Device Off-Event Detection: I’m one of Sense’s biggest fans, however - I’d like to get the community’s individual feedback/input on the below-listed observation. This is a base-function, cornerstone to Sense’s level of accuracy and touches every customer - regardless of their local configuration (solar or non-solar) or available powered device types (EV, Smart Home, or standard appliances).
Conditions: Always On is (87w), quiet electrical signatures, and strong Google WiFi mesh network signal of 361 Mbps (up) / 34 Mbps (down)
Observation: I opened the microwave door and the internal light came on. After about 10 seconds, I simply closed it - without using it to cook/warm anything. Sense indicated that my microwave’s natively-detected (29w) appliance light-bulb was on from (3:53am - 5:47am) and again from (6:39am - 7:05am).
Results: Obviously, these are two clear indications that Sense successfully identified the On-Event - triggered by the light turning on (door open). Unfortunately, Sense missed the device’s final “off-event” triggers - light off (door closed). Subsequently, the resulting “Event Timeout” function engaged to act as the default Off-Event . Sense shows me various examples like this, throughout any standard week, involving various types of devices - despite active electrical activity levels (quiet or noisy).
Questions:
What are your individual conditions and experiences?
If this occurs for “detected” devices, does it happen to non-detected “Other” devices?
Are programmatic adjustments made to adjust “Always On” and if so, how fast are the adjustments made?
Is the Off-Event detection information availability anywhere or is it encompassed within the Sense area of proprietary information?
My most recent occurrence was my kitchen lights. I have a bank of recessed can lights (LEDs) that Sense has accurately detected. Been very solid with on and off events. Earlier this week we turned the lights off and Sense didn’t detect the off event. Said they had been on all night.
My salt water generator on my pool has been identified by Sense and is mostly solid. There are times where it doesn’t recognize when it turns off.
My TP-Link smart plugs/strips have been acting up recently. Sense is not detecting on and off correctly with them, two in particular.
My Other fluctuates widely too, but I think it has to do with the smart plugs.
I have tickets opened with Support. They think there are at least three issues at play here.
The “pattern establishment” possibility of any relatively low wattage device is going to be lower than a higher wattage device with the same signature physics and usage patterns. Another way of viewing that is “the lower the wattage the higher the noise” (in the on/off detection and the usage pattern).
As one would then expect, Sense is good at large kW water heaters because the ons and offs are more likely to be distinct. Rough analysis on my system, using dedicated Sense CTs on my HW tank circuit vs Sense-native detection on two Senses, is that it’s around 94% accurate, by some measure, on my fairly simple electrical network. That’s an interesting number because “95%” is a typical confidence that you might work with before “moving on”.
Yes, I have seen lost off-detections in a number of detected devices. In most cases I don’t see them as harbingers of some thing bad.
I think about Sense in these terms: The Sense monitor is tuned to “listen” for and report on on and off transitions within a sub-second time window. You see these transitions of interest that are tagged in Power Meter in the iOS/Android app.
So there are some transitions the Sense monitor never “hears” at all (slow ramps like EVs, mini-splits) and Sense sorts out later at the mothership, and others in the Sense monitor sweet spot that get lost to noise.
If a transition is reported by the monitor, the monitor takes a shot at classifying based on a set of models associated with that monitor. I’m assuming that classification of negative going transitions are also conditioned on the device, as Sense views it, being on. I also assume, based on what I have seen, that Sense doesn’t associate a model until it has “found” both an on and off signature for a prospective device.
So your missing off signature could be a transition that was lost to noise and never reported by the monitor, or one that couldn’t be classified by the model as the off-signature for an on-device.
Here’s a case where blinking Christmas tree lights were enough to obscure off-signatures of a washing machine, in a house where the blinking lights weren’t on during the learning for detection.
As for Always On, it is purposefully slow to responding to changes in power usage. More here:
Props on your usage:generation ratio BTW. Close to 1:10 is impressive.
There was a blog entry a while ago … I have an image of the kids playing with the bubbles: “Daddy, how do we switch the sun off it’s using so much electricity?!”
