Skip to content

About

Distributed EV charging simulation with Python services, Kafka telemetry, PostgreSQL, Docker Compose and a web dashboard.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

Distributed EV Charging Platform

A local simulation of an electric vehicle charging network, based on the final sd_pract2 implementation. It began as a Distributed Systems project at the University of Alicante.

Python services register a charging point, coordinate driver requests, send encrypted telemetry through Kafka, store records in PostgreSQL, update weather conditions, and provide a web dashboard.

Architecture

flowchart LR
    D[Driver simulator] -->|authorization request| C[Central service]
    CP[Charging point] -->|registration| R[Registry]
    R --> PG[(PostgreSQL)]
    CP -->|login and callback URL| C
    C -->|authorize, start, stop| CP
    CP -->|encrypted telemetry| K[(Kafka)]
    K -->|telemetry| C
    C --> PG
    W[Weather service] -->|temperature and alerts| C
    O[OpenWeather API, optional] --> W
    F[Web dashboard] -->|status and controls| C
    Z[ZooKeeper] --> K
Loading
  • Registry (ev_registry) registers charging points, issues tokens, and stores their records.
  • Central service (ev_central) manages charging point sessions, driver authorization, dashboard controls, tickets, audit logs, weather alerts, and Kafka telemetry.
  • Charging point (ev_cp) registers as MADRID-01 by default, receives central service callbacks, simulates charging, and publishes encrypted telemetry.
  • Driver (ev_driver) makes simulated authorization requests.
  • Weather (ev_weather) polls active locations and sends temperatures and low temperature alerts. It uses simulated values when no OpenWeather key is set.
  • Dashboard (ev_front) shows charging point status, weather, tickets, and controls through the central service.

PostgreSQL stores charging points, tickets, and audit logs. Kafka carries charging telemetry; ZooKeeper supports the Kafka broker. The central service also keeps active sessions and current weather in memory.

Run locally

You need Docker and the Compose plugin. From the repository root:

cp .env.example .env
# Set a unique local DB_PASSWORD in .env
docker compose up --build

On Windows PowerShell, copy the example with Copy-Item .env.example .env. Open http://localhost:5000 after startup. The Compose file starts one charging point (MADRID-01) automatically. The driver starts only when you run a manual command. Allow time for the database, Kafka, registration, and login to become ready; the charging point retries registration and login. Driver requests authorize a connection; starting and stopping a charge are separate dashboard actions.

For one manual driver request, use:

docker compose run --rm ev_driver python app.py --user Alice --cp MADRID-01

That command waits for Ctrl+C after a successful request and then asks the central service to disconnect. To stop the stack, run docker compose down. To reset its database, run docker compose down -v, which discards local simulation data.

The dashboard, central API, registry, Kafka, and PostgreSQL are bound to 127.0.0.1 on the host. No private LAN address is needed for this Compose workflow.

Manual multi-point workflow

The original sd_pract2 workflow started charging points and drivers in separate terminals. The public Compose file starts one default point for a quick demonstration. Run drivers manually when you want a charging request, and add more points while the base services are up.

In a new terminal, start a second point. Keep this terminal open:

docker compose run --rm --no-deps --name ev-cp-sevilla -e CP_CALLBACK_HOST=ev-cp-sevilla ev_cp python app.py --id SEVILLA-03 --city Sevilla --price 0.45

The container name and CP_CALLBACK_HOST must match so the central service sends commands to this point, rather than to the default ev_cp. Give every additional point a unique container name and --id. The point registers itself, stores its location, and logs in automatically. Its city then appears on the dashboard and in the weather service's location list.

In another terminal, simulate a driver requesting that point:

docker compose run --rm --no-deps ev_driver python app.py --user Ana --cp SEVILLA-03

The request authorizes a connection. Use the dashboard at http://localhost:5000 to start and stop charging and inspect the ticket. The manual driver waits for Ctrl+C to disconnect.

The weather service normally polls registered cities automatically. To force one temperature without it being overwritten by the next poll:

docker compose stop ev_weather
docker compose run --rm --no-deps ev_weather python app.py --city Sevilla --temp 18

A temperature below 0 °C sends a low temperature alert and may stop charging. Run docker compose start ev_weather to resume automatic updates. These commands use Docker's internal network, so they do not require a private LAN address or host-side Python setup.

Configuration

.env.example contains placeholders only. Set DB_PASSWORD in your ignored local .env. OPENWEATHER_API_KEY is optional; leave it empty to generate simulated temperatures. With a key, failed provider requests do not generate substitute readings. Do not reuse a key that has previously been exposed.

Charging point tokens are generated during registration. Local token_*.json files, .env, local databases, Python caches, and virtual environments are ignored by Git. The container demo does not preserve its generated token across a fresh charging point container; resetting only the container while retaining the database may require a database reset or re-registration.

Run docker compose config --quiet after setting DB_PASSWORD to validate the Compose configuration.

Repository structure

Path Role
docker-compose.yml Local service topology
db/init.sql PostgreSQL schema
ev_common/ Shared configuration
ev_registry/, ev_central/ Registration and coordination APIs
ev_cp/, ev_driver/ Charging point and driver simulators
ev_weather/ Weather polling and alerts
ev_front/ Flask dashboard
tests/ Focused weather behavior tests

Scope and limitations

This is a local educational simulation with one charging point in the default Compose setup. Startup depends on several services becoming ready, and the Kafka consumer has limited recovery if the broker is unavailable at its initial connection. Active sessions and weather state are held in memory. The HTTP APIs and local database configuration are intended for development and should not be exposed to the internet.

About

Distributed EV charging simulation with Python services, Kafka telemetry, PostgreSQL, Docker Compose and a web dashboard.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages