Brazil – October 27, 2026 -- Brazilian flat glass maker Vivix expects to cut furnace energy consumption by 3%-5%, extend furnace lifespan by one year and shorten its data collection and analytics cycle by 80%, based on a pilot of an executable digital twin (xDT) built with Siemens, Snowflake and NVIDIA.
Vivix targets blind spots in 1,550°C-1,600°C float-glass furnaces
The furnace is one of the most critical and hardest-to-observe assets in float-glass production. At operating temperatures of roughly 1,550°C to 1,600°C, teams have limited direct visibility into how combustion gases, molten glass and refractory walls interact. Vivix supplies glass and mirrors to the construction and furniture markets.
The xDT initiative initially covers three connected areas: refractory health, combustion efficiency and glass-melting quality. It combines engineering simulations, observed furnace conditions and AI models so teams can estimate conditions they cannot measure directly and test operating changes.
Siemens models run as lightweight executables inside Snowflake
High-fidelity Siemens Simcenter multiphysics models are converted through Simcenter reduced order modeling into portable executables. These are deployed natively in Snowflake's AI Data Cloud. The Simcenter Executable Digital Twin Manager handles the xDT lifecycle in a single solution.
Vivix brings IT and operational technology data, including physical sensor readings, together in a unified namespace in Snowflake. Engineering and operations teams can then compare observed furnace conditions with digital twin predictions rather than reviewing simulation results in isolation.
NVIDIA PhysicsNeMo drives a virtual sensor on AWS
Vivix uses NVIDIA PhysicsNeMo to build AI surrogate models that combine physics-driven causality with simulation and observed data. Teams review results through technologies built on NVIDIA Omniverse libraries. Deployed on Amazon Web Services, the hybrid system works as a continuous virtual sensor. It reconstructs volumetric heat distribution, gas velocity fields and thermal fluxes across refractory walls in near real time, without a constant connection to heavy simulation environments.
Twin targets refractory wear, gas consumption and defect rates
- Refractory preservation: The twin fuses real-time thermocouple streams with refractory thermal profiling to calculate localized wear rates and detect emerging hotspots hours in advance. The aim is to extend furnace campaign life before cold repairs are needed.
- Combustion efficiency: Airflow and thermal models evaluate burner power levels and heat distribution to optimize setpoints, lower natural gas consumption and reduce emissions.
- Melting quality: Models of thermal gradients and fluid flow in the molten glass bath show residence time distribution and melting completeness, which is meant to improve homogeneity and reduce defect rates.
Natural-language agents let operators run what-if scenarios
A Snowflake AI multiagent interface lets operators query the twin in natural language. Engineers can ask about optimal air velocities, evaluate what-if operating scenarios and receive root-cause diagnostics without putting the physical furnace in unknown or potentially unsafe conditions. The architecture was developed by Snowflake and Siemens DISW, combining Simcenter physics-based modeling with Snowflake Cortex AI analytics.