Losing device detections

I’ve had my sense installed since Oct 2017, what puzzles me is device that have been detected regularly suddenly stop detecting Ex. my shop hot water heater which had been detecting suddenly stop, this has happened with other devices more then once since installed prompting reporting to sense technical who generally response in a timely manner for that one instants but seems to cause another device to stop detecting, I wouldn’t consider my home to be unusual has all the normal things you would find. can anyone explain why this happens?

You don’t give enough detail for us to compare against our experiences. Maybe if you share:

  • What did Sense suggest that seemed to fix the disappearing device detections ?
  • Were there any changes in your house prior to this issue ? Did you add a well pump or a plasma TV or other “noisy” devices ?
  • What other devices have stopped being detected ?

Many things could cause a device to disappear and later be re-discovered.

  • A change in device behavior - my AC compressor was detected as different devices during different periods in the cooling season
  • New Sense algorithms and models - typically this results in the device being lost for a bit, then re-emerging later on.
  • Changes in you house as I previously discussed
    .

There is the possibility of a failing element, thermostat or sediment buildup on your water heater.

2 Likes

I would agree with you on the elements being dirty except for the fact that the hot water heater is a 50 gallon gas fired with a power vent, was used to heat my in-floor heat and domestic hot water for my shop. I know the hot water tank is still operating as always nothing has changed it it’s operation.

well sense said they would look into the problem then send a email saying it should be working fine now and it did in most cases. when the warm season arrive and I turn on my window box air conditioner for the season it was detected as my air compressor that was corrected by sense, when I turned on my other window box unit for the season which had been detecting for the previous season, turn out to be a no show this season, sense suggested deleting it and let it find it again after several weeks it still hasn’t found it again, I don’t have a well pump or added anything new to the house, to my knowledge we have no noisy devices. our tv’s are led and only detect occasionally and then only briefly when the turn on, it was detecting my laser printer but that’s a hit and miss thing. as I said in the beginning there is nothing unusual about the devices in my home and it would seem that once devices are detected they would remain detected and not drop out for periods of time or disappear all together. kind of takes away the ability of finding out what’s in your home that may expose problem devices, the only reliable reading has been the incoming power, that seems to follow with the utilities monthly bill plus or minus a few KWH, so back to my original question why does these things happen, why do devices disappear from time to time and some times reappear or not at all.

So Sense was once detecting the power vent or pump, not the actually heating, right ?

that’s correct and was identified as being the hot water heater at least that’s what I labeled it. to add to this, this not my first time at monitoring my electric, I have a efergy hub with 5 ct scanners and had been in use long before installing the sense. it is still in use today. it’s set up is as follows, 1 ct monitors the main 1 monitors a sub panel of the main and 1 monitors a sub panel feed from the 1st sub panel the other 2 monitor at my digression and can be switch from device to device. so far the efergy has been consistent in it’s recording and matches the utility bill within a kwh either way, I purchased the sense in hope of replacing the efergy. but I can not see doing that in the near future

I’m going to give you a probably unsatisfying answer. Sense uses machine learning, which is one of the best automated tools for learning and detecting patterns and images, but also has a challenge - it really can’t diagnose why it didn’t detect something it presumably once “learned”. But here are some possible reasons using human vision and recognition as an analogy. I probably should have used hearing as the analogy, since that is closer to Sense in the type of analysis it uses, but we’re more self-aware of why we didn’t spot something than we are about why we misheard something.

Some examples of why people miss seeing a person they are looking for (and the Sense comparable):

  • He/she was wearing a hat - sometimes the Sense on-pattern might come adorned with an additional flourish. For instance, motor patterns look different depending whether they start with a light load vs. a heavy load.
  • I only met him/her once or twice so my picture of him/her was slightly different - machine learning needs lots of repetition. Sometimes, early in the recognition cycle, Sense might not have enough training to really recognize a pattern solidly.
  • It’s been a long long time, so my memory of him/her has faded - Machine learning has a challenge in what is called “concept drift” - Over time, as Sense learns how to remember more and more things, some of it’s “memory” for little-used patterns erodes. That’s what I think is what is partially to blame for why detection of my 2 AC units got conflated after a long idle period in the winter. Concept drift might be why your AC units don’t reliably get identified from season to season. My experience here:

https://community.sense.com/t/sense-vs-ecobee/3409

  • I just got a new pair of glasses (this example is a bit of a push) - Sense will occasionally update their models based on new training. This will typically result in some new detections, but might also make some of the old patterns that were reliable “look” a little different to Sense.
2 Likes

just to be clear on my part, I like the concept of sense monitoring. I have devices that have been constant since they have been detected, I have 2 refrigerators and 1 freezer that are constantly detected I have shop lighting that is detected regularly although since retired I don’t use it as often as I did, but shows up when ever it’s turn on no matter how long it’s been since the last time, it senses my boiler HSI when it the boiler is turned on with no trouble, but oddly it never has detected the boiler pump, my boiler feeds a hot water tank in the system so it comes on several times a day as hot water is used. it took awhile in the beginning to detect my washing machine, but once it was discovered, it has detected it consistently since same goes for my dish washer, my oven, it has occasionally detected my gas dryer, but not consistently. it appears that 2 devices with approx the same kwh tends to confuse the sense machine learning (example a 120 volt 12000 btu window box AC unit and a 220 volt air compressor.) I don’t seem to be the only one who has problems with device detection as your blog has pointed out, as I pointed out I think sense is a good concept but has a long way to go yet, I truly hope it gets there some day

1 Like

based on your last statement in your analogy " * I just got a new pair of glasses (this example is a bit of a push) - Sense will occasionally update their models based on new training. This will typically result in some new detection’s, but might also make some of the old patterns that were reliable “look” a little different to Sense." you have provided me with the answer I was seeking, it appears that it was not what was happening in my house but what was happen at sense and their updates that caused the loss of devices, it appears we can expect that to happen every time there is a new up date to machine learning, I and many others certainly hope not, I truly do hope it improves as time goes by. Like I have said I think it’s a good concept.

I’ve had the same thing happen to me in the last couple months. The first was due to a power outage and reset everything. Yesterday sense was still recording on my phone but lost WiFi connection with the sense team…made no sense, excuse the pun!