SOLARIS is the academic operating prototype of a 10 MW agrivoltaic facility - bringing IoT / Simulation Layer telemetry, an AI Decision-Support Module, a carbon marketplace and field operations together under one auditable record. The AI module does not replace the human operator; it supports decision-making with human validation on sensitive operations.
SOLARIS is built around a simple invariant: every reading, every decision and every credit can be traced from the sensor in the field to the auditor's screen - without manual hand-off.
SOLARIS can ingest field telemetry such as irradiance, soil moisture, panel temperature and yield. The academic prototype uses a simulation layer to represent IoT telemetry and device states, while the architecture remains compatible with future real-device integration.
The AI module does not replace the human operator. It supports decision-making by analyzing the available context and generating explanations or recommendations. Sensitive operations remain subject to user permissions and human validation.
Modeled credit and production indicators are recalculated from available measurements, confidence values and audit records, without presenting modeled estimates as live market facts.
Remote-sensing vegetation indices and on-site soil moisture are reconciled against PVsyst P50 models, so the yield curve matches what is actually growing under the panels.
Marketplace orders flow through a deterministic state machine - DRAFT, PENDING, CONFIRMED, PROCESSING, DELIVERED, CANCELLED, REJECTED - with payment-provider hooks enabled only when configured.
A Flutter app for Android and iOS gives field engineers the same data the operator sees in the browser: alerts, drone events, irrigation control, and the operations assistant - all under the same RBAC.
The same platform adapts to operators, investors, customers and field teams. Every data source and provider is labeled so users know what is operational, modeled, or not configured yet.
Track plant status, IoT devices, water, grid events, hydrogen assets and governed actions from one command center.
Review high-level production, modeled carbon indicators, marketplace activity and audit-backed operational progress.
Use marketplace orders, support tickets and assistant guidance without seeing internal security or governance controls.
Use mobile alerts, camera security mode and drone simulation with clear labels when a provider is simulated or not configured. The academic prototype uses a simulation layer to represent IoT telemetry and device states.
Four stages, each with its own data contract. No silent steps, no manual uploads, no spreadsheets in between.
The 10 MW Tajerouine agrivoltaic plant produces clean energy while olive cultivation continues under the panels.
Sensor and weather records are ingested through the backend telemetry model. The academic prototype uses simulated operational control with clearly labeled sources so modeled data is never shown as live production data.
The verification workflow reconciles measured output against PVsyst models and NDVI satellite data to compute carbon avoidance.
Approved listings and orders remain traceable through marketplace statuses, audit logs and optional payment-provider integration.
Battle-tested components, picked for operability and long-term maintainability rather than novelty.
SOLARIS is developed in partnership with BBC Group, with technical advisory from Feki Kacem. The codebase, the data contracts and the policy engine are all open to inspection inside the platform.