Porosity in aluminum is always hydrogen..... No matter if GMAW, GTAW, or any other process.... Always hydrogen
How the hydrogen gets into (or more accurately doesn't get out of" the weld is the big mystery we each must solve.
Oxides can also sometimes be misconstrued as porosity on RT... especially when the highest standards for defect size are in place,,,, But it really doesn't matter as best practices will eliminate both oxides and hydrogen from the finished weld.
I very strongly but respectfully disagree with Mr. Hughs... Preheat should be avoided in all but the most special cases in aluminum welding...
Molten Aluminum is "thirsty" for hydrogen and will drink it up any way it can (laymens terms here) Meaning the longer the molten pool is exposed to any kind of atmosphere, the more hydrogen the weld drinks up.... When the weld cools the hydrogen is forced out, unless there is so much of it that the pool freezes before the hydrogen escapes.
So with GTAW a slow travel speed often produces a weld that has a "grainy" bumpy surface on the crown and especially at terminations... That is the little hydrogen bubbles that were racing to the top but didn't make it.... Those little bits of roughness can be through the entire thickness of the weld if the operator is going too slowly. In the 21st century, welding power supplies are available that can bring heat to the weld zone fast, so the operator does not have to "wait" for the base metal to heat up.... If you cannot do this, you have the wrong equipment to do work that has this kind of high inspection criteria.
As Al mentioned, there can be other contaminations that result in indications that appear like poroisity,,,, Aluminum oxides from abrasives, hydrocarbons in solvents, paints etc. Castings that have not had adequate prep will offer indications like this as well, especially if the surfaces are not dealt with.
The Cliff notes: Clean HOT FAST
No matter which process you choose... Those three things are the key to excellent RT results in an aluminum weld.
We have not even scratched the surface of how preheat can expand the degraded performance zone in heat treatable aluminum alloys and the very real possibility that an acceptable RT on an aluminum weld that was produced with too much heat input and preheat can have only 50% of the strength that was designed for the joint and proven in the PQR. If there was no preheat in the PQR (meaning a real engineer did not demand preheat) don't do it on a weldment that has a high standard for inspection criteria or performance.