Skip to content

optimizer.mdx

Latest commit

 

History

History
85 lines (58 loc) · 3.32 KB

optimizer.mdx

File metadata and controls

85 lines (58 loc) · 3.32 KB
sidebar_position 12

import Screenshot from "../../../../../src/components/Screenshot";

Optimizer 🧪

:::warning Experimental The optimizer is in an early stage of development. The displayed data is currently informational only. Control actions will follow in future versions. :::

The optimizer analyses forecasts, consumption data, and the current state of your energy system to make cost-optimal decisions. It complements evcc's rule-based control with predictive optimisation.

Why an Optimizer?

evcc works rule-based and deterministically. This works great for many setups. E.g. a solar system, battery, and one vehicle.

More complex scenarios push this approach to its limits:

  • Multiple vehicles: Which one should be charged first?
  • Battery or vehicle: Where should the available energy go?
  • Dynamic tariffs: Is it worth charging from the grid tonight, or will there be enough solar energy tomorrow?

The optimizer can answer these questions. Currently, you set values like price limits or battery priorities yourself. In the future, you can let the optimizer handle these decisions. It will then automatically find the optimal values. We're working on this step by step.

What Does the Optimizer Do? {#how-it-works}

The optimizer collects various data:

  • Forecast data: Solar yield, electricity prices, feed-in tariffs
  • Historical data: Your household's typical consumption profile
  • Current state: Battery state of charge, connected vehicles, heating demand

Based on this data, an optimisation algorithm calculates the expected behaviour of your energy system. It identifies cost-optimal actions: efficient charging plans, battery control (hold, grid charging). It also predicts when the home battery will be full or empty.

Goal: Minimise energy costs.

Using the Optimizer {#setup}

Enable via the user interface:

  1. Configuration → Experimental → enable
  2. Configuration → Optimizer 🧪 → enable

The optimizer requires an active sponsorship. For new installations, it can take up to 24 hours to collect enough data to show first results.

  • Menu → Optimizer: Graphs showing how your home battery and charging points should behave over the coming hours and how much money can be saved.
  • Main view → Energy flow (expand): Shows when the home battery is expected to be full or empty.

Current State & Next Steps {#status}

The optimizer is currently informational only. It shows forecasts and potential savings but does not yet actively control anything.

Next steps:

  • Integrate actions: actively control the home battery
  • Let the optimizer optimise charging plans
  • Enable users to hand over specific settings to the optimizer

Technical Background {#technical}

The optimizer is Python-based and leverages the strong ecosystem for mathematical optimisation and statistics. It is not part of evcc itself but a standalone service.

When enabled, the cloud service optimizer.evcc.io is called. The service is stateless. No data is stored.

Like evcc itself, the optimizer is open source: github.com/evcc-io/optimizer. You can also install the Docker image evcc/optimizer locally. Use the OPTIMIZER_URI environment variable to point evcc to your own endpoint.