Skip to content

Latest commit

 

History

History
266 lines (184 loc) · 11.5 KB

README.md

File metadata and controls

266 lines (184 loc) · 11.5 KB
Raw

evcc

Build Status

EVCC is an extensible EV Charge Controller with PV integration implemented in Go.

Features

  • simple and clean user interface
  • multiple chargers: Wallbe (tested with Wallbe Eco S), Phoenix controllers (similar to Wallbe), go-eCharger, openWB slave, Mobile Charger Connect (currently used by Porsche), any other charger using scripting
  • more chargers experimentally supported: NRGKick, SimpleEVSE, EVSEWifi
  • different vehicles to show battery status: Audi (eTron), BMW (i3), Tesla, Nissan (Leaf), any other vehicle using scripting
  • integration with home automation - supports shell scripts and MQTT
  • status notifications using Telegram and PushOver
  • logging using InfluxDB and Grafana
  • soft ramp-up/ramp-down of charge current ensures contactor only switched at minimum current
  • electric contactor protection
  • REST API

Screenshot

Index

Installation

EVCC is provided as binary executable file and docker image. Download the file for your platform and then execute like this:

evcc -h

or to run EVCC with given config file and UI on port 7070 using Docker:

docker run -v $(pwd)/evcc.dist.yaml:/etc/evcc.yaml -p 7070:7070 andig/evcc -h

To build EVCC from source, Go 1.13 is required:

make

Note: EVCC comes without any guarantee. You are using this software entirely at your own risk. It is your responsibility to verify it is working as intended. EVCC requires a supported charger and a combination of grid, PV and charge meter. All components must be installed by a certified professional.

Configuration

Charge Modes

Multiple charge modes are supported:

  • Off: disable the charger, even if car gets connected.
  • Now (Sofortladen): charge immediately with maximum allowed current.
  • Min + PV: charge immediately with minimum configured current. Additionally use PV if available.
  • PV: use PV as available. May not charge the car if PV remains dark.

In general, due to the minimum value of 5% for signalling the EV duty cycle, the charger cannot limit the current to below 6A. If the available power calculation demands a limit less than 6A, handling depends on the charge mode. In PV mode, the charger will be disabled until available PV power supports charging with at least 6A. In Min + PV mode, charging will continue at minimum current of 6A and charge current will be raised as PV power becomes available again.

PV generator configuration

For both PV modes, EVCC needs to assess how much residual PV power is available at the grid connection point and how much power the charger actually uses. Various methods are implemented to obtain this information, with different degrees of accuracy.

  • PV meter: Configuring a PV meter is the simplest option. PV meter measures the PV generation. The charger is allowed to consume:

    Charge Power = PV Meter Power - Residual Power

    The Residual Power is a configurable assumption how much power remaining facilities beside the charger use.

  • Grid meter: Configuring a grid meter is the preferred option. The grid meter is expected to be a two-way meter (import+export) and return the current amount of grid export as negative value measured in Watt (W). The charger is then allowed to consume:

    Charge Power = Current Charge Power - Grid Meter Power - Residual Power

    In this setup, residual power is used as margin to account for fluctuations in PV production that may be faster than EVCC's control loop.

Charger configuration

When using a grid meter for accurate control of PV utilization, EVCC needs to be able to determine the current charge power. There are two configurations for determining the current charge power:

  • Charge meter: A charge meter is often integrated into the charger but can also be installed separately. EVCC expects the charge meter to supply charge power in Watt (W) and preferably total energy in kWh. If total energy is supplied, it can be used to calculate the charged energy for the current charging cycle.

  • No charge meter: If no charge meter is installed, charge power is deducted from charge current as controlled by the charger. This method is less accurate than using a charge meter since the EV may chose to use less power than EVCC has allowed for consumption. If the charger supplies total energy for the charging cycle this value is preferred over the charge meter's value (if present).

Implementation

EVCC consists of four basic elements: Charger, Meter, SoC and Loadpoint. Their APIs are described in api/api.go.

Charger

Charger is responsible for handling EV state and adjusting charge current:

  • Status(): get charge controller status (A...F)
  • Enabled(): get charger availability
  • Enable(bool): set charger availability
  • MaxCurrent(int): set maximum allowed charge current in A

Optionally, charger can also provide:

  • CurrentPower(): power in W (used if charge meter is not present)

Available charger implementations are:

  • wallbe: Wallbe Eco chargers (see Hardware Preparation for preparing the Wallbe)
  • phoenix: chargers with Phoenix controllers
  • simpleevse: chargers with SimpleEVSE controllers connected via ModBus (e.g. OpenWB)
  • evsewifi: chargers with SimpleEVSE controllers using SimpleEVSE-Wifi
  • nrgkick: NRGKick chargers with Connect module
  • go-e: go-eCharger chargers
  • mcc: Mobile Charger Connect devices (Audi, Bentley, Porsche)
  • default: default charger implementation using configurable plugins for integrating any type of charger

Wallbe hardware preparation

Wallbe chargers are supported out of the box. The Wallbe must be connected using Ethernet. If not configured, the default address 192.168.0.8:502 is used.

To allow controlling charge start/stop, the Wallbe physical configuration must be modified. This requires opening the Wallbe. Once opened, DIP 10 must be set to ON:

dip10

More information on interacting with Wallbe chargers can be found at GoingElectric. Use with care.

NOTE: Opening the wall box must only be done by certified professionals. The box must be disconnected from mains before opening.

openWB slave

EVCC can be used to remote control an openWB charger using openWB's MQTT interface. Here is an example for how to use the default charger for controlling the first loadpoint:

chargers:
- name: openwb
  type: default
  status:
    # with openWB, charging status (A..F) this is split between "plugged" and "charging"
    # the openwb type combines both into status (charging=C, plugged=B, otherwise=A)
    type: openwb
    plugged:
      type: mqtt
      topic: openWB/lp/1/boolPlugStat
    charging:
      type: mqtt
      topic: openWB/lp/1/boolChargeStat
  enabled:
    type: mqtt
    topic: openWB/lp/1/ChargePointEnabled
    timeout: 30s
  enable:
    type: mqtt
    topic: openWB/set/lp1/ChargePointEnabled
    payload: ${enable:%d}
  maxcurrent:
    type: mqtt
    topic: openWB/set/lp1/DirectChargeAmps

Meter

Meters provide data about power and energy consumption:

  • CurrentPower(): power in W
  • TotalEnergy(): energy in kWh (optional)

Meter has a single implementation where meter readings- power and energy- can be configured to be delivered by plugin.

Vehicle

Vehicle represents a specific EV vehicle and its battery:

  • Title(): vehicle name for display in the configuration UI
  • Capacity(): battery capacity in kWh
  • ChargeState(): state of charge in %

Optionally, vehicles can also provide:

  • CurrentPower(): charge power in W (used if charge meter not present)
  • ChargedEnergy(): charged energy in kWh
  • ChargeDuration(): charge duration

If vehicle is configured and assigned to the charger, charge status and remaining charge duration become available in the user interface.

Available vehicle implementations are:

  • audi: Audi (eTron)
  • bmw: BMW (i3)
  • nissan: Nissan (Leaf)
  • tesla: Tesla (any model)
  • default: default vehicle implementation using configurable plugins for integrating any type of vehicle

Plugins

Plugins are used to implement accessing and updating generic data sources. When using plugins for write access, the actual data is provided as variable in form of ${var[:format]}. If format is omitted, data is formatted according to the default Go %v format. The variable is replaced with the actual data before the plugin is executed.

EVCC supports the following read/write plugins:

  • mqtt: this plugin allows to read values from MQTT topics. This is particularly useful for meters, e.g. when meter data is already available on MQTT. See MBMD for an example how to get Modbus meter data into MQTT.

    Sample configuration:

    type: mqtt
topic: mbmd/sdm1-1/Power
timeout: 30s
payload: ${var:%.2f}

    For write access, the data is provided using the payload attribute. If payload is missing, the value will be written in default format.

  • script: the script plugin executes external scripts to read or update data. This plugin is useful to implement any type of external functionality.

    Sample read configuration:

    type: script
cmd: /bin/bash -c "cat /dev/urandom"
timeout: 5s

    Sample write configuration:

    type: script
cmd: /home/user/my-script.sh ${enable:%b} # format boolean enable as 0/1
timeout: 5s

  • openwb: the openwb plugin is used to convert a mixed boolean status of plugged/charging into an EVCC-compatible charger status of A..F.

    Sample configuration (read only):

    type: openwb
plugged:
  type: mqtt
  topic: openWB/lp/1/boolPlugStat
charging:
  type: mqtt
  topic: openWB/lp/1/boolChargeStat

Background

EVCC is heavily inspired by OpenWB. However, I found OpenWB's architecture slightly intimidating with everything basically global state and heavily relying on shell scripting. On the other side, especially the scripting aspect is one that contributes to OpenWB's flexibility.

Hence, for a simplified and stricter implementation of an EV charge controller, the design goals for EVCC were:

  • typed language with ability for systematic testing - achieved by using Go
  • structured configuration - supports YAML-based config file
  • avoidance of feature bloat, simple and clean UI - utilizes Bootstrap
  • containerized operation beyond Raspberry Pi - provide multi-arch Docker Image
  • support for multiple load points - tbd