I have a home with multiple heat pumps controlled by about 10 ecobee thermostats. A backup power system with 2 LG batteries and 2 SolarEdge inverters can supply emergency power to at most one heat pump at a time or the inverter will trip offline.
By offering options to command heat or cooling to one zone at a time, using small temperature offsets, my hope would be to manage the load to stay within the limits of the inverters.
I was attracted to openHab because of available bindings for both ecobee thermostats (to control the thremostats) and ModBus/Sunspec (to read the output of the inverter and detect a grid outage). Most of the tedious work is there, leaving only the high level logic to be implemented, plus perhaps an Alexa control interface.
If this is my only goal, do folks think openHab and its learning curve is overkill, or maybe just about right? I’m a software engineer and could hack something much smaller together perhaps from an MQTT and Modbus library and try to keep it simple, but it would likely be less extensible and maintainable.
I do something like this with my OH system. I have an off-grid home with 10kw of solar and a 5kw Vicron inverter. Once the battery bank is full, I dump excess solar into a waterheater element. I have a set of openHab rules that calculate excess solar potential and total inverter load. OH then turns a sonoff POW on or off to dump load into the waterheater. It detects the total inverter load and reacts to other loads coming on and off to avoid overload issues.
This has been working well for over a year with no inverter overload cut-outs.
I’m happy to share the code if it helps.
No, the Victron gear uses MQTT.
Before changing to Victron I used (awful) Morningstar gear - that was all Modbus, it was not very reliable. I now use MQTT wherever possible.
openHAB is rarely the “best” or most efficient tool for a specific task. In many cases, we don’t care really. Would it matter if your logic took 100 or 500mS to make a decision?
Ask instead, what if I later decide to fine tune the management based on occupancy? Or weather forecast?
There will be an investment in something new to learn, the payback is longer term.
This is to approach a custom specific optimum solution. Clearly this is about trying various strategies, varying parameters, and also to take yet more inputs into your calculation such as the thermostats, heat pump data, weather forecasts, eventually lux sensors to calculate solar gains and more. Alexa all by itself also is hard to integrate.
You need a solid basis to talk to all of thes devices of all sorts (and you don’t have a finite list of those yet), so yes you will need a versatile tool like openHAB to provide all the comms and conversions.
So yes it’s worth to take on the learning curve.
Coding on top, there’s multiple options, rules DSL or Python being the prime candidates.
Gavin, I am forced to used Modbus with SolarEdge, unfortunately. Just to double check, you don’t mean to imply that the openHAB Modbus bindings are buggy, right? If so, I could use a separate RPI and convert Modbus to MQTT for and use MQTT on openHAB. But that adds an additional layer, so I’m thinking I’ll try the serial Modbus binding.
