Vivix, a Brazilian flat‑glass maker, is piloting an executable digital twin of its high‑temperature furnace. The twin combines Siemens Simcenter physics models, NVIDIA AI surrogate models and Snowflake’s data platform, running on AWS. It predicts heat distribution, gas flow and refractory wear in near real time, letting operators test “what‑if” changes without stopping production. The company says the pilot could reduce furnace energy consumption by 3‑5%, extend the furnace campaign by one year and shrink data‑to‑insight time by 80%.
Why it matters
If the savings hold, Vivix can lower production costs and emissions while keeping glass quality high.
