Not sure it is that simple to auto change water using a sump due to the large changes in water levels encountered between when running and when "failed".
I have seen systems that used an extra section in their tank weir overflow to allow water to overflow into their drain. You have this drain weir slightly higher than you main overflow so most of the water drains into the sump, but some overflows into the drain. Replacement water was added via float valve in the main tank.
Need to also consider how you are dosing de-chlorinator if you are using tap water to auto water replace.
A failure of the float valve means excess water flows into the drain overflow or if really bad overflows into the sump overflow drain.
Failure of downpipe, means sump pump empties the sump pump compartment and overflows tanks into drain, but at least not losing a large volume of working water.
You need to consider what happens when you have a failure, blocked pipes, failed pumps, reduced flow in pipes, blocked filters etc and how it can auto-recover after say a power failure.
Another sump fail I read about, was a guy came home and his carefully designed baffled and weir'ed sump was empty and tank had overflowed making a nice wet mess. He had purchased a nice custom sized acrylic tank for use as sump BUT he had siliconed in the baffles/weirs making what he thought was a wonderful fail safe/auto restarting sump. However silicone does not in fact bond very well to acrylic and after a year odd, the silicone started coming away, eventually allowing the return pump to pump the entire contents of the sump into the tank, causing the overflow.
I'd be using RO, so I'd fill a header tank, auto-dose a remin salt solution via a dosing pump, and then water change with that. I'd have chlorine and chloramine pre-filters so I wouldn't be too worried on using a dechlorinator (its what I do currently as well so I know it works).
What I do currently on both my tanks is a separate plastic water tank fills with water change water (using electronic ball valves or solenoids on the main supply inlet), the fresh water is then pumped or solenoid released (via gravity) into the main tank. Waste waste then overflows out of the tank via a waste over flow pipe as the level in the tank increases. I'm at risk if the overflow pipe blocks, but the chances of that are very slim.
A sump is so more complex though. I guess the key would be to replace smaller quantities of water at a time, and to have a sump with sufficient spare capacity designed in.
The only way I have thought to do it is as follows:
First you have a header tank (higher than the sump), which fills with water change water fed by the RO unit, with its outlet the the sump on a float switch in the sumps final chamber with the float switch set at the working water level - similar to a ATO tank, so working by gravity. This header tank would have it's own overflow and a float switch on the RO inlet to turn off the RO unit once filled.
On water change you then have a pump in the first sump chamber, on a timer, and pump waste water out into a waste container with a float switch that turns the pump off when the container is full - just so you can get a measured waste amount each time. As the water level in the sump decreases, the float switch will let water from the water change header tank into the sump to replace that being extracted.
There would then be a pump in the waste water tank (also on a timer, set to a later time) to empty to waste water to the drain (or garden etc).
For additional fail safes I could add an overflow to the waste water container. I could also add a low level float valve in the sump to switch off the waste water extract pump if the level gets too low. The timer would also turn that pump off in any case to prevent it over running. That should prevent the pump draining the sump. I could also have a solenoid operated by a second float valve or level sensor on the fresh water tank outlet, to act as a failsafe in case the first float valve fails and get stuck open, with the sumps main overflow outlet acting as the final failsafe.
As far as I can tell with that system, if the waste extract pump fails to come on, no fresh water gets added eliminating that risk of sump overflow. If the waste extract pump gets stuck on, the timer should turn it off, or the low level sensor on the sump should turn it off if that fails. The only remaining risk, as far as I can tell, is that in the event of a failure, scheduled water changes aren't happening as planned.
That being said, there are likely risks and failure points I've missed - do you see any
@ian_m (assuming you understand my rambling explanation).