This is a really interesting concept. It would be great if you could pull it off.
Just thinking through the potential points of failure:
1) main panels - bottom, back, sides, front that holds glass/acrylic
Your suggestion is to make these light weight for portability. Seems like this should be possible. From another post on fiberglassing I was surfing the West Systems site and found this video -
http://www.westsystem.com/ss/building-a-carbon-fiber-skimboard If you are going for ultimate strength and low weight you could consider carbon fiber or kevlar over foam. I was amazed they could just cover this 1/2" foam and it would now support a persons weight - certainly a wall of water has more force than a persons body, but this shows how you can increase the strength. I would think using very thin plywood (1/4") as a stiff lightweight cost effective base and then layering with fiberglass or carbon fiber/epoxy would be a good way to go.
2) Seams - connecting points for all panels
This is the biggest challenge in any build and tends to get much attention. To me - this would make or break (every pun intended) this project. I think you need to get creative here and prebuild corners into the panels. They make pipe out of fiberglass (Fiber Reinforced Plastic = FRP) and a common way they fit these together is through the use of flanges & gaskets. Basically as they built the sections of pipe they flair the ends out to a flat reinforced perpendicular surface with bold holes. They align the ends of two pipe sections with a gasket between and torque down all the bolts to seal the joint (google "FRP flange" to see pictures). Generally - you don't see too many seams/joints in Fiberglass (FRP) tanks and when they use them they are body flanges away from "corners". So if you built your panels to incorporate the corners (extend around the corner) and then had a flange type connecting joint in the straight side, this might work. It could be ugly though and you would really need to think about where to locate the seams. I have seen lots of FRP tanks in chemical industrial use - some over 40 ft high made completely of FRP. None of these tanks have sealing joints at a corner - they flare all joints to flat flanges and use gaskets and bolts to seal these surfaces.
3) Sealing to glass/acrylic viewing panel
This will be a challenge as well. The viewing panel tends to be very heavy and so you would not normally want to have any extra weight on it. You could permanently seal the viewing panel into the front panel which would make this panel heavier, but might lesson worries about the joint between the viewing panel and the wall panel leaking. You could also similarly use flange type connections and a gasket to seal the viewing panel into the front wall section. Many concerns with this - appearance, bolts exposed to water in tank, etc. You would need to modify the flange design for this where you would embed the bolt into the wall (likely need a metal reinforcing strip to hold bolt inside fiberglass panel) protruding into the tank around the outside of the viewing panel. You could then have a fiberglass backer plate with matching bolt pattern (likely need a steel plate inside) to press against the glass - this would hold the glass in place and apply force to compress the gasket between the glass and the front wall panel. I have seen this used to attach glass windows to a tank (sight glass).
4) Gaskets for joints
This is another tricky part. You don't want seams in your gaskets as the seam will leak water. You can use silicone as your gasket material, but this makes it very difficult to disassemble. Honestly, I can't think of how you would use gaskets in this design - I think you would be better off just using silicone. To maximize the strength of the joints, you want to have more surface area that the silicone can adhere to. If you can make this strong enough, you wouldn't need to have as large of bolts to hold the panels together as the silicone both seals and holds things together (like an all glass tank).
In a plant I worked in we had FRP boxes that were part of the process. These boxes were arranged vertically - 2' wide, 2' long, and 8' high. We had to be able to remove the long side to maintain the stuff inside so this was attached with a gasket and bolts as I described. They were not under pressure other than the weight of the liquid inside - they ran completely full of liquid. So this can be done - not sure how practical this all is?
I hope this helps...
Patrick
