Hi everyone,
Very cool thread, I have a long term plan to replace my APEX with HA. In the meantime, I have brought my apex data into HA and started doing some basic monitoring of a small tank in my office.
I wanted to know if anyone is using Bayesian to create new monitoring sensors. Idea - Low sump level
Context:
For example, my large planted aquarium has a automatic continuous water changer that uses an overflow drain in the sump to ensure the sump cannot overflow.
That said, when I have a power outage, some water drains out of the sump. In addition, the continuous water changing system runs for a couple minutes each night to refill evaporation and on saturdays does a 50% water change. This week I had an issue with my RO unit and forgot to reopen a bypass valve after fixing a leaking RO line. Several days later, I noticed that the water level was low in my sump.
A Bayesian sensor could look at the values of PH and ORP and given a sudden and large variation identify that the sump level is below the probes.
I just posted in another thread my issues with the Bayesian YAML code, but thought it might be interesting for this thread too. Would also love to know if anyone is using ML or advanced analytics on the aquarium. It seems like we deal with many stochastic situations that would be well served by applying a statistical control framework.
Thanks