Born on the Discharge Chute of Plant #3
In February 2024, our founders—materials engineers from HUCE and AI researchers—stood under the cold Hanoi rain as an 800m³ continuous foundation pour cracked from undetected aggregate moisture swings.
The industry was still relying on Abram's slump cone invented in 1918. Technicians manually measured one cone out of twenty trucks, while 28-day cylinder compressive strength tests meant you only knew if the concrete failed weeks after the skyscraper had already advanced three stories.
We asked a radical question: What if concrete could report its own rheology in 300 milliseconds before leaving the plant? We mounted ruggedized 120fps high-speed cameras on the discharge chute of Trạm Trộn Bê Tông Số 3, wired acoustic transducers to transit drums, and wrote the first neural models for fluid aggregate dynamics.
Conventional Construction
- Slump cone test samples only 0.05% of concrete volume
- 28-day hydraulic press wait before knowing true strength
- Static batching recipes unable to adapt to wet sand variations
- Thermal cracking risks on thick foundation mass pours
XN Betong AI Platform
- 100% continuous chute video telemetry in 350ms
- Physics-Informed Neural Network predicts 28D curve in real-time
- Dynamic batch moisture compensation saving 15-20% clinker
- Digital-twin maturity tracking eliminates formwork guesswork