Equacionamento de rendimentos em fornos de pirólise

Abstract

In order to maximize financial results Braskem uses optimization software to generate monthly plans for raw material purchases, production and meeting demand. The software models the main equipment’s and flows of all its olefin and aromatics units (1st generation) and resin plants (2nd generation). The quality of the model is essential for the quality of the generated plan, and one of the main points in it are the pyrolysis furnaces since their output will feed all the rest of the operation. The equations in these furnaces are based on linear regressions with product yields as a function of naphtha qualities (olefin content, aromatics content, density, etc.) and furnace operating conditions (hydrocarbon recycle flow, temperature, pressure, etc.). Thus, when major changes occur in the raw materials received or in the furnaces, it is necessary to update the equations to maintain their representativeness. The study occurred because since the last revision of the equations Braskem adopted a policy to increase the use of imported raw materials that have distinctive characteristics from the domestic ones and, therefore, generate a gap in the region of validity for which the equations were initially modeled. The objective of the study was to find a new set of yield equations for the different olefin and aromatic products generated in the cracking process considering in its modeling the changes that occurred in the raw materials processed and in the characteristics of the furnaces of the unit in question. During the study, the following steps were conducted: obtaining the quality of the naphtha in the laboratory; definitions of the Naphtha to be simulated and the operational conditions of the plant; simulation of the yields by the SPSL SPYRO software. Determination of the equations by the SCMart Suite software and their implementation in the same. With this study it is expected that the new equations found will better represent the real yields obtained in the plants in comparison with the previous equations, thus guaranteeing more confidence and quality in the monthly optimization plans conducted.