I've seen a clever setup with the sensors in a dry container above the water tank. There is a hole in the bottom.
Before testing, a pump fills the container up with the tank water, flooding the sensor probes. When the pump stops, the water drains back out into the tank.
You’d need to wash the sensors and return their caps with protective fluids. It would be totally possible to automate, but perhaps the same overall cost as buying industrial grade sensors which can handle long term submersion.
You’d also need to ensure the caps contained enough storage solution at the right concentration. Over time the probes would introduce drops of nutrient solution (unless you rinsed them with distilled water, in which case you’d dilute the storage solution), and you’d need to replenish it.
