I understand
I studied the link and the image posted, and the writings of the author. I understand how it works in principle, though I known not the degree of control the user has over the water level between tide transitions. Since a computer can be used to control this, I suspect that with experimentation, the tide could be made to gradually rise over time, and recede likewise by toggling valves on & off repetitiously when they are not entirely on or off. Regardless, here's what it appears to be, and if it is not, here's how I'd build it.
Notice in the left half of the picture identified as the shallow end, a large cavity wherein water may be stored. If this resevoir is full and at the same level as the rest of the tank, we're at low tide.
The drain in the deep end mentioned would be near the bottom of the tank (somewhere on the right), and would allow water to simply re-enter this resevoir freely, thus acheiving low tide. This drain as mentioned is controlled by a solenoid valve.
The second drain into the sump is at a higher level (on the slope) and is also controlled by a solenoid valve. If the lower valve in the deep end were closed, and the second drain were open, the water level would rise to some point beyond the second drain, as the resevoir would become somewhat depleted.
If both valves were closed, the water level would rise further, as the sump is likewise futher depleted, acheiving high tide, where the resevoir is at its lowest level, and the majority of the water is in the 'rest' of the tank.
The resevoir of course must not run dry, so there may be an (unmentioned) overflow in there somewhere to allow some water to re-enter the sump when both valves are closed. If not, I'd add one
It would appear that the hang-on filter is all that is responsible for pumping water out of the sump and back into the rest of the tank.
I commend the designer of such a system due in part to the fact that it is self-contained, so it won't spew all over the floor. Brilliant.
HTH, DHM