Calculating the best Time Of Use Tariff

curioustony

Last year I had solar panels installed. Brilliant decision. Delighted. Got in before the rush, and things got bad. Eagerly awaiting a Feed In Tariff (FIT).

Then I got a smart meter. Was not so sure about that one. Stayed on my existing plan.

The inverter app gave me loads of data on my use. Invaluable. I have reduced my consumption by about 20% without impacting the way I live.

The kids went from being at home (pandemic) to being in college, usage dropped again. Slightly over provisioned solar now. Never mind, more FIT 🤞.

The annual shop around time came in March, so I used the data that the inverter gave me and a spreadsheet to figure out the best of the available supplier and price plan for my usage. I wanted to take into account the FIT, no data available (there is another thread for that), so I made the best decision that I could.

Then I figured other folk might like to do the same thing. So I wrote a not so simple app that will do the calculation for you. Turns out to be a lot faster than a spreadsheet. I put it on github as free software (with no warranty 😁).

It's a bit mad because you need to know when you are using electricity, when the battery is charged, when the sun is shining, and what you are being charged at that specific time. The suppliers have lots of data, so they are at an unfair advantage.

"What about the comparison sites?", I hear you ask. Great sites, but (1) Solar & battery not considered (2) My load/usage pattern is odd -- isn't everybody's?

The app has evolved (and continues to, whenever I have time on a rainy weekends). It is now quite functional. I'm sure there are bugs, but I know it works in a straight line too.

I have not included the supplier rates as (1) they will soon go out of date and (2) I was afraid I might get in trouble.

I urge you to spend an hour of your time putting your own data into the app (all offline)

• If you have a smart meter and are considering a Time Of Use Tariff (TOUT) -- just ignore the solar stuff.
• If you are considering solar and want to estimate what it will save you -- solar data from European tool, PHOTOVOLTAIC GEOGRAPHICAL INFORMATION SYSTEM
• If you want to see the difference a battery will make -- add some scenarios
• If you have solar and want to sweat the asset to reduce the monthly bills

If you have an AlphaESS inverter (like me), you can download your data from within the App.

The app is available as python code and there is also a windows installer (msi). The readme (scroll down past the code) has a load of instructions.

There is a quick start with demo data included.

You can find it here: github com / Tonyslogic / tout-compare

Maybe somebody with more posts can reply with the link...

Good luck & enjoy.

the_amazing_raisin

You should copyright it and sell it to bonkers or switcher 😁

curioustony

It is copyrighted. It is also free.

The odds are stacked in favour of the suppliers. This might even the playing field a little bit.

graememk

Not an endorsement from me but here's the link ( I haven't tried it yet)

Looks promising, and as always the more information you feed it, the more accurate it is.

curioustony

The link to GitHub: https://github.com/Tonyslogic/tout-compare

graememk

https://www.boards.ie/discussion/comment/119265019#Comment_119265019

I meant to put the link in.. must have forgot

curioustony

The link posting rule confuses me. Sometimes I can, sometimes not. There is a 50 post rule but it is not enforced consistently.

curioustony

I've been working away on this over the summer. It's more functional and with less bugs.

One of my working assumptions in creating the tool is that the standard load profile (SLP) does not work for a lot of people (like me😀).

TLDR: IMO, the SLP is a good approximation of load.

The CRU mandates that price comparison sites must use the SLP. Suppliers must use the Estimated Annual Bill (EAB) which is based on the SLP and an average annual usage. https://www.cru.ie/document_group/smart-metering-comparison-tools-for-time-of-use/

The SLP can be found here: https://rmdservice.com/standard-load-profiles/. There are three variations (rural, urban and non-metered) for smart meters (which are based on D/N meter data as there were not enough deployed by Nov 21).

I've added the ability to use the SLP in the tout comparison application https://github.com/Tonyslogic/tout-compare

Here's how it stacked up with 6200 kWh (my usage). First the rural (the monthly generation is from PVGIS for my location and panel configuration):

This is a little different to the urban -- which is a lot flatter in terms of hourly of use:

Then I compared the simulations for my own load profile (inverter data) with the rural and urban. The supplier rates are correct for Oct 2022 based on published data to 3rd Sept.

My load profile:

The rural load profile:

And the urban load profile:

Finally the simulation graphs with seasonal load shifting (the spike @ 2 &3 a.m.) -- No EV, No diverters. First for my profile:

And then for the rural:

I need to fix the time zone difference in the PVGIS data as it is shifted ~ 2hrs to the right (later) -- next week maybe.

Anyway, the point being for the level of accuracy I am hoping for, the SLP is the way to go. The CRU was correct.

That of course confirms the approach for comparison sites, but does not help with smart plans that deviate from the 'SST'.

Having your own data is of course better, but unless it is wildly out, then stick with the SLP!

Oh, and best of luck if you are brave enough to try it out!

poker--addict

This looks really great.

Can my lowly macbook run this?

curioustony

https://www.boards.ie/discussion/comment/119572776#Comment_119572776

Yes, but you will need to install python 3+ and pip first. Then install manually. Either copy the archive from the latest release page or clone the GitHub repository. There are some needed modules listed in requirements.txt, use pip to install those

I don't have a MacBook.. so I cannot help so much on the install of python or pip, but should be ready enough.

curioustony

The latest release is updated to deal with the AlphaESS anti-crawling measures. See https://github.com/Tonyslogic/tout-compare/releases/tag/v0.0.27

It also includes Solis cloud (m.ginlong.com) as a data source since v0.0.26 -- A friend was kind enough to share credentials.

Manion

This is excellent, thanks for adding Solis support, I'll make good use of this once my installation is complete.

curioustony

I just published a new release: https://github.com/Tonyslogic/tout-compare/releases/tag/v0.0.32

I kept my life simple -- once you have registered and downloaded the HDF file (https://www.esbnetworks.ie/existing-connections/meters-and-readings/my-smart-data), there is an import/load option in the 'Load profile' menu.

Note that if you already have solar, this data won't be of any great value in the application as it will be very skewed by you solar generation.

My meter has been activated since March and this is what I get:

This looks very much like the output from the simulator/comparison, but shifted by an hour:

I will read the HDF FAQ when it is not broken... and likely update. ---> Now fixed.

Alun

Can you tell me if this would be of any use for someone like me who doesn't have any solar (yet), just to analyse my current usage patterns (obtained as CSV files from my efergy energy monitor) and see what effect various smart tariffs would have on my bill? Can I also experiment with switching around blocks of usage from one time period to another?

curioustony

@Alun, one of the main reasons for this it to help people without solar understand the their current use and see the impact solar would have. If you have solar, it can help you optimized your use of that given the insanely complex set of suppliers and plans available. The challenge has always been to get the input data that is relevant to the individual user.

I would love to integrate energy monitors as a source of load data. I don't have any monitors. I don't have any data. There are so many sources of data that I cannot do them all myself. However the project is open source so if you have one or other of these it is possible to integrate (and hopefully contribute). The DB schema is simple enough and documented in the readme. There are now 6 examples of how to do this too 😀. Feel free to PM me if you want to know more about that.

The application does not allow you to 'switch blocks of usage', say cooking dinner in the middle of the day or changing how you use a heat pump. It does have (very rough) support for hot water diverter, I think good support for EV charge scheduling and load shifting (this one I use a lot).

I recently looked for libraries of load profiles for cookers, electronics, fridges, washing machines, dryers -- like EI do after a few months of analysis, quite cool. Sadly this is a long way off, but if you are aware of any open source libraries please let me know. For now the closest thing is the 'awkward' sliders to create the hourly/daily/monthly distributions, but you need to manually translate the output from the monitor(s). The loadProfile.json schema is very simple:

The sum of the values in each section needs to add up to 100. So in theory you could create something to generate this from 'blocks'. Something to think about.

Alun

Thanks Tony, I'll take a closer look later. The only significant chunk of usage I'm happy with shifting to night rates is EV charging. I've got a PHEV at the moment and my current guesstimate is that it counts for around 15-20% of my usage so would like to be able to see what impact that would have. Should be pretty easy to spot as a load profile too if that facility were provided, other things like washing machines or dishwashers would be quite tricky I think and would depend heavily on what program is used.

DC999

https://www.boards.ie/discussion/comment/119901214#Comment_119901214

You might be surprised with the PHEV if it runs from battery all the time until the battery is used up (as in you maximise the battery side which will cost you a lot less of course). Our fully electric EV uses close to as much as our house and that’s ~15k per year (so low enough mileage). Our EV uses about 150-170kWh per month and we drive like Ms. Daisy 😊. So if you've a high % from charging the EV, then the night rates start to make sense. Never did for us pre getting the EV this year.

Tis roughly 1-2kWh per device per wash (washing machine and dishwasher). Depends on the program used of course as you say. That will be close enough for you for now. We use about 800kWh a year between them, so that's €350 give or take at a 45c tariff.

curioustony

For anybody that found the above python app useful, and more for those that did not or could not install, there is now an android version of the app.

It is more usable, but there is still a pretty big learning curve for anybody new to solar PV.

It has most of the functionality of the original (no inverter integration yet...). Just started BETA testing. You are invited!

You can find documentation for the android app here:

https://github.com/Tonyslogic/comparetout-doc

You can get it from the play store here (from android) or using a browser to join the BETA testing, here. Not listed publicly yet.

This is a free app with no ads, subscriptions or in-app purchasing. The code is not open source this time. I reserve the right to monetize if there is a market 😁

Give it go, and let me know how you get on!

Note when you define panel strings, you need to fetch data from PVGIS for each string separately --- seems to have caught a few Alpha testers out. The app will not simulate if there is no data. The presence of data is indicated by a sun on the panels with a green tick.

To get started quickly, you can use the download/fetch button with the default url to get a load profile (don't forget to FETCH data from PVGIS) and ~12 supplier plans from each of the USAGE(S) and COSTS tabs.

Feel free to pass it along too.

Now I need to figure out how to do a youtube video.

curioustony

And here is a video to get you started:

https://youtu.be/vUdPhGrInnM

I will add more on the details in other videos, a lot skimmed over in that 'short' video.