Analyzing My Electricity Consumption

> the organization managing the national French power grid (Enedis) has been in the process of replacing the legacy “dumb” electricity meters by “smart” meters, known under the brand name “Linky”

My jurisdiction, Ontario Canada, mass implemented these and found that people didn’t really shift their peak demand use much and it’s debatable if the amount spent was worth it. Costs just as much to implement for light users but they’re diminishing returns for any habit changes.

https://www.auditor.on.ca/en/content/annualreports/arreports...

Also went all-in on submetering small units like condominiums+apartments where there’s not much tenant/occupant ability to control their usage/equipment. If anything, just further encourages owners/builders to install the poorest efficiency stuff since they’re not paying the usage bill.

Not surprising that other tactics are more effective (dirt cheap LEDs, renovation programs, improving appliance efficiency… my TV and computers use a fraction of what they used to, wall warts are all SMPS now).

Still can’t find a “smart fridge” that runs in ultra-chill mode when rates are cheap and backs off when they’re high. Kinda antithetical to use more electricity to save but it’s true.

I really wish they could push the telecoms to focus on CPE efficiency since I’m forced to use their equipment.

I did a project with a french Linky a few years ago that used an ESP8266 to grab the serial data stream from that exposed port on the bottom and log the interesting line items it prints out to a InfluxDB timeseries database running on a local server and use Grafana to get a dashboard of various metrics.

I remember having to do some sleuthing with the serial port as it wasn't something standard - it was like 7 bits, no parity, 1 stop bit, or something like that. And a small circuit to do level shifting for the ESP8266.

The result was data on wattage draw approx every 1 second. This allowed for some very, very fine analysis in the data.

I tell people about this when I talk about various sidechannel attacks on things that use data like this- for instance, after observing the data for a while, I could tell which floor of the house someone was on by watching transients of the wattage and the quantized changes of various devices turned on and off - e.g. the kitchen has a kettle that uses 1220 +/- 5 watts, every time, and nothing else in the house results in a step change of that value. The top floor bedroom has lights that use exactly 42 watts. The laundry machine starts up with a very specific ramp and then oscillates between two levels of power with a period of 30 seconds, etc.

You can figure a lot out about the movements of people and what they are doing by simply having fine enough time resolution on a single number - the total power consumption of the entire house.

In Finland (and I guess at least other Nordics as well) we are starting to get neters with so called P1-port, which is open for the user to get their data. All I needed was a wire with RJ26 and an ESP8622 with some esphome definitions and now I get all kinds of data from the meter every 4 seconds. Total consumption, consumption, amperage and voltages per phase etc.

I have my dumb electricity meter sending my electric usage to google sheets every one minute

I used an ESP32 with a very basic light sensor, the sensor was blue tacked over the 1000th/Kwatt LED (this blinks 1000 times per unit of electricity used) the info is sent it over my wifi to google sheets

it also records the sum per hour indefinately, or until google sheets complains

The UK has smart energy meters for a lot of customers. Octopus are an energy supplier which is using them to great effect for data nerds and people who want to save money.

To get the data out, you can just call their API to get daily consuption figures, split into half hourly blocks[0]. They also allow you to just download a CVS file directly from the energy use section of their website.

The data in the API is only update daily, but they can give you a tiny addon device[1] for free which will allow you to access live usage in realtime (updates ever 10 seconds I think).

I use a Home Assistant integration to collect all this data live there.

To make use of the smart meter, they also have a series of tarrifs which change price dynamically. The simplist is Tracker[2] which charges you a different price per unit every day based on the wholesale price. There's also Agile[3] which is the same concept but the price changes every 30 minutes with higher highs and lower lows. That one is great if you can shift your energy usage outside of peak times.

They also have "intelligent" tarrifs where you allow them to control your car charger[4] or home battery[5] so it charges when it's cheapest.

Octopus are doing really great things in the UK and part of their business is that they sell the backend as a service to other energy companies who were previously stuck in the stone age as Kraken[6].

[0] https://developer.octopus.energy/docs/api/

[1] https://octopus.energy/blog/octopus-home-mini/

[2] https://octopus.energy/smart/tracker/

[3] https://octopus.energy/smart/agile/

[4] https://octopus.energy/smart/intelligent-octopus-go/

[5] https://octopus.energy/smart/intelligent-octopus-flux/

[6] https://kraken.tech

I would love a way to turn off my electric hot water heater between the hours of about 10pm and 7am.

I see no reason to keep the water hot during the time absolutely nobody needs it, and we know it will take less energy to heat it back up again rather than keep it at temperature.

Also, yes, I have turned the temperature down as low as it will go.

I have a very nice Graphana-based UI showing electric consumption data from Tesla power wall:

https://github.com/jasonacox/Powerwall-Dashboard

Together with temperatures in the rooms and AC activity:

https://github.com/cdzombak/ecobee_influx_connector

I plan to add a vehicle charging date to it (via Chargepoint Home).

I think Graphhana+Influx is a way to go for this type of projects.

A few years back I had a Sense power meter installed and its been great. It took some time to configure and get used to, and for it to detect various devices, but now I can reliably predict my upcoming power bill. And make adjustments as needed.

Whats helpful is it shows the top power consumers. And by paying close attention to that I've cut my bill by at least 1/3.

Few notes:

- fear and polemics about connected meters was not about "surveillance" by some giant but more about ability to disconnect or throttle a customer from remote and third party monitoring to elicit family habits for instance to organize a burglary;

- reading the meter without the need to get inside the home was already a solved problem with "telereport", a small local dedicated bus with wires going in some public places (like the hall of a condo, a small box near the portal of a home etc) who allow local reading, this could be used to send data remotely WITHOUT adding extra abilities like throttling or disconnecting;

- polemics exists for similar reasons for IVRE norms (BEV domestic charging) where the norms demand a kind of "remote control" to allow disconnecting the load from the grid when needed by the grid, without considering the few but not so few with domestic p.v. in self consumption that might get disconnected as well because there is no analysis of the grid absorption load just a "run this if the grid is strained";

- `Enedis website is nice` is well... debatable...

Beside this small notes NOT intended to diminish the author creation, IMVHO monitoring is useful only if you can act, for instance with p.v. and controllable appliances it's a good thing, and unfortunately there is next to none appliance really designed to run on p.v. when available, even 99% of those who claim the contrary. The first are domestic water heaters who actually could be piloted to shift load with HA/Shelly/* sometimes just abruptly cutting the power and giving it back when there is enough p.v. but not much more, secondary running dishwashers and washing machines left loaded and ready in a "run before 'given date and time' preferably on p.v. power" and so on. Technically such automation are damn simple and cheap if designed as part of every appliance but so far next to no one seems to be doing so, while mandatory monitoring and published monitoring solutions are more and more pushed. Personally having p.v. I do my best to shift loads and it's damn hard, mostly limited to smart switches connected to HA, because there is not much more to do. I can "hack" my dishwasher or washing machines, are dumb and basic enough to be easy just seeing their small controls and replicate them with a some GPIO on a raspizero and alike but that's a long uncomfy process since every machines have it's own controls and you have to redo when you change it. There is no standard.

One project I worked on was to track indoor temperature and AC settings (via a Nest thermostat) and outdoor temperature (via weather report). I found that in the evenings I was often still running the AC even though outdoor temperature had dropped below my target. This alerted me that I should instead open my windows. I didn't have a way to track energy usage changes, but I was really happy to get so much more fresh air.

It stopped working when Google bought Nest and then changed their API authentication method. I had a newborn and just couldn't find the time to support the new API. Maybe I'll revive it this summer.

Wouldn't it be cool if every wall and ceiling socket could report its electrical usage? I wonder if that's possible via powerline networks, or if it would work over something like Zigbee?

Not sure I really agree with all these design choices. The color scale on the hour/day grid initially seemed backwards to me. The daily total takes up so much space that it would be better as a normal line or point graph with scaled axes, instead of trying to guess the correspondence with the color.

Aside from that I am curious about the discontinuity at 22h. How many and what kinds of loads automatically respond to these price signals? Just heaters and water heaters?

Question for experts: I'm on SDG&E (North County San Diego) and apparently have a smart meter that supports HAN at 900 MHz, but at the end of last year SDG&E declared "no new devices [HAN devices] can be connected". Does this mean that if I want to monitor my electricity use that I’m out of luck?

If you live in the UK, you can sign up to https://glowmarkt.com/ and they'll give you the same 30 minute time slices, or you can buy one of their in-home devices for £70 (https://shop.glowmarkt.com/products/display-and-cad-combined...) and it'll push your usage every 10 seconds to MQTT to give you very interesting if completely useless graphs in Home Assistant.

(Not sponsored, I just use it)

How is this sustainable as the gap between self generated and stored energy and the grid price continues to diverge?

If present trends hold it won't be long before a few years worth of energy bills even in Europe finance a full off grid deployment. There doesn't seem to be much room to cut on the energy companies pricing which remains variable and pointlessly complicated vs "I bought this off grid system and now I never think about energy anymore".

Just people in high density housing, people in a lease, and those who can't afford the up front cost will constitute the majority of residential grid energy usage and purchasing?

Anyone here use the NiceGUI Python web app framework the OP used? There is mention of Electron-like cross-platform ability, which I assumed meant that one could generate a standalone app with it, but I did not find mention of that in the online docs.

You can get your data from the edf.fr site. In your account, under "Ma consommation d'électricité" there is a button "Télécharger mes données". It comes as a zip of csv files with daily and monthly values.

Has anybody figured out a way to do this with coned in the us?

Alternative:

Build glow [1,2]for $10. Hook it up to homeassistant for the price of an old laptop. Measure any and all interesting plugs using an athom [3] smart plug.

Zero cloud. Good data & charts.

Good fun.

[1] https://glow-energy.io/

[2] https://github.com/klaasnicolaas/home-assistant-glow

[3] https://www.aliexpress.com/item/1005006078553282.html

As a former electrician turned software engineer and serial entrepreneur I have always been conscious of my utilities consumption. Per Lorde Kelvin one can only improve what one measures and there are many things most people in society complain about yet those individuals could improve if they just measured; the data doesn't lie, people do.

Around about 2006, after being in my newly built home for a few years, I began to ask the questions that many more are just starting to ask now about their utilities. Having had an extensive background in residential and commercial electrical work I knew the options then were limited to measure home consumption digitally but I stumbled upon TED, The Energy Detective. In 2007 I installed TED in my panels and it has been running ever since. I initially only used it as a curiosity item, much akin to many Tesla Solar+PW users today in checking the app for production and consumption data, never before available in such a simple near real-time form of access. TED did however earn it keep in helping me deduce electrical draw issues when systems within the house were changed alongside just other general consumption curiosity.

In 2016 I got serious and started logging all of my utilities at the first of every month directly from my meters and I also cross collected utility bill data into a spreadsheet. That spreadsheet is now large but the information is invaluable in providing me an aggregate collection specifics to me and my family only over that time. Another side story in itself are the errors I have discovered and reported, some of those errors have benefited me greatly, once again you can only know such things through measuring and monitoring. All of this effort was my LONG GAME preparation for going near fossil fuel free and energy independent at my home when I found the correct solution for my needs.

In 2018 we went fully electric with our personal road vehicles. I installed the largest wall charger at the time which supported 220V up to 70A. This of course spiked our home usage at times however free supercharging was a great bonus and still is. Yes, we were early adopters of a fully electric vehicle family now for over 6 years and have driven hundreds of thousands of miles combined across the U.S. and Canada. I am still amazed at how ignorant people are about charging infrastructure since the fossil fuel industry has brainwashed people's expectations with gas stations on every corner. I have referred many an EV owner and while the program no longer exists I still refer people to electric, just referred a CT a few weeks back.

As time passed so too did my roof age and those here aware of asphalt shingle degradation know builder selections rarely last. In 2020 I started down the solar roof planning journey which had me upgrading all aging appliances to EnergyStar and in August of 2021 the solar roof was installed with two PWs. This led me to refactor my spreadsheet to now include my personal generation and consumption which the new system was providing me. Here I mention the TED again because this old still functioning device had provided me with years of historical data and now I would use it to compare the numbers from the new system. Yes, I found discrepancies that no one had found before and those issues had to be resolved within the new system; a new more accurate CT device was required as well as firmware changes. You may be able to hear me still smiling from this near energy independent choice.

My choices have guided my curiosity which has guided my choices and this is still so extremely relevant today in that I have a unique energy loss problem of which I intend to solve through my own energy storage device I am building. Everything provable in life boils down to either science or math so if the science and math check out then it is only the implementation that matters to reach one's goal. I am applying my software, hardware, electronics, and electrical experience as a unique subject matter expert to build something that solves my problem through the dog-food method. Once I have hashed out all the basic issues with this simple "complex system" I am building I will then apply my entrepreneurial talents once again, that whole ‘serial’ thing. You likely know what happens next given the world's growing awareness of a pursuit that I have long had devoted interest in represented by the choices I have made solely from monitoring and measuring. Thanks Kelvin!

Stay Healthy!

So you have to go to a third-party to get your data from a public service. Absurdity is the word that defined our era it seems.

> "Depuis le passage à l'OAuth2.0, il vous faut obligatoirement avoir une entité juridique afin de signer un contrat avec Enedis. Pour avoir une entité juridique, il faut obligatoirement être une société ou une association."

> "Since the transition to OAuth2.0, you must have a legal entity in order to sign a contract with Enedis. To have a legal entity, you must be a company or association"