SA335-P22 vs P91 2

[rls911]

Dear Members,

I am looking for your thoughts and comments on the following:

We are in the process of constructing a new boiler with steam outlet design conditions of 1750 psig @ 950 - 1000 deg. F. The boiler manufacturer will be supplying the outlet header (boiler external piping) as SA335-P22. Our normal standards for this application would be to continue on with P/F22 for the pipe, fittings, and valves.

But we have been presented a more economical alternative of SA335-P91 (lighter wall thickness + smaller header diameter = less cost). The design would include F91 fittings but would use F22/WC9 for large bore valves and F11 for small bore valves (socket welded valves).

Other than a potential capital cost savings, are there any other benefits to using the P91?

After reading some of the previous posts on P91, I am not so sure we are goiing down the right path. ie.

- disimilar weld joints.
- transition pieces at the valves.
- PWHT everywhere.
- Critical QA/QC of welding process.
- potential code changes on the material specifications.
- lack of clarity/consensus between the piping codes and metallurgists on handling the PWHT.

 

[abbver98]

P22 is completly diffrent to P91. P91 is a ferritic alloy steel with 9 Cr, 1 Mo and same grades with V. You can reduce your P22 existing piping by 60% by using P91. But it requires special care when welding (preheat 390°F, welding max 570°F, slow cooling and PWHT 1350°F). Welding Tests were made with dissimilar material like X20 and P22 that show promising result.
Note that P91 (P-N 5B)requires more care than P22 (P-N 5A). In some case Welding could be performed with P22 only with Preheat and max interpass temperature of 570°F. It clear that welding with P91 is more expensive than with P22.
All parameters had to be taken into account to evaluate the use of P22 or P91.

 

[DSB123]

abbver98,
Suggest your quoted PWHT for P91 is a bit out (1350 F = 732 C) normally PWHT target temp is around 760 Deg C.

rls911,
You seem to have done homework on P91 and have identified potential problems with it. One of the main concerns is the need to take care in welding it if Type IV failures are to be avoided. The preheat, welding and PWHT need to be controlled to give good results. Also the other major concern is even with good control of the welding the welds are weaker than the base material and as such weld stregth reduction factors need to be employed in the design assessment for creep. (Note :- These reduction factors are given, by all accounts, in the long awaited 2004 edition of ASME B31.3.

 

[metengr]

I can tell you from experience that we have used T91 material in our boilers with good success. The current project that I am working on is replacement of a series of SSH outlet headers for two of our supercritical boilers.

The original header material was SA 335 FP22, which ended up having a shortened design life because of the 5.8” wall thickness required to meet supercritical design pressure. After considerable thought, we made a deliberate decision to go with P91 material to reduce wall thickness to 3.1”. The reduced wall thickness will increase resistance to bore hole cracking from thermal fatigue (this is what had caused us to replace the header after a thru-wall leak had developed).

All of the reasons cited in the above posts were considered during design and fabrication. Yes, P91 material must be carefully handled during fabrication and in-service repair. However, we felt that the risk was manageable and we have moved forward with installation.

The terminal tubes that will be welded to the SSH outlet header were safe ended during fabrication of the P91 header, and were supplied as T23 material to provide for improved design.

Type IV cracking is indeed an issue, but is driven by thermal/mechanical bending stresses. We took this into account by carefully re-designing the header support system to reduce bowing and torsional loading on the header during start-up and shutdown conditions.

 

[davefitz]

A few years ago, there was a paper by Combustion Engineering - Alstom on the economic selection of P22 vs P91. The break was at 975F- use )22 below 975 F, based on first cost comparison only .

The Thinner wall P91 not only is lighter ( and has fewere,lighter supports) , but would require fewer flexibility loops due to its smaller momentof inertia.

The main technical reason to use P91 is to reduce teh thermal stress on the thick headers and valves during cold startup. If the sizing of the piping at 1750psig / 1000F leads toa P22 wall thickness of less thatn 2" AW, I would not even consider P91. Likewise, if the plant is base loded and the wall thickness is less than 3" , I would stick with P22. IN any case, if you will be using WC9- F22 valves, then these will have the governing thicknesses anyhow, and nto the headers, so stick with P22.

There is a lot of re-education needed for correctly impletmenting P91, and the investment in education needed probably is not worth it for only one small plant that is on the borderline of using P22. If you do select P91, be sure that the deisgner + fabricator + erector each be totally experieinced and qualified with the correct means of implementing P91- including all the new proposed rules that ASME is now floating.