Bitumen Penetration Prediction

about
Our client is the largest petrochemical company in Turkey, operating four crude oil refineries.
Problem
  • Heavy vacuum gas oil (HVGO) is a valuable end-product where bitumen penetration impacts the yield of the product
  • Accuracy of the current statistical model to predict bitumen penetration is not satisfactory to increase HVGO yield
Action
  • Analyzed the crude oil distribution network to identify the main source of complexity and mathematically described the network
  • Developed an advanced analytics model to more precisely predict bitumen penetration level
  • The model was integrated into client’s system and process with dashboard visualization
Tool Stack
  • Phyton

20

K Ton/Y

Additional HVGO produced

2

M $/Y

Additional revenue