Force = pressure times area. Not much hydraulics knowledge required.

Note that flat plates tend to bend, turning the force from a pure tensile pressure force in the glue to a peel strength problem. Many adhesives have low peel strength.

2000 psi * 144 square inches ~ 288,000 pounds force. If your glue is only capable of 1900 psi you need at a minimum more than the area exposed to water - and about 3X-5X for a safety factor.

While such items don't typically rupture explosively, small leaks will provide potentially lethal jets of liquid that can certainly cause injection injuries and slicing injuries.

to carry on from above,
288,000 lbs / perimeter = 288,000 lbs / (4 * 12 in) = 600 lb/in along the edge (ignoring stress concentrations at the corners, etc.)
assuming the border plates overlap the Al plates on each side by 0.5 inch,
shear stress on bond = (very approximately) 600 lb/in / 0.5 inch = 1200 psi.
put a factor of safety of at least 5 because you have a pressure vessel and the shear stress = 6000 psi.
then account for the fact that the plates will likely bend some under pressure, creating more stress concentrations,
so in conclusion, the adhesive joint is no where near adequate.

PS chatGPT is completely stupid

Are there important electrical (or presentational) reasons why the capacitor elements need to be flat squares with those dimensions? If not:

If you could make the plates the same area but circular instead of square, you could eliminate the corners on your border pieces (which will otherwise be a weak point).

If you were able to reduce the thickness of the water layer (if you had dielectric strength to spare), you could achieve a corresponding reduction in required surface area for the same capacitance with similar reductions in the end loads and in the forces on your edge joint.

If you were able to reduce the thickness of your aluminium plates, then you could reduce the thickness of your assembly, making it much easier to replace the fabricated border pieces with a single piece of material. Are you expecting the plates to have to withstand significant electrostatic forces? You didn't mention any structure for keeping them apart.

How important is it for the capacitor to be flat plate? As Dave mentioned, flat walls are difficult to keep flat when you pressurise them. Could there be scope for a coaxial design instead (tubular electrode with a rod running down the middle)? This could then fit inside a cylinder - a much easier shape to make pressure-retaining.

2000 psi is well into the territory where working out how to contain it needs to be part of the initial design process, rather than a detail to be sorted out at the end.

A.

Thank you all for your replies. Based on your words I think I am going to scrap the water capacitor idea.

In further reading of the DoD paper on water capacitors in my attempt to see if I should just not pressurize it, it indicated that gas bubbles was a problem with or without pressurization. They said the dielectric strength would almost double with 2000 psi applying pressure to the water versus no pressure. But it was only good for single shot applications and it would take time for the water dielectric to heal itself, for the gases to recombine or pumped out of the system.

I think I am going to try building a capacitor with PVDF instead. That has its own problems in bonding due to its low surface energy and the need for a bleeder resistor but I am hoping mixing powdered PVDF resin or using a coffee grinder on PVDF pellets and mixing it with DMF solvent will allow me to use it as a potting compound to pot any air gaps in the PVDF/aluminum capacitor rather than use epoxy which wouldn't work well.