LONG-RANGE PLANNING FOR THE BRAZILIAN PETROCHEMICAL INDUSTRY USING MIXED-INTEGER PROGRAMMING

Abstract

A mixed-integer multiperiod model is used to propose a long-range planning for the Brazilian petrochemical industry. The specific aims of this work are to present a model to select the petrochemical processes capable of increasing over the next 20 years and carry out a sensitivity analysis of this selection if there are changes in demands for products or in the purchase prices of chemicals. The optimization model seeks to minimize the cost and the constraints related to demand, supply, and plant capacities in the petrochemical industry. The long-range planning was suggested for six scenarios, which differ according to the investment constraints, demands for products, and chemicals purchase prices imposed. Among the 295 processes considered, 26 were selected for investment in new process units through the optimization procedure in all scenarios, such as the processes for the production of polyurethane and acrylic fibers. These results are consistent with previous studies of investment assessment of the Brazilian petrochemical industry.