A Cost-Effective Model for the Gasoline Blend Optimization Problem

JAIME CERDÁ; PEDRO C. PAUTASSO; DIEGO C. CAFARO

AICHE JOURNAL

JOHN WILEY & SONS INC

Lugar: New York; Año: 2016 vol. 62 p. 3002 - 3019

0001-1541

Gasoline blending is a criticalprocess with a significant impact on the total revenues of oil refineries. Itconsists of mix-ing several feedstocks coming from various upstream processesand small amounts of additives to make different blends with some specifiedquality properties. The major goal is to minimize operating costs by optimizingblend recipes, while meeting product demands on time and qualityspecifications. This work introduces a novel continuous-time mixed-integer linearprogramming (MILP) formulation based on floating time slots to simultaneouslyoptimize blend recipes and the scheduling of blending and distributionoperations. The model can handle non-identical blenders, multipurpose product tanks,sequence-dependent changeover costs, limited amounts of gasoline components,and multi-period scenarios. Because it features an integrality gap close tozero, the proposed MILP approach is able to find optimal solutions at muchlower computational cost than previous contributions when applied to largegasoline blend problems.