PIL’s Innovative AI-Driven Real-Time Optimisation at Chinese Refinery

Background: A Chinese refinery processing 4 million tons of crude per year, sought to enhance its crude distillation unit (CDU) operational efficiency and increase its economic benefits. The existing systems could not perform real-time data reconciliation and optimisation across the entire production process.

Challenges: This project presented a great challenge due to the large variety of crude oils processed, significant fluctuations in the blended crude oil properties, and frequent changes in the operation scenarios.

In 2022 alone, 36 types of crude oil were imported and processed by the CDU unit. This variety of processed crude resulted in two scenarios of high-sulphur and low-sulphur as the operational conditions needed to be changed drastically to accommodate the different sulphur-content crudes. For the side products, each side draw had an average of 2 to 3 different downstream units, with a total of 24 operation scenarios.

Additionally, the changes in processing load, equipment performance and unit operation conditions and performance throughout the operation cycle needs to be considered.

Our Solution: The refinery implemented PIL’s AI- driven CDU product (i-CDU) along with an intelligent control system (iES). The system utilised artificial neural networks (ANN) modelling to create a digital twin of the CDU process for real-time data reconciliation and optimisation. Our system can accurately describe the dynamic production process in real-time and detect operational changes. It runs operational optimisation in a closed-loop settings, ensuring each unit will run steadily, product yield be enhanced and energy consumption be reduced.

Economic & Energy Saving Benefits:

  • Product yield Optimisation: After the product being deployed, the yields from the first and second side-draw of the atmospheric column has been increased by 0.78%, and 0.61% respectively, whilst for the vacuum column the yield of the first side-draw increased by 0.2%.
  • Reduction in Operating Costs: The optimisation reduced the atmospheric furnace outlet temperature by 1.07°C and vacuum furnace outlet temperature by 0.85°C. This led to a reduction in fuel gas consumption and lower utility costs, decreasing the unit’s energy consumption by 0.18 kg of standard oil per ton of crude oil.
  • Overall Energy Savings: Steam consumption reduced by 0.5 kg/ton and fuel gas consumption reduced by 0.4 kg/ton, providing an additional benefit of 1.27 GBP /ton.
  • Over seven months, this resulted in a reduction of 2700 tons of CO₂
  • The optimisation provided an annual economic benefit of 2.8 million GBP/year