(443f) Revamping a Multivariable Predictive Control Application of a Gas Condensate Fractionation Unit

The oil refining industry relies on the multivariable predictive control (MVC) applications to improve stability and efficiency of their process units. The effectiveness of the applications can degrade over years as the process and economic conditions change. This paper illustrates how an existing MVC application was revamped to regain its effectiveness and recapture the economic return in a gas condensate fractionation unit of Aramco’s Ras Tanura oil refinery.

Ras Tanura Refinery is the largest refinery inside Aramco with a crude distillation capacity of 550,000 BPD. Its gas condensate distillation unit processes 225 BPD Khuff gas condensate.   The original MVC application, using Honeywell RMPCT technology, was successfully designed and implemented in 2004 to maximize more valuable diesel product. Over years, the effective utilization and economic return had gradually faded away as process and economic conditions changed. A revamping project for revising and updating the existing application was executed in 2012. This paper reviews the improvement made as result of this revamping effort.

Following a brief overview of the process, base level control loop configuration and special consideration are reviewed. Major process constraints and economic considerations are discussed. Steps for collecting plant data, and identifying inferential and control models are outlined.  The control strategies, including control and handling of major constraints and key economic variables, are present. Finally, operating data about product yield and quality before and after re-commissioning of the revamped application are compared to demonstrate the economic benefit.

The primary objective of this presentation is to share our experience in applying the MVC techniques to optimize unit operation and improve reliability in oil refining industry.