Will the Tesla App get Appliance Detection?

[RKCRLR]

My understanding is that the Gateway uses a Neurio PB1 for power monitoring. Some Neurio home monitoring models have an app that can detect and learn to monitor appliance power consumption.
I'm wondering if Tesla has any plans to incorporate appliance detection into the app or if the Neurio app can be used with the PB1 hardware.
Anybody know?

 

[jjrandorin]

I doubt it will get that feature. I think the one for tesla is not a "stock" one. I really wanted that feature when I read about it, and was considering getting a sense device until I read that they dont work very well when you have solar + powerwalls.

The good news is, you dont really need that feature, really. The amount of power used by your home in the tesla app will change fast enough that you can easily just use stuff while looking at the app, and figure it out for anything that actually matters as far as energy usage.

Microwave, coffeemaker, toaster, AC, etc. Just look at your tesla app home consumption, note the number, turn the device on, watch the number instantly jump to a new value, and note the difference, and thats your appliance. Thats how I now know that my AC units each use approximately 4 kW of electricity when running, and I have two of them. My keurig coffee maker uses about 1.9 to 2.1 kWh of electricity when it is turned on and heating up, my microwave uses about 3 ish kWh when it is running heating up a dinner plate, and my toaster uses 2.5 kWh of electricity when it is making toast.

EDIT: ... and my bosch double wall oven uses between 3.7 and like 8.5kWh while running, depending on if we are using one oven or two... I never knew that my wall oven used as much energy as my AC, but it was easy to figure these out doing the above.

EDIT 2.. as pointed out by @BrettS , the numbers above should be in kW not kWh.. and I just checked my keurig again and its the total home load I was looking at. it just used .9 kW to warm up. Most of those numbers above should have 1kW subtracted from them, as my average home load is somewhere between .9 and 1.2 kW.

Thanks for the call out, @BrettS , I dont want to be providing information that is incorrect if I can help it.

 

[BrettS]

@jjrandorin, I think you meant kW in those examples. I also question those numbers. A normal 15A 120V outlet, such as the one your keurig, microwave, and toaster are plugged in to can only provide a max of 1.8kW of power. So it is very unlikely that your toaster is using 2.5kW or that your microwave is using 3kW.

 

[jjrandorin]

@jjrandorin, I think you meant kW in those examples. I also question those numbers. A normal 15A 120V outlet, such as the one your keurig, microwave, and toaster are plugged in to can only provide a max of 1.8kW of power. So it is very unlikely that your toaster is using 2.5kW or that your microwave is using 3kW.

Yeah you are right.. let me see if I can change it.... I also am likely remembering what the total home number is using, because I didnt write them down, I just eyeballed em while it was running when I first got the powerwalls for the same reason the OP is talking about.

 

[BrettS]

As far as the original question goes, I don’t see Tesla adding appliance detection anytime soon. Frankly the tesla app seems designed to provide the minimal amount of necessary information. For example, with the solaredge app you can see panel level production data, but the tesla app doesn’t let you see that level of detail.

As far as appliance detection in general goes, I’m not sure the technology is really there yet. I’ve used a brultech energy monitor for a good 8 or 10 years now. This device allows me to monitor individual circuits by placing small CT’s on the circuits that you want to monitor. This is great for monitoring large appliances that have a dedicated circuit, like an air conditioner, or a range or a washing machine.

But it doesn’t really work for smaller devices that just plug into an outlet like a coffee maker. You can put a CT on the circuit that feeds your kitchen outlets, but that will get everything that plugs into your kitchen outlets... your coffee maker, your toaster, your phone charger, your mixer. And a lot of times the smaller circuits are feeding completely unrelated things. I have one circuit that feeds the lights in my bedroom and bathroom as well as the outlets in my bathroom.

Because of that I don’t monitor the smaller branch circuits at all. But really, just by monitoring the large appliances I can capture a good 80 or 90% of my home’s power usage to see where that power is going.

For example, I know that over the past 7 days my house has used 298kWh. And of that I know that my heat pump has used 139kWh. I have a large fish tank with a dedicated circuit and I do monitor power going to that circuit. In the last 7 days the fish tank used 72kWh. I have a small pool in my back yard and the pool pump has used about 47kWh. My water heater has used about 15kWh. I didn’t do any laundry this week, so nothing for my washer and drier.

It’s one thing to know that my heat pump uses 2.8kW when it’s running, but another to be able to see that my heat pump is taking 50% of the power that my house uses.

Before my solar/powerwall install I removed my energy meter because I didn’t want all of the individual CT’s to get in the way of the install or cause problems with the inspection or anything. Then after my system was installed I spent quite a while trying to decide if I wanted to put my old system back in or upgrade to something newer like one of the sense devices. In the end I decided that the problem with the sense devices is that they are just hit or miss. Maybe it will detect my heat pump, maybe it won’t. Maybe it will detect my water heater, maybe it won’t. Maybe it will detect a random light bulb in my closet that I don’t really even care about. The ability to install it without installing a lot of individual CT’s was tempting, but in the end it just seemed like there was too much uncertainty in what it might be able to detect.

So that wound up being a really long post, but I guess what I’m trying to say is that I think there is definitely a benefit to being able to monitor more than just the whole home’s power usage, but I’m not sure that the Sense type devices are really the way to do it. There are a number of devices out there that can monitor multiple circuits with individual CT’s and I think that if you really want to get into power monitoring like that then that’s the way to go.

 

[bob_p]

It's not that difficult to determine how much power an individual device is using with the current app, by having the app display the current house power usage and then toggling the device on and off. It was pretty easy to identify the largest power users in our house - and helped to remind us of which lights we often leave on that were good targets for LED bulbs.

 

[wwu123]

It's not that difficult to determine how much power an individual device is using with the current app, by having the app display the current house power usage and then toggling the device on and off. It was pretty easy to identify the largest power users in our house - and helped to remind us of which lights we often leave on that were good targets for LED bulbs.

Just want to add here that if you have a smartmeter in CA PG&E territory, the Rainforest Eagle is a reasonable $100 option for those without Powerwalls, to provide net consumption reporting, as well as a real-time graph that works for this toggling estimation as well. I've validated my hardwired devices like whole house fan 280W, furnace blower ECM motor 90W low, 400W high, etc. Uses Zigbee to talk directly to the smartmeter.

 

[jjrandorin]

To revisit this for myself, I was disappointed in myself for mis remembering the numbers and then posting them, so I checked again this morning to try to be more accurate.

Keurig k-Cafe Coffee Maker, from off position, during inital warm up = 1.5 to 1.6 kW
Microwave = 2.2 to 2.3 running on 100% power re heating a beverage
Toaster =1.5 to 1.6 kW while toasting

So, it looks like I was remembering the total consumption number as I mentioned, so anyone reading that first post which I can no longer edit, please subtract 1 to 1.1 kW from the numbers I posted there as I mentioned. sorry for the mis information!

 

[RKCRLR]

Yea, I agree that appliance detection is more of a novelty than a necessity. And, yes, I can identify the power draw of appliances by monitoring the change when they go on or off.
But it would be nice to have a cumulative history of the usage over time. And it seems that the hardware would be capable of recognizing and monitoring the larger appliances. It would just require some enthusiasm on Tesla's part to incorporate it in the software.
Has anyone tried to see if the Neurio app could be tricked into monitoring the PB1 hardware?

 

[woferry]

If one were installing a Neurio to measure their household usage, they'd connect the CT's to the house loads. Tesla uses the Neurio for different purposes, only measuring net grid current and solar current. So in Tesla's case the home load isn't directly measured at all, it can only be inferred from what the Neurio is measuring, plus or minus whatever the PWs are doing. This leads to a less accurate home load reading, so it would make it pretty useless for what the OP is trying to do. A second Neurio (or similar solution) properly-connected to measure the loads one cares about seems like the right thing to do if that's what you want.

But Tesla's Neurio itself definitely couldn't even guess at the home load, since it has no information on that or the Powerwall input/output, so it's missing too many terms to balance the equation. Only the Gateway has the data in Tesla's setup.

History is available from the app and the owner API it uses. Though I'll note if you get the every-5-minute 'daily power' data from the API there is no home data, again it has to be derived from the grid/solar/battery(/generator/grid_services) data that is provided.

 

[BrettS]

If one were installing a Neurio to measure their household usage, they'd connect the CT's to the house loads. Tesla uses the Neurio for different purposes, only measuring net grid current and solar current. So in Tesla's case the home load isn't directly measured at all, it can only be inferred from what the Neurio is measuring, plus or minus whatever the PWs are doing. This leads to a less accurate home load reading, so it would make it pretty useless for what the OP is trying to do. A second Neurio (or similar solution) properly-connected to measure the loads one cares about seems like the right thing to do if that's what you want.

But Tesla's Neurio itself definitely couldn't even guess at the home load, since it has no information on that or the Powerwall input/output, so it's missing too many terms to balance the equation. Only the Gateway has the data in Tesla's setup.

I think this is how any Neurio or Sense would need to be installed. There is nowhere that a set of CT’s could be installed to directly measure the home load. In a ‘normal’ non solar install the CT’s would be installed on the grid connection and obviously in that case any power coming from the grid would be used by the home. With solar a standard Neurio or Sense would have two sets of CT’s. One for the grid and one for the solar production and it would need to do the same math to figure out how much power the house is using.

The only place where there could be some confusion is when your house was running on your powerwalls. I believe that the powerwalls report their own production and consumption data directly to the gateway and the Neurio would have no idea how much power was coming from or going to the powerwalls.

If you were operating in backup only mode this probably wouldn’t be a big problem as it would only be an issue when the grid was down. However if you were operating in a mode where your house regularly draws power from the powerwalls then there would be an issue as there would be long periods of time where the data would be wrong.

 

[woferry]

There is nowhere that a set of CT’s could be installed to directly measure the home load.

That's certainly not true for my house. All of my loads come from a single panel, since I have whole-house backup. Granted there is a PW breaker and a solar breaker that feed into this panel, arranging the wires such that a single pair of 'home' CTs measured everything entering the panel minus whatever comes in from the solar and PW would tell you what the house is consuming. It would certainly be harder without whole-house backup as the loads would be in multiple panels with different sources, probably easier to just ignore the non-backed-up loads (or have dedicated sensing just for them).

 

[BrettS]

That's certainly not true for my house. All of my loads come from a single panel, since I have whole-house backup. Granted there is a PW breaker and a solar breaker that feed into this panel, arranging the wires such that a single pair of 'home' CTs measured everything entering the panel minus whatever comes in from the solar and PW would tell you what the house is consuming.

That’s exactly my point. Because the powerwalls and solar feed into that panel you can’t directly measure what the house is consuming. You can measure what’s coming in to that panel by putting CT’s at the grid connection or at the point where it enters that panel, but the power coming from the solar will make that reading inaccurate. That’s why it would also need to have another set of CT’s to measure the solar production.

 

[woferry]

That’s exactly my point. Because the powerwalls and solar feed into that panel you can’t directly measure what the house is consuming. You can measure what’s coming in to that panel by putting CT’s at the grid connection or at the point where it enters that panel, but the power coming from the solar will make that reading inaccurate. That’s why it would also need to have another set of CT’s to measure the solar production.

No, you seem to have missed my point, or perhaps how the CTs work. By putting multiple wires through the CTs you can perform simple addition/subtraction. So what I described above would get you exactly the home load, removing the extra solar and battery currents that were not going to home loads. So it would be a direct measurement of what the house is consuming. It wouldn't tell you anything about what the solar was producing or what the battery was producing/consuming, but the GW already has sensors for those.

 

[BrettS]

No, you seem to have missed my point, or perhaps how the CTs work. By putting multiple wires through the CTs you can perform simple addition/subtraction. So what I described above would get you exactly the home load, removing the solar and battery levels no matter where that current was going. So it would be a direct measurement of what the house is consuming

Maybe we’re saying the same thing in different ways, but I was replying to what you said below:

If one were installing a Neurio to measure their household usage, they'd connect the CT's to the house loads

All I was saying is that there is nowhere to put a set of CT’s to measure *just* the house loads. You would also need CT’s to measure the solar (and potentially PW) production as well so you can do the math to calculate the house loads.

 

[RKCRLR]

Here is an installation diagram for the Neurio W1 with solar:
pkb
My system appears to be installed the same way as the first figure other than the Neurio solar CTs appear to monitor both the Powerwalls and the solar production.
By the way, my Neurio looks just like the W1.

 

[RKCRLR]

Here is an installation diagram for the Neurio W1 with solar:
pkb
My system appears to be installed the same way as the first figure other than the Neurio solar CTs appear to monitor both the Powerwalls and the solar production.
By the way, my Neurio looks just like the W1.

I was incorrect about the solar CT. There is only a single solar CT and it only monitors the solar production, not the Powerwall output.