Virtuel
We took over an unfinished system from several previous vendors, brought it into stable production over three years and then handed it over to the client's in-house development team. Virtuel now controls decentralized energy sources and optimizes the operation of aggregation blocks.
Project description
Virtuel is software for controlling and managing decentralized energy sources. It monitors the performance of energy-producing assets, resolves outages before they affect supply, and forecasts future production for more reliable planning.
Starting point
We took the system over from several previous vendors. Each had left behind different approaches, incomplete documentation and parts of the solution that were not ready for live operation.
What we did over three years
Over three years we consolidated the codebase, stabilized the architecture and steadily developed the core capabilities: dispatch control, preparation and activation of sources in aggregation blocks, availability forecasting and reactive outage handling. The system communicates with sources over the industrial Modbus protocol and uses machine-learning models.
Outcome
An unfinished solution became a production system the client relies on in the day-to-day operation of its energy assets. We then handed it over to the client's in-house development team, which continues to develop it further.
Tech stack
The platform combines a real-time control layer with a data and ML layer that optimizes aggregation blocks.
- Python for data processing and machine-learning models
- Node.js / TypeScript and React for the control interface
- Modbus protocol for communication with energy sources
- PostgreSQL and time-series storage for measurement data
- Docker and Kubernetes for deployment