Otimização da programação de curto prazo de duto bidirecional de derivados de petróleo. / Short-term scheduling optimization of derivative petroleum bidirectional pipeline.
AUTOR(ES)
Marcelo Kenji Hassimotto
DATA DE PUBLICAÇÃO
2007
RESUMO
Pipeline systems play a major role in the supply chain of the petroleum industry. These systems are responsible for the transportation of most of the crude oil and petroleum derivatives. Pipeline systems transfer large amounts of different petroleum types and their products at a lower cost than any other transportation mode. Pipelines interconnect oil fields, ports, refineries, distribution centers (or depots), and consumer markets. The problem addressed is this work is based on a system that is composed by an oil refinery that must transfer multiple products through a single pipeline connected to one depot. The products are a set of petroleum derivatives that must be either transported from the refinery to the depot or from the depot to the refinery. Both depot and refinery also connect other refineries as well as other depots and customers, thus forming a complex transportation network. On the other hand, there are several demands that arise either from external customers or from refineries. The latter demand is due from the need of processing intermediate streams with components mixtures such as diluents, propane and diesel. In order to take advantage of the structure of the transportation network, it becomes beneficial and even necessary to operate the pipeline in both directions so that internal and external demands are satisfied. The objective of this work is to develop a mathematical model for the short term scheduling of a multiproduct pipeline system. The scheduling formulation must account for the bidirectionality of the multiproduct pipeline. In this context, the scheduling a multiproduct pipeline involves the from-to decision, the product amounts, their sequence and timing, in the optimal sense, considering all constrains on demands, production rates, inventories, and costs. The scheduling model is based on a discrete time representation and is composed by one refinery tank farm, one depot and one multiproduct pipeline. Moreover, the pipeline is divided into segments of equal volume, as in Rejowski Jr and Pinto (2003). The main decisions variables are the directions of transfer (refinery to depot or depot to refinery) and the types of products at each time interval. These decisions are formulated with a disjunctive representation. The disjunctions are represented in mixed integer formulation based on the convex-hull approach. The objective function involves inventory, transfer and product interface costs. The model is first applied to a prototype case and after applied to a real-world system that is composed of the São Sebastião and Guararema depot and the OSPLAN pipeline. Overall four families of products are transported: gasoline, kerosene, naphtha and oil diesel. These are scheduled over a period of one week.
ASSUNTO(S)
mathematical programming otimização matemática pipeline transportation programação mista programação matemática optimization mixed-integer programming transporte dutoviário dutos pipeline
