sounds like you are on the right track with removing the dissimilar metal ,make everything the same if possible , you probably could get away with just the pad in 800
 dont try to PWHT unless you have a good reason, its not a straight forward process with this metal
check for local high stress, maybe the agitator is setting up reaction stresses and you are getting fatigue cracking

Is this vessel insulated with calcium silicate insulation by any chance?

The APIGUY....

No... fiberglass

A couple of questions:

1) what weld filler metal are you using?
2) what kind of temperatures are you seeing?  (are they cyclic?)

Also, I agree with Aybee and your approach of making it all similar material.  But, it would be interesting to note the other environmental variables...

~NiM

Don't know the filler metal since the vessel was built in 1995... it's a P1-P45 joint (cs-alloy 800).  Inject live 200 psig steam so the vessel heats to around 350F.  Not sure about cycling.  Vessel is at a remote location and don't have all the operating history.

I think I'll go back with what was already stated.  The best thing is to get rid of the dissimilar metals.

The comments from the others are right on.  And remember, without any of us knowing a lot more about your situation, none of us can make an accurate judgement.  (Referring to Aybee's comments that the weld may be subject to fatigue cracking.)

There are a lot of other things to consider when doing the replacement, including asking yourself if Alloy 800 is the right material for the job.  (As I'm sure you are aware.) When I'm specifying new equipment replacements (due to failures) I want to know EVERYTHING.  This includes the total environment within and outside, any forensic metallurgical evidence of failure, potential failure points, potential damage mechanisms, upset conditions, etc.

Good luck!

~NiM