AN MINLP FORMULATION FOR PLANNING QUALITY-SENSITIVE CRUDE OIL SUPPLIES THROUGH LONG-DISTANCE PIPELINES

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

Bahía Blanca

Congreso; IX Congreso Argentino de Ingeniería Química (CAIQ 2017); 2017

Asociación Argentina de Ingeniería Química

In this work we present a novel mixed-integer nonlinear programming (MINLP)formulation for planning crude oil supplies to an oil refinery. We precisely monitor keycomponent concentrations keeping oil contamination below threshold values. The proposedMINLP formulation is based on a hybrid approach, combining the potentials of slot-basedand general precedence continuous-time representations. General precedence sequencingvariables are used to coordinate incoming/outgoing flows to/from every tank, and wesimultaneously address the scheduling of a long-distance pipeline sequentially transportingcrude oil batches from storage to charging tanks, following a slot-based scheme. Wedetermine the best sequence and timing of transportation operations in pipeline and tanks.Basic input data consist of the amount of oil and composition in arriving vessels and tanks,the total demand of crude oil to be satisfied, flow-rate limits and cost coefficients. Ournovel approach is able to precisely calculate inventory holding costs while it constitutes, tothe best of our knowledge, the first method proposed in the literature that does not requirerigorous inventory tracking over time for so doing. We propose an MINLP solutionalgorithm that sequentially solves a couple of mathematical programming formulations: (a)an MILP continuous-time model that is a tight approximation of the MINLP, and (b) anNLP model that includes the nonlinear terms arising in the key component mass balancesand inventory carrying cost calculation. Compared to previous contributions, up to 28%cost savings are obtained in much shorter CPU times.