Defesa de Tese de Doutorado – Leonardo Salsano de Assis – 28/5/2020

26/05/2020 17:15
Defesa de Tese de Doutorado
Aluno Leonardo Salsano de Assis
Orientador Prof. Eduardo Camponogara, Dr. – DAS/UFSC
Coorientador Prof. Ignacio Grossmann, Dr. – Carnegie Mellon University (EUA)


28/5/2020  (quinta-feira) – 8h30

Videoconferência (




Prof. Eduardo Camponogara, Dr. – DAS/UFSC (presidente);

Prof. Leandro Magatão, Dr. – DAMEC/UTFPR;

Prof. Erlon Finardi, Dr. – DELT/UFSC;

Prof. Werner Kraus Junior, Dr. – DAS/UFSC;

Prof. Laio Oriel Seman, Dr. – UNIVALI.

Título Operational Management of Crude Oil Supply: models and solution strategies
Abstract: The supply of crude oil from offshore platforms to refineries is an important problem faced by vertically integrated oil companies which control production, transportation, storage and refining. In deep-water offshore oilfields, Floating, Production, Storage and Offloading units (FPSOs) produce and store crude oil which is transferred to an oil terminal by a fleet of shuttle tankers. Upon arrival at the terminal, a shuttle tanker unloads crude oil through a pipeline into Storage Tanks (STs). The crude oil is then pumped through a pipeline from the storage tanks to Charging Tanks (CTs), and subsequently sent to Crude Distillation Units (CDUs) at the refinery. This dissertation advances the state of the art on the management of crude oil supply by proposing models and algorithms to consider elements of the operational decision level in an integrated fashion, which leads to the Operational Management of Crude Oil Supply (OMCOS). OMCOS comprises both the upstream (i.e., platforms, vessels and terminal) and the midstream (i.e., CDUs at the refinery) segments. In relation to the technical literature, OMCOS combines elements of Maritime Inventory Routing (MIR) with Crude Oil Scheduling (COS) by considering decisions at the operational level (i.e., scheduling and crude oil blending) and tactical level (i.e., inventory control and resource allocation). Such an integration leads to non-convex Mixed Integer Non-Linear Programming (MINLP) models that are addressed in this dissertation. The main contributions are the following:

  • Chapter 2. An iterative two-step MILP-NLP decomposition algorithm, which implements a domain-reduction strategy for handling bilinear terms in the scheduling of crude oil operations (COS).
  • Chapter 3. A non-convex MINLP model for OMCOS that brings elements of the operational level into the management of crude oil supply, thereby incorporating elements of maritime inventory routing and crude oil scheduling. Further, an iterative MILP-NLP decomposition is presented to tackle the MINLP problem that relies on bivariate piecewise McCormick envelopes (to yield an MILP relaxation), domain reduction (to reduce complexity), and a NLP solver (to reach feasible solutions).
  • Chapter 4. A Mixed Integer Linear Programming (MILP) clustering formulation for OMCOS that offers the following benefits: (a) reduces the number of routes for the vessels; (b) simplifies offloading and unloading operations; (c) imposes rules for crude mixtures in clusters of storage tanks that minimize property variations; and (d) produces bounds on crude properties inside storage and charging tanks that are used to linearize the bilinear terms in blending constraints. Through the combination of clusters and a MILP-NLP decomposition, good solutions were obtained for a set of representative instances of OMCOS at a reduced computational cost.