I live in South Africa, where we have power outages for between 2 and 8 hours per day. While most of my Z-Wave devices are on a solar-based battery backup system, but underfloor heating is not. This means all of my UF heating Z-Wave controllers lose power at regular intervals throughout the day.
This plays hell with the Z-Wave routing as al lthe devices that stay on suddenly lose their routes (the UF heating controllers) and noting works reliably.
Is there a recommended approach to ensuring a large chunk of my mesh can fail without affecting the other nodes? Manually setting primary routes/return routes, etc?
That’s a very good question. We have periodic weather induced power outages and while I have a backup generator that comes on in 30 seconds, it does cause routes to change (and not for the best.). I can imagine in your situation it’s absolutely horrible. For now I go and manually repair/heal the nodes that get affected.
But given it seems to take 12-24 hours for the network to stabilize after the events - I’m not sure what you can do outside of getting some of the mains powered devices in a UPS.
I have thought about creating a script to automate healing devices. There should be a way of doing it via the Invoke CC API.
I was wondering if adding more repeater nodes to the network would convince all my devices to use them as a preferred route over the thermostats and minimise the impact of the thermostats going off during the power failures.
Triggering a re-heal each time the power goes out using the API may be worth looking at. But as you say - healing takes a while and our outages are in 2 or 4 hour stints through the day.
Requiring the house to put the UF heating on the battery is not really viable. Neither if requiring each flooring circuit to separate the thermostat relay from its power source.