SKY | Cloud detection for remote PV sites

CALIBSUN

Services

SKY: Cloud Detection

Cloud detection for isolated industrial sites with PV

FOR ISOLATED INDUSTRIAL SITES

SKY optimizes the incorporation of photovoltaics for self-consumption in isolated industrial sites

Anticipating power drops due to cloud passages above the plant is essential to maximize solar energy penetration while maintaining a stable electrical grid.

Hybridization, which involves mixing energy from a photovoltaic plant (PV) with other sources (generators, batteries, hydrogen, the electrical grid, etc.), poses many technological challenges to ensure a constant and stable electricity supply. SKY, the cloud detection tool through sky observation, anticipates power drops associated with cloud appearances. It optimizes the integration of photovoltaic energy into the grid and the management of storage units, particularly important for isolated industrial sites such as mines or island areas.

TO ANTICIPATE POWER DROPS

SKY models clouds in 3D to anticipate power drops in hybrid isolated sites

SKY anticipates cloud passages, even those not detected by satellite imagery, to anticipate the activation of alternative energy sources to PV.

Since PV production is variable and directly linked to available solar resources, it can experience significant production drops during cloudy periods. In the case of an isolated industrial site powered by a self-consumption solar plant, it is essential to predict short-term PV production variations to reactively anticipate the activation of alternative energy sources.

Discover our project in Mali

From Sky observation to Energy management

HEMISPHERIC SKY OBSERVATION CAMERAS

Illustration de balisage d'une zone avec des caméras hémisphériques

Real-time cloud formation detection with multiple sky observation cameras

SKY offers a unique sky observation technology that detects and captures cloud images from different angles through triangulation.

CalibSun installs a network of several hemispheric cameras at strategic locations in the photovoltaic plant. The cameras capture real-time images of approaching cloud cover.

ARTIFICIAL INTELLIGENCE SIMULATION

Illustration d'une chute de puissance de tension électrique

3D cloud modeling and real-time simulation of their impact

An innovative algorithm using a geometric approach, statistical learning, and artificial intelligence simulates the speed, altitude, and opacity of cloud formations in real-time by modeling them in 3D.

These data are transmitted to algorithms combining geometric analytics and machine learning capable of determining the speed, altitude, and shape of clouds to simulate shading and deduce the timing and intensity of the photovoltaic energy production drop. The AI also enables real-time processing of hemispheric sky observation images, via a deep neural network, to associate them with a minimum photovoltaic production level.

GRID MANAGEMENT

Illustration du démarrage ou de la décharge des sources alternatives

SKY automatically informs the EMS and optimizes grid management for self-consumption

The service informs the EMS (energy management system) to activate alternative sources (batteries, generators, etc.) necessary to maintain industrial activity.

SKY predicts the decrease or increase in photovoltaic yield minute by minute. The energy management system (EMS) can then send an automatic command or issue an alarm to anticipate the start of alternative sources or the discharge of storage. With up to 30 minutes anticipation of shading on the plant and their consequences on the production drop, SKY is used in installations with or without storage.