(72m) Optimisation of Refinery Diesel Blending | AIChE

(72m) Optimisation of Refinery Diesel Blending


Jiang, S. - Presenter, Center for Process Intergration



Shixun Jiang, Nan Zhang

Centre for process integration, University of Manchester

Oxford Road, Manchester M13 9PL, UK

In recent years, modern refineries have been confronted with higher prices of petroleum, more stringent environment standards and stricter requirements of quality specifications. To achieve the objective of maximizing the profit, it brings great challenge toward planning and scheduling for refinery operators.

Diesel, one of the main petroleum products, has a growing market proportion because of its higher energy utilization ratio and therefore lower carbon dioxide emissions compared with gasoline. Diesel is widely used in the industry and transportation. Only high quality diesel product can survive in the more and more competitive market. Refiners haven’t stopped improving the methodology in diesel production and management.

To maximize the profit or minimize the cost is the main driving force in refinery diesel management. It is difficult to achieve the optimal recipe towards the site level as an optimal recipe for one single process may be suboptimal for other processes or even infeasible for the overall refinery. A multi-period planning model for refineries, which was based on the decomposition method, was proposed by Dave in 2004, to optimise decisions at different time intervals. Furthermore to optimise the recipe, refinery operators have to deal with volatile prices of crude oil and products, fluctuant market demand, uncertainty of feedstock arrivals and other abrupt changes. Sourabh(2008) demonstrated an “Operating Window” algorithm, which decomposed an MINLP problem into an NLP model and an MILP model.

In this work, a model for management of refinery diesel streams has been developed to optimize the diesel production of a refinery. A modified property estimation model for diesel, which is more accurate than traditional linear models, is employed for off-line diesel blending processes. Intricate environment regulations and product specifications, such as cetane number, sulphur content and pour point, etc. are taken into account. A novel scheduling method, which involves an MINLP model, is optimized by a robust solving algorithm. By adopting the method, the optimal solution of schedule a refinery diesel streams can be reached. Case studies have been carried out to demonstrate the effectiveness of the developed method.


Dhaval Dave, 2004. Planning and scheduling of refinery operations. Ph.D Thesis, University of Manchester, Faculty of Engineering and Physical Sciences.

Sourabh Gupta., 2008. Planning and scheduling of refinery operations. Ph.D. Thesis, University of Manchester, Faculty of Engineering and Physical Sciences.