A - Your idea makes sense (no pun intended), in theory, but suffers from 3 practical technical barriers
-
Sense firmware and software doesn’t support it today - Sense is set up to treat the solar CTs as an energy source, which means the power arithmetic / accounting is different than if it were another device.
-
Sampling rate / provisioning - I don’t know how Sense has provisioned sampling of the main CTs and voltage vs. Solar, but there is a limit to how much data the current Sense probe can sample (4M samples /sec). I think the sampling is asymmetrical, with much of the bandwidth going to the main CTs and main voltages, with the Solar CTs likely receiving lower rates of sampling, occasionally interleaved with mains sampling. The sampling would likely need to be reprovisioned if the solar CTs and voltages were to be used for identification vs. pure monitoring. But more than just a firmware issue - a basic design issue fitting into sampling rate constraints.
-
Sense isn’t using “ground truth” (measurements directly on a single device) for training yet, though they are in the process of setting that up for single devices given their own smartplugs. Conceivably, Sense could do the same for the solar CT input, but that’s probably more work than integrating the smartplugs, because:
- The smartplugs are already setup to do the power reporting, and are cheap and easy to set up.
- Sense would need to support an alternate configuration of the solar CT which is an expensive support consideration