I’ve recently installed a Aeotech HEM Gen 5 for a real purpose: I want to monitor the amp draw through my anemic 100-amp panel to see if I need to upgrade to a 200-amp panel in order to add some HVAC units to my house.
I’m mostly getting reasonable data out of the HEM; however, there are occasional spurious data points that make me skeptical of the readings, along with throwing off the scale of my graphs:
As you can see, between midnight and 4:30 this morning, I apparently consumed 1.55 terawatt-hours of electricity, or about the same as 140 households average over the course of a full year. The spike in current earlier in the week isn’t graphed at its actual value due to the chart resolution, but history tells me I drew over 16,000 amps at one point Monday afternoon (maybe I don’t need to upgrade that panel after all!).
What’s up with this? Is anyone else seeing strange/preposterous data out of this device? Other than the bizarre outliers, can the output be trusted?
Yeah i remember having that problem also. Check https://community.home-assistant.io/t/utility-meter-with-zwave/197494 as an example. You get these spikes sometimes. Which zwave integration are you using? I remember using outlier filter platform to get rid of these spike values.
Thanks for the solution, as those spikes are driving me crazy (the Energy consumption is all over the place). One quick question: are those filters applied directly to those entities values, or it creates a new entity or something?
P.S. Also, what values would you recommend for kWh, considering I’m using around 1000kWh/monthly, so one device would be impossible to increase more than 30kWh or so daily (I see now even thousands of kWh spikes - strange is there are also negative spikes)?
So, my sensor.monthly_home_energy gives me the monthly consumption in kwh, which resets every 21st of the month (invoice period). Once you use outlier sensor you won’t need to put any more checks about spikes. Try it and see if it stabilizes your readings. If not, then there is some problem else going on.