I'll preface this with a few things:
1) I am not a structural engineer, but I am an engineer, PE, but electrical
2) I have a pretty fair amount of acrylic experience, but not with anything close to this size
3) I have gathered a lot of knowledge about acrylic fabrication
4) I would classify myself as highly knowledgeable in the aquarium building world, but I hesitate to label myself an "expert". So I am totally open to being called wrong.
With that said...
Looking at that massive rip at the seam would make me think about additional support against the outward water pressure.
This has significantly less to do with it than does the vertical height of the tank. Water pressure is a function of depth, not length, of the water column behind it. If that was the case, one would expect the failure to occur closest to the right side of the joint (with 16' of perpendicular water column behind it)
One could have a 200' long tank and the water pressure at the bottom is no more than a 2' long tank. There are other factors that come into play when dealing with
widths, but not
perpendicular to the panel/joint.
If anything, the width of the tank has more to do with it, but then mainly with respect to bowing. The longer a panel is, the more it will bow (all other things equal). Bowing also is a function of the thickness of the bottom panel and the top euro and configuration of the bracing. A wider perimeter brace and wider and higher # of crossbraces will contribute to the tanks propensity to bow (more or less bowing) over time.
Bowing is also a factor of time (and material thickness). Generally speaking, after 16 weeks or so of water in the tank, it's going to have bowed as much as it ever will as acrylic is part water and therefore will absorb a very small percentage of water over time.
This is also why you do not want to put water into a freshly bonded tank for any longer than maybe an intial fill test, because the freshly bonded joint is more porous than the material, and will absorb water at a (fractionally) higher rate. This can lead to whiting of the joints over time and may lead to premature failure.
While it sounds quite ridiculous, a good rule of thumb is that for every 1/4" of material thickness, wait 1 week before filling. This makes for a not so good business model as one would have to tell their customers to wait 2 months after their 2" thick walled tank is complete to fill it. But that's what you really should do. This is obviously not the case with this tank I might add, plenty of dry time.
The front panel that failed is only 1" and it has 16' of water pushing against it. We didn't have any type of wave device that would add undue stress.
I'm not an acrylic bonding expert, but it appears that the seam was faulty as there was no physical incident with the are.
CONFIGURATION
Display Tank
The acrylic tank is L shaped with the following dimensions: Length is 8 feet plus 16 feet: Width is 3 feet: and Height is 2 and one half feet. (24'X3'X2.5') I'm not sure what the metric equivalent would be. The tank is at least 1" thick all round except the 16 ft panel which is 1.5 "œ.
The tank is accompanied by a custom tubular steel stand fitted to the tank by ATM.
For a 30" tall tank, 3/4" is usually the recommended minimum. So you are double that for the 16' panel and still over that for the 8' panel.
The factor that I cannot knowledgeable shed light on here, but I think is worth considering, is the shape of the tank. The L shaped tank has some properties that are a bit beyond my reach, but I know who to ask and I might. So here is where some educated guessing comes into play so take this with a grain of sea salt.
If you look at the 2end point of the panel failure, you will see something that supports my guess here
Do you see it?
If you extend the panel on the other side of the 16' panel so that it intersects with the end panel - that's where the failure cracked the panel and went up.
So my thoughts here are that the additional (special) stresses involved inherent to an L shaped tank maybe have somehow focused a bit higher, somewhere near the middle of the 8' panel. Keep in mind - just an educated guess.
Now normally, if the tank was built to the minimum, this might cause an issue. Overbuilding the tank, I would think, would alleviate that stress.
Also, I know that ATM uses Polycast so we are talking about the best material, and actual 1" material not the "metric equivalent" that it would be if it were Plexiglas/Arkema or Acrylite/Cyro.
In addition to that, I mentioned earlier about solvent vs 2-part (40) and it was commented that ATM does both. I've seen this too, especially one long tank with a big coral insert where they solvent welded the bottom on after the insert was placed on it and the rest of the tank lowered over it. It also occurred to me that it is likely that all bottom joints are more easily and likely solvent welded, because 1) the only joints most people care about being glass-clear are the vertical ones 2) a 2-part joint requires a space between the panels to fill the joint (either a fixed gap or a beveled edge) which would be difficult to do on a top/bottom joint. But that doesn't mean it's not done and I'm just speaking from common sense here, not direct knowledge or personal experience. If the joint looks flawless and clear, it's 40. If you can see what I would compare to a "thermocline" when diving, it's definitely not 40 - it's solvent. That's how you can tell.
Now I've seen tanks with bottom seam separation when they showed up. One is in a restaurant in town owned by a friend where I got called to look at a scratch. I wasn't concerned about the scratch. But what can happen is that when a builder sets a joint and doesn't let it cure long enough before routing, handling, shipping, etc is that the joint can cure up poorly. Again, we're on this issue of # of days before even touching the build being related to the thickness of the material as thicker material requires longer. And the result may not show up for a long time - a bad section of a joint would be totally imperceptible until it failed.
Again, I'm not saying that is what happened - just a possibility. I'm typing out loud. Because I don't want to work right now.
If you do opt to remove/replace the existing stand I would be curious to see if the floor is still level. The house itself was still fairly new when this tank went in and I'm curious if the foundation has settled a bit during the intervening years leading up to this terrible disaster.
Dave.M
Certainly a possibility also...
The other possibility is that there was a material defect. I just remembered that in the tank at our LFS that blew, the insurance company did not call that a manufacturing defect - they were somehow able to determine that it was in fact a material defect in the bottom panel. In that case, the point of failure was in the bottom panel, near the back edge. The failure actually ripped through the joint, and the joint was largely in tact.
Peter's tank however clearly failed at the joint. This seems to point to a manufacturing defect at the outset, but that's not to say that there wasn't actually some kind of material defect near the joint that started the whole process. Unless you had a camera on it, you'll likely never know.
But I have investigated tank failures myself and I've been able to directly pinpoint the cause. One was when someone decided to use a 70" and 2" panel and cut them at 45s and weld them together instead of using a solid 72" piece for the bottom of a sump. They used a ton of 40 on the bottom joint but the panel was not flush at the bond point (there was a ridge) and the result was that the vertical panel failed at the point of the bottom panel butt-joint. It happened within hours of a test fill.
It also could have been a combination of many factors: L shape stresses, panel thickness, joint quality, settling, material defect - any combination of a few of these could have piled up and just been too much.
I will say this: the tank appears to have been built well from a material thickness standpoint. I'll have to dig for euro pics (may be a factor). They didn't skimp on the bottom panel and I'm guessing not the euro either (I've not seen ATM cut a corner on these like some others do).
I'll show off my humble 93 cube (if its up in time).
