Yes Donk, I have dealt with this problem many times.
Trying to trace down the source original source of the leak is difficult.
Many times it seemed that the primary cause for the leak would be either porosity or in complete fusion in the interior tank boundry weld, which would allow fluid to travel inbetween joined plates that were welded with fillet welds on each side of the plate. Then, of course, if the fluid finds any type of opening in the outer boundry weld, it will come through. Sometimes the fluid may travel great distances in this manner. Additionally I've found that Diesel fuel will leak through holes that are too small for water to escape from. Maybe it's because diesel is thinner, or perhaps the diesel can dissolve solids such as trapped slag that water does not. For this reason, I always do a very close bubble test inspection with air on a diesel tank, prior to hydrotesting with water.
Having to drain, clean, and ventilate diesel form a leaking tank sure is a time consuming chore...
It seems that the most probable place for a leak to originate from is a corner where two bulkheads and a deck plate intersect. Often, a Welder will start his bead in the corner and then move out. This has a bad tendency to create a difficult to detect incomplete fusion defect right in the corner where the three separate welds did not fuse properly, mostly due to a cold start and trapped slag. Therefore, I will not pass an interior tank inspection unless all of the corners have been "double wrapped". By "double wrapped" I mean that the corner is reinforced with two separtate welds, one above the other, which start about 1 1/2" away from the corner on one plate, and continue through the corner to about 1 1/2" along the other plate. This helps alot. Also I always instruct Welders to never start a bead in a corner, meaning that they should start the bead a couple inches away from the corner and continue through. Also, it's a good idea to place the vertical weld first, and then cap over it in the corner with the flat weld.
As for porosity in tank perimeter welds, either inside or outside, even though the Code may allow some porosity before it is considered a defect, porosity cannot be allowed on tank boundary welds. The potential for a leak either right away or one that develops months away is too great.
Performing a tank inspection using the 2 psi holding time period method only, on any tank larger than 50 gallons, is asking for trouble. This method is only good for finding large leaks and will not help with pinhole leaks. Additionally any differences in temperatue of the tank or it's surroundings, such as ther sun going down or rising, will affect the pressure reading. Personnaly, I only use this method while hydrotesting piping systems, and then only to reinforce a thorough visual inspection.
Your Level 3 sounds like he is good buddies with the Production people....
Good Luck!
Tim