I have to eat a bit of crow here. In the initial comment, I was assuming (you know what happens whe you assume) a pressurized pipe carrying process water. For a high-strength weld, you'd want to back-gas it, so water... probably isn't your friend.
A sewer pipe has essentially no pressure load to fatigue the weld, so based on the comments above you're probably fine as long as you understand the loading and the are able to quantify the degraded strength of the weld.
Not exactly the same situation but when I had a leak in my gas tank, they filled it with water prior to welding. But the gas tank is made of sheet metal and any water pressure buildup was vented through the open fill tube. Interesting process, though.
Welding on the water filled pipe that is 36 inches in diameter and has a 1/2 inch wall thickness can be performed; however the type of base metal is going to influence the mechanical properties of the weld and adjacent heat affected zone.
If the pipe is low carbon steel the results should be acceptable, but as the carbon equivalence increases, the hardness of the heat affected zone will increase from the rapid absorption of heat by the water which will result in increased amounts of martensite.
You can minimize the amount of untempered martensite in the HAZ by using a butter layer technique. The technique employs a layer of low hydrogen weld deposit on the 36 inch pipe as the first step. The actual weld deposited to join the members is deposited in a manner that will temper the HAZ of the initial weld beads deposited, i.e., the butter layer. It is imperative that the last weld beads deposited do not impinge directly on the wall of the 36 inch pipe to prevent the formation of untempered martensite once again in the 1/2 thick base metal.
In the natural gas pipeline industry, it's fairly common to weld branch connections to live lines while flowing gas. The joint configuration we use is in API-1104. The electrodes normally used for live line welding are E-7018. Take a look at all of this then consult a welding engineer with all the variabls of your project.
Olefazz is exactly correct. Welding on a wet line is regularly done. If the fluid is flammable, assure adequate flow -- API-570 and 1104 will give guadance on that.
You only have [filthy] water. Just ensure that pressure cannot accumulate.
Then conduct a Ultrasound survey of the proposed weld location to ensure sound metal thick enough to be weldable. I use 3/16" for a minimum.
Weld using E7018 electrodes ONLY. The celoustic rods that Eddycurrentguy warned against are 6010 and a few others. Use 7018 only. Stay with 3/32" diameter rods to control the heat input, even though it takes a little longer to finish.
That's it! Again, this is well-tested, and documented in API.
Other suggestions include using E8018-C2 electrodes which have good impact properties. These electrodes operate really well, have a low affinity for hydrogen, and the final weld, if properly applied via a qualified welding procedure, will yield positive results.
The butering technique suggested by GTAW is a great one...can be labor intensive, but, if properly applied, is well worth the efforts.
For in-service welding, I suggest the wall thickness be 0.250" minimum....thats the standard that several pipeline companies have used in the past, althought the material grade, OD, and flow conditions will need to be factored as well...
