Spark Your Imagination: Eneco's Virtual Power Plant • Wesley Coetzee & Sebastian Du Rand • GOTO 2024

Discover how Eneco built a Virtual Power Plant to manage 200k+ energy assets using microservices, event-driven architecture & smart optimization for grid balancing

Key takeaways
  • Eneco’s Virtual Power Plant (VPP) integrates and manages multiple electricity assets (solar panels, wind parks, electric vehicles, greenhouses) to help balance the power grid

  • Implementation uses microservices architecture, with key components:

    • VPP Core for asset management and dispatch
    • Aggregation Layer for pooling assets
    • Planning Domain for schedules/forecasting
    • Behind-the-meter optimization for household level control
  • Key technical solutions:

    • Outbox pattern for data consistency
    • Leader election pattern to prevent duplicate processing
    • Asset identifier as Kafka partition key
    • Caching of static asset data to reduce I/O
    • Azure Data Explorer (ADX) for time series data
  • Architecture choices prioritize:

    • Low-code approach for faster development
    • Fit-for-purpose technologies
    • Scalability for handling 200,000+ customers
    • Cost optimization
    • Reliability due to grid-critical nature
  • System capabilities:

    • Smart scheduling of asset charging/usage
    • Real-time telemetry processing
    • Weather forecast integration
    • Multiple solver implementations for optimization
    • Market participation management
    • Customer energy cost savings
  • Challenges addressed:

    • Data consistency across services
    • High volume telemetry processing
    • Duplicate message handling
    • Long-running workflow management
    • Asset data duplication