Looking for some collective brainstorming on the topic of using Homey as an Energy Management System. With the latest updates, Homey brought the Energy tab to provide insights into energy usage. And of course you can utilize flows to automate things, but I’m not entirely sure whether this is going to fully support the usecase I’m looking for.
My current set-up:
Solar PV
Home battery for energy storage
Heat pump with boiler
EV with home charger
All of these systems are NOT integrated with each other. The battery monitors solar and is capable of charging on surplus solar power, but that’s it. But all of these component do support integration options and most of them I’ve already got hooked up to Homey. Just using that for insights at the moment.
I’m looking into ways of making this set-up smart. Some requirements that come into mind:
Use predictions (weather) to estimate whether the battery can be charged using solar power
Use dynamic tariff rates to see which times are most cost effective for charging / discharging the battery
Take into account my kalender to see when the car needs to be charged and what the most cost effective time to do so is (including battery stored energy and solar when available)
Use weather predictions to estimate heat pump use and plan heating the boiler at a cost effective time of day
Basically I want the whole to be as cost effective / power efficient as possible. Which will only be the case when everything starts working together. The EMS should be aware of potential solar power, the battery capacity and SOC, predictions on usage, all of it.
Now I could go into flow building, but the feeling creeps up to me that I will end up in a complex world in which flows might prove to be out of their league. I also considered going to specialized EMS options which I can then link to Homey to have all insights in one place.
Wondering whether others have similar set-ups and how you’re all managing this?
I think it depends on the algoritme you want to use, and how sophisticated you want it to be. The more sophisticated, the more probability will be involved which will lead to less understanding of what is happening.
To give you an example: when I want to use the washing machine, I always switch it on at 11:00, no matter when it’s sunny or not, because the probability that sunpower is available is highest. And the machine is ready at 13:00, always. Such an algorithm is easy to implement in daily life.
Maybe an algorithm based upon sun prediction will be more efficient, but only a few percent compared to the simple algorithm.
And keep in mind that when you want several optimation rules, the last one will not have so much effect as the first one. The Pareto principle.
Agreed. Great example of the washing machine btw, I do have a smart washing machine which already takes into account dynamic pricing for it’s optimal start time. But once again; this system is on it’s own and does not take other components into account. For the washing machine specific I (and my wife) are usually flexible in when it runs, so we’re used to that already. Other loads like cooking are or course much more tied to a specific time of day. Which is fine, the battery is capable of serving those loads when electricity is expensive. On the other hand, in summer when there’s an abundance of solar in the afternoon I don’t need the battery to pitch in that solar energy might be offloaded to the grid at night, or used to partially charge my car.
In my opinion this puzzle for home use is not that difficult. 100% optimization isn’t needed, 95% is good enough. But when these components don’t talk to each other this is hard to achieve. Thus the need for some kind of HEMS.
When this components have each individual time control, you can make an overall time schedule without an HEMS. Like setting the boiler temperature higher from 13:00 till 16:00.
Yes, I have already done so, but such a schedule does not align with the real world. Availability of things like solar power and energy tariffs are different over the course of days and months. The amount of energy needed for heating differs and I do not use my car every single day. A fixed schedule brings you places, but is nowhere near optimal. And it should not be that hard to get close to optimal, the situation for a single home is not that complex. With the current issues surrounding energy usage, we should strive to get close to optimal when possible. That’s at least my personal opinion. And I’m a nerd who likes to automate things, so this simply needs to get done
Interesting brainstorming here!
I am in the middle of a similar project for my home. For now it is split around devices:
Heat pump
I use the power level of my east-directed solar panels to lower the water temp, to surpress the overshoot when the sun is shining. My south- and west-directed panels also influence the water temp. Homey is in the middle of this process.
EV charging
I use the smart control of the Wallbox Plus (using surpluss only, except when I need the car fully charged) and control the home battery to not discharge while charging the car (due to the inefficiency of the home battery). I only listen to the Wallbox with Homey.
Home battery
I have 2 of them, one for zero on the meter and one to optimize for dynamic tariff. As said I block discharging when the car is loading. But often I still have a surplus of solar energy while the car is charging and this is allowed. Homey is doing a great job here.
I am looking for a mechanism to dynamically change the function of each battery (zero power, max earnings due to dynamic tariffs) depending on the euro’s.
First I build flows for each device, than I connected them “smart”. But there is much more to do to reach the 95% you are talking about.
Nice! So have you done everything using just flow? From what I gather from your post I suspect your battery itself determines the best time to load?
I already found a few drawbacks of the system I have in that it does allow me to schedule based on dynamic tariffs, but it will not calculate what the optimal time is to charge; it will simply charge as soon as tariff drops below a specific threshold. And since tariffs differ day by day this is less optimal.
The idea to stop the discharge when the car is charging I like as well. My energy company (Tibber) currently handles scheduling the charging of my car on the most optimal time, so that also means that at that specific time it doesn’t make much sense to discharge the battery as the tariff will be higher later in the day / night. So discharging the battery makes more sense at times when energy is more expensive. That said; if I know the battery will have a surplus getting through the night, it might still be useful to discharge it for a bit. Which again is also linked to the expected gains of solar the next day It’s quite a fun puzzle to work with, need to start by getting all of the pieces in place first.
It’s quite a fun puzzle to work with, need to start by getting all of the pieces in place first.