If you are going to 3D print them yourself, then printing complete spheres might make more sense. Most 3D printer plastics are much less flexible than poly, I think, so squishing might result in cracking.
Measure the inside dimensions of your tank and work out a packing scheme that maximises ‘snugness’. If the inside of the tank is exactly 1000mm(w) x 1000mm(l) x 1200mm(h), for example, then balls exactly 200mm in outer diameter would probably fit through your hatch and could be packed in 6 layers, each with a 5x5 array of baffle balls. More balls in total (120), but no assembly, no chance of falling apart, and perfect packing.
[Edit: Actually, alternating 5x5 and 4x4 layers would work better. See comment further down.]
You don’t need specific “baffle ball” models, either. Sure, a bit of engineering can be used to optimise resistance to fluid flow, but you could literally take a model of a soccer ball, rip out one-third of the faces, and use that.
Since your use case is “domestic consumption” your extraction rate is likely to be quite low, so you don’t need baffle balls that offer maximal resistance to lateral flows (sloshing) and, at the same time, minimal resistance to vertical flows (draining). Any roughly spherical shape would do. Heck, it doesn’t even need to be spherical — as long as it packs well.
Finally, if your shape needs a little bit of internal scaffolding, then you can probably just leave it in place. Internal scaffolding will strengthen the walls of the balls, and offer a little bit extra resistance to flow — neither of which are a bad thing.
PS: Make sure that whatever material you plan to print with won’t break down in water and cause health issues.
