(640f) Refinery-Wide Optimization by Using Novel Model Reduction Techniques

Abstract: The oil refinery remains a key component of the global oil production and distribution process. It consists of an industrial process plant where crude oils are processed and refined into more valuable petroleum products. These products should meet certain commercial specifications before being shipped to the domestic and international markets. The oil refining industry is currently facing new challenges. In fact, there is a change in fuel demand towards lighter products, as well as a shift in the quality of crude oils towards more sour and heavy crudes. Furthermore, environmental restrictions are leading to tighter fuels specifications that require refiners to either invest in new refinery technologies, usually very expensive, or use better crude feedstock to improve the quality of the products downstream. Heavy sour crudes are more difficult to process and yield more heavy products which need additional processing. Challenges from the operational point of view include the selection of crude oils to be fed to the refinery in order to achieve higher margins and to meet the final product specifications, impacting the cost per barrel. Oil refineries range from a simple topping refinery to a deep conversion refinery integrated with a petrochemical plant having a combination of various technologies for heavy ends conversion, product quality blending as well as efficient fuel usage etc.

The complexity of oil refineries operations and configurations makes the decision making process an extremely difficult task, especially when uncertainty in feed specifications, product demand and economic parameters is involved [1]. Refinery wide optimization is a key requirement to overcome these challenges to achieve a high standard of performance and to stay competitive within the market. Through refinery-wide optimization optimal refinery crude feed mixture as well as gasoline and distillates blending can be achieved. It can also be used to minimize fuel production and quality giveaway whilst ensuring a better energy management.

The oil refining industry mainly uses Linear Programming (LP) modelling tools for refinery optimization and planning purposes, on a frequent basis. LPs are attractive from the computational time point of view; however these models have limitations such as the nonlinearity of the refinery processes is not fully taken into account. In addition, building the LP model can be an arduous task that requires collecting large amounts of data. In this work, approximate nonlinear models are developed in order to replace the rigorous ones providing a good accuracy without compromising the computational time, for refinery optimization. The data for deriving approximate models has been generated from rigorous process models from commercial software, which is extensively used in the refining industry. In this work we present novel model reduction techniques based upon optimal configuration of artificial neural networks [2] to derive approximate models and demonstrate how these models can be used for refinery-wide optimization.

Acknowledgment: Financial support from EPSRC (EP/G059195/1) is gratefully acknowledged.

References: [1] Slaback, D.D. and Riggs, J.B. (2007) The inside-out approach to refinery-wide optimization, Industrial and Engineering Chemistry Research, 46, 4645-4653. [2] Dua, V. (2010) A mixed-integer programming approach for optimal configuration of artificial neural networks, Chemical Engineering Research and Design, 88, 55-60.