Optimal Synthesis of Heat Exchanger Networks Using Enthalpy-Temperature Functions to Describe Streams

This work presents a mixed integer non linear programming (MINLP) model for heat exchanger network synthesis including a detailed description of the process streams resorting to a mathematical function that matches temperature with enthalpy. This allows the model to account for non linear behavior of streams by taking into consideration their composition. To achieve the objective, the energy balance equation proposed in the "synheat" model by Yee and Grossmann (1990) is modified but equations involving temperature are kept the same. A new equation is added to that model to describe the relationship between temperature and enthalpy by correlation. As the original model, this modified MINLP model provides the network structure that minimizes the total annual cost. The mathematical program has non-convex equations, and only locally optimal solutions can be guaranteed. The approach between cold and hot streams has more detailed information because describes the real behavior of streams, and avoids matches that a linear functionality or a classical model may allow. Correlations between temperature and enthalpy are a function of the components present in a particular stream. The required information can be collected from bibliography or public databases as those provided by NIST (National Institute of Standards and Technology). The proposed methodology is a useful tool when the minimum temperature approach between streams is small. Examples are presented and discussed to compare results when streams are described by linear and non linear functionality between temperature and enthalpy.