An improved mathematical programming model for generating optimal heat-exchanger network (HEN) designs

The traditional model-based approach to synthesize the heat-exchanger networks (HENs) was developed on the basis of several unrealistic assumptions. An improved mixed-integer nonlinear programming (MINLP) model is thus developed in this work to circumvent these drawbacks. Specifically, a well-established empirical relation is introduced in the improved model formulation to account for the variation in heat-transfer coefficient with respect to flow rate, and, furthermore, the concept of heat-exchanger efficiency is incorporated to produce an accurate estimate of the heat-transfer area in each exchanger. As a result, it is possible to produce more cost-effective HEN designs with this model. The optimization results obtained in various case studies also show that the process conditions of individual exchangers in the optimal network can generally be tuned simultaneously to achieve a proper compromise between high efficiency and low irreversibility.