons. I check and see if I have a reference to it.
If you decide on welded pipe just make sure it's welded as I've heard a little about some poor quality pipe floating around.



 
The issue related to seamless header versus a seam welded header is complex in the Power industry. For P22 header material, we know from years of expierence (40+) and post service creep testing of seam welded pipe/header material that the seam welded components do not last as long as seamless components because of creep damage either confined to the fusion zone of the weld OR creep damage confined to the base metal HAZ (Type IV damage). Studies have shown all weld metal creep rupture strengths and all base metal creep rupture strengths are comparable given compatible chemistry. However, the life limiting factor is the weld fusion zone or base metal HAZ.

I have seen seam welds that were either subcritically post weld heat treated that developed Type IV creep damage in the fine grained HAZ of the header base material, and I have seen creep cavitation damage in the weld fusion zone related to clusters of non-metallic inclusions from the use of submerged arc welding or flux core arc welding processes to produce thick-walled header seam welds. This is not to say that the Code was wrong or was in error. The use of Code approved materials and design are based on lab testing AND expierence. Most of the boiler superheater and reheater seam welded headers that I have dealt with have been in service well over 30 years without serious incident. At some point however, the remaining life will be dictated by the weld fusion zone or base metal HAZ for these components.

Switching over to P91 material. You basically have the same issues where the creep life of the header will be limited by the weld fusion zone or if subcritically post weld heat treated - the fine grained HAZ. To assure long creep rupture life, seam welded P91 headers should be specified as normalized and tempered to assure complete removal of the weld HAZ. Field weld repairs of P91 components (welded or seamless)have to be carefully performed to avoid under or over tempering the fine dispersion of carbides in this material.

As a side note, any time we replace boiler headers or external boiler piping, we specify seamless material to avoid the future expense of in-service monitoring of seam welds to check for creep damage - plain and simple.
 
I would agree with metengr. Unless you want to inspect all your long-seam welds on a regular basis, you're better off using seamless. Also, any field weld made on P91 material will create an unavoidable zone of weakness adjacent to the weld where the martensite has been altered by the heat of welding. The magnitude of this weakness is controllable to some extent, but you will always have some degradation of the martensite microstructure what will be the point where creep damage will form first.
 
I think the section I code requires that a 0.60 weld strength reduction factor is to be applied for a longitudinally welded P91 header. This essentially counters the approx 30% reduction in creep life caused by the soft section of the HAZ.

The 30% reduction in creep strength does not directly affect the wall thickness of girth welded P91 welds, as long as you do not locate the girth weld in a region of "multiaxial stress", normally considered to be high bending moments ( from piping flexibility ) or high shear stresses caused by locating the weld within 1.5 wall thickness away from a major change in wall thickness , such as at a steam turbine stop valve. This is because the axial membrane stress on a girth weld is only 50% of the circumferential stress of a longitudinal weld.

There are available seemless pipes up to 48" OD in P91; only need for longitudinally welded P91 might be hot reheat transfer pipes for large fossil fired Rankine units over 500 MWe
 
Davefitz;
Not sure I agree with your statement regarding the affects of using a 60% strength reduction factor of the seam weld to offset the 30% reduced creep stength of the HAZ. It is not that simple. The only benefit derived using a stength reduction factor is to increase the wall thickness of the pressure retaining item which reduces service stress. You still have metallurgical issues related to creep strain rate differences between the weld metal and base metal that will ultimately limit the creep rupture life of the seamed component.
 
meteng:
first , I need to make a correction: the wording should be creep strength and not creep life.

second, there is general consensus ( and sect I recognition) that the creep strength is decreased in the HAZ by about 30% at typical service temperatures near 1050 F for P91. The 30% reduction is directly measurable from elevated temp creep tests, and already includes the effects of the interaction of the stiffer coarse carbide zone vs softer zone, provided one is only addressing steady membrane stresses.

In the case of more complex stressses,such as local to a partial weld repair or close to a sharp change in wall thickness or in locations with high local stress raisers, it is true that the interaction of the weak vs stiff zones can further lower creep life, and some estimates can be had using nonlinear finite element analysis. But to date such nonlinear analyses are only fodder for PhD theses and are not yet commonly used in designing weld joints.

For a girth weld at least 1.5 wall thickness away from a thickness change and not subjected to external bending moments , the weaker HAZ can provide normal service life ( assuming all code approved weld details are followed, proper preheat, interpass temp below 300 C,control heat input so intercritical temp never exceds 775F, exact PWHT temp and hold times, hardness traverse check after PWHT). But I must admit such an assumption is not always valid.

For a longitudinal weld, the creep strength reduction factor together with exquisite QC at the root pass and zero ovality should meet normal service life expectations , assuming torsion is not applied at elevated temperatures.
 
One well known Boiler Manufacturer was required by its customer to manufacture a P91 welded header for a boiler in Australia and nearly doubled the wall thickness to assure long term safety. Theoretically, normalizing and tempering after making the long seam should restore most of the creep strength, but long term creep studies were not available to support the theory. A number of attendees at the P91 EPRI conference thought that a significant safety factor would also be required.

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor