Otimização da programação das limpezas em redes de trocadores de calor utilizando técnicas de programação matemática

Abstract

Heat exchanger networks are widely used in industry to promote heat recoveryby making hot and cold streams to exchange heat, however these networks are subjected to fouling, which causes an increase in heat transfer resistance, jeopardizing it. One of the main techniques used to diminish the effects of fouling is to perform periodical cleanings in the heat exchangers. This work s goal is to develop a new method to find the optimal cleaning schedule for a heat exchanger network. The developed method uses a sliding horizon concept associated with a mixed integer linear programming problem. This MILP problem is capable of finding the optimal set of heat exchangers that must be cleaned at a certain instant (first instant of the sliding horizon), considering its influence on the further instants (the rest of the sliding horizon). The MILP problem consists of constraints corresponding to energy balances, heat exchanger equations and maximum number of cleanings.The problem objective function represents the energy consumption that must be minimized. The optimal cleaning schedule is formed by the combination of the results to each one of the time instants. The proposed approach was evaluated through its application in several cases presented in literature, including a large size example from a Brazilian refinery. The results show that the proposed approach was capable of finding similar and, sometimes, superior results compared to the literature ones, concluding that the developed method was capable of providing good results with low computational efforts