Dynamic MinSOC Based on Solar Forecast – Help Bring This Feature to EVCC

[MartijnvanGeffen13]

Hi everyone,

I recently submitted a pull request that introduces Dynamic MinSOC based on Solar Forecast, and I'd love your feedback and support.

👉 Pull Request: #32688

The feature is already available for testing on Ubuntu using the build attached to the pull request.

Why is this needed?

Many EV owners face the same challenge, especially during the darker months of the year.

You arrive home from a trip with a low battery and plug in your vehicle. The question becomes:
How much should EVCC charge now, and how much capacity should be reserved for today's or tomorrow's solar production? If i fast charge how will i stop it when the sun starts shining.

After all, solar energy is typically the cheapest energy available.

Today, EVCC uses a static MinSOC value. While this works well in summer, it can be less efficient during winter:

• A high MinSOC may cause unnecessary grid charging when you arrive home.
• A low MinSOC may leave the vehicle insufficiently charged for an unexpected trip.
• Solar production later in the day may go unused because the battery is already full.
• Family members and other drivers often don't want to manually adjust charging settings after every trip.

Introducing Dynamic MinSOC

The concept is simple:

EVCC automatically adjusts the effective MinSOC based on the solar forecast.

This allows you to keep your vehicle in Solar mode while Dynamic MinSOC automatically determines how much fast charging is required.

Example Scenarios: Vehicle returns with a low battery.

☀️ High Solar Forecast

• Significant solar production is expected.
• EVCC fast charges only to a low MinSOC level (for example 5% SOC).
• Maximum battery capacity remains available for solar charging.

🌤️ Medium Solar Forecast

• Moderate solar production is expected.
• EVCC fast charges to a medium MinSOC level (for example 40% SOC).
• The vehicle is asap ready for medium-distance trips while still reserving enough capacity for solar energy.

☁️ Low Solar Forecast

Little solar production is expected.
EVCC charges more aggressively from the grid to ensure the vehicle is ready when needed. (for example 80% SOC).
Only a small amount of battery capacity is reserved for available solar generation.

The Result

A smarter balance between:

• Cost of Fuel
• Vehicle readiness
• Self-consumption of solar energy
• Reduced manual configuration
• Better charging efficiency throughout the year

Key Benefits

✅ No need to manually adjust charging settings after every trip

✅ Optimizes charging behavior during winter and low-production periods

✅ Maximizes the use of available solar energy

✅ Works with existing weather forecast integrations

✅ Configurable solar thresholds (Low, Medium, High)

✅ Configurable MinSOC targets for each forecast level

✅ Existing MinSOC behavior remains supported and can override dynamic settings if required

✅ Configurable per vehicle, allowing adjustment for different battery sizes and driving patterns

How You Can Help

If this functionality would be valuable for your setup, please:

• Test the build attached to the pull request
• Share your feedback and real-world experiences
• Upvote This discussion to help demonstrate community interest

Community feedback and adoption are important factors in determining which enhancements are prioritized, so every test and every vote helps.

Thank you for taking the time to review and test the feature. My goal is to make EV charging a little smarter, more automated, and more solar-friendly for everyone.

Screenshots:

Extra configuration options in weather prediction settings:

Car settings for overwrite or active setting:

[mfuchs1984]

Have you tried the experimental optimizer? It will automatically find the most economic solution.

[MartijnvanGeffen13]

it wont take into account minsoc, / switching fastcharge / solar charge as there is no basis as the user does not feed it in. that is the hole challange. :)

If the hole family would enter every trip as a plan than optimizer would work, but that not something anyone will accept in the family

[mfuchs1984]

I don't fully get why you need xx% asap when no sun will not shine but it's ok to charge later when the sun will shine. Either you need xx percent for the next potential trip or not, isn't that independent from the weather?

[MartijnvanGeffen13]

but thats the hole point you dont know when the next trip is going to be can be 10 min, can be 1 hour can be next day.

Only thing you can do is fastcharge to the average minimum you need and than let solar finish it to 100%. But when sun does shine i dont need to fast charge as PV delivers 9-11 KWH allready directly from the pannels so fast charge would use 2 kwh mains.

[andig]

The assumption seems to be that one should fast-cache more now if the sun doesn't shine later. This implies some kind of demand consisting of (dynamic) minsoc + (later) solar. If that demand was specified as a charging plan, the optimizer would make this economically perfect, the planner already economically viable.

[MartijnvanGeffen13]

If the hole family would enter every trip as a plan than optimizer would work, but that not something anyone will accept in the family, thats the core of the problem,

Having like a 18 Y old or even the wife (hope she does not read this :) ) they dont care about a app and loading and cost

[andig]

You've just automated minsoc. You can already have a recurring plan that you don't even need to automate. That is a none-argument to start with?

[MartijnvanGeffen13]

@andig @mfuchs1984

Thanks everyone for the feedback and for challenging the proposal. My goal here is not to push a specific implementation, but to solve a real-world problem that I encounter frequently.

I also want to better understand your point of view and whether there are existing ways to achieve the same behavior using the scheduler, optimizer, or another mechanism that I may have overlooked.

I've been running a Dynamic MinSOC setup outside EVCC for almost 2 years now, and it has not let me down yet. That doesn't automatically mean Dynamic MinSOC is the correct solution for EVCC, but it does suggest that the underlying use cases are real and useful in practice.

I'm completely open to alternative implementations if they provide the same behavior. So rather than focusing on how it should be implemented, I'd like to focus on what behavior we're trying to achieve. If we can agree on that, the implementation can always be discussed afterwards.

For me, the distinction is:

The scheduler/optimizer answers: "How do I reach a target SOC?"
Dynamic MinSOC answers: "What should my reserve SOC be today based on expected solar availability?"

To illustrate what I mean, here are a few examples.

Example Scenarios (Bear in mind this is for the Family car and NOT the car you take to work 900: - 17:00)

Assumptions

Battery capacity: 60 kWh
Fast charging: 11 kW
Sunny day PV production available for EV: 8 kWh/hour
Cloudy day PV production available for EV: 3.5 kWh/hour
Grid electricity price: €0.25/kWh

Case	Situation	Dynamic MinSOC	Available Energy When Trip Occurs	Cost Impact
1	Car arrives at 08:00 with 5% SOC. Cloudy day. Unexpected departure at 09:00.	60% (bad forecast)	Vehicle is below MinSOC, therefore fast charging starts immediately. 1 hour × 11 kWh = 11 kWh charged. SOC increases from 5% → ~23% before departure.	No savings. Grid charged 11 kWh × €0.25 = €2.75. Vehicle was made as ready as possible given the available time.
2	Car arrives at 08:00 with 15% SOC. Sunny day. Unexpected departure at 10:00.	20% (good forecast)	Vehicle is only 5% below MinSOC. EVCC charges approximately 3 kWh to reach 20% (~16 minutes at 11 kW). The remaining time is solar charging: ~1h44m × 8 kWh ≈ 14 kWh PV. SOC increases from 15% → ~43% before departure.	Grid cost only €0.75. Static 60% MinSOC would have purchased 27 kWh (€6.75). Savings ≈ €6.00.
3	Car arrives at 08:00 with 1% SOC. Partially cloudy day. Unexpected departure at 13:00.	40% (medium forecast)	Fast charging runs until MinSOC is reached. Going from 1% → 40% requires approximately 23.4 kWh, taking about 2h08m at 11 kW. Charging reaches 40% around 10:08 and then switches to solar-only. Remaining ~2h52m × 3.5 kWh ≈ 10 kWh PV. Departure SOC ≈ 56%.	Only 23.4 kWh bought from the grid (€5.85). Static 60% MinSOC would require 35.4 kWh (€8.85). Savings ≈ €3.00.
4	Car arrives at 07:00 with 20% SOC. Sunny day. Unexpected departure at 14:00.	20% (excellent forecast)	Vehicle is already at MinSOC, therefore no fast charging occurs. 7 hours × 8 kWh = 56 kWh PV available. Battery reaches 100% SOC using solar energy alone well before departure.	Dynamic MinSOC buys 0 kWh from the grid. Static 60% MinSOC would immediately purchase 24 kWh (€6.00). Savings ≈ €6.00.
5	Car arrives at 08:00 with 8% SOC. Cloudy morning followed by sunny afternoon. Unexpected departure at 16:00.	40% (mixed forecast)	Vehicle is 32% below MinSOC. Going from 8% → 40% requires 19.2 kWh, which takes approximately 1h45m at 11 kW. At around 09:45 MinSOC is reached and charging switches back to solar-only. With more than 6 hours of sunshine remaining, approximately 48 kWh PV becomes available. Vehicle departs close to 100% SOC.	Dynamic MinSOC purchases 19.2 kWh (€4.80). Static 60% MinSOC would purchase 31.2 kWh (€7.80). Savings ≈ €3.00 while still departing with a nearly full battery.

In all 5 cases we also if departure is unexpected even earlier we almost always achieve maximum charge to a baseline minimum of 60. so car is as ready as it can be for unexpected departure. And in all cases we did NOT TOUCH or made any changes to any plan as a end user as we also could not have seen we do not know departure times.

Why I Think This Is Different From a Charging Plan

In all examples above, the optimizer can absolutely determine how to get to a target state.
However, the question I'm trying to solve is:

"What should that target reserve actually be today?"

When the forecast predicts poor solar production, I want the reserve to increase automatically.
When the forecast predicts strong solar production, I want the reserve to decrease automatically and let PV do the work.

So I'm not attached to Dynamic MinSOC as the only implementation. I'm interested in any solution that provides the same adaptive reserve-management behavior. If there's a better architectural fit for EVCC, I'm happy to discuss and help explore that direction as well.

What would you consider the cleanest way to achieve these scenarios with the current EVCC architecture?

[andig]

So neither of that will acomplisch anything of the goals nor will it charge with the cheapest source at the right time, it will only garantee a X amount at 7:00 clock.

It ensures cheapest way to achieve X amount at Y time.

Car will be charged by 7:00 yes, but for unplanned departure it will use am ore expensive power source to charge, SAme counts for setting the minsoc fixed will charge allways from expensive source.

It seems what you're trying to do is to minimize the cost for unplanned departure, and doing so taking solar only into account (and entirely missing dynamic tariffs).

I guess I need to give up. I'd suggest that we'll not merge your PR as it's too specific but you are of course welcome to use any automation you seem fit for your purpose. I've not managed to understand the generalized use case I'm afraid.

[andig]

You're basically saying "if I can't charge cheaply later I might just do so immediately". And that is exactly a planned amount of energy you need and what the planner will do. If all things are equal, the planner will do exactly the same. Only difference today: planner would prefer charging as late as possible. You behaviour would be easily achieved by adding an "early charging" option, in combination with optimizer.

/cc @mfuchs1984

[andig]

Only difference today: planner would prefer charging as late as possible. You behaviour would be easily achieved by adding an "early charging" option, in combination with optimizer.

Need to correct myself: planner alone doesn't take expected PV into account, only hard costs. I'll see if we can prototype the expected "full auto" optimizer and see if that would help.

[MartijnvanGeffen13]

You're basically saying "if I can't charge cheaply later I might just do so immediately". And that is exactly a planned amount of energy you need and what the planner will do. If all things are equal, the planner will do exactly the same. Only difference today: planner would prefer charging as late as possible. You behaviour would be easily achieved by adding an "early charging" option, in combination with optimizer.

/cc @mfuchs1984

correct and there in also is the problem, you dont know when the vehicle returns or needs to go away again, hence why a dynamic minsoc is sutch a nice solution,

if there is pv dont use minsoc, if there wont be pv, than charge now upto a fixed % so the car is ready to go again no matter if there wont be sun for next 3 days.

I can keep the fork open and maybe maintained for now so people with interest can still use it ,

[MartijnvanGeffen13]

It seems what you're trying to do is to minimize the cost for unplanned departure, and doing so taking solar only into account (and entirely missing dynamic tariffs).

Precisely, but we don`t have dynamic tariffs , we have fixed tariffs so Solar is 100% preffered and even if solar will be the next day i still prefferd solar over netuse. And thats the hole why we need solar forecast integration with minsoc.

Dynamic tarifs don`t make sense when you consume 18 MW a year and have for 13MW a year of PV , when prices are cheap i have a surplus of PV and when prices are expensive i am massively using with workloads that i can not controle .