This flags an error if the following conditions are met:
electricity consumption is more than 200W
heat output is 0W
DT between flow and return is more than 1K
there is at least 60s of error state in each error period.
flow temperature is at least 30C
Error flag based on heatmeter error register
The axioma heat meter that we now supply as standard provides an error register that we can use to detect this issue. The error register is recorded to a feed called heatpump_error which is then configured for use in the app:
The axioma is a little more conservative in when it flags the error, e.g it didnt flag the error until a few minutes later compared to the automated conditional detection method above.
Thanks @matt-drummer. I checked for this on Colin’s and couldnt find good examples, thanks for noting it on yours! I’ve change the conditions to require:
DT: 1.5K
At least 2 minutes of error state.
That seems to have cleared it on your system. Will refine until I get this right.
Love this - we’ve never looked at the time lag between detection and flagging the status in district heating applications, and now are the criteria documented - thanks!
@TrystanLea thanks for this- a great feature. Does it also exclude the erroneous data from the overall COP? I constantly have this problem at the end of my hot water cycle despite trying to clear the system of air. Think I may need a deaerator.
No not yet. While I’m pretty confident in the error detection capabilities , I’m not 100% confident yet to the point that I’d want to start automatically excluding electricity data at those points. It’s probably best to use it as a flag to notify of the issue for now.
Errors can now be highlighted on the daily summary view, and interestingly I have 14minutes in my 4 and a bit years of data with the conditions met to trigger this flag, even though I have a sontex heat meter.
I’m a little confused by this.
Expanding out to 1week shows “air issue detected for 5mins”, but when I try to zoom in (or out) I can’t find where this is being flagged. Is this “spurious” error detection going on or is there really something I should be looking out for in more detail?
Probably spurious, as you zoom out, interval of each data point is longer and that seems to be circumventing my checks. Will test on your dashboard and see what’s needed to avoid.
