Otimização do planejamento do transporte de derivados claros de petróleo em uma malha dutoviária

Abstract

Solution approaches for the planning module of the optimized scheduling of a real-world pipeline network to transport light oil derivatives are presented in this work. The proposed model is the input block of a decomposition strategy which aids the scheduling of products (derivatives with high earned value) transportation. The referred network is composed of 14 areas (4 refinaries, 2 harbors, 2 final clients and 6 depots) and 30 bidirectional pipelines which interconnect those areas transporting more than 35 light oil derivatives. The related problem is complex, thus the proposed model is part of a decomposition strategy, formerly proposed by Boschetto et al. (2010), based on the integration of distinct modules composed of Mixed Integer Linear Programming (MILP) models and heuristics. This integration is able to solve the aforementioned problem, considering all the particular network characteristics. Briefly, among the developed modules are the: (i) Planning, (ii) Assignment and Sequencing and (iii) Timing blocks. Solutions are obtained in a reduced computational time (seconds to few minutes) for real scenarios. This work is an improvement of the MILP model introduced by Magatão et al. (2012), considering new operational characteristics, and it aims to detail the Planning Model which is responsible for determining total volumes that will be transported in the network and the routes that will be required for this transport. Additionally, a division of the new proposed model in distinct stages is made to investigate the control and influence of each variable in both planning and scheduling solution scopes. Therefore, the proposed approaches are defined here as the single-phase approach and the three-phase approach. The continuous-time MILP model gives global decisions at a tactical level taking into account operational characteristics for a 30-day programming horizon. It is intended to reach an optimized mass balance considering operations such as: pipeline reverse flow, contamination between products, surge operation/temporary stock, tank maintenance and degradations. The approaches are compared and results such as computational time reduction, lower inventory violations at the end of the scheduling and fewer pipeline reverse flow operations are presented.