Energy Distribution
An operational platform for energy distribution - real-time dashboards, alerts, remote commands for field devices, and ANEEL compliance (DRP/DRC) treated as a first-class datum.
Operating an energy grid means seeing the field in real time and acting on it with confidence.
Energy distribution in Brazil is a regulated, continuous operation. Devices spread across the grid report state all the time, and the operations center needs to see it now - not in tomorrow's report. When voltage drifts out of the adequate range, the distributor has an obligation to ANEEL and the consumer gets automatic compensation on their bill.
The engineering challenge is twofold. Get live state from many devices onto a screen without choking the browser, and let the operator send a command back to the field with the certainty that it arrived, arrived exactly once, and was recorded - because here the consequence is physical on the grid.
Full-stack engineer: backend services in NestJS and the React operations console.
- - Backend services in NestJS for telemetry ingestion, indicator computation, and alarm dispatch.
- - A two-way WebSocket channel between the operations center and field devices, with reconnection and state reconciliation.
- - React operations console: real-time dashboards, alarm list, and remote command dispatch with confirmation.
- - Modeling the ANEEL indicators (DRP/DRC, DEC/FEC, DIC/FIC/DMIC) so compliance was a first-class datum, not an end-of-month report.
Understand the regulatory vocabulary before modeling any data.
DEC, FEC, DIC, FIC, DMIC, and especially DRP/DRC are not internal metrics - they are ANEEL obligations with legal limits. Getting the domain wrong here costs fines and consumer compensation.
Define the telemetry contract between field and backend.
Field devices emit state continuously. I decided the message format, the aggregation window, and what becomes an alarm event versus what is just a time-series reading.
Stream live state to the console without choking the browser.
An operations center with many devices produces more updates than a screen can take. Backpressure, reconnection with backoff, and state reconciliation after a drop became day-one requirements, not a detail.
Treat remote commands as a serious operation, not a button.
Sending a command to field equipment needs confirmation, idempotency, and an audit trail. A lost or duplicated command has a physical consequence on the grid, not just an ugly log line.
The operator now sees the grid in real time and acts on the field from the screen, with ANEEL compliance built into the flow instead of computed at month's end.
- - Live telemetry gave the operator the real state of the grid in seconds, not in a spreadsheet the next day.
- - Alarms derived from DRP/DRC anticipated voltage transgressions before they turned into consumer compensation.
- - Backpressure and reconnection kept the console stable even with many devices reporting at the same time.
- - Remote commands with confirmation and auditing gave operations the confidence to act on the field from the screen.