[mardifoufs]

From my limited knowledge[0], it does not seem like this would be an easy structure to scan for inspection. Acrylic can even used for the wedges that interface between the phased array (or single element) probe [1] because it's very good at not interfering with ultrasounds. Water is also a very good "coupling" material, which makes it even harder to get a useful scan since there would be no "backwall" reflecting back the ultrasounds on the other side of the acrylic.

There are other ways to scan for defects, but I'm mostly familiar with ultrasound phased arrays. Other NDT methods are also hard to use in the field after manufacturing.

[0] I work in NDT (non destructive testing), but I'm not a physics engineer and (work further up the stack from the actual probes/scanners). So this is an approximation, and I might be wrong.

[1] This random link is pretty useful to get an idea I think, but I'll also ask around our inhouse scientists if I get the occasion today!

https://www.olympus-ims.com/en/ndt-tutorials/transducers/pha...

[bombcar]

So it sounds like you'd need something on both sides of the "glass" kind of like this aquarium magnet but much larger/complicated: https://www.amazon.com/JRing-Aquarium-Cleaner-Aquariums-Clea... and slowly move it around the whole tank, repeatedly scanning as it goes.

So I'm going to put it in the "theoretically possible but nobody will do it" category for now. ;)

[jacquesm]

It's a much harder problem than that. Even if you did not find the crack just before taking the structure into use a crack could develop at any point in time after that and you're definitely not going to be emptying this tank on a regular basis for inspection.

So you either detect it before commissioning or it will eventually fail if there is a flaw.

[chmod775]

They had divers regularly going inside to clean it anyways. Doesn't seem like such an unlikely thing.