Case Study: UNIPETROL achieves real-time quality prediction and millions in annual savings with STATISTICA

UNIPETROL - Customer Case Study

Unipetrol RPA, a leading Czech refinery and petrochemical producer, needed to maintain strict product composition during manufacturing but relied on time‑consuming, costly laboratory sampling and analysis. The company sought a “virtual sensor” to predict quality parameters in real time from routine process data so operators could act immediately instead of waiting for lab results.

Unipetrol deployed STATISTICA Enterprise with Automated Neural Networks and a Code Generator to build models from historical data and export them to Excel for live use on the plant floor. The solution was straightforward to implement after brief training, letting operators get immediate composition estimates, reduce sampling frequency, stabilize quality, minimize production losses, and save several million Czech crowns annually.

UNIPETROL