O Professor LEONIDAS SAKALAUSKAS (Vilnius University, Lituânia) irá proferir palestra com o tema "OIL PRODUCTION OPTIMIZATION UNDER UNCERTAINTY".

Local: Auditório da COPPE/UFRJ - Sala G-122

Data: 28 abril 2017 (sexta-feira) - 10 horas

Abstract

The oil production optimization model is developed seeking for profit maximization. Costs related to gas lift and investment decisions are included in the model, making it more realistic. Decisions are made by varying gas lift flow rate, well routing and water/gas injection flow rate. The uncertainty related to demand and oil price as well as that with technological parameters of gas lift are also included in the mathematical model, by introducing stochastic programming approach. A stochastic formulation is highly relevant when maximizing profit, because market and technological parameters are uncertain, and, thus, are implemented as stochastic parameters to obtain a reliable model. The solution method consist of joint modification of Benders decomposition and Branch& Bound method (Branch & Price), which is able to guarantee global optimum. The resulting mixed integer nonlinear problem (MINLP) is implemented by means of GAMS, the solver CPLEX and an user interface, which allows us to combine the simulation of data and the running of an optimization model into one program. Issues of practical implementation of the model in Brazilian corporation Petrobras are discussed.