SolarGrid
Solar farm, Romania
What needed
solving
A 5MW solar farm in Romania was relying on daily manual performance reviews of inverter output data to identify underperforming panels and faults. On average, faults went undetected for 2–3 days before a field visit was scheduled, during which time the affected string was generating minimal or no power. Given the revenue model of the farm (feed-in tariff), each day of reduced output had a direct financial cost. The operations team also had no predictive visibility into maintenance requirements.
What we built
We built a panel-level IoT monitoring system that integrates with the farm's inverter network and string-level monitoring hardware. Performance data is collected every 5 minutes per string and compared against the expected output model based on current irradiance and temperature conditions. Deviations trigger automated fault classification (soiling, shading, inverter fault, panel degradation), and a maintenance alert is dispatched to the field team with exact string location, estimated output loss, and recommended action.
The numbers
Concrete, measurable outcomes delivered in production.
From brief to production
Discovery & scoping
We map the problem, understand the existing stack, and write a fixed-price proposal. No surprises.
1–3 DAYSBuild & weekly demos
You see working software every week. Feedback is incorporated immediately — not at the end.
2–6 WEEKSTesting & QA
Every integration is stress-tested before go-live. Edge cases handled, fallbacks in place.
3–5 DAYSGo live & handover
Full documentation, team training, monitoring setup, and 30-day post-launch support.
1 WEEKBuild something similar?
Book a free 30-minute call. We'll assess your situation and tell you exactly what we'd build — and what it would cost.